Stp1021-eb.25232.pdf

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep

STP 1021

Turbine Oil Monitoring

William C. Young and Reginald S. Roberton, editors

ASTM 1916 Race St. Philadelphia, PA 19103

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductio

Library of Congress Cataloging-in-Publication Data Turbine oil monitoring/William C. Young and Reginald S. Roberton, editors. (STP ; 1021) Papers presented at the symposium held in Orlando, FL on 8 Dec. 1987 and sponsored by ASTM Committee D-2 on Petroleum Products and Lubricants. "ASTM publication number (PCN) 04-01021012"--T.p. verso. Includes bibliographies and index. ISBN 0-8031-1252-1 1. Turbomachines--Lubrication--Congresses. I. Young, William C. II. Roberton, Reginald S. III. ASTM Committee D-2 on Petroleum Products and Lubricants. IV. Series: ASTM special technical publication; 1021. TJ266.T873 1987 621.406--dc19 Copyright 9 by

89-301 CIP

AMERICAN SOCIETY FOR TESTING AND MATERIALS 1989

NOTE The Society is not responsible, as a body, for the statements and opinions advanced in this publication.

Peer Review Policy Each paper published in this volume was evaluated by three peer reviewers. The authors addressed all of the reviewers' comments to the satisfaction of both the technical editor(s) and the ASTM Committee on Publications. The quality of the papers in this publication reflects not only the obvious efforts of the authors and the technical editor(s), but also the work of these peer reviewers. The ASTM Committee on Publications acknowledges with appreciation their dedication and contribution of time and effort on behalf of ASTM.

Printed in Baltimore, MD April 1989

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authoriz

Foreword This publication, Turbine Oil Monitoring, contains papers presented at the symposium of the same name held in Orlando, Florida on 8 Dec. 1987. The symposium was sponsored by ASTM Committee D-2 on Petroleum Products and Lubricants. William C. Young, General Electric, and Reginald S. Roberton of Princeton, New Jersey presided as symposium cochairmen and were coeditors of this publication.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. N

Dedication This book is respectfully dedicated to Andrew N. Smith and Reginald S. Roberton in recognition of their years of effort as, respectively, chairman and secretary of Subcommittee D02.C on Turbine Oils. In their offices they promoted, both in the United States and internationally, greater knowledge of turbine lubrication and the use of ASTM standards to improve turbine lubrication performance. ASTM D 4378, "Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines," the focus of this symposium, was the result of their dedication.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No f

Contents Overview Background and Development of ASTM D 4378: Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines--R. S. ROBERTON

A Turbine Lube Oil System Monitoring Program~D. BONETT

3 19

A Supplier's Approach to Turbine Oil Monitoring---G. R. TRABERTAND G. J. SCHREUDERS

25

The Reliable Control of Oil Quality in Japanese Turbine Units--R. OHGAKE, M. SUNAM1, T. YOSHIDA, AND H. WATANABE

Monitoring Practices for Phosphate Ester Fluids for Turbine Lubrication and Control---J. F. ANZENBERGER,SR.

32

44

Experiences with an Analytical Monitoring Program Designed for Commerical Flight Tests of Advanced MiI-L-23699 Turbine Fluids-D. G. PACHUTA, J. A. THICH, R. W. KNIPPLE, AND D. A. STEPHANIC

54

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductio

STP1021-EB/Apr. 1989

Overview

ASTM D 4378, Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines, was developed by Subcommittee D02.C on Turbine Oils in response to the needs of the electrical power generation community. The organizations for which this document is intended are the utilities and industries which generate their own electrical power. The purpose of D 4378 is to recommend: 1. Which tests to run. 2. The frequency of testing. 3. The interpretation of the results. 4. Corrective actions. 5. Limiting values. This symposium brought together suppliers and consumers to present their views on turbine fluid monitoring. An overview was presented by Roberton giving the user needs that generated D 4378. The history of the document was discussed with the reasons for having the tests which are included, the basis for interpreting the results, and the appropriateness of the limits. The necessity for a monitoring program was brought forth by Bonett, who presented the experiences of an American utility. These show that a program can identify problems so that corrective action can be taken in time to prevent a major shutdown. The views of an American lubricant supplier were given by Trabert and Schreuders, who stressed the need for close cooperation between the turbine operator, the turbine manufacturer, and the lubricant supplier to ensure the successful operation and effective monitoring of a turbine circulating system. Ohgake and his associates reported on the technological revolution in Japan since 1955. Maximum unit capacity has increased from 66 M W in 1955 to 1000 MW in 1987. A t the same time, steam pressure of 60 kg/cm 2 in 1955 has risen to 246 kg/cm 2 with a corresponding rise in steam temperature from 480~ to 566~ Turbine oil operating temperatures increased by 10 to 15~ thereby potentially more than doubling the oxidative degradation rate of the lubricant. Ohgake confirmed the use of ASTM D 2272 in their monitoring programs and for the calculation of turbine severity, and also indicated that the concept of regular makeup (replacing part of the oil charge) to compensate for oil degradation was still considered a useful practice to ensure long oil service life. In Japan and in the United States, increased demands have been placed on turbine lubricating oils by changes in the design of the turbine generators. The increased use of combined cycle stations has added the requirement that the same lubricant be used in both gas and steam turbines. (In a combined cycle installation,

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed Copyright9 byby ASTM International www.astm.org Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No furth

2

TURBINE OIL MONITORING

heat recovered from the gas turbine exhaust is used to help generate steam for driving a steam turbine.) These new demands on the oil have made monitoring programs even more important than previously. Although D 4378 addresses the more commonly used petroleum oil turbine lubricants, fire-resistant fluids based on phosphate esters are increasingly being used for electrohydraulic control (EHC) system lubricants. They are also used in some applications to lubricate steam and gas turbines. Documents similar to D 4378 but covering phosphate ester fluids are being developed both in ASTM and the International Standards Organization (ISO). Anzenberger covered the program of a manufacturer of such fluids. This manufacturer provides sample bottles in which fluid can be sent to the supplier's laboratory for testing. A report sent to the user gives results, recommended limits, and suggested corrective action if appropriate. Although the gas turbines covered in D 4378 are heavy duty, that is, not aircraft, the paper by Pachuta et al. demonstrates the desirability of a monitoring program in testing lubricants for aircraft jet engines. The methods of storing and plotting data are of interest to any monitoring program. With increased use of aircraft derivative gas turbines for ground applications monitoring programs for these engines can be expected in the future. A previous symposium, "Aspects of Lubricant Oxidation," published as STP 916, discussed methods of measuring oil oxidation and the significance of oxidation on remaining oil life. It would be well to review STP 916 in conjunction with this present document, STP 1021, since oxidation is of major concern in oil monitoring. Many of the observations of Dr. Pachuta and his colleagues on oil oxidation, particularly those on total antioxidant capacity (TAC), are relevant to the previous STP. It was the intent of the symposium to review the current status of turbine oil monitoring. It emphasized the importance of a monitoring program to maintain power generation equipment in good operating condition. It is hoped that the readers of this STP will give the editors their suggestions on improvements to D 4378 as well as their thoughts on what other monitoring documents appropriate for power generation should be prepared under the aUspices of D02.C. D 4378 is the basis for an ISO document on the same subject which should be issued in about one year. As mentioned previously, similar documents are being prepared in both ISO and ASTM on phosphate ester fluids for electrohydraulic control (EHC) systems. As experience accumulates with users of these documents, an appropriate subject for a future symposium would be a review of these and user experience with them as well as an update on D 4378. We wish to thank the authors for their contributions, the paper reviewers for their time and thoughts, the ASTM personnel whose efforts have resulted in this publication, and the members of ASTM Subcomittee D02.C who prepared and reviewed D 4378.

William C. Young General Electric, Schenectady, NY 12345; symposium cochairman and coeditor.

Reginald S. Roberton Princeton, NJ 08540; symposium cochairman and coeditor.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

Reginald S. Roberton I

Background and Development of ASTM D 4378: Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines REFERENCE: Roberton, R. S., "Background and Development of ASTM D 4378: Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines," Turbine Oil Monitoring, ASTM STP 1021, W. C. Young and R. S. Roberton, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 3-18. ABSTRACT: Several years ago a study group was set up at ASTM (American Society for Testing and Materials) under Subcommittee C (Turbine Oils) of Committee D-2 on Petroleum Products and Lubricants to develop a standard for in-service monitoring of steam and gas turbine oils. The study group included well qualified and experienced representatives from turbine equipment builders, oil producers, the power generation industry, and consultants among its members. The objective was achieved and an ASTM standard D 4378, entitled "Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines" was issued as a single standard in 1984 and is now included in the Annual Book of ASTM Standards. In developing the standard, it was recognized that turbine oils are unique lubricants requiring the highest degree of oxidation stability and resistance to sludge formation and are expected to give years of service. Also, the high cost of oil changes and unplanned shutdowns had to be balanced against the cost of laboratory manpower and equipment for analytical testing. This paper covers the results of an initial study of monitoring practices within the power generation industry and discusses in some detail why some methods in use for many years are of doubtful significance and should be eliminated. It also describes the development of sampling and testing schedules. To follow the oxidative degeneration of the oil, the well-established TAN (Total Acid Number) test is retained and used in conjunction with the Rotary Bomb Oxidation Test, (RBOT) (D 2272) 2 to provide a much more reliable indication of remaining service life. Other properties of the oil which must be retained in service are discussed, and major significance has also been given to the interpretation of test data and appropriate action steps. The technical merit of D 4378 and its acceptance by the international community is confirmed by the fact that it has been used by a joint working group of ISO (International Standards Organization) and IEC (International Electrotechnical Commission) as a basis for an international maintenance and use guide. KEY WORDS: oxidation, antioxidant, turbine severity, oxidation stability reserve, degradation, oil service life, contamination, oil make up rate, warning limit

The involvement of A S T M in the preparation of an in-service monitoring practice for turbine oils grew out of the activities of D-2 Subcommittee C, which supplements its w o r k on standard methods with discussions on turbine oil problems and their resolution. The n e e d Consultant, 497 Herrontown Road, Princeton, NJ 08540. Test Method for Oxidation Stability of Steam Turbine Oils by Rotating Bomb (D 2272).

3 Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Copyright9 by ASTM International www.astm.org Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

4

TURBINEOIL MONITORING

for a review of lubricant test methods in use and a more systematic approach to monitoring practices was dear. The high capital cost of equipment and the need to avoid costly unplanned shutdowns were important considerations. As a result, a direct request was made to A S T M by a section of the power generation industry to develop a standard monitoring practice for steam and gas turbine oils. In addition to the needs of utilities, both the turbine manufacturers and the major oil producers expressed an interest in this venture. Many of the turbine builders had experienced problems of oil degradation and difficulties associated with existing oil maintenance and handling practices. Also, most of the major turbine oil suppliers, having for years provided their customers with technical service, had developed some expertise in this area. Subsequently, as a further encouragement to this effort, it became clear that the International Standards Organization and the International Electrotechnical Commission were becoming engaged in the preparation not only of turbine oil specifications but of an international Maintenance and Use Guide for turbine oils. This acted as a spur to complete our work promptly so that the ASTM practice could be a possible basis for their standard. In order to carry out this work, a new study group, Section 9, was formed in Subcommittee C, under the chairmanship of the author, Reginald Roberton. This paper describes how the standard was developed and highlights the areas of major significance.

Purpose and Objectives The main purpose was to provide a monitoring practice for use by those responsible for the operation and maintenance of power plants at major utilities and other industrial plants. The objective was the preparation of a fairly comprehensive guide to the following: 1. An understanding of what turbine oils are, their properties, their sensitivity to contamination, and how they degrade. 2. Identifying operational factors, apart from oil quality, that affect oil performance and service life. 3. The handling of new oil deliveries and tests to run to safeguard the quality of the oil charged to the turbines. 4. The setting up of sampling and testing schedules in order to follow changes in the oil's condition. 5. The importance of proper sampling techniques. 6. The interpretation of test results. 7. Appropriate action steps.

Turbine Oils--Their Nature and Composition Steam and gas turbine oils are unique lubricants. In steam turbines they must operate in wet conditions and show exceptional stability over long periods of service. In gas turbines they must withstand contact with very hot surfaces, often with intermittent operation and periods of nonuse. Therefore, to be effective, both types of oil must have, in addition to good corrosion protection and demulsibility, outstanding resistance to oxidation, which includes a minimum tendency to form deposits in critical areas of the system. To achieve these desired properties, it is necessary to formulate these oils from specially refined base stocks of the highest quality plus a carefully balanced additive package. It takes many years to develop and field test a new turbine oil formulation. The nature of these fluids makes them very susceptible to contamination, particularly from other lubricants and additives. A relatively small degree of contamination can markedly affect the properties and Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreemen

ROBERTON ON D 4378

5

expected service life of these lubricants. This becomes especially important at all stages of their manufacture, transportation, and use. O n e of the objectives of the standard was to make the user aware of the sensitive nature and application of turbine oils.

Study of Existing Practices In commencing our work, it was felt that a good basis to start would be to m a k e a study of test methods and practices in use in the p o w e r generation industry. With assistance from the Edison Electric Institute ( E E I ) , who graciously allowed us to use the results of a questionnaire to utilities, and information supplied by oil producers, turbine builders, and consultants, a valuable insight was obtained on industry monitoring practices. It was obvious from the results of this study that test methods and the interpretation of test data were far from standardized and confirmed the need for a m o r e effective and consistent approach to monitoring oil condition. A summary of the results of this survey are shown in Table 1. The most frequently used tests are shown at the top of the table. It would be useful at this stage to study the relevance of some of these tests in greater detail. TABLE 1--Preliminary study of test methods in common use. Test

% of Users Using Test

Comments

Total acid number

91

Viscosity

87

Moisture

66

Color

60

Sediment

45

Interfacial tension

42

Percent antioxidant

19

Rotary Bomb Oxidation Test (RBOT)

17

Gravity Rust test

17 17

Silicon

17

Steam emulsion no.

11

Flash point Oxidation by infrared

11 9

This is the most frequently run test. It is a very valuable test, but wide variations were noted in the controlling limits used. Widely used. Useful but not critical. Most useful to detect contamination. Lack of agreement on level which could be harmful. Opinions vary as to significance of color. Sudden darkening is a cause for alarm. No universal standard in use but most often associated with cleanliness level and contamination by particulate matter. Still fairly widely used. Erroneously used to detect degradation. Test is out of date since introduction of additized oils which often have low initial values. Considerable controversy over usefulness or validity of this test. No recommended. Not yet widely used but usage is growing. Shows significant promise to follow degradation of oil in service. Useful as a predictive tool. Often used to detect contamination. Used rarely since there is usually little change in service if original oil is OK. Indicative of dirt contamination. Spectrographic method not accurate or reliable. Still used by a few power plants although no longer an ASTM test. ASTM D 1401 demulsibility is generally preferred. Only of value to detect contamination. Used by only a few companies. Method offers advantages over the Total Acid Number. Accepted as a necessary requirement for new oil but used very little for monitoring purposes.

Water Separability (D 1401)

9

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No furthe

6

TURBINE OIL MONITORING

As expected, the Total Acid Number (D 974) 3 is recognized by most utilities as a useful test for determining the oil's oxidative condition.We agree with this consensus despite the fact that its precision is not too good at low values. Most plants have the facility for running this test. Surprisingly, very few utilities use an infrared method for determining the oil's condition. This may be due to lack (or cost) of equipment or to lack of a specific test method. For many years analytical testing services laboratories and oil suppliers have used a method based on absorption in the infrared (at 1710 cm-~). The absorbance is directly related to the concentration of the carboxyl ( - C O O H ) functional group. As would be expected, this also implies a correlation with Total Acid Number (TAN) results. A T A N of 0.1 is approximately equivalent to 1.0 Absorbance. The use of long (500 I~m) cells lends greater accuracy to the IR method when applied to turbine oils. The IR method is not included in the standard mainly because most plants do not have the equipment. It was interesting to note the growing use of the Rotary Bomb Oxidation Test (D 2272). In our opinion this test represents a most useful and reliable tool for determining the oxidative state of the oil and therefore its remaining service life. The use of R B O T will be further discussed under the section, "Quality Characteristics Which Must Be Retained In Service." The survey also revealed the use of two tests which are unreliable and often provide misleading information. One test is Interracial Tension (IFT) (D 971), 4 its use being based on the assumption that oxidation of the oil is identified with the lowering of the IFT. When the test was first used, possibly over 40 years ago, this assumption was probably correct for the type and composition of the turbine oils on the market. A t that time new turbine oils had very high I F T values (>40 dynes/cm), and relatively small changes in the oil's oxidative condition resulted in a substantial drop in the IFT value. However, additive technology has changed and now many first class turbine oils, when new, have low IFT values (around 20 dynes/cm) and do not exhibit a rapid drop in IFT when the oil starts to degrade. The other test is the Percent Antioxidant. There is still some controversy on the subject, but the facts are these. Historically, this is an infrared method for the determination of DBPC (ditertiary butyl paracresol) by measuring the absorbance due to the phenolic functional group. On the surface there appears to be some validity to this test if the turbine oil antioxidant system relies entirely on DBPC. A drop in DBPC concentration is construed as consumption (or loss) of additive and that when the concentration reaches zero the oil would be unprotected and would degrade very quickly. However, two additional facts should be considered. Firstly, evidence has accumulated which shows that many oils based on DBPC have not shown marked instability when the apparent concentration of DBPC drops to zero. Many oils have continued to give satisfactory service for years after this point. To explain this, it has been suggested that the DBPC has entered in reactions, possibly involving trace constituents in oil and oxygen, to form an intermediate. In this case, absorption in the infrared due to the phenolic functional group no longer occurs, while the species retains, to some degree, its properties as an antioxidant. Secondly, and of even more importance, is the fact that many excellent present day turbine oils no longer rely fully or even in part on the use of DBPC. Some oils now contain more than one antioxidant, usually of different chemical types. It is usually possible to develop a specific test method for each antioxidant, but this requires the knowledge of the formulation. This is outside the scope of the present standard practice.

a Test Method for Neutralization Number by Color-Indicator Titration (D 974). 4 Test Method for Interracial Tension of Oil Against Water by the Ring Method (D 971).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

ROBERTON ON D 4378

7

Format and Contents of Standard Practice The table of contents of the new "Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines," D 4378-84a, is as follows: Introduction Scope Referenced Documents Significance and Use Properties of Turbine Oils Operational Factors Affecting Service Life Sampling Examination of New Oil on Delivery Deterioration of Turbine Oils in Service How Turbine Oils Degrade Properties of Oils Which Must be Retained Other Properties of Used Oils Monitoring Program Sludge and Deposits Test Schedules Steam Turbines New Oil Installation of New Oil Charge In-Service Testing Gas Turbines (nonaircraft derivative) New Oil Installation of New Oil Charge In-Service Testing Interpretation of Test Data and Recommended Action Steps The following section discusses some of these topics in detail, explaining the basis for choice of methods and concentrating on critical areas relating to oil degradation in service.

Operational Factors In order to make the standard a self-contained guide to ensure trouble-free operation of the turbine lubrication system, operational factors other than oil quality that affect oil service life have been covered in some detail. These factors are: 1. 2. 3. 4. 5.

Type and design of system. Condition of system at startup. System operating conditions. Contamination. Oil makeup rate.

The severity level at which a turbine operates is a combination of all the above plus original oil quality. It is identified with the expected service life of the lubricant. The more severe a turbine, the shorter the service life of the oil. Turbine severity has been defined as the "Per cent of fresh oil oxidation resistance lost per year due to oil reactions in the turbine."

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement

8

TURBINEOIL MONITORING

An equation for calculating turbine severity is included in an Appendix to the standard. The equation was developed by DenHerder and Vienna [1]. It is not the purpose of this paper to cover, in detail, the factors described above. This has been done in the standard. However, the oil makeup rate deserves further comment since it plays such a significant part in determining the life of the oil charge. In the United States the annual makeup varies from a low of 1 to 2% to a high of 30% in extreme cases, with an average probably around 7% per year. In turbines where makeup is low, a fairly true picture of oil degradation is obtained based on system conditions. However, under the same system conditions at high makeup, the oil service life will be considerably extended. A point can be reached where makeup more or less compensates for the degree of degradation and the oil has an indefinite service life. This is shown graphically in Fig. 1. Quality Characteristics Which Must Be Retained In Service Turbine lubrication systems are particularly sensitive to changes in oil condition, and there should be a clear understanding by the power plant operator of what quality characteristics of the oil are most important and how these are measured. We have identified the following oil quality characteristics which must be retained to ensure safe and continuous operation of the turbine: (a) viscosity; (b) oxidation stability reserve; (c) freedom from sludge; (d) freedom from abrasive contaminants; (e) anti-corrosion protection; (f) water separability. These are discussed in some detail in the standard but merit further comment in view of their importance in developing the standard.

a = ZERO H A K E - U P b = 15~/YR.HAKE-UP c = 30S/YR.'HAKE-UP

LLI ._J

I00.

~> (:3 rn

n-

75,

_J Z L9 n-"

o

50-

LL 0 C

I-Z W r.J

25. t.d EL

I~0 YEARS

OF

1'2

SERVICE

FIG. 1--Effect of makeup on service life. Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authorized.

ROBERTON ON D 4378

9

Viscosity Change in viscosity even by as much as plus or minus 20% would have very little effect on lubrication. However, such an order of change usually signifies contamination, possibly due to the use of the wrong oil as makeup. For this reason, viscosity should always be included in a monitoring schedule. ASTM Method D 445 is normally used (ASTM Test Method for Kinematic Viscosity of Transparent and Opaque Liquids and the Calculation of Dynamic Viscosity).

Oxidation Stability Reserve This is the most significant property relating to its condition and its suitability for further service. In the development of improved high quality steam turbine oils, ASTM Method D 943, 5 known as the Turbine Oil Stability Test or TOST, has played a very significant role despite many attempts to discredit it. Over 25 years of experience has confirmed its usefulness and validity. In the view of our panel, it is a necessary and meaningful specification test for new oils. Unfortunately, it is a long test (over 2000 h for a good turbine oil), making it unsuitable for a monitoring program. Oxidation of hydrocarbon oils proceeds by a chain mechanism initiated by a peroxide radical. Intermediate compounds formed include aldehydes, ketones, and fatty acids. A measure of the changes taking place can be followed in various ways, the simplest being to monitor the change in acidity or more reliably to follow the decrease in the oxidation stability. The acidity is measured using ASTM D 974 Total Acid Number (often called Neutalization No.) and the oxidation stability by ASTM D 2272 Rotary Bomb Oxidation Test (RBOT). The Total Acid Number (TAN) is easy to run but lacks precision at values between 0.1 to 0.5 mg KOH/gm, which is the range we are interested in. Therefore, we strongly advocate that the TAN test be augmented by running the RBOT at regular intervals, particularly when the TAN values begin to increase. Experience has shown remarkably good precision for the RBOT on both new and used turbine oils. It has been used successfully and endorsed by many investigators of turbine oil performance in the United States [1] and Japan [2]. The significance of actual values and trends is discussed later in this paper under "Logging and Interpretation of Data."

Freedom from Sludge Although for many types of lubricants this may seem a normal requirement, there is a need to emphasize its importance with respect to steam and gas turbine oils. First, there are many parts of turbine oil systems where even small amounts of sludge can lead to problems and malfunction. Second, there is an unusually large amount of oil involved (up to 7000 gal in some turbines), and, although the sludge content of the oil may be hardly detectable, it can accumulate in sensitive areas of the system. What would be of minor significance in other types of equipment might be catastrophic in a steam or gas turbine. It must be borne in mind that turbine oils are highly paraffinic in nature and therefore exhibit very poor solubility for their own oxidation products. This often leads to the phenomenon of "varnishing" of metal surfaces, which occurs before the onset of visible sludge. The oil is degraded to a point where sludge precursors are present in a quasisoluble state. The oil should be discarded before this condition exists. Unfortunately, there are no direct methods to measure and monitor the concentration of these sludge precursors. The condition s Test Method for Oxidation Characteristics of Inhibited Mineral Oils (D 943).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

10

TURBINEOIL MONITORING

of the oil has to be determined from its oxidation stability reserve, as measured by TAN and RBOT.

Freedom from Abrasive Contaminants The presence of abrasive contaminants in the oil cannot be tolerated since they promote scoring and damage to bearings and journals as well as cause malfunction of control mechanisms. The possible source of contaminants is varied. Major contamination can occur during installation of the turbine and when it is opened for maintenance and repair. In very dusty areas where units may be out of doors, some solids may enter through improperly installed vents. There is also the possibility of the buildup of tramp wear metal due to wear in journal and thrust bearings, gears, and pumps or as the result of rusting and corrosion. Possibly the most common source is through the makeup oil when it is added to the system. Whatever the source, these contaminants must be removed from the system by the use of filters or centrifuge or both. In a properly maintained system with low makeup, this presents no problem. With high makeup, careful filtration of the oil added to the turbine is essential to minimize the risk of any possible buildup of contaminants. The oil in a wellmaintained system tends to improve in cleanliness during operation. Cleanliness can be determined gravimetrically using ASTM F 313 (Test Method for Insoluble Contamination of Hydraulic Fluids by Gravimetric Analysis), but the use of electronic particle counters is becoming much more widespread. These latter instruments are good, but in the hands of inexperienced personnel, they are subject to a wide range of errors. Sampling is critical. Frequent standardization is required and care must be taken to ensure that the sample contains no trace of free water, since minute water droplets are measured as particles.

Anticorrosion Protection Protection against rusting is very important. However, present additive technology is extremely effective and very few turbines experience rusting problems. Antirust inhibitors maintain their effectiveness for long periods of time, and an occasional check is all that is necessary. ASTM Method D 665, 6 Procedure A is a satisfactory rust test for turbine oils. Rarely is there a problem of corrosion due to acidity arising from oil oxidation since the oil charge is removed from the system before the acidity reaches a high enough level to be corrosive.

Water Separability--During the operation of steam turbines, separation of water from the oil must be effective and clean. In practice this is rarely a problem since experience h a s shown that, if the original oil charge had good demulsibility characteristics, the oil will maintain good water separability over long periods of time. The exceptions to this are contamination with highly emulsive oils or serious oxidative degradation. ASTM D 1401 (Test Method for Water Separability of Petroleum Oils and Synthetic Fluids) can be used to check the oil, when necessary. 6 Test Method for Rust-Preventing Characteristics of Inhibited Mineral Oil in the Presence of Water (D 665).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

ROBERTON ON D 4378

11

Other Properties Although not considered an essential property, color does have some significance and will be considered here. There is a diversity of opinion as to the significance and importance of color degradation. In Japan many utilities include a tight color condemning limit in their monitoring schedules, and the oil is changed merely on this basis. However, it is our experience that a gradual darkening of color is normal with many excellent fluids and is of little concern if other tests indicate a satisfactory oil condition. On the other hand, sudden changes in color are a warning that something is wrong and needs investigation. In this case, if the cause is not found and remedied, accelerated oil degradation will most likely follow.

Development of the Monitoring Program The main purpose of a monitoring program is to protect the equipment and ensure long, trouble-free utilization of the oil in service. In general, the same type of tests can be used for both steam and gas turbine oils, although the test frequency needs to be increased substantially for high-severity gas turbines. The life of a turbine oil may be as long as 10 to 20 years in a steam turbine and may possibly reach or exceed 100 000 h in a low-severity gas turbine, whereas a range of 10 000 to 30 000 h would be considered satisfactory for a high-severity gas turbine oil. A monitoring program should not include unnecessary testing since this is expensive, but it must ensure protection in cases of unexpected changes in operational factors which might lead to rapid oil degradation and consequent danger to the equipment. The following points were considered important in choosing the test methods and determining the frequency of testing. 1. The first twelve months of operation of a steam turbine, or 4000 h for a gas turbine, should be used to assess the severity of the turbine oil system. This period must be covered by a sufficiently detailed testing program in order to obtain reliable information on turbine oil performance. 2. Following this initial period, a comparatively long period should ensue when sampling and testing can be reduced to a minimum. 3. As the oil approaches the end of its service life, increased frequency of testing is again required in order to determine, with some precision, the point at which the oil should be discarded. 4. A t all times unusual results in any one area suggesting either degradation or contamination have to be followed up quickly with investigation and, if necessary, increased testing to determine the cause of the problem.

The Program The monitoring program consists essentially of three parts, all of which are of equal importance. 1. The sampling and testing schedule. 2. Logging and interpretation Of test data. 3. Taking necessary action steps.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No

12

TURBINEOIL MONITORING

Sampling and Testing Schedule The tests included in the schedules (Tables 2 and 3) are run to determine the extent of (a) contamination and/or (b) degradation. Generally we relate viscosity, appearance, water content, and cleanliness to contamination, whereas TAN, color, and RBOT are related to degradation. Viscosity is not included as an indicator of degradation since TAN and RBOT are more sensitive indicators. Oils are normally changed out before any significant viscosity increase occurs. The use of RBOT in addition to TAN is highly recommended, and these two tests are linked in interpreting the condition of the oil. TABLE 2--Sampling and testing schedule--steam turbine oils. Frequency Test

First 12-Month Operation

Normal Schedule

Viscosity Total acid no. Appearance Water Color Rust test Cleanliness RBOT

1 to 3 months Monthly Daily Monthly Weekly 6 months 1 to 3 months 2 to 3 months

3 to 6 months 1 to 3 months Daily 1 to 3 months Weekly 1 Year 1 to 3 months 6 to 12 months

TABLE 3---Sarnpling and testing schedule--gas turbine oils. Frequency Test

First 6-Month Operation, h

Normal Schedule, h

Viscosity Total acid no. Appearance Color Cleanliness RBOT

500 500 100 200 500 500 to 1000

500 500 to 1000 100 200 1000 1500 to 2000

Logging and Interpretation of Test Data The importance of keeping an up-to-date record of results needs to be stressed. In addition, graphical representation of properties which change as the oil degrades is highly recommended so that any trend can be carefully watched. Possible errors in test results can often be detected by reference to the graphs and new samples taken for check testing. This applies particularly to the TAN test and the RBOT. For easy reference the standard contains a table, reproduced as Table 4 in this paper, which gives guidance on the interpretation of test results and warning limits. Most of it is self-explanatory. However, the rationale behind the choice of certain warning limits and data interpretation deserves further explanation. With respect to oxidation, not all turbine oils show the same behavior mainly due to differences in the base stocks and additive packages used. This is illustrated in Fig. 2. Oil A shows a low TAN value for a long period, then shows a distinctive break when the TAN value increases very quickly. Oil B, however, shows a fairly regular increase in TAN value

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreem

ROBERTON ON D 4378

13

during the life of the oil. Comparison of the condition of the oils at the half life point would suggest that Oil A is the better oil, which, in fact, is not so. This is one reason why regularly scheduled testing is necessary in order that trend changes can be detected. It should be noted that although Fig. 2 shows an oil life of around 14 years, this is not to suggest that this is a normal or expected lifetime for steam turbine oils. Oil life is a function of many factors, the most important being turbine severity, which can vary widely. Oils which last for 14 years or more in one turbine may be spent in three to four years in another turbine. Since the oxidative condition of the oil is the most important single factor is determining if an oil is likely to throw down sludge or take off in an accelerated oxidation mode, considerable thought has been given to establishing safe warning limits. When the T A N value increase is in the range 0.3 to 0.4 mg K O H / g m , the oil may be at or approaching the end of its service life. This is based on observations of oils in service and on laboratory studies based on the D 943 Turbine Oil Stability Test. It is significant that in the D 943 test most oils show an accelerated oxidation rate when the TAN exceeds 0.3 mg K O H / g m . Recent studies by Stringer, Burton, and Brown at the National Research Council, Canada [3] have confirmed this for a number of turbine oils. Therefore, when the T A N value reaches 0.3 mg K O H / g m , the operator should be aware that the oil may be in a marginal condition. It is at this point that several steps should be taken. Testing frequency should be increased and, if current R B O T data are not available, regular RBOT tests should be included in the schedule. Both the TAN and RBOT should be utilized in the final analysis to determine the point at which the oil should be changed. An RBOT of less than 25% of that of new oil represents a danger point even if TAN values remain consistent within the 0.3 to 0.4 range. Cleanliness level as determined by particle counting is becoming more widely used, and some turbine manufacturers are setting limits. As already discussed, considerable care is

._.1

'n,. I.-t.9

hi t..) Z N

~

~,

"8

1"o

1"2

14

YEARS OF SERVICE

FIG. 2--Degradation patterns of two different turbine oils.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. N

0.3 to 0.4 mg KOH/gm

Less than half value on original oil

Less than 25% of original.

Exceeds 0.2%

Total Acid No. Increase over new oil

RBOT

RBOT

Water Content

Up to 20 000 h Up to 3000 h At any time

At any time

S, G

S

At any time during life of oil charge

Up to 20 000 h Up to 3000 h

S G

S, G

S G

Steam (S) or Gas Turbine (G)

Together with high TAN indicates oil at or approaching end of service life. Oil contaminated. Potential water leak.

Above normal degradation.

This represents above normal deterioration. Possible causes are: .(a) system very severe; (b) anti-oxidant depleted; (c) wrong oil used; (d) oil contaminated. Oil at or approaching end of service life, (c) or (d) above may apply.

Interpretation

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authorized.

0.1 to 0.2 mg KOH/gm

Warning Limit

Total Acid No. Increase over new oil

Test

Oil Life (Running Hours)

TABLE 4---Interpretation of test data and recommended action.

Investigate and remedy cause. Clean system by centrifugation. If still unsatisfactory consider oil change or consult oil supplier.

Resample and retest. If same, consider oil change.

Look for signs of increased sediment on filters and centrifuge. Check RBOT. If RBOT less than 25% of original, review status with oil supplier and consider oil change. Increase test frequency if left in system. Investigate cause. Increase frequency of testing.

Investigate cause. Increase frequency of testing--compare with RBOT data. Consult with oil supplier for possible reinhibition.

Action Steps

T~ fi3

O

-4

O z

I""

o_

z Ill

---I c

4x

Light fail

Light fail

Hazy

Unusual and rapid darkening.

-+20% from original oil viscosity.

Drop 300F or more compared to new oil. Exceeds following limits Tendency---450 Stability--10

Rust Test D665A

Rust Test D665A

Appearance

Color

Viscosity

Hash Point

S, G

S, G

S, G

S,G

S,G

S, G

S

S, G

At any time

At any time

At any time

At any time

After 20 000 h during life of oil charge At any time

Up to 20 000 h

At any time

Possibly contamination or antifoam depletion. In new turbines residual rust preventives absorbed by oil may cause problem.

Oil contains water and/or solids. This is indicative of: (a) contamination or (b) excessive degradation. (a) Oil is contaminated or (b) Oil is severely degraded. Probably contamination.

Source of particulates may be: (a) make-up oil; (b) dust or ash entering system; (c) wear condition in system. (a) The system is wet and/ or dirty. (b) The system is not maintained properly (e.g. water drainage neglected, centrifuge not operating.) Normal additive depletion in wet system.

Rectify cause. Check with oil supplier regarding inhibition. Note: Plant problems often mechanical in origin.

Determine cause. Check other quality parameters. Consider oil change.

Determine cause. If viscosity is low determine flash point. Change oil, if necessary.

Investigate cause and remedy. Filter and/or centrifuge oil. Determine cause and rectify.

Consult oil supplier regarding reinhibition.

Investigate cause and make necessary maintenance and operating changes. Check Rust Test. Consult oil supplier regarding reinhibition if test result unchanged.

Locate and eliminate source of particulates. Clean system oil by filtration and/or centrifuging.

NOTE 1: Limits will depend on method used (consult equipment builder or oil supplier), Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authorized.

Foam Test D892 Sequence I

Exceeds accepted limits (Note 1)

Cleanliness

ol

..g

.,4 O0

Z

--I O z O

ITI

O IgJ

16

TURBINEOIL MONITORING

required as well as some expertise to get accurate results. Using the National Aerospace Standard 1638 of measurement, a maximum rating of 8 is considered satisfactory, although, in tight systems with low makeup, ratings of 6 or 7 can be obtained. Warning limits and interpretation of other test data are more or less self explanatory. Table 4 does include flash point and foam tests, which do not appear in the testing schedules. These need only be used if contamination is suspected.

Taking Necessary Action Steps Because the change in the condition of a turbine oil in service is usually very gradual, immediate and drastic action is rarely necessary unless it is a case of contamination. There is usually time to follow a trend and plan actions in advance. The use of the term "warning limit" is used rather than "condemning limit" in order to alert the power plant operator that some action is required even though the oil may still be in a usable condition. Along with the interpretation of test data in Table 4 are recommended action steps. The interpretation includes possible causes of the condition, which will assist the user to rectify the situation. If the oil is reaching the end of its service life, this has to be confirmed. The user is warned against taking drastic action or making important decisions based on one sample. In the case of oxidative degradation, the use of graphical plots for TAN and R B O T are reiterated. The question of reinhibition often crops up. The term applies to the addition of an additive concentrate to remedy a deficiency or loss of a certain characteristic of the system oil. For turbine oils we consider this appropriate for reestablishing good rust protection or improving the antifoaming qualities. In some rare cases the use of a demulsifier has proved successful.To be effective such additions have to be carefully dispersed and the user needs to obtain the advice and help of the oil supplier. Reinhibition with antioxidant is not usually productive. In the author's experience, loss of oxidation stability cannot be effectively remedied in this way. Usually the R B O T is hardly affected by extra antioxidant, or, if an initial boost is obtained, it is not sustained. It appears that once the oil itself has undergone substantive oxidative changes, additional inhibitor has only a small and temporary effect. This has been substantiated by studies in actual service. This is shown graphically in Fig. 3, which follows the R B O T of an oil in steam turbine service, with addition of antioxidant (DBPC) at intervals. Turbine severity was rather high in this case since the oil life was only of the order of five years. Of course, it is possible that reinhibition may be more effective in less severe turbine service (for example, low severity gas turbines). Should reinhibition with antioxidant be considered, it should be reviewed with the oil supplier taking into full account the chemistry involved. Only if an initial check shows a good RBOT response and this is maintained in service is reinhibition likely to be beneficial.

Other Areas of Significance

Sampling Because of its importance, sampling is covered in some detail in the standard. Poor sampling techniques can invalidate test results and can often result in the wrong action steps being taken. In accordance with standard D 4047 on sampling, 7 guidelines are given for sampling from tanks, drums, lines, and reservoirs. 7 Standard Practice for Manual Sampling of Petroleum and Petroleum Products (D 4057).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

100

150

20O

250

2

I

6

I

I

I

I

I

I

I

I

\

I

I

I

I

I

I

I

I

I

I

I

AntJoxidant Added

I

I

I

I

I

I

I

OF

SERVICE

FIG. 3--Effect of reinhibition with antioxidant.

MONTHS

10 14 18 22 26 30 34 38 42 46 50 54 58

I

,L

Anlm~dam Added

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authorized.

R B 0 T

oI\

4O0

I

"4

,...t

O~

0 z

z

0

:73

m

0

18

TURBINEOIL MONITORING

New Oil Handling The proper handling of turbine oil during manufacture, transportation, and storage before it is utilized has been recognized as being of fundamental importance. Turbine oils are much more susceptible to changes in properties due to contamination than most other lubricants. What would be considered a minor contamination for other oils could cause loss of oxidation stability, demulsibility, and risk of deposit lay down in a turbine oil. As a result, reputable suppliers of turbine oils take special precautions within their own manufacturing facilities to segregate the product, usually by the use of dedicated tankage and transfer lines. Transportation to the user also poses a problem and the use of dedicated road tankers and rail cars is not always feasible, and improper cleaning procedures can lead to serious problems. The standard gives guidance to power plant operators on the handling of new oil receipts and for the verification of oil quality on delivery. Conclusion ASTM standard D 4378 is the result of detailed investigation, much discussion, and careful reviews by many experienced personnel in all phases of the industry. It should be looked upon as a guide and not as a strict set of test schedules. Information we have received indicates that a number of utilities throughout the country are using it and have found it useful to their operations. It is significant that the Joint I S O / I E C International Working Group on Turbine Oils gladly accepted the standard as a basis for an international maintenance guide for turbine oils. In the future, it is to be expected that modifications and additions will be made on the basis of experience and need. References [1] Denherder, M. J. and Vienna, P. C., "Control of Turbine Oil Degradation During Use," Lubrication Engineering, Vol. 37, No. 2, 1981, pp. 67-71. [2] Watanabe, H. and Kobayashi, C., "Degradation of Turbine Oils---Japanese Turbine Lubrication Practices and Problems," Lubrication Engineering, Vol. 34, No. 8, pp. 421-428. [3] Stringer, P. L., Burton, G. E, and Brown, K. J., "Some Properties of Canadian Steam Turbine Oils," Lubrication Engineering, Vol. 43, No. 4, 1987, pp. 283-289.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

Donald Bonett I

A Turbine Lube Oil System Monitoring Program REFERENCE: BoneR, D., "A Turbine Lube Oil System Monitoring Program," Turbine Oil Monitoring, ASTM STP 1021, W. C. Young and R. S. Roberton, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 19-24.

ABSTRACT: Con Edison has 32 steam turbines and 100 gas turbines in many locations in New York City and Westchester County. A monitoring program is administered from a central organization, but tested and tracked on site. The test results, when trended graphically, are used to predict system problems and to plan corrective action with the intent to eliminate catastrophic unit failures. The monitoring program as outlined can be used as a model to plan a specific program. KEY WORDS: turbines, monitoring programs, lube oil system Historically, utilities sought the guidance and support of oil suppliers and turbine manufacturers for the recommendation of test and monitoring requirements for use with the turbine lubrication system. These test requirements, methods, frequencies, and limits were based on the system design parameters. At Consolidated Edison we modified these requirements to develop and implement a rigorous monitoring program that minimizes turbine lubricant and system problems.

Basic Monitoring Program From the 1950s through the 1970s, the Con Edison turbine lubricating oil monitoring program consisted of weekly field tests for water content and monthly tests for neutralization numbers. Samples were submitted to the main laboratory every three months from all turbine reservoirs. Table 1 indicates the tests performed. In the late 1970s the Interfacial Tension Test was discontinued and field laboratories began to determine viscosity by using a "Visgage." The final changes were made in 1979 when we began to concentrate our efforts on monitoring turbine lubricating oil and system cleanliness. Samples from the main reservoir were submitted to the main laboratory at least twice per month for particle count. The current turbine oil monitoring program and the limits are listed in Table 2. Con Edison has traditionally tested all deliveries of turbine lubricating oil made to the generating stations. This is done at the main laboratory. The oil is tested and compared to the purchase specification for ISO VG 32 (ISO V G 68 at one location) turbine grade oil. The delivery truck is routed to our main laboratory, where the oil is sampled and tested per the parameters listed in Table 3. A further safeguard used to maintain turbine lubricating oil cleanliness at the generating

Senior engineer, Consolidated Edison Company of New York, New York, NY 10003.

19

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Copyright9 by ASTM International www.astm.org Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions aut

20

TURBINE OIL MONITORING

TABLE 1--Original test program. Viscosity Neutralization Number Rotating Bomb Oxidation Stability Rust Test Interfacial Tension station is through the use of 5-1xm filters (absolute). The delivery truck pumps the lube oil through the filter into the receiving tank. This maintains the integrity of the lubricating oil and the oil system with regard to cleanliness. The filtration will reduce the particles below Class 4 for new oil as indicated in Table 4. Turbine lubricating oil is purchased using these stringent specifications, safeguards are taken to assure that the specifications are met, and, finally, a monitoring program was established and implemented. What is done with all the data collected through the monitoring program? The first line of defense is field testing, which is done by the station chemist. He is responsible for all sampling and performance of the designated testing of the turbine oil. The chemist will inform the operating personnel of all test results within the work day. If the results indicate an out-of-limit condition, the chemist will immediately inform the operating supervisor. He will continue to inform various levels of management at the generating station and within the central organization. The course of corrective action to be taken is dependent on the parameter and the level. As an ongoing effort to predict when corrective action will be required, the chemist will graphically track the results obtained at regular intervals. This method can be used to predict water buildup, acid formation, viscosity changes, and any other parameter of concern. An example of trend charts is shown in Fig. 1 (viscosity) and Fig. 2 (neutralization number). Typical Conditions and Corrective Action High Water--Immediate action is taken if the water level is in excess of 0.1%; the recommended action is to put all available oil conditioners in service: centrifuge and/or Bowser 2.

TABLE 2--Current test program. Test

Method

Limit

Viscosity Water Neutralization number Rust test RBOT Particle count

ASTM D 445~ ASTM D 1796b ASTM D 974' ASTM D 665(A)d ASTM D 2272e Electronic

+ / - 10% 0.1% max 0.3 mg/g KOH max Pass 100 minutes min Class 6

" Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and the Calculation of Dynamic Viscosity) (D 445). h Test Method for Determination of Water and Sediment in Fuel Oils by the Centrifuge Method (Laboratory Procedure) (D 1796). ' Test Method for Neutralization Number by Color-Indicator Titration (D 974). d Test Method for Rust-Preventing Characteristics of Inhibited Mineral Oil in the Presence of Water (D 665). ' Test Method for Oxidation Stability of Steam Turbine Oils by Rotating Bomb (D 2272). -' Bowser is a manufacturer's name for an oil conditioner that has the capability to remove free water and particles in the 40 to 100-~m (nominal) range.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep

.'-7:' ~

,

1 i

i

31 ,...=-, --a

i

!

i

I

2

./

2

i .3

',

! ~.~

*'

~ 8

!

.~,

",,

,,,,"

i

!

!~,

1--Viscosity trend.

I':-,

FIG.

9

...,=

!

i~:i

7

!

~

i

!

1,7.

!i

i

J

-:",,..

~ P'.I ! i

;'

14-

,., ,"

:=,.._.,

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authorized.

+J

b,.

9-:~.---'

.$a,-5

,~.. , ~ ,~, ~

"..

i 18.

.2"

.,J

i ~,.,

..r"

! I ~._.

.,..

~9

~-...'~"'

I

!

i

......~i

~

w

E

O "n

"u "n

z

...-4 O

z

O

E

z

O

ill

z

O

!

i

t

I

',

':7,.02 ~.

'C:'04.

0/->6 !

0.1 l O.,:>.~,

.'12- I

':?'.~,4 J

r !g J

i

0',2 ~

0.24-

2

ZZ-.

,,,"

,~

..."

/

,I-

':

~

., /

,.~=

I

6

i

FIG.

I

1 ',

# ',

I"1

v=

I

/

!

~':~', I I

I

?

/ /

M ,7.ri'.~~

~.,

i

iI

~__.--+

I !'1

'i

i II

"

'i /,~

I

77

12

1

,~

2--Neutralization number trend.

--r..

!

~'

/ I

: 1

/I /

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authorized.

.2.

0",,i-"

..~

0.Z6 .~i E

,-.,

I

[]

TRE

~,3.. i 4

i

,.

N E L I T R : A L ! - 4.,~.., ' "- t,,..~;',.~ ""' ! ' , J~U M ~'-", L'-F R

1.5.

i

i

~,E

'~,

N D

,r

17

I ~.

t I "=.. Z':':'

,_..

Q

z

0 z -I 0

r

o

Z m

"-4 C

BONETT ON MONITORING PROGRAM

23

TABLE 3---Acceptance test criteria. Method

Limit

TESTS PRIOR TO OFF LOADING

28.8 to 35.2 cSt 0.2 (max) 200 (max) Pass Report

ASTM D 445 ASTM D 974 ASTM D 157-51T ASTM D 1401~ Electronic Clear and bright

Viscosity, centistokes at 100~ Neut. No., mg KOH/gm of oil Steam emulsion number, s Emulsion characteristics Particle count Visual appearance

TESTS DONE THE FOLLOWINGDAY

<200 minutes min Pass

ASTM D 2272 ASTM D 665(A)

RBOT Rust test

Test Method for Water Solubility of Petroleum Oils and Synthetic Fluids (D 1401).

Identify the source of the water contamination and plan for prompt repairs. If the water content is greater than a preestablished amount, which is determined on a unit-by-unit basis (free and/or emulsified), the unit should immediately be removed from service. Viscosity change > + / - 1 0 % - - I f the viscosity change is < - 10%, investigate the possibility of solvent contamination. If the viscosity change is > + 10%, this should be correlated to the increased acidity and low RBOT 3 results, which can be indicative of product oxidation, possibly caused by overheating. A n increase in water level with the conditioners in service signifies a continuous leak. Neutralization increase and viscosity increase may be due to oil degradation, oxidation taking place, and/or hot spots in the turbine oil system. All out-of-limit parameters should be investigated immediately. Trending the results will target potential problem areas that will be investigated immediately to prevent catastrophic failures. Turbine lubricating oil cleanliness--monitoring and trending--The turbine lubricating oil is sampled at least twice per month and sent to the laboratory for particle count. The oil is passed through the oil conditioners continuously. Samples are checked before and after the conditioners to ascertain the integrity of the conditioners. If the level of contamination continues to increase while the conditioners are operating satisfactorily, the operators should investigate the source of the contamination. If a large or concentrated sample of the contaminants can be obtained, laboratory analysis of the material will help in the investigation. The amount of makeup oil added to the system should be tracked. This will also help in analyzing the oil system. New oil added to the system changes the characteristics of the oil

TABLE 4---Particle count specification. Micron Size Range, Number of Particles per 100 mL

Class 4 (new oil) Class 6 (in service)

10-25

25-50

50-100

100-250

>250

10 700 42 000

1510 6500

225 1000

21 92

0 0

3 ASTM Test Method for Oxidation Stability of Steam Turbine Oils by Rotating Bomb (D 2272).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further repr

24

TURBINE OIL MONITORING

and should be noted on a trend chart. Use of the cleanliness data can determine the extent of the flush. In summary, testing, trending, followup, and knowledge of the turbine lubricating oil system will avoid catastrophic turbine lubricating oil and system problems. We should not be lulled into a false sense of security if the oil and system conditions are consistently within limits; we should continue the established program and improve the tests, instrumentation, and awareness.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. N

George R. Trabert I and Gerard J. Schreuders 2

A Supplier's Approach to Turbine Oil Monitoring REFERENCE: Trabert, G. R. and Schreuders, G. J., "A Supplier's Approach to Turbine Oil Monitoring," Turbine Oil Monitoring, ASTM STP 1021, W. C. Young and R. S. Roberton, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 25-31. ABSTRACT: To ensure the successfuloperation of a turbine circulating oil system, the turbine manufacturer, the turbine operator, and the lubricant supplier must cooperate by furnishing their technical expertise. Examples of turbine oil problems show that effective sampling procedures and monitoring of turbine oils is required to assure a long service life. ASTM Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines (D 4378) is an excellent guide to accomplish this. KEY WORDS: turbine circulating oil system, turbine oil monitoring, quality control, turbine lubricant, turbine oil analysis, warning limits, turbine oil additives

The Parties Involved Three parties--the turbine manufacturer, the operator, and the lubricant supplier--must work closely to ensure the successful operation of a turbine circulating oil system. Each brings his own expertise into the process so that all aspects and details are covered. The lubricant supplier brings expertise in petroleum handling, analytical techniques, and product knowledge to the attention of the operator and manufacturer. Together they work to ensure that appropriate action is taken. Figure 1 outlines the essential elements provided by each participant to ensure long turbine oil life.

The Importance of Good Sampling ASTM standard D 4378 [1] is a useful guide to assist those involved in the monitoring of mineral turbine oils, and we encourage all turbine operators to use it in reviewing their procedures. One of the key aspects of a successful monitoring program is sampling. We would like to add the following comments to those included in Section 6, "Sampling," of ASTM standard D 4378. In earlier times when economic conditions were different, oil company representatives were often present at the time of sampling and were able to ensure that samples were Product quality manager, Product Engineering, Chevron U.S.A. Inc., San Francisco, CA 941052856. 2 Senior research engineer, Industrial Oils, Lubricants Division, Products Research Department, Chevron Research Co., Richmond, CA 94802-0627.

25

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Copyright9 by ASTM International www.astm.org Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions au

26

TURBINE OIL MONITORING

I Tu, ,.e ..n.ta .,rar I Design Manu,actur:N~ QualityControl Installation TechnicalServic;

I o,, Su..iie. I

-\

I Tu..,n. Opera,*, I Purchasing OperatingDemand:~_ MonitoringD~k Operationand Maintenancet_ ~ LaboratoryFacility

-\

.ro.,,cfDe.,o..en,. / Blending

LongOil ]I ~,JTurbine Life J

.u.,,oo.o/,

Oil DeliverySystem~ / / TechnicalServiceD/ LaboratoryAnalysis /

FIG. 1--Essential elements for long turbine oil life.

representative and not contaminated in some unusual way. Now, we usually see a sample for the first time when it arrives at the laboratory. Therefore, it is more important than ever that the operator obtain samples from his turbine oil system using correct procedures. Another area of concern, particularly when systems are either filled for the first time or drained and refilled, is to get enough samples to ensure establishing a proper baseline on which to compare future samples. We recommend samples be taken from the following locations: 1. From the bulk truck or drums prior to being put into customers' storage tanks. 2. From intermediate storage tanks if used. 3. From the unit itself after filling and circulation for sufficient time to ensure homogeneity. It is not always necessary to analyze all the samples. If the system sample is satisfactory and appears normal, there is no need to look at other samples. However, if the system sample is suspect, the other samples can potentially provide a clue as to the cause.

Turbine Oil Testing: Who and What The role of the oil supplier in the monitoring process has changed due to changing economic conditions and product distribution processes in the oil industry. This role still varies widely depending upon the area of the country, historic customer relationships, and market conditions. The majority of the larger utilities are doing most of the testing themselves. Some utility companies use local independent laboratories; we recommend and support this use for timely assistance in routine analysis of system samples. The utility testing facilities or independent local laboratories are usually able to run several tests, such as appearance, color, total acid number, water content, viscosity, and cleanliness. To assess remaining service life, more

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreem

TRABERT AND SCHREUDERS ON SUPPLIER'S APPROACH

27

extensive testing is required. For example, the oxidation inhibitor content is often measured using an infrared technique that quantifies composition in combination with an oxidation performance test such as the Rotating Bomb Oxidation Test (RBOT), ASTM Test Method for Oxidation Stability of Steam Turbine Oils by Rotating Bomb (D 2272). Utility or independent laboratories usually do not have the equipment or expertise to run these and other tests of interest, such as rust prevention, wear, and metal contaminant levels. When routine analyses begin to approach warning limits, signalling the need for more intensive examination, or when some plant upset occurs and the oil becomes contaminated, the oil supplier should become more actively involved in the analysis process. Most major oil suppliers have all the laboratory equipment required to do an extensive laboratory investigation on any turbine oil sample. Often, this equipment is utilized to provide service to customers when an extensive analysis is required because of an actual or potential turbine oil problem. The amount of testing is dependent on the desire of the customer and the seriousness of the turbine oil problem. The first test results may indicate a possible problem and more sampling may be required to have that problem analyzed further.

Typical Turbine Circulating Oil Problems Here are three examples of the kind of turbine circulating system problems which customers requested the oil supplier to assist in resolving: 1. A 13 800-gal turbine oil system became contaminated with water. A centrifuge operation revealed sludge as well as water. Samples of the system oil and sludge were submitted for analysis. 2. A power plant operator complained that after a slug of water entered a turbine oil reservoir, the oil became hazy and formed stringy, insoluble material. Oil samples from the unit and a used-oil storage tank were received for inspection. 3. A company observed that the color of their turbine oil changed from yellow to dark green in three months of service. A laboratory inspection of a sample of the oil was performed.

TABLE 1--Analysis of system oil and centrifuge deposit from example No. 1. Sample Source Oil service, Year Viscosity at 40~ cSt" Viscosity at 100~ cSt" TAN, D 974, mg KOH/g Pentane insolubles, Mass % Toluene insolubles, Mass % Water, vol % Ash, mass % Spectroscopy analysis, metals, ppm" Appearance Oxidation inhibitor, % of new oil"

Turbine Systems 4 31126 5.31 0.09 . . . . . .

Typical New Oil

Centrifuge Deposit

Fresh 29.3 5.13 0.15

... ...

71()

<0~2

<1'0

iii

70~2 58.5 Wet 16.5 ...

Milky 36

100

Black powder ...

" After drying sample.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep

28

TURBINE OIL MONITORING

TABLE 2--Analysis of oil from the unit and from the storage tank from example No. 2. Sample Source

Unit, Used

Storage Tank, Used

Viscosity at 40~ cSt TAN, D 974, mg KOH/g Water, vol % Principal oxidation inhibitor remaining, %a Calcium, ppm Color, D 1500

33 0.11 < 0.1 54

32 0.14 0.4 37

5 4.0

3 4.0

~ Approximate value based on infrared scan.

The analyses results and recommendations developed were as follows: 1. Table 1 shows the detailed analyses of the system oil and centrifuge sludge deposit. The turbine oil analysis shows that 36% of the oxidation inhibitor remained in the oil. Also, 7.0 vol % water was found. Other characteristics, such as viscosity and total acid number (TAN), were normal. Wear metals were not found. Based upon these analyses, the oil was judged to be suitable for continued service after centrifuging to remove the water. However, close monitoring of the oil was recommended. The centrifuge deposits were lumpy black powder. Optical microscopy identified asbestos fibers. Spectrochemical analysis of the deposits indicated high concentrations of iron, copper, tin, and silicon. In addition, mass spectroscopy and gas chromatography identified coal. In summary, the centrifuge deposits consisted mainly of coal, dirt, wear metals, and asbestos fibers. Recommendation was made to take steps to prevent excessive ingress of dirt into the oil circulation system. 2. Inspection of the samples, the oil from the unit, and of the used oil from the oil storage tank are listed in Table 2. Inspection of the used oil from the unit itself shows that the oil is in good condition and suitable for further use. Analysis of the used oil from the storage tank shows the presence of 0.4 vol % water. Also, the oxidation inhibitor was depleted by 63%. It was recommended to heat, centrifuge, and filter the oil in the storage tank and then resample to ensure suitability before adding to a turbine oil system. 3. Inspection of the "green" used turbine oil sample is listed in Table 3.

TABLE 3--Analysis of used turbine oil from example No. 3. Sample Source

Used Turbine Oil

Viscosity at 40~ cSt Viscosity at 100~ cSt TAN, D 974, mg KOH/g Appearance Filter residue, mg/100 g Spectroscopy analysis, ppm Aluminum Iron Phosphorus Silicon

30.5 5.2 0.11 Dark green 12.6 48 36 26 5

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep

TRABERT AND SCHREUDERS ON SUPPLIER'S APPROACH

29

TABLE 4--Recommended warning limits for in-service steam and gas turbine oils. Oil Property Appearance and odor Viscosity change at 40~ % max Total acid number, mg KOH/g, max Water, vol %, max Calcium content, ppm, max Oxidation inhibitor level, % of new oil, min Rust inhibitor additive content RBOT

ASTM Test Method

Chevron Warning Limits

D 445 D 974

Rapid change 20 0.3

D 95 ...h ...'

0.2 20 30

0.2

D665, Procedure A

Fail test

Light fail Less than half value or less than 25% of original oil" Exceeds accepted limitsd Unusual and rapid darkening Drops 30~ ~ or more compared to new oil Tendency--450 max Stability--10 max

D2272

Cleanliness Color

D1500

Flash point

D 92

Foam test, Sequence I

D 892

ASTM D 4378 Warning Limits Hazy 20 0.1 to 0.4 increase"

" Depending on oil life (running hours) and type of turbine (steam or gas). b Atomic absorption procedure. c Infrared procedure. d Limits will depend on method used.

Laboratory inspection showed that the dark green oil contained a considerable a m o u n t of sediment consisting of aluminum and lesser amounts of iron, phosphorus, and silicon (quartz). A f t e r filtration, the oil was yellow ( A S T M D 15003 color of 2.0), typical for a used turbine oil. The color change might be explained by the combination of the primary colors, yellow from the turbine oil and the bluish tinge of contaminants, producing a green color. D u e to the massive concentration of metallic residue, it was r e c o m m e n d e d to change the oil and to d e t e r m i n e the cause of the rapid wear. The above three examples are typical of the kind of turbine oil problems we see. T h e y show that continuous effective monitoring of mineral turbine oils in service is r e q u i r e d to maintain effective turbine lubrication and to assure a long service life of the turbine oil.

Turbine Oil Warning Limits O v e r the years, our company has established warning limits for in-service steam and gas turbine oils. W h a t is a warning limit? It's a flag that shows an oil is getting " t i r e d " but has not reached the point where it should be " r e t i r e d . " Warning limits simply mean, " L e t ' s watch this condition a little m o r e c a r e f u l l y - - l e t ' s sample and test the oil a little m o r e frequently." Table 4 shows " R e c o m m e n d e d Warning Limits for In-Service Steam and Gas Turbine Oils." N o t e the similarities with the " W a r n i n g Limits" listed in Table 3 of A S T M D 4378 [1]. 3 Test Method for ASTM Color of Petroleum Products (ASTM Color Scale) (D 1500).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

30

TURBINE OIL MONITORING

Monitoring of Newly Developed Products in Turbine Oil Systems The oil supplier is certainly involved in oil system monitoring in cases where a newly developed turbine oil is to be field tested. A turbine oil consists of a base oil and several types of additives. The additives usually are antioxidants, antirust agents, metal deactivators, antifoam agents, and possibly others. If a product reformulation takes place, for instance to improve the quality of a turbine oil, the new product will be first extensively tested in the laboratory. If all results are satisfactory and the new product shows customer benefits and market potential, the oil company will proceed with a field test. Thee field test will be conducted by collecting performance data from the new product in actual service in a steam or gas turbine. Good communication with the customer is important on all details. He must be confident that the oil supplier will make every effort, including even abortion of the test, to assure that his operation will not suffer because of the field test. To assure the maximum value of the test, all details must be agreed to. During the field test, the oil supplier does most of the testing for in-service monitoring. An example of tests which are performed on a monthly basis and additional tests done on a yearly basis is given in Table 5.

TABLE 5--List o f inspections run by an oil supplier on monthly samples taken from a gas turbine oil development product during a field test.

9 Appearance 9 Viscosity, ASTM D 445, cSt --40~ --100~ 9 Flash point, COC, ASTM D 92, ~ 9 Neutralization no., ASTM D 974 --Total acid no., mg KOH/g 9 Oxidation stability by rotating bomb, ASTM D 2272, min 9 Trace metals by spectrochemical analysis, ppm --Silver --Aluminum --Boron --Barium --Calcium --Cobalt --Chromium --Copper --Iron --Magnesium --Manganese --Molybdenum --Sodium --Nickel --Phosphorus --Lead --Silicon --Tin --Titanium --Vanadium --Zinc Additional tests on a yearly basis 9 Color, ASTM D 1500 9 Trace sediment, ASTM D 2273, % by vol 9 Rust-preventive test, ASTM D 665, Procedure A, 24 h 9 Oxidation inhibitor content, infrared analysis

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep

TRABERT AND SCHREUDERS ON SUPPLIER'S APPROACH

31

The duration of a turbine oil field test can vary from less than a year to perhaps two years. Occasionally, the oil company may wish to continue routine monitoring of field units for extended periods of time. If during the field test the oil performs well in the turbine oil system and meets the performance targets indicated by the laboratory bench test program, the new turbine oil will probably be introduced into the marketplace. Conclusion Close cooperation between turbine manufacturers, turbine operators, and oil suppliers is a very important step for good functioning of the turbine and for a long service life of the oil. Proper sampling of the oil in the turbine is a key to a good evaluation of the turbine oil condition. Monitoring of the oil in service by the A S T M D 4378 schedule and applying the required action steps is necessary for a long trouble-free operation of turbines. Further cooperation of the three p a r t i e s - - t h e turbine manufacturer, the operator, and the lubricant supplier--will lead to improved ASTM procedures and A S T M test methods. ASTM standard D 4378 has proven to be a useful vehicle, and further cooperation will lead to refinement and improvement of this method. Reference [1] ASTM Practice for In-Service Monitoring of Mineral Turbine Oils for Steam and Gas Turbines (D 4378), 1987 Annual Book of ASTM Standards, ASTM Volume 05.03, Philadelphia, PA, 1987, pp. 505-518.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

Ryoji Ohgake, a Motoshi Sunami, t Toshio Yoshida, 1 and Harumichi Watanabe 1

The Reliable Control of Oil Quality in Japanese Turbine Units REFERENCE: Ohgake, R., Sunami, M., Yoshida, T., and Watanabe, H., "The Reliable Control of Oil Quality in Japanese Turbine Units," Turbine Oil Monitoring, ASTM STP 1021, W. C. Young and R. S. Roberton, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 32-43.

ABSTRACT: The reliable operation of Japanese steam, gas, and water power turbines using a newly developed trouble-resistant oil in association with a successful maintenance service system is discussed. A classic steam turbine oil formula containing a phenolic oxidation inhibitor and a succinic acid-type rust inhibitor made from low-aromatic base stock has been used for trouble-free maintenance. In Japan, the oil supplier and the user jointly administer quality control programs for turbine oils. The accumulated data of the oil supplier are utilized for estimating the optimum oil makeup for individual turbines. Since the installation of several computerized systems responsible for maintaining the quality of steam turbine oil in 165 power stations, the remaining life of the oil has been maintained over the warning limit by estimating the optimum oil's makeup. Remote control systems have been employed at many small hydroelectric power plants in Japan to save manpower. Such modernization has brought about an additional lubrication problem--severe valve sticking. More than 100 water turbine generators have suffered from this problem since around 1965. Maintenance-free type oil was required for these units. A new-type turbine oil for which a sludge dispersant is employed has solved this valve-sticking problem. The evaluation method and in-service performance are provided. The largest combined-cycle power station in the world (output 2000 MW) was recently installed at Futsu Power Station of the Tokyo Electric Co. (in 1985). The lubrication of 14 heavy duty gas turbines (GE9001E, 113 MW) is performed by. a new high-temperature gas turbine oil. This oil has done well, providing more than 8000 trouble-free hours of lubrication. The oil quality monitor and control systems are discussed. KEY WORDS: quality monitor, turbine oil, steam turbine, water turbine, gas turbine, oil makeup, valve sticking, combined-cycle power station, lubrication and severity constant

The monitoring of turbine oil is vital for maintaining and improving the reliability of turbine lubrication systems. Improving turbine oil formulations is important, of course, but it is also important that the monitoring be optimal for the operating conditions of the various types of equipment. Steam turbines exhibit relatively m o d e r a t e oil deterioration conditions. H o w e v e r , there is a demand for a high degree of reliability when turbines operate for long periods of time at high capacities, which necessitates the creation of a m e t h o d for evaluating not only normal deterioration but also unusual deterioration. Water power turbines operate at relatively low utilization and essentially have no oilNippon Oil Co., Ltd., Tokyo, Japan.

32 Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Copyright9 by ASTM International www.astm.org Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

OHGAKE ET AL. ON OIL QUALITY

33

cleaning equipment. Since trace amounts of sludge occurring in base stocks sometimes cause valve sticking, dispersant type turbine oils are necessary. Gas" turbines are exposed to rather severe conditions and sludge formation is a problem. These conditions require the use of heat-resistant antioxidants and the establishment of an appropriate oil-monitoring method to determine any required reconditioning and future oil replacement.

Steam Turbines with Proper Supply Monitoring Features of the Turbines During the 1950s, electric power generation in Japan was almost evenly divided between hydroelectric and thermoelectric generation. During the rapid economic growth of the 1960s, the use of thermoelectric power increased sharply, reaching 80% of the total. In the 1980s, nuclear power, which also uses steam turbines, came to account for 20% of the total electric power generation. These changes in the type of power generation were accompanied by an amazing technological revolution in the equipment used. Around 1955, the maximum capacity of one unit at a thermoelectric power station was 66 MW; that value has now reached 1000 MW. A t the same time, steam pressure has increased from 60 to 246 kg/cm 2, steam temperature has risen from 480 to 566~ and oil temperature in the tank has increased by 10 to 15~ In this situation, with turbine oil being used for ten years or more, it is vital that the oil quality be superior and that maintenance methods be established for the oil being used. Uniform standards have not been established in Japan for these maintenance methods. Instead, it is necessary to consider in detail the specific operating characteristics for each piece of equipment. Fortunately, the conditions of turbine oils used in power stations are analyzed, as a rule, every six months, and, based on these results, the remaining life of the oil is predicted and steps taken with regard to oil makeup, replacement, etc. [1-5].

Oil Design Philosophy Steam turbine oil is normally used for over ten years with repeated makeup. However, during this time problems caused by oxidative degradation, sludge formation, and foam may make it necessary to shut down the equipment. Oxidation stability is particularly important. The Japanese quality standards restrict the total acid number to below 1.0 mg K O H / g for the Turbine Oil Oxidation Stability (TOST) test [ASTM Test Method for Oxidation Characteristics of Inhibited Mineral Oils (D 943)] equipment after 1000 h. ASTM recommends that D 4304 (Specification for Mineral Lubricating Oil Used in Steam or Gas Turbines) the total acid number be below 2.0 mg K O H / g after 2000 h for TOST, and that the Rotary Bomb Oxidation Test (RBOT) life [ASTM Test Method for Oxidation Stability of Steam Turbine Oils by Rotating Bomb (D 2272)] be under 200 min for new oil. Table 1 shows the characteristics of typical turbine oils used in Japan. The TOST life and R B O T life are at the high levels of over 2000 h and 200 min, respectively. In addition, it should be noted that the amount of sludge formation after a TOST 1000 h (in accordance with A S T M D 4310 (Test Method for the Determination of the Sludging Tendencies of Inhibited Mineral Oils) and MIL-F-17331F) is kept very low, below 10 mg. Turbine oil is designed to meet the performance levels described above since it is used in equipment which plays a vital role both socially and economically in supplying electricity. For the base stocks used in turbine oils, efforts are usually made to reduce the amount of polar compounds containing sulfur, nitrogen, aromatic compounds, etc., which tend to have Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement.

34

TURBINE OIL MONITORING

TABLE 1--Typical properties of new turbine oils in Japan.

Specific gravity, 15/4~ Kinematic viscosity, mm2/s 40~ 100~ Viscosity index Total acid number, mg KOH/g Color, ASTM Pour point, ~ Aniline point, ~ Refractive index, 20~ Fraction Saturates, % Aromatics, % Resinous, % Copper corrosion, 3 h at 100~ Forming characteristics, mL (D 892) Sequence 1 Sequence 2 Sequence 3 Air release property, min, (D 3427) 25~ 50~ Rust preventing characteristics, D 665, Procedure B Oxidation stability TOST h RBOT min Sludging tendency, mg (sludge for TOST after IOO0h)

A

B

C

D

0.8587 32.43 5.607 111 0.10 L0.5 - 12.5 110.4 1.4717 94.1 5.7 0.2

0.8638 32.05 5.458 105 0.09 L0.5 -20 111.1 1.4728 99.5 0.4 0.1

0.8618 32.02 5.555 111 0.16 L0.5 -20 108.5 1.4734 86.1 13.8 0.1

0.8637 31.65 5.402 104 0.09 L0.5 -20 110.8 1.4726 98.8 1.1 0.1

la 20/0 20/0 10/0 3 1 pass

lb <10 20/0 <10 19.5 4 pass

2a 10/0 20/0 10/0 16.5 3.5 pass

lb 100/0 <10 10/0 3.5 1 pass

>2000 230 3

>2000 287 3

>2000 200 6

>2000 388 3

a negative effect on the oxidation stability of the product. This reduction is done by: (a) solvent refining and hydrofinishing or (b) hydrocracking. Turbine oils that can endure long use are produced from highly refined base stocks and well-balanced additive packages. The additives used for turbine oil in Japan are combinations of phenolic antioxidants (di-tertbutyl p-cresol, etc.) and succinic acid-type rust inhibitors. Although, to obtain longer life oil, amine antioxidants and metallic antioxidants are also used in some cases. Care must be taken with these additives, as incorrect use or concentration can cause sludge formation and early discoloring of the lubricant. The problems of foaming during actual use often appear as excessive amounts of surface bubbles. Thus, the air release performance of the oil is important. At present, the principal defoaming agents are nonsilicone types.

Oil Monitoring Measurement of the oxidation life and other characteristics of turbine oil is essential for understanding how the oil will actually perform. It is important to correctly learn the actual environment and operating conditions for each turbine in which the oil will be used. It could be said that by combining the characteristics of the turbines and the knowledge about the oil used, the actual performance of turbine oil can be learned in advance. The authors' company has more than ten years experience monitoring with computerized systems some 200 steam turbines operating in Japan. Specifically, we have introduced a turbine severity concept defined as "severity constant" for the characterization of these turbines in the field. Here, "severity constant" refers to the factors affecting the actual performance of the turbine Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Ag

OHGAKE ET AL. ON OIL QUALITY

35

oil, that is, the average value of decrease in the R B O T value per unit of time (4000 h) during normal operation. For statistical purposes, 4000 h is taken to be half of a year. The severity constant is different for each turbine; it varies widely between 5 and 30. The distribution of the severity constant for the commercial plants monitored by Nippon Oil Co. in Japan is shown in Fig. 1. The number of turbines has increased from 70 to 90 since 1980. The analysis for 1980 shows that the average severity constant for all the turbines was 13.7; the value for the current analysis, 13.4, is nearly the same. However, the distribution has widened as it used to be centered around 5 to 15, but now there are some moderate turbines below 5 and some very severe turbines over 20. In fact, the turbines exceeding 20 account for 14% of the total. This change in distribution is due to the improved precision of the severity constant resulting from an increase in the amount of monitoring data for the oil used. This range shows that each turbine has its own individual characteristics. The differences among equipment manufacturers and electric power companies have decreased sharply since the 1980 analysis because of improvements in operating conditions (the decrease of the residence time at high temperature spots, the increase of oil spray rate to the turning gear, etc.). The important point for optimal monitoring of each turbine (that is, for correct makeup) is to keep a proper balance between economy and reliability. In this case, economy is determined by deciding what level the target monitoring value should be. The margin of safety must be determined separately for each plant. In moderate turbines, this margin is small and favorable. However, in severe turbines, it is important that this margin be wide enough. As a result, an appropriate method for determining the margin of safety is to use the severity constant for each individual t u r b i n e . Figure 2 shows an example of the oil in Turbine 2, which was used for 150 000 h at the Y Power Station of the T Electric Power Co. The calculation for 1980 showed this turbine's severity constant to be 10.0, the average value. T Electric Power's target monitoring value was a RBOT value of 100 min. Adding to this value the safety margin of 10 rain obtained from the severity constant, the appropriate supply amount for a safe target monitoring value was obtained. As a result, the makeup required to maintain at least this value was calculated to be about 5% for each half year. The current calculation shows the severity constant to be 10.3, nearly the same as seven years earlier. The average supply rate per half year was 5.12%, and the turbine has been operating smoothly during this time without the targetmonitoring value ever dropping below 100 min. Makeup is performed once a year, so some

30 1980 1987 .~_ 20

10

!UUUUUUUUm-n O

2.5

5

7.5 1O 12.5 15 17.5 20 22.5 25 27.5 30 Severity Constant

FIG. 1--D~tribution of severity constants.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions authoriz

36

TURBINEOIL MONITORING

1980 estimate

300 250 65 200 o m Q~

15o IOO 50 ,

o

,

i

,

~

3

% i 1987 estimate

300 250 E 200 I-,-

_~

~

~

~

~

o

~

15o

lOO

50

~

o

0

.

.

,

.

2

4

6

8

.

.

10

.

.

12

Runninz Time,

.

-"3% \ o ~

14

16

16

.

20

22

24 x 104

h

FIG. 2--Actual and estimated R B O T value as a function of make-up rate. variation in the RBOT value is observed. However, during the entire seven-year period, the value agreed almost exactly with the calculations. As a result, we can see that it is sufficient to monitor turbines while keeping in mind a margin of safety adequate for the intervals between makeup. Recent P r o b l e m s with Steam Turbines

Steam turbines for commercial electric power generation have long operated at heavy base loads, with hydroelectric power being used for load adjustment. However, with the increase in nuclear power generation facilities, the two-shift method has been initiated for fossil fuel thermal electric stations. With this method, some steam turbines are used for load adjustment and operate only in the daytime when the demand for electric power is high. As the oil load on the turbine oil increases during startup, there is a strong possibility that the tendency toward degradation will increase. For these turbines, we analyze the data taking into account changes in the operating mode. With nuclear power generation, a phenomenon has been occurring that was not observed previously with steam turbines. It had been predicted that the maximum oil temperature would be 10 to 20~ less because of large steam turbines' supercritical pressures and that, as a result, the progress of oxidative degradation would be extremely slow. Contrary to this prediction, however, it was discovered that the antioxidants, which have a close connection to the life of turbine oil, were disappearing. Even when the phenol antioxidants and their degradation by-products in the oil were analyzed with gas chromatography and mass spectrometry, no difference was discovered Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No

OHGAKE

ET AL. O N OIL Q U A L I T Y

37

Number of Investigated Power Station (145 PS)

I

I

t

Trouble Free (50PS)

Trouble Experienced ( 95 PS)

I

J

1

Mechanical Trouble(35PS) Deposit TroublelOOPS) (37% of TroubleExperienced) (03~ of TroubleExperienced)

FIG. 3--Analysis of malfunction on spool valve at water power stations.

from the case of other steam turbines, nor were any hydroperoxides detected when there were equipment problems such as local heating. Finding solutions to this problem will be an important task, which we are pursuing and hope to report on in the future. Monitoring of Water Power Turbines Remote control systems have been employed at many small hydraulic power plants in Japan to save manpower. Deterioration uf Turbine Oil ( Oxidation of Hydrocarbons)

Deterioration of Suffer Compounds ( Sulfur Compounds in Base Oil, Additive, Coating Material )

RH ~ R . + - H O.-t- 02 ~ R 0 2 "

R-S-O . Oa

l 'RSO, oso

RO= + RH~ROOH4- O.

so,"

02

Contaminants Organic Acid Type Rust Inhibitor ( R'COOH)

/ oust,

Wear\

~Coatin| Material) x Oust /

RCOOH

Ester (R~OR')

Metallic Salt of Oroanic Acid

Sulfate Other Contaminants

(MeS04)

(RCOOMo)

Deposit R ; Hydrocarbon Me~Metal

FIG. 4--Valve deposit formation process.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

38

TURBINE OIL MONITORING

In spite of the use of inhibited oil similar to steam turbine oil, 95 water turbine generators have suffered from valve-sticking problems in Japan since 1965 (Fig. 3). About 37% of the valve-sticking problems were caused by mechanical factors, but the remaining 63% were caused by the adherence of sludge to the surface of the valves. Although the inherent origin of this problem is the lack of oil-cleaning systems, it is judged economically impossible to install a cleaning unit in many small hydraulic power plants. This malfunction of the hydraulic control valves in water turbines has been a serious problem for both the electric companies and the equipment manufacturers. In the past it has been difficult to reproduce this phenomenon in the laboratory, so a new test had to be devised. Valve sticking is caused by various contaminants such as rust, coating materials, wear particles, and deposits formed by deterioration of the oil. The mechanism for deposit formation is believed to follow two possible major courses (Fig. 4). Valve deposits are composed in most cases of carboxylic acid ester metallic salt. The sources of carboxylic acid are (a) oxidation products and deterioration of the base oil and (b) rust preventative additives. The next major group of the deposit is composed of metallic salt and sulfate. The sources of sulfur are the base oil and additives. This case is noticed if, after use of straight mineral oil for a long time, there is a change to an inhibited oil. A laboratory testing apparatus was developed by combining a solenoid valve and a vane pump. It has been used to effectively determine valve malfunction and valve sticking with satisfactory repeatability.

filter (~ ammeter

~_va~

(

pu y

flow meter

otor [

I

(

~

switch

J..

switch solenoidvalve FIG. 5--Laboratory testing apparatus for reproducing valve sticking.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions author

OHGAKE ET AL. ON OIL QUALITY

39

A testing apparatus is assembled with six solenoid valves and a vane pump, which is the same arrangement as a typical hydraulic system in a large power station (Fig. 5).

Test Method Place 10 L of the sample oil and an appropriate amount of neat reference sludge (20 to 100 mL) into the tank. Circulate the oil at a pressure of 20 kg/cm 2 and temperature of 60~ adjust the quantity of oil leakage from the solenoid valves to 6 to 8 mL/min/unit in order to simulate the quantity of leakage in an actual unit. At this time read the ammeter and record it as the initial value. To actuate the solenoid valve, the ammeter is increased gradually by operating the slidetype transformer. Thereafter, at the specified interval, actuate the solenoid valve and determine sticking and seizure of the solenoid valve in comparison with the initial value.

Prevention of Valve Sticking After having evaluated several types of sludge, one was selected as the reference, which was obtained by the modification of the TOST test (ASTM D 943) using a highly refined oil and without the presence of a copper catalyst for 500 h. With this reference sludge, a number of additive formulations were tested. As a result, a new "dispersant" type turbine oil has been formulated (Fig. 6) which prevents valve sticking and has superior oxidation stability and rust preventative and antiemulsifying characteristics (Table 2).

Monitoring of Gas Turbine Oil

The lubrication conditions of the General Electric large-capacity gas turbine are more severe than those of steam turbines [6, 7]. With the temperature surrounding the bearings over 260~ (MS7001 and 9001 model), G E has established the standard, GEK32568, requiring nonvolatile antioxidants. To meet these needs, a new oil formula has been developed specifically for gas turbines. It contains aromatic amine and a sulfur-containing hindered phenol. This oil offers long service life and has excellent sludge prevention.

~

~4 ~ 3

Sticking

turbine 0il

0

9=--

1

eE E

0

,,=

<~

--0.

J7

New dispersal type turbine oil

0

~

~

'

'

'

10

20

30

40

50

Time,

~J

'

I00

h

FIG. 6--Test results of valve-sticking preventative characteristics of turbine oils. Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions

40

TURBINE

OIL M O N I T O R I N G

TABLE 2--Typical properties of the water turbine oils. ASTM No.

Dispersant Type

Conventional

Specific gravity, 15/4~ Viscosity, mmZ/s 40~ 100~ Viscosity index Total acid number, mg KOH/g Color

D 1298 D 445

0.8617 44.79 6.880 109 0.01 L0.5

0.8620 44.81 6.897 110 0.11 L0.5

Copper corrosion, 3h at 100~ Rust preventing characteristics, Procedure A Emulsion characteristics, 54~ minutes to 3 mL emulsion Oxidation stability TOST, h RBOT, min Sludging tendency, mg Valve sticking resistance

D 130 D 665 D 1401

la pass

la pass

10

3

1600 220 3 pass

2500 240 3 fail

D 445 D 974 D 1500

D 943 D 2272 D 4310 ...

Performance Data for Gas Turbine Oil at Futsu Thermal Power Station o f the Tokyo Electric Power Co., Inc. In December 1985, the world's largest combined gas turbine cycle generator burning liquefied natural gas (LNG) was installed at the Futsu Thermal Power Station. A combined cycle is the use of heat recovered from exhaust of a gas turbine to help generate steam for steam turbines at the same facility. This station consists of two groups of seven General Electric MS9001Es for a total of 2 million KW. The station has operated satisfactorily to date with the one group which has been in commercial operation. All units were expected to be in full operation by November 1988.

54~

P

~

120

42~

3

4

100

80 60 40

nil i l,

1, n, I i ~m

Bearing Turbine 0il

FIG. 7--Outlet temperatures of turbine oil at seven bearings.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

OHGAKE ET A L ON OIL QUALITY

~ L4 L3

_.-o----o- . . . . . o . . . . . . .

"~

L2

,.o- - - - ~

o'-

o..

Make-up(5.8%)

2000 o

'i,~

o-. . . . . . . o

/

L1

I00

41

I

I0-- . . . . L

(2~''''"0"

....

0

0

~-0~ -0 . . . . . .

0 ...~

.........

Inhibitor

"" . . . . . "(3

8O

1

5

~

-- 70 bo

/

0o

g. 5O

1000

o ~ 9 9

Stage Stage Stage Stage

No.1 No,2) No.3 No.4

= =

Warning

5OO

Use Limit

0

0

I

I

I

3

6

9 xlO3

Running Time, h FIG. 8--Changes in operating conditions of gas turbine.

In the single-SHAFT MS9001E, the gas turbine, generator, and steam turbine are all along the same axis and are supported by seven bearings (one on each side and one in the center of the gas turbine; one on each side of the generator; and one on each side of the steam turbine). The temperature of the supply oil for the gas turbine is about 55~ the temperature of the bearing drain oil is under 70~ on each side and over 80~ in the center (See Fig. 7). At another temperature sensor location, the supply oil temperature of the bearing is under 40~ and the bearing drain oil temperature is only about 60~ In this case, the severity of operation condition on the turbine oil is low. All 14 turbines are used for the two-shift load operation. After the turbines began operation, the oil used was sampled every 1000 to 2000 h and its degradation checked. The following results appeared (Fig. 8): 1. The drop in the RBOT value was low: In one year (7500 h) of operation, it decreased 10% (200 min) from the initial value.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No

42

TURBINE OIL MONITORING

FIG. 9--1nside the bearings and the oil tank (after one year operation). Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further r

OHGAKE ET AL. ON OIL QUALITY

43

2. The residual rate of the additives was high: After one year the residual concentration relative to new oil was 85 to 90%; this corresponded well to the rate of decrease in the RBOT value. 3. All the parts were free of sludge: When the turbines were opened after one year, almost no sludge was found in the oil tank or inside the bearings (Fig. 9). There was also very little dirt in the oil filters. 4. The increase in viscosity and acid number was low, and the color increased only slightly. Future Trends in Monitoring At present, a regular inspection of each stage is performed once a year. (In Japan, regular repairs, etc., at commercial power stations are performed at times determined by the Ministry of International Trade and Industry.) However, an extension of the period of these government inspections closer to the two years used for steam turbines is being studied. Since these turbines have no reserve tanks, two methods are being studied: (1) replacing all of the turbine oil after a fixed period of continuous use (for example, once every four years); or (2) supplying the appropriate amount of oil during the regular repairs described above. A t present, the compensation for each year is estimated for the supply at the scheduled maintenance time each year, but in the future it will be necessary to select the best of the above two methods while taking into account the degradation conditions of the oil. Conclusion

In Japan, a classic steam turbine oil formula containing a phenolic oxidation inhibitor and a succinic acid-type rust inhibitor made from low aromatic base stock has been used for trouble-free maintenance by the installation of the computerized systems responsible for maintaining the remaining life of the oil over the warning limit by estimating the optimum oil's makeup. A new-type turbine oil for which a sludge dispersant is employed has solved the valvesticking problem caused by the lack of oil cleaning systems at many small hydroelectric power plants in Japan. Evaluation method and in-service performance are provided. The lubrication of 14 heavy-duty gas turbines (GE9001E, 113 MW) is performed by a new high-temperature gas turbine oil which has done well, providing more than 8000 troublefree hours of lubrication with a monitor system for oil quality control. References

[1] Watanabe, H. and Kobayashi, C., "Degradation of Turbine Oils; Japanese Turbine Lubrication Practices and Problems," Lubrication Engineering, Vol. 34, No. 8, 1978, pp. 421-428. [2] Itabashi, S., Oba, T. and Watanabe, H., "Computerized System for Analyzing Steam Turbine Oil in Power Stations," Lubrication Engineering, Vol. 37, No. 5, 1981, pp. 279-285. [3] Den Herder, M. J. and Vienna, P. C., "Control of Turbine Oil Degradation During Use," Lubrication Engineering, Vol. 37, No. 2, 1981, pp. 67-71. [4] Wilson, A. C. M., "Problems Encountered with Turbine Lubricants and Associated System," Lubrication Engineering, Vol. 32, No. 2, 1976, pp. 60-66. [5] Den Herder, M. J. and Bennett, H. O., "Long-life Turbine Oils for Power Generating Equipment, ASLE preprint No. 69AM-6B-1, 1969. [6] Den Herder, M. J., "Mineral Oil Lubrication of Large Gas Turbines," Lubrication Engineering, Vol. 33, No. 6, 1977, pp. 303-307. [7] Smith, A. N., "High-Temperature Lubricants for Gas Turbine Service," Lubrication Engineering, Vol. 32, No. 2, 1976, pp. 66-69. Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further

Joseph F. Anzenberger, Sr.1

Monitoring Practices for Phosphate Ester Fluids for Turbine Lubrication and Control REFERENCE: Anzenberger, J. F., Sr., "Monitoring Practices for Phosphate Ester Fluids for Turbine Lubrication and Control," Turbine Oil Monitoring, ASTM STP 1021, W. C. Young and R. S. Roberton, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 44-53. ABSTRACT: In-service monitoring of the customer's phosphate ester turbine oils has been practiced for over 15 years. This service originated with users of electrohydraulic control (EHC) fluids and gradually expanded to include specialty-type synthetic triaryl phosphate ester lubricants for large and medium-sized gas turbines and turbine-compressor operations. In-service monitoring is intended to assist the user in caring for the fluid to obtain maximum service life. Users who avail themselves of this service have their used fluids examined for various properties, which could include as many as seven measurements: viscosity, acid number, specific gravity, water content, mineral oil content, fire properties, and stability. Corrective actions are suggested if a fluid exceeds certain property measurements.

KEY WORDS: turbine lubrication, electrohydraulic control (EHC) system, EHC fluid, Alcor Deposition Test, fluid resistivity, phosphate ester

Fluid-related hydraulic system problems represent a potentially critical situation in the proper control of turbine operations. The high-pressure hydraulic operations of an electrohydraulic control (EHC) system consist of the hydraulic power unit, the emergency trip system, and the steam valve control system. The nature of the system components dictates the necessity of a high degree of cleanliness and maintenance of certain physical and chemical properties of the EHC fluid. Because of the high pressure, the use of synthetic phosphate ester fluid as a working medium is recommended by certain steam turbine manufacturers to enhance the safety of operations. As a producer of approved phosphate ester fluids for EHC systems, we have a long-standing reputation for assisting EHC fluid customers in assessing the overall condition of their in-service fluid. This paper will review the monitoring practices for phosphate ester fluids for turbine lubrication and control. Discussion

Most users of phosphate ester hydraulic control fluids rely on fluid analyses to monitor hydraulic system and component conditions. As a producer of these types of fluids, we offer to analyze the customer's phosphate ester fluids on a periodic basis. The turbine engine manufacturers have always recommended that regularly scheduled fluid care is needed to ensure long, reliable operations of the hydraulic system. Hydraulic fluid analyses are intended to supply this information, Senior research chemist, Stauffer Chemical Co., Gallipolis Ferry, WV 25515, now retired. He lives at 18 Mark Drive, Detmont, PA 15626.

44

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Copyright9 by ASTM International www.astm.org Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions au

ANZENBERGER ON PHOSPHATE ESTER FLUIDS

45

TABLE 1--VACWAFS analysis coverage. V A C W A F S

= = = = = = =

Viscosity, m2/s at 37.8~ Acid number, mg KOH/g Contaminants, mineral oil, % Water content, % Autoignition temperature, ~ Foaming tendency, mL Specific gravity, 60/60~

The frequency of fluid testing usually follows the original e q u i p m e n t m a n u f a c t u r e r ' s ( O E M ' s ) recommendations. W h e n we first began to offer technical service to customers, a V A C W A F S analysis was run (Table 1) to judge the fluid's condition. Triaryl phosphate esters are excellent solvents. It didn't take long to realize that the customers had to be supplied with clean sampling containers that were inert to the solvent action of the fluid. Also, o t h e r tests were soon introduced and became standards for establishing the condition of the hydraulic fluid. Thus, the measurements in Table 2 are now m a d e with all incoming samples of E H C fluid.

Results Viscosity Viscosity affects fluid flow in pipes and is important to p u m p life and system response time. O u r experience has shown that when the fluid is out of specification it has probably been contaminated with a mineral oil, fuel, or water. The source of the contamination should be located and corrected. Fluid which is out of specification should be brought into specification by fluid m a k e u p as soon as possible. A scheduled maintenance down-time for changeout in the near future should be considered. Acid Number As an E H C fluid is used in service, an increase in the acid n u m b e r indicates that the fluid may be starting to deteriorate. Maintaining a low acid n u m b e r greatly extends the life TABLE 2--Standard tests run with all EHC fluids. Property

Test Method

Viscosity, mVs, at 37.8~ Acid number, mg KOH/g Mineral oil content, % Water content, % Resistivity, G ohm-cm Chlorine content, ppm Particle count per 100 mL

ASTM D 4451 ASTM D 9742 ASTM D 13993 ASTM D 17444 ASTM D 1169, 5 D 2576 ASTM D 808, 7 D 13178 SAE-A-6D

Test Method for Kinematic Viscosity of Transparent and Opaque Liquid (and the Calculation of Dynamic Viscosity). 2 Test Method for Neutralization Number by Color-Indicator Titration. 3 Test Method for Unsaponifiable Matter in Alkyd Resins and Resin Solutions. 4 Test Method for Oil Separation from Lubricating Grease During Storage. 5 Test Method for Specific Resistance (Resistivity) of Electrical Insulating Liquids. 6 Test Method for D-C Resistance or Conductance of Insulating Materials. 7 Test Method for Chlorine in INew and Used Petroleum Products (Bomb Method). 8 Test Method for Chlorine in New and Used Lubricants (Sodium Alcoholate Method).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No fur

46

TURBINE OIL MONITORING STANDARD

[--0, -STANDARD 0 . 5 ~ FILTER FLUID RESERVOIR

FULLERS' EARTH

n

,r + FIG. 1--Use o f fihers in an EHC system.

expectancy of the E H C fluid. Fuller's earth has been used for many years to neutralize acidic components in E H C fluids. There are two schools of thought by utility station operators concerning the use of fuller's earth. There are those who use the earth filters only when the acid number increases. They are in the majority. The others keep the fuller's earth filters in place all the time. My own preference is for the use of a fuller's earth filter system (Fig. 1) that is in constant use. Thus, when the acid number begins to increase you know that you have exhausted the acid-neutralizing ability of the adsorbent so you remove the spent cartridge and replace it with a new one and resume the use of the fuller's earth on a fulltime basis. This practice should result in keeping the E H C fluid at a low acid number and thus prolong the life expectancy of the fluid. Mineral Oil

Concentrations of mineral oil greater than 4% may attack the accumulator bladders and certain other seal materials in the system and will reduce the fire resistance of the hydraulic fluid. Concentrations below 1% are generally the result of hydrocarbon extraction from pipe dopes, seals, and packings. When the mineral oil concentration exceeds 1%, it may indicate contamination from an external source. Mineral oil concentrations above 2% may be in-

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No furt

ANZENBERGER ON PHOSPHATE ESTER FLUIDS

47

soluble in the E H C fluid, resulting in a hazy mixture. If the mineral oil concentration exceeds 3%, it may jeopardize its qualification for a fluid repurchase program.

Water Content New phosphate ester hydraulic fluid typically contains about 0.01 to 0.03 weight % water. A water content in used fluids of 0.15% means water is entering the system from outside, such as from a heat exchanger or an ineffective air dryer. Corrective measures should be taken to remove the excess water and to eliminate the source of the leak. The presence of water will result in hydrolysis, and the concentration can affect the rate of hydrolysis (Fig. 2). Fuller's earth filtration should effectively reduce the water contamination if it is in the range of 0.1 to 0.3% water. A t higher water levels, fuller's earth will still help to reduce the water content, but the acid number of the fluid may increase during an extended period of filtration. Vacuum dehydration via a'portable filtration unit or complete fluid changeout are alternatives that can be taken.

Resistivity Fluid resistivity is measured to determine its resistance to conduct an electric current and thus protect the valves from electrical erosion. Fuller's earth filtration will maintain electrical resistivity at acceptable levels. If resistivity decreases, the fuller's earth should be changed and resistivity again measured after several days exposure to the new filter's earth. If the E H C fluid is not responsive to new fuller's earth filtration, new fluid should be added to the system as soon as possible.

A 16

12

50

100

150

C

B

200

250

HOURS AT 200~ F I G . 2--Stability o f phosphate esters as a function of water concentration [ASTM Test Method of Hydrolytic Stability of Hydraulic Fluids (Beverage Bottle Method) ( D 2619)]. A = fluid containing 1.0 wt% water; B = fluid containing 0.50% water; C = fluid containing 0.03% water.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

48

TURBINE OIL M O N I T O R I N G

TABLE 3--Society of Automotive Engineers, Inc. hydraulic fluid contamination study of contamination

levels, SAE, ASTM, and AIA tentative standard for hydraulic fluids (per 100 mL). Contamination Class Size Range, Ixm 5 to 10 10 to 25 25 to 50 50 to 100 >100

0

1

2

3

4

27 000 670 93 16

4600 1340 210 28

9700 2680 380 56

1

3

5

24 000 5 360 780 110 11

32 000 10 700 1 510 225 21

Chlorine T h e c h l o r i n e c o n t e n t of E H C fluids is m a i n t a i n e d at u n d e r 100 p p m to g u a r d against serious valve erosion. N e w E H C fluid usually c o n t a i n s less t h a n 35 p p m chlorine. T h e c h l o r i n e c o n t e n t is m o n i t o r e d to e n s u r e t h a t c h l o r i n e - c o n t a i n i n g c o n t a m i n a n t s do n o t get into the system.

A 4" x 6" card is included sample bottles to the fluid

plus bag,

supplier.

with two one-pint

in a m a i l i n g The customer

polyethylene

kit which supplies

is returned

the f o l l o w i n g

information.

Technical Service Laboratory Program Information S h e e t Company:

I IZi0C~

Station: Address: City, State: Name or title: Telephone No. 5.

Unit Number

6. Unit MImufacturer 7. Sample DIte 8. Name of fluid usecl g. Fluid makeup Date/Quantity 10.

FIG. 3--Customer information card.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

ANZs163

ON PHOSPHATE ESTER I=I_UIDS

49

TABLE 4--Typical analytical technical service report sent to the customer.

TECHNICAL SERVICE REPORT -COMPANY.......

SAMPLE PRODUCT C U S T NO. U N I T NO. TYPE RECEIVED

: : : : : :

E2631 EHC 1"295 2 GE 8-14-87

ANALYTICAL TEST RESULTS TEST VISCOSITY, SUS, I00 M I N E R A L OIL, W T % W A T E R CONTENT, W T % ACIDITY, m g K O H / s C H L O R I N E CONTENT, P P M RESISTIVITY, G-Ohm-cm PARTICULATE MATTER/100 ML S I Z E (Micrometers) 5-10 10-25 25-50 50-100 >100

210.0 0.30 0.11 0.03 11. 94.0

200-230 4.0 0.2 0.2 100. 5.0

MAX MAX MAX MAX MIN

9194 296 34 7 S

24000 5360 780 110 11

MAX MAX MAX MAX MAX

TECHNICAL SERVICE SUPV. All a n a l y s e s are b a s e d u p o n s a m p l e m a t e r i a l f u r n i s h e d and t e s t e d a n d u p o n use o r a p p l i c a t i o n of s u c h m a t e r i a l as k n o w n

Particle Count Fluid cleanliness is essential for smooth, trouble-free operation of the E H C servovalves. All new fluid must meet the SAE-A-6D Class 3 particle count specification (Table 3). This measurement can be made with a H I A C particle counter and uses 100 mL of fluid per test. If the particle counts exceed these recommended limits, the filters must be checked for correct micrometre sizes and alignment. Attention must also be given to the possibility of dynamic seal failure, water in the system, a rupture of a fuller's earth cartridge, or contamination with silica gel from the reservoir air breather.

Customer Reports The customer is provided with two pint-size polyethylene bottles plus cap, data sheet (Fig. 3) to indicate the company, location, unit, unit manufacturer, fluid type, date sample drawn, and running hours with fluid. The customer also receives a mailing kit with a preaddressed mailing label attached and a polyethylene bag into which the two pint sample bottles are sealed. The samples are usually analyzed within four to five days after their receipt. Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement.

50

TURBINE OIL MONITORING

The analytical results are then mailed to the customer in as short a time as possible. All analyses are recorded in a computer system, and individual sample results are printed as shown in Table 4 and sent to the customer. The frequency of sampling is dependent upon several conditions. A fluid that has been in service and is performing well may be sampled only once a month or once every three months depending upon the station manager and the condition of the fluid. A system that starts out with a new fluid will submit samples of (1) new fluid put into sump, (2) fluid after 24, 48, and 168 h, (3) fluid monthly thereafter if all fluid properties are well within fluid specification. The Technical Service Laboratory Report (Table 4) shows the seven physical and chemical properties discussed previously, the results of the properties measurements, and the recommended limits for each property. In the event one or more of these properties exceeds the limits shown, the result is starred and a footnote added suggesting an appropriate action to be taken. For example, should the mineral oil or water content values begin to increase, the customer is requested to check for the potential source of leakage or contamination that caused the increased amount of water or mineral oil. If the acid number exceeds 0.20 mg K O H / g , the customer is advised to inspect and/or change the fuller's earth cartridges. Likewise, should the resistivity value fall below 5 G ohm-cm, the customer is asked to change the fuller's earth filters in the hope of increasing the resistivity value and maintaining it above the 5 G ohm-cm minimum value. If the fluid does not respond to one or two changes of fresh fuller's earth treatment within 24 h after a filter change, the customer can either replace some of the in-use fluid with new E H C fluid or plan to change-out the whole system during turbine shut down. An increase in chlorine content usually is the result of a contaminant getting into the system. Chlorinated solvents are usually responsible. Although a direct connection between chlorine content and servovalve erosion has never been fully demonstrated (General Electric), it is advisable to limit chlorine to 100 ppm in new and operating fluid. New E H C fluid must meet the SAE-A-6D particle count specification for a Class III fluid (Table 3). When a clean E H C system is filled with new E H C fluid which is circulated for 24 to 48 h, the refiltered fluid will frequently show a 5 to 10 ixm particle count of less than 1000. Because of the ease of contamination of the fluid, special care is exercised in handling TABLE 5--Computer printout of the history of fluid analyses from a given customer. CII'tP4~NYNO. REGION,OH~ UNIT

TYPE

SI~ 2~, 1987 DATE

VIS SUS

6E

HINERO/.k~T~

ACIDCHL~ I~'SIST.

OIL CI~TrEHT +,'Q. Pin

1- 7-G4 219.0 4-1(rBq 218,0

0.20 0,20

8-15'84 220,0

0.20 0".09 0.0i

1- 7 - ~ 3-2%95 8-19-~ 12-12-4~ 4-11-86 8-13-85 1- 7-97

21%0 218,0 215.0 209,0 207,0 219,0 ~8,0

3'-1B-97 207.0

0,10 0,40 0.,,.0 0,20 0,I0 0,10 0,20

0,06 0,05 0,04 0,09 0.08 0,06 0,06 0,09 0,05

O.Oi 0,03

O-ohs-c

II. )!00.0 34, )100,0

31. )lO0.O

0,(~4 19, )tO0,O 0,03 11, )100,0 0.03 26. )100.0 0.02 15, 8q,O 0,01 14, )100,0 0,03 21, )!09,0 0,01 29, )100,0

O.lO 0.06 0.01 26. )lO0.O

---P~TIIIA.~TE~TT~,/lOOml

-

5-LO 10-~ 25-50 50-100 )tO0 1034 679 421 1175 204 1185 2625 2~.Ob 635 1418

lOB 120 161 146 122 70" 22~ 211 160 10~

1483

69

12 47

3 19

2 4

,'~

2

l

17 1'2 13 49 63 37 34

7 9 3 II 17 8 13

6 3 ! 3 1 1 2

16

3

1

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No furt

51

ANZENBERGER ON PHOSPHATE ESTER FLUIDS VENT ~

TEST SECTION SCHEMATIC

C. MAX.tEMP.

r TRANSFORMER SPENT

|

I

T.C.TUBE TEMP.CONTROL

COOLANT

COOLER AND SUMP

AIR

---H

PUMP

9

_1

PRE$SURi-. . . . GAGE

T.C.

| FILTER

AIR COOLANT

COKING HEAD WITH%" DIA. RESISTANCE HEATEDTUBE

-I

I

I

. . . . . . . . . . . . . . . . .

j c

T.ET

-

FIG. 4--Schematic of Alcor deposition tester.

and drumming the E H C fluid by the manufacturer. This is also another reason why clean sample bottles are supplied to the customer for returning used fluid samples for analysis. Storing the test results in a computer and furnishing the customer with a computer printout report also allows for a rapid retrieval of the history of analyses of a given fluid with each customer. Thus, if a potential fluid problem is occurring, a print out summarizing the fluid property measurements can be obtained with the samples from a given turbine operation as shown in Table 5. Having the information summarized in this manner often allows for a more rapid understanding of why the problem may occur. Other Tests In addition to the standard seven analyses run with each customer's fluid (Fig. 2), there may be an occasional need to run additional tests to help a customer understand why a problem occurs such as filter plugging, rapid rise in acid number, gel formation, a reduction in resistivity, change in specific gravity, and an increase in chlorine, water, or mineral oil TABLE 6--Alcor Deposition Test conditions. Fluid volume, sump, mL Fluid flow rate, mL/min Fluid temperature: in Out

Lower tube temperature, ~ (~ Upper tube temperature, ~ (~ Volume of air, CC/min Test time, h

250 300 325 450 288 (550) 315 to 427 (600 to 800) 1000 24

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions au

52

TURBINE OIL MONITORING

TABLE 7--Effect of starting acid number on stability of fluid in Alcor Deposition Test. Deposits, mg Acid Number of Test Fluid

Test Time, h

A Viscosity, %

A Tan mg KOH/g

Filter

Rod

Overall Rating

0.23 0.02

8.5 24

2.7 10.3

1.23 0.92

10.0 0.6

298 68

219.3 69.3

content. These tests may include the use of separation technology to isolate different components, followed by the use of: 1. 2. 3. 4. 5. 6. 7. 8.

Metals analyses. Infrared spectra. Liquid chromatography. Mass spectra. SEM-EDAX. Gas chromatography. NMR spectra. Other tests to understand and resolve the customers' problems.

Occasionally a customer may submit a sample of a solid or semisolid residue that was found in a sump or in holding tanks. He wants to know what it is. This is when it becomes necessary to revert to some of the above-mentioned tests to determine the composition of the unknown material.

Gas Turbine Lubrication

Most of the foregoing discussion has dealt with EHC-type phosphate esters used as hydraulic fluids in steam turbines. Phosphate esters are also used in some large, industrial gas turbines and in medium-sized gas turbines and turbine-compressor operations. These customers are also invited to submit samples of their fluids for periodic inspection at certain time intervals. Usually these fluids are examined for acid number, viscosity, specific gravity, water content, mineral oil content, fire properties (flash point, fire point, and autoignition temperature), and foaming tendency. Occasionally, a customer may ask for an evaluation of the remaining life expectancy of a phosphate ester in service. Depending on fluid type, kind of service, and history, the fluid may be examined in an oxidation-corrosion test, a hydrolysis rate study, or in an Alcor Deposition Test. The latter test is usually used only with the so-called synthetic triaryl phosphate esters and not with trixylyl phosphates, which are sometimes referred to as natural phosphate esters. The fluids are also neutralized, usually with activated alumina treatment, then filtered before being examined in any tests. In the Alcor Deposition Test (ADT), 250 mL of fluid at 163~ (325~ is circulated at 300 mL/ rain by a high-temperature pump. Air is injected at 1000 cc/min into the fluid just prior to entering the heater section. After leaving the heater section, the air-fluid mixture discharges into the sump, which is cooled by air passing through a jacket surrounding the oil chamber. The air that was injected into the test fluid leaves the reactor through a discharge tube. Any volatile by-products formed from oxidized test fluid can be recovered in a receiver at the end of the discharge tube. The test is run for 24 h, and the test rod is maintained at 288~ (550~ The top part of the rod can vary from 316 to 427~ (600 to 800~

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No furth

ANZENBERGER ON PHOSPHATE ESTER FLUIDS

53

When the test is completed, the fluid is examined for changes in viscosity and acid number. The tube deposits are weighed and classified into types such as flaky, crinkled, smooth, sludge, and varnish. Filter deposits are also weighed. An overall rating is determined based on the amount and type of deposits formed. The lower the rating value, the more stable the fluid. Figure 4 shows a schematic of the Alcor Deposition Tester and Table 6 lists the test conditions used. Table 7 shows the effect of the starting acid number of the stability of a used fluid when examined in the A D T test. The used fluid as received had an acid number of 0.23 mg K O H / g and failed to survive the 24-h test. By neutralizing the fluid with activated alumina and then examining it in the A D T test, it gave a relatively low overall rating which demonstrated excellent thermal-oxidative stability. In this case the customer was advised to reduce the acid number of his fluid and to maintain the acid number at below 0.10 mg KOH/g. Conclusion

Most customers of phosphate ester fluids used in turbine lubrication and control are dependent on the fluid supplier for guidance in determining how long the fluid should remain in service. By monitoring critical physical and chemical properties of the fluids, the supplier is able to offer guidance for maintaining the fluid in good condition. If the user heeds the various warnings and suggestions offered by the supplier, he should realize a longer service life for the fluid.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep

Donald G. Pachuta, 1 John A. Thich, 2 Ray W. Knipple, 3 Deborah A. Stephanic 4

Experiences with an Analytical Monitoring Program Designed for Commercial Flight Tests of Advanced MiI-L-23699 Turbine Fluids REFERENCE: Pachuta, D. G., Thich, J. A., Knipple, R. W., and Stephanic, D. A., "Experiences with an Analytical Monitoring Program Designed for Commercial Flight Tests of Advanced Mil-L-23699 Turbine Fluids," Turbine Oil Monitoring, A S T M STP 1021, W. C. Young and R. S. Roberton, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 54-76. ABSTRACT: Aviation turbine engine manufacturers possess different philosophies concerning the qualification of new turbine fluids for commercial aircraft. Most require a flight test employing a monitoring program that was designed to measure only basic fluid parameters such as viscosity, total acid number, etc. Warning limits for changes in these parameters were generally established from the performance characteristics of earlier generations of turbine fluids and older engine designs which imposed less operational stress upon the fluid charge. Recent flight test experience with a third generation Mil-L-23699 fluid suggests that revisions to currently recognized programs and parameter limits should be made in order to properly reflect modern engine hardware features, airline operating practices, and advanced additive behavior. It is recommended that analytical techniques such as plasma-based atomic emission spectrometry, size exclusion chromatography, gas chromatography, and thermal analysis be integrated into existing flight test programs. A "two-stage" monitoring scheme consisting of standard physical-chemical tests and these analytical techniques provided essential data about the fluid's response to thermooxidative and operational stress. Compositional information needed to detect and identify contaminants such as hydraulic and other foreign fluids was also obtained. Specific examples are cited about the diagnostic abilities of this analytically oriented monitoring program. KEY WORDS: commercial aircraft, aviation turbine engines, oil monitoring program, MilL-23699, flight test, analytical methods, thermal analysis, gas chromatography, liquid chromatography, spectrometry, metals, hydraulic fluid, water, contamination, deposits, thermal oxidative degradation, additives, antioxidants

Obtaining engine manufacturer approval for a new aircraft turbine fluid is an extensive, multistage process involving an array of physical and performance " b e n c h " test requirements, followed by a series of mechanical rig and engine test-stand studies. U p o n successful completion of these stages, several engine manufacturers require and grant approval for a

Manager, Analytical Chemistry Department, Royal Lubricants Co., Inc., E. Hanover, NJ 07936. 2 Director of Operations, Royal Lubricants. 3 Consultant, Royal Lubricants. 4 Environmental analyst, Royal Lubricants.

54

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Copyright9 by ASTM International www.astm.org Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproductions

PACHUTA ET AL. ON MONITORING PROGRAM

55

flight test. Successful completion of all flight test requirements allows the new fluid to become qualified for commercial use in the engine type(s) specified. A flight evaluation program is independently arranged and conducted between a fluid supplier and a cooperating airline. All program requirements such as the minimum number of engines, hours of flight operation, engine-fluid monitoring tests, and criteria for final approval of a new fluid are clearly established by the engine manufacturer. Detailed test engine teardown inspections of all fluid-wetted parts by the engine manufacturer are principally used to evaluate the new fluid's flight performance. Engine manufacturer requirements for qualifying aircraft turbine fluids and fluid-change guidelines for their in-service engines (see Table 1) were developed many years ago when the predominant engine design employed labyrinth-type bearing seals which resulted in corresponding fluid consumption rates on the order of 0.5 L/h. Since that time, carbon seal based engine designs have become predominant, which in turn have reduced turbine fluid consumption rates to approximately 0.02 to 0.20 L/h. This decrease in consumption rate was soon accompanied by reductions in cycle times and operational temperature increases in some higher thrust engines. For example, transient fluid exposure temperatures can reach approximately 370~ in some cases. Long-standing maintenance practices of some airlines involving no fixed fluid-change intervals have also continued. Such factors have imposed increased thermooxidative and operational stresses upon aviation turbine fluids. Most often the net effect of overstressing a turbine fluid can be visually observed during engine maintenance overhauls as increased levels of sludge, varnish, lacquer, corrosion, and coke deposits in the fluid-wetted components of the engine. Such conditions require extensive shop time and parts-cleaning operations to correct. They can also shorten the lifetime of expensive engine parts by preventing sufficient cooling of critical engine components, the proper operation of carbon seals, and subsequent increases in wear rates and wear-related damage to gears, pumps, and bearings due to circulating carbonaceous debris. In other more visible cases of fluid stressing, flights can be delayed or inflight engine shutdowns performed due to fluid-filter clogging. New generations of turbine fluids have been and will continue to be required to provide the levels of engine protection and service lifetime needed for increasingly severe operational regimes [1]. Advances in lubricant formulations have been greatly aided by analytical methods which were developed to investigate the mechanisms of antioxidant behavior and TABLE 1--Typical engine manufacturer program requirements and parameter limits for flight tests or in-service monitoring of MiI-L-23699 type fluids. Test Viscosity change [mmZ/s at 37.8~ (100~ (ASTM D 445)~ Total acid number [mg KOH/g, (ASTM D 664)b] Foam [mL, (ASTM D 892)c] Gravimetric solids [mg/L, 1.2 v.m paper, (FTM 3010)a]

Used Fluid Maximum Limits +25 to +35% 2.0 300 40

" ASTM Test Method for Kinematic Viscosity of Transparent and Opaque Liquids (and the Calculation of Dynamic Viscosity). b ASTM Test Method for Neutralization by Potentiometric Titration. c ASTM Test Method for Foaming Characteristics of Lubricating Oils. Federal Test Method for Solid Particle Contamination in Aircraft Turbine Engine Lubricants (Gravimetric Procedure).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

56

TURBINEOIL MONITORING

basestock degradiation. While much of this work remains proprietary, there are literature reports which feature some of these methods as well as the types of synergistic and degradive mechanisms associated with Mil-L-23699 type fluids [2-6]. Many of these works and others [7-21] have also described the application potential of various analytical methodologies to turbine engine monitoring programs (with most involving military aircraft). Published reports about the incorporation of analytical methods into monitoring programs for commercial aircraft are scant. The details of many analytical methods used for such applications are often not available, primarily because of proprietary considerations. Unlike applications involving electrical power generating steam and gas turbines [22], there are no currently recommended practices for performing commercial aircraft turbine engine monitoring offered by ASTM, SAE, or any other professional standards organizations. Routine monitoring of qualified in-service turbine fluids are also not currently mandated for commercial aircraft by either the engine, airframe, or accessory [for example, constant speed drive (CSD) or auxiliary power unit (APU)] vendors or any government agency. Historically, commercial airlines have depended primarily on engine manufacturer bulletins which list a monitoring program and associated fluid parameter limits for both flight tests and routine in-service operations. For flight tests involving new generations of fluids, however, the basic physical-chemical tests that are prescribed may not provide sufficiently detailed information about thermooxidative conditions or insure the detection of contaminants that could seriously degrade fluid performance and affect the flight test. The analytical methods which were included in this program were designed to provide information that would complement and supplement data obtained from standard physicalchemical tests. Their most important function was to provide unique and essential chemical information about the fluid's response to thermooxidative stress and to detect the presence of likely contaminants. The overall monitoring program consisted of two major stages. A listing of the tests employed in each stage is given in Table 2. Stage II tests were performed periodically on all monitored units or if warranted by trend-line behavior measured from Stage I tests. This strategy minimized the number of tests routinely performed yet provided for rapid confirmation and follow-up actions on any atypical result. A listing of all acronyms used to describe these analytical techniques and associated special terms used in this paper is given in the Appendix. Case studies were selected to illustrate some of the problems encountered during a comTABLE 2--Flight test monitoring program: two stage structure. Stage I Tests

Stage II Tests

Kinematic viscosity [at 37.8~ (100~ ASTM D 445)]

Kinematic viscosity [at -40~ and 200~ (ASTM O 445)] Foam test (ASTM D 892) Specific heat (ASTM D 3947)"

Total acid number (ASTM D 664) Water content (ASTM D 1744) ]automated titration] Size exclusion chromatography [Royal automated method] Trace metals and total phosphorous [Royal automated ICPES method]

Thermal analysis [Royal DSC method] Capillary gas chromatography ]Royal automated method] Scanning electron microscopy X-ray fluorescence spectrometry ]Royal method]

" ASTM Test Method for Specific Heat of Aircraft Turbine Lubricants by Thermal Analysis.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

PACHUTA ET AL. ON MONITORING PROGRAM

57

mercial flight evaluation of a new turbine fluid and to demonstrate some of the capabilities of the analytical monitoring program employed. The interpretation of each featured problem was either supported by the airline or strongly correlated to information obtained during subsequent engine teardown inspections.

Procedure

Monitoring Program Overview The main objectives of this analytically oriented program were to insure the safety of engine operations and overall integrity of the flight test so that the performance capabilities of the new fluid could be accurately evaluated. This program included all tests mandated by the engine manufacturer as well as additional ASTM-based physical and analytical tests. Chromatographic methods developed in-house were used to determine proprietary additive concentrations, monitor fluid degradation patterns, and detect the presence of likely contaminants. A differential scanning calorimetric method was used to measure the relative thermooxidative stability of used fluids. This parameter was considered to be critical for assessing the effectiveness of the new fluid's antioxidant system under actual flight conditions. An in-house developed emission spectrometric method was used to follow engine component wear metals, likely contaminating metals, and phosphorus (which was inherent to some additives in the formulation). Sampling was performed by the cooperating airline according to procedures and schedules mandated by the engine manufacturer. Throughout this study, test results and data interpretation were communicated to both the cooperating airline and the engine manufacturer for timely review.

Analytical Methods (a) Gas Chromatography (GC). A Varian 6000 gas chromatograph equipped with an automatic sampler, split-splitless injector, dual capillary columns, flame ionization (FID), and nitrogen-phosphorus (NPD) selective detectors was used for all gas chromatographic analyses of flight test samples. The columns were 30-m fused silica types (J & W Scientific) which had a 0.32 mm ID, a 0.25-1~m film of DB-5 stationary phase, and a 350~ upper temperature operating limit. These columns were commonly connected to the injector (via a two-holed ferrule) and separately to the individual detectors. Both detector signals were simultaneously displayed on a Varian 401 data station printerplotter. This instrumental arrangement and temperature program from 150 to 350~ provided for selective monitoring of most major and minor nitrogen phosphorus containing additives present in the formulation. It also presented a general high resolution view of most volatile and thermally stable carbon hydrogen containing basestock, additive, and degradation products present. The method used was designed to be capable of potentially detecting low levels of jet fuel, petroleum based hydraulic fluid, phosphate ester based aircraft hydraulic fluids, mixtures of aircraft turbine fluids from different manufacturers, and other foreign fluids. A major limitation of this GC method was that nonvolatile or thermally unstable chemical components could not be detected directly. (b) Size Exclusion Chromatography (SEC). A Waters liquid chromatograph (LC) equipped with a Varian LC autosampler, automatic Rheodyne loop injector, Waters column heaters, 254-nm fixed wavelength absorbance detector (UVD), and refractive index detector (RID) was used for all SEC analyses of flight test samples. Two high-resolution 30-cm, 500-,~ (Waters Ultrastyragel) SEC columns (connected in series) with a reported molecular

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. N

58

TURBINEOIL MONITORING

weight separability range o f between approximately 100 to 10 000 (relative to polystyrene standards) were connected to a serial arrangement of the detectors. Both UVD and RID signals were simultaneously displayed on a Waters data module. This instrumental system (which performs separations according to effective solvated size of a molecule) was primarily designed for the selective monitoring of most major UV absorbing additives and their UV active degradation products. Degradative changes in the molecular weight distribution of non-UV absorbing basestock components could be monitored by the RID. The major advantages of SEC are that it is a rugged, reliable technique capable of rapidly separating and detecting the presence and buildup of many nonvolatile, high molecular weight sample components, some of which can potentially be sludge precursors. It also complements GC data and extends the range of chemical information obtained from the sample. Another practical advantage is that one simple and common sample preparation (dilution-filtration) scheme is required for both GC and SEC analyses. A major disadvantage of SEC is its inherently lower resolving power and selectivity relative to high-temperature capillary GC techniques. (c) Differential Scanning Calorimetry (DSC). A Dupont Model 1090B thermal analysis system and DSC cell was used to obtain all thermograms of flight test samples. Air at 1 bar pressure and 25 mL/min flow was used as a cell gas. Sample weights of 2.0-mg, open, flatbottom aluminum sample pans and 60~ heating rates from 25 to 400~ were other essential method parameters. With this system, the temperature at which a large, sharp exothermic signal appeared (oxidative "breakpoint" temperature) was taken as a measure of the "total antioxidant capacity" (TAC) of the used lubricant sample. While there is no formally accepted definition of this term, it is intended to describe the relative capability of a lubricant to continue tolerating normal thermooxidative stress loads without losing its ability to perform protective functions within an engine system. Based on developmental data obtained with this DSC method, new (unused) fluid exhibited a maximum "breakpoint" temperature of 320~ A "worst case" sample taken at the end of a prolonged (280 h) Coordinating Research Council (CRC) type oxidation-corrosion test (run at 190~ had a corresponding minimum "break-point" temperature of 275~ This "worst case" sample displayed antioxidant exhaustion behavior (that is, TAC values reached a relative minimum prior to 280 h which coincided with enhanced viscosity and TAN slope increases, etc.). In this manner, the analytical working range between relative maximum and minimum TAC levels was established. Experimental "breakpoint" temperatures for this new fluid were observed to decrease in a generally linear manner with increasing CRC test hours. It must be emphasized, however, that just as viscosity and TAN curves exhibit formulation-dependent behavior during CRCtype tests for various MiI-L-23699 products, dynamic DSC curves also exhibit formulation dependency [16]. Other investigators have also reported formulation-dependent TAC curves when employing pressure differential scanning calorimetry (PDSC) under isothermal analytical conditions [3,4, 7, 9,15]. DSC methods have several advantages over various chromatographic methods that are used to estimate the TAC behavior of a lubricant. A fundamental advantage of DSC is that it provides a measure of a fluid's total resistance to massive, rapid oxidation. The antioxidant properties of a turbine fluid are dependent upon the nature and concentration of all the chemical components that exhibit antioxidant behavior. This includes not only the antioxidant additives but also secondary antioxidant species formed by in-situ thermooxidative reactions. Often these synergistically formed antioxidants are more effective than the original formulation additives. Because of these factors, chromatographic methods which simply measure the remaining formulation antioxidants can provide misleading information about a lubricant's ability to withstand continued oxidative degradation and thermal stress.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further

PACHUTA ET AL. ON MONITORING PROGRAM

59

An advantage of this dynamic DSC melthod is that it is rugged, operationally simple, reproducible (approximately 1.5~ repeatability), rapid (analysis time less than 10 rain), and data interpretation is straight forward. One of the disadvantages inherent to apparently all DSC methods concerning Mil-L-23699 type fluids is that the formulation dependency of TAC curves complicates the establishment of universally applicable "warning limits" on TAC values. (d) Inductively Coupled Plasma Emission Spectrometry (ICPES). A ThermoJarrelt Ash Plasma 300 sequential-scanning instrument equipped with an autosampler, air path, and vacuum monochromators was used to monitor the levels of wear metals, potential contaminant metals, and total phosphorus content in all samples. A particle size dependent method involving simple dilution of the used fluid (50 wt/wt percent sample load) with kerosene was routinely employed. Table 3 lists the 21 elements that were monitored, which include wear metals as well as potential contaminant metals from sources such as dirt, cement dust, rust-inhibited or shock-strut type hydraulic fluids, and other foreign fluids. Emission spectrometry has been applied to monitor wear metals in turbine fluids for many years, and these applications have been well documented. Most of the application literature, however, has involved electrode-based instrumentation, which is widely used in military programs. Recent and apparently future trends concerning aircraft wear metal monitoring suggest that ICPES-based instrumentation may soon become predominant for both military and commercial applications, especially when widespread incorporation of "fine filtration" technology [19] begins to occur in engine designs. The advantages of the ICPES instrumental system employed in this study are its ease of sample preparation, high sample load capability, automated operation, flexibility of elemental selection, excellent sensitivity, low detection limits, and good overall reliability. The disadvantages of the method employed are its particle size dependence and need for sample

TABLE 3--Elements, transition types, and analytical wavelengths monitored by an ICPES test method. Element

Transition

Wavelength, nm

Silver Aluminum Barium Boron Cadmium Calcium Chromium Copper Iron Magnesium Manganese Molybdenum Sodium Nickel Phosphorus Lead Silicon Tin Titanium Vanadium Zinc

Atomic Atomic Ionic Atomic Atomic Ionic Ionic Atomic Ionic Atomic Ionic Ionic Atomic Ionic Atomic Ionic Atomic Atomic Ionic Ionic Atomic

328.07 308.22 233.53 208.89 228.80 317.93 267.71 327.40 259.90 285.21 294.92 202.02 589.59 231.60 253.57 220.35 251.61 284.00~ 336.12 292.40 213.86

" Interelement correction for Cr is required when using this wavelength.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

TURBINEOIL MONITORING

60

dilution. Analysis times of approximately 5 min per sample did not pose any problems for this study but would be a consideration for programs involving very high daily throughput. Results and Discussion Case I: Insuring the Integrity o f the Engine System's Fluid Charge Case l(a) Determination o f Foreign Turbine Fluid Contamination. Most of the engines, CSD, and A P U systems employed in this flight test had been previously operated on one or more Mil-L-23699 formulations from other suppliers. Changeover to the new turbine fluid typically consisted of a simple drain-flush-fill operation. Of initial interest in this study was the completeness of fleet changeover and the amount of residual foreign turbine fluid remaining in each unit. Both viscosity and TAN data are not capable of supplying such critical information. In Fig. 1, curves for Engine 324 and A P U 236 demonstrate that due to the inherently lower phosphorus content of this new fluid compared to many other commercial turbine fluids, the presence of these other fluids could be detected via an ICPES method. Figures 2 through 4 demonstrate that SEC-UVD, GC-NPD, and GC-FID methods were capable of detecting and potentially identifying the intermixed foreign turbine fluid by the differences in their additive and basestock compositions. Quantitative data concerning fluid intermixing are difficult to obtain without a sample of the formerly used fluid just prior to changeover, but a minimum dilution can be established if one assumes the formerly used fluid had remaining additive concentrations that were close

Turbine Fluid C h a r g e

Moniforing

Total P Levels va. Fluid H I 1 . 1.5 1.4 1.3 1.2 1.1

~,~

1.0 0.9

,g Jl~ 9c

0.8

F ~

0.6 0.5 0.4 0.3 0.2 0.1

0.0

I

"1

0 D

3 2 4 Eng

2 +

450 s

I

(ThousQnds) Fluid Hrs. o Z 3 6 APU

I

4 A

461

r'ng

FIG. 1--[~ Engine 324 found to contain residual amounts of a formerly used foreign turbine fluid (refer to Fig. 3). + Engine 450 suspected of experiencing water contamination at approximately 1802 fluid hours (refer to Fig. 6). ~ APU found to contain a significant amount of foreign turbine fluid (refer to Fig. 2). /~ Engine 461 found to be contaminated with hydraulic fluid at 4541 fluid hours (refer to Fig. 5A).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further r

PACHUTA ET AL. ON MON{TORING PROGRAM A

61

GC-F}D Profile New Fluid

Polyot Ester Baeeatock

Components .

L

]. . . . . . . .

. . . .

>

Areu I Time ( m l n . ) . . . .

>

GC-FID Profile APU 0.1 h Sample Polyol Ester Basestock Components . . . . . > Antloxldant from o Foreign Turbine Fluid

. . . . . . . .

>

Area 1 I Time (rain.). . . . > FIG. 2--(A) Reference profile of new MiI-L-23699 turbine fluid. (B) 236 APU sample. Comparisons with Fig. 2A (that is, Area 1) confirm the presence of a formerly used turbine fluid.

to formulation values. Based on this assumption, typical residual levels of foreign turbine fluid were found to be between a minimum of 1 to 10% just after the initial fill of new fluid. Throughout the course of this 15-month evaluation study (which involved a total of eight twin-engine aircraft), engine, CSD, and A P U samples were monitored for fluid charge integrity. ICPES and SEC data were routinely used to screen samples for possible intermixing with foreign turbine fluids or other contaminating fluids. Use of unauthorized "make-up" turbine fluid in some engines, APU, and CSD units was quite infrequent, but proved to be quickly and easily detectable using this combination of analytical methods. Improper aircraft servicing at a distant airfield by contract maintenance crews from another airline was eventually traced as the apparent problem source.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No fur

62

TURBINE OIL MONITORING

A

GC-NPD Profile New

Fluid

Major N,P Bearing Addlfivee - - >

j

II ....... Tlme (mln.) . . . .

>

B

GC-NPD Profile Eng. 324 116h Sample

HoP Bearing Additives - - >

Antloxldanl From a Foreign Turbine

Fluld . . . . . . . . . . .

>

Time (mln.) . . . .

>

FIG. 3--(A) Reference profile of new Mil-L-23699 turbine fluid. (B) Engine 324 after 116fluid hours. Comparisons with Fig. 3A confirm the presence of a formerly used turbine fluid.

The presence of intermixed turbine fluids in routine commercial fleet operations is currently not a major concern and is probably not as frequent an occurrence as possibly exists in military operations. However, problems related to turbine fluid intermixing are expected to become of increasing concern in the future as airlines begin to convert part or all of their fleets to advanced generations of low-coking, highly thermally stable fluids. Since the performance desired from these premium-cost fluids can be degraded by intermixing with earlier generations of turbine fluids, periodic monitoring for fluid intermixing via analytical methodology may become necessary. Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No fur

PACHUTA ET AL. ON MONITORING PROGRAM

63

Another reason for possible concern about potential fluid intermixing problems is that future engine designs employing "unducted fan" technology could require specialized types of turbine fluids which may not be totally compatible with current MiI-L-23699 type fluids. Other trends toward the introduction of rust-inhibited versions of MiI-L-23699 and the development of specialized fluids for gas turbine helicopters also indicate that analytical methods capable of diagnosing fluid intermixing may eventually be added to some routine monitoring programs. Case l(b) Contamination from Fire-ResistantHydraulic Fluid. On one occasion, increases in total phosphorus levels above normal "steady state" levels and abnormal viscosity decreases were found in both engines of an aircraft. Examples of this behavior are illustrated in Figs. 1 and 9 for Engine 461. Subsequent follow-up analyses by GC-FID and GC-NPD indicated that these atypical trends were not totally due to a possible increase in fluid consumption. Figure 5 illustrates the presence of three prominent additional peaks in the GC profiles of samples taken from these engines. Such peaks are typical for the major components of phosphate-based hydraulic fluids. Comparisons of retention time and relative peak ratio data against commercial hydraulic fluid GC profiles were successful in identifying both the manufacturer and specific type of fluid present. Confirmation of the hydraulic fluid's identity was obtained from the cooperating airline, who reported that this specific hydraulic fluid was used exclusively in their fleet. The levels of hydraulic fluid contamination found in both engines indicated possible erroneous addition during routine fluid "make-up" operations, although this fact was never clearly established. Standard addition techniques which involved spiking known amounts of the suspected hydraulic fluid contaminant into new turbine fluid indicated that this contaminant was present at a minimum of approximately 2%. This is close to a 3 to 6% level, which would be expected from typical 1 to 2-L make-up additions to the main fluid tank. Rather than risk the possibility of continuing to operate this aircraft with both engines contaminated, complete drain-flush-fill operations were performed before resampling could confirm fluid charge contamination. Subsequent engine samples did not show any further evidence of this hydraulic fluid, which tended to eliminate any remote possibility of leakage between the gear box and hydraulic pump system. Case l(c) Contamination of CSD and Engine Systems with Water. A series of water contamination problems was experienced in the test fleet during a two-month period beginning in October 1986, primarily involving several CSD units (Fig. 6), with indirect evidence of engine fluid contamination. This incident demonstrated how complementary data from various program tests were sometimes required to indirectly detect and solve external contamination problems. The presence of water in CSD units is apparently a relatively common problem in both commercial and military aircraft. However, apparently few commercial airlines routinely monitor CSD units for water [24], and appropriate monitoring has just started to be more widely implemented by the military [25]. Due to the fact that CSD units are sealed and do not "breathe" as compared to engine systems, which are vented, external moisture, or water contaminated, "make-up" fluid which enters a CSD system tends to remain, accumulate, and can eventually lead to mechanical damage. Visual detection of water-contaminated CSD fluid via sight gauge examinations cannot reliably prevent subsequent CSD damage as they can only reveal a grossly contaminated essentially biphase fluid system. It must be recognized that synthetic polyol esters which are basestocks for Mil-L-23699 type turbine fluids are much more hydrophilic than petroleum oils. Because of this property, these ester-based fluids are capable of carrying up to approximately 4000 ppm of water at 20~ [as measured by ASTM Test Method for Water in Liquid Petroleum Products by Karl Fischer Reagent (D 1744)] before becoming visually hazy.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No fur

64

TURBINE OIL MONITORING

A

SEC-UVD Profile New Fluid

Molor UV Aellvo Additives . . . . .

>

High M.W.

Low Id.W.

Exelullon

~,./ ...~ Exoluslon

Limit.--> I

J

9"

Tlmo (mln.) . . . .

",~,t --->1

>

SEC-UVD Profile

Eng. 481 1.4h Sample Molor UV Aoflv, Additives . . . . .

>

Antloxldon! From o Foreign Turbine Fluid . . . . . --->//~

Low M.W.

High M.W. Exr

Llmlt,--> I

Ex 9

-.J~Umlt --->1 Time (mln.) . . . .

>

FIG. 4--(A) Reference profile of new MiI-L-23699 turbine fluid. (B) Engine 481 at 1.4 fluid hours indicating low levels of a formerly used turbine fluid which contains a high molecular weight (UV active) antioxidant. (C) Engine 469 at 105 fluid hours illustrates a normal engine-use pattern involving the insitu formation of oligomeric antioxidants.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduction

PACHUTA ET AL. ON MONITORING PROGRAM C

65

F-SEC-UVD Profile Eng. 469 105h Sample

UolorUVActive ..... > /

Additives

AHIngh:~dld.anY~.;, 0lelg.o_m_e r_l0_; HighM.W. Exclusion Llmlt.--> I

/

H M.W. Exclusion ~./~mlt --->l Low

Tlmo (mln.) . . . .

>

FIG. 4--Continued. Elevated water levels can lead to partial hydrolysis of such turbine fluid basestocks. Hydrolysis is a direct consequence of the fact that water is a reaction product of ester synthesis from polyhydric alcohols and fatty acids. This reaction is reversible, therefore if CSD water levels become high and are allowed to remain in the unit, hydrolysis apparently soon follows. Contamination-induced hydrolysis is evident in Fig. 7 from the sharp TAN increases observed in CSD 450 samples taken one month after water contamination was first detected at 1641 fluid hours. The TAN curves for both CSD units shown in this figure also indicate that hydrolysis reactions were relatively slow in these systems because TAN increases were not coincident with sharp rises in water content. An important observation made was that if CSD water levels began to exceed approximately 1000 ppm, corrosion rates (as measured from ICPES-based Mg levels) associated with the magnesium CSD casing would soon begin to increase significantly (see Fig. 8). As was observed with TAN response, Mg level maxima do not necessarily coincide with the onset of water contamination. Therefore, use of Mg or TAN monitoring data can indirectly diagnose water contamination, but only after some degree of damage has occurred. Detection of engine contamination from water is often not as direct as are incidences involving CSD units. Most of the contaminating water is expected to escape from the engine during warm-up operations due to the much higher temperatures involved (that is, typical main fluid sump temperatures are approximately 200~ and because engines are designed to breathe as well as being mechanically vented. In one engine suspected of being contaminated by water at approximately 1500 fluid hours (see Fig. 6, Engine 450 example), measured increases in engine water content were minimal, although the CSD associated with that engine had been seriously contaminated by water at that time (see Fig. 6,450 CSD example). A slight increase in TAN was subsequently observed for this engine (see Fig. 7, Engine 450 example), which could have indicated some fluid hydrolysis. Because of the expected rapid venting of water, tank dilution effects and loss of some volatile free fatty acids generated, such TAN increases would tend to be minimal and not be as pronounced as was observed in sealed CSD systems (see Fig. 7, 450 CSD example).

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No fu

66

TURBINEOIL MONITORING GC-NPD Profile Eng. 461 4541h Sample 3.

2.

H,P Bearing Additives - - >

Peaks 1 - 3 Phosphate Ester-

\ Time (mln.) . . . .

>

GC-FID Profile Eng. 498 894h Sample Polyol Ester Basntoek Components . . . . . > 2.

3. Peaks 1 - 3 Phosphate Estero

~ m . (mln.) . . . .

>

FIG. 5--(A) Engine 461 at 4541 fluid hours indicating contamination (Peaks 1 to 3) with a phosphate ester-based hydraulic fluid. (B) Engine 498 at 894 fluid hours indicating contamination (Peaks 1 to 3) with a phosphate-ester based hydraulic fluid.

In some cases of engine contamination by water, it is known that main fluid filter deposit levels can shortly increase to the point of complete plugging. The cause is thought to be due to a "steam-cleaning" effect by engine-heated water, resulting in the partial loosening to rapid, extensive shedding of internal carbon deposits. Significant transient increases in carbonaceous filter debris did accompany these suspected water contamination episodes. Visual examination of these carbon deposits revealed a "rustCopyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No

67

PACHUTA ET AL. ON MONITORING PROGRAM

Turbine Fluid Charge Monitoring ASTM D - - 1 1 4 4 W a t e r Content vs Fluid Hrs. 2.6 2.4 2.2 2.0

E Q,

1.8 1.6 1.4 1.2

L.I-. eV

G

1.0 0.8 0.6 0.4

0.')

+-....._+

0.0

I

i

o a

4 5 0 CSD

1 +

4 5 0 Eng

I

I

2

i

Thou.andw) uld H o u r i <> 4 1 4 CSD

i

i

3

4 A

7 0 3 CSD

FIG. 6--[S] CSD from Engine 450. The sampling date associated with a sharp rise in water levels at

1641 fluid hours was 11/10/86. A fluid change was performed on 1/20/87 due to excessive water contamination. + Engine 450. The sampling date associated wth the slight apparent rise in engine water levels at 1543 fluid hours was 10/31/86. ~ CSD from Engine 414. A sample taken on 10/27/86 with 1362 fluid hours contained 6800 ppm of water, was opaque, and displayed free water. The pilot reported a "low-oil-pressure" warning light on the 10/27flight which initiated this sampling. A CSD from Engine 703. A n immediate fluid change was performed at 1653 fluid hours when 1100 ppm o f water contamination was found.

colored" coating on some of the carbon particles. Elemental analysis by scanning electron microscopy-X-ray fluorescence spectrometry (SEM-XRF) determined that the "rust-colored" coating consisted of submicron "dustlike" material that was comprised primarily of Fe. Phosphorus was the only other predominant element found via SEM-XRF. A more detailed morphological analysis was not pursued. Wear metals in the used fluid samples, especially Fe and total phosphorus levels, underwent slight coincidental upward jumps in this engine between approximately 1500 to 2000 fluid hours during this episode (see Figs. 8 and 1, Engine 450 examples). One possible explanation for the slight increase in phosphorus levels is that degradation products from phosphorus-containing additives are known to be associated with stationary carbon deposits within the engine. The weight percent of elemental phosphorus in some deposits can also be enriched over its corresponding percent in the fluid formulation. Therefore, it is surmised that certain incidents which cause accelerated rates of carbon deposit mobilization can be accompanied by observed increases in total phosphorus concentrations in fluid samples. Increases in Fe content can be caused by a number of mechanisms including corrosion and accelerated wear rates due to circulating carbon debris. While it cannot be positively established that either the "rust layer" or shedding was directly or solely caused by water contamination, the fact that significant increases in water

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. N

68

TURBINEOIL MONITORING

Turbine Fluid Charge Moniforlng ASTM D - - 6 6 ~ ~o0

TAN v= Fluid Hr=.

--

4.0

Z

2.0

1.0

0.0 ~[-': -0

,

~ 2

i

i

4

FThOu=mnd ) D

4 5 0 CSD

+

450

Eng

uld Houm # 414

CSD

A

481

s

FIG. 7--[5] CSD from Engine 450. Samples taken approximately one month after a sharp rise in water levels was observed at 1641 fluid hours (refer to Fig. 6A) suggest the onset o f basestock hydrolysis reactions due to entrained water. + Engine 450. The relative TAN maximum seen at 1802 fluid hours was from a sample taken approximately one month after a maximum in engine water content was observed (refer to Fig. 6B). 9 CSD from Engine 414. An immediate fluid change was performed at 1362 fluid hours when 6800 ppm o f water contamination was found (refer to Fig. 6C). A Engine 461. This curve generally represents typical TAN behavior for test fleet engines up until approximately 3800 fluid hours when engine starter problems were encountered.

levels also occurred in the CSD units attached to these engines during the same time period was considered to be strong circumstantial evidence, A subsequent investigation by the cooperating airline into these incidents revealed that at a distant airfield a contract maintenance group from another airline had improperly transferred the new fluid from hermetically sealed quart cans into a large capacity portable fluid tank. This tank was reportedly stored outside a hanger and contained visual amounts of water contamination. Following the correction of this improper service practice, no similar episodes were detected. Case II: Confirmation o f the Establishment o f an Engine-Fluid "Steady-State" C o n d i t i o n

A "steady state" condition within a turbine engine's fluid system has been described as an operational period during which the rate of antioxidant loss due to thermooxidative stress is closely matched by the rate of antioxidant addition into the engine via "make-up" fluid [5]. The kinetics of these processes can be studied either individually by plotting quantitative data obtained for each antioxidant in the formulation, or collectively via TAC profiles. Establishment of an engine-fluid "steady state" normally requires the addition of between one to four lubricant charge, volumes of "make-up" fluid. For engines equipped with carbon

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement.

PACHUTA ET AL. ON MONITORING PROGRAM

69

seals, this usually represents between approximately 640 to 2500 fluid hours of flight operation. These calculations are based on the assumptions that a typical engine fluid system capacity is approximately 32 L, fluid consumption is 0.05 L per hour, and that a "stirred flow reactor" model [5] adequately describes the kinetics of antioxidant degradation in an aircraft turbine engine. Once a stable "steady state" has been established, relative changes in the fluid's physical and chemical properties should become minimal. If such stable conditions are not observed within the calculated period of operation or if significant departures from established "steady state" conditions occur, the presence of abnormal engine stress, contamination, or formulation inadequacies should be investigated. Among the elements of performance that were desired to be monitored during this flight test was whether or not the new turbine fluid's chemical and physical properties would exhibit normal "steady state" behavior within engines, CSD, and A P U systems of the test fleet. Frequent sampling and inspections of fluid filters and breather pressure checks were among the procedures used to determine this. Graphing of viscosity, TAN, and engine data (that is, total fluid added since zero-hours-on-fluid), as well as decay curves of total phosphorus levels, formulation additive levels and TAC data provided detailed information which helped to verify the establishment of a "steady state" condition. As expected, profiles of remaining formulation antioxidant levels (which were determined via GC-NPD and not shown for proprietary reasons) as well as TAC curves tended to be inversely associated to the fluid's viscometric profile. A good agreement between TAC and viscosity trendlines was considered to be an essential indication that "steady state" conditions with respect to the overall rates of antioxidant loss

Engine Hardware Condifion Moniforing Mg end

Fe L e v e l l vs. Fluid H i s .

150 140 130 120 110 "~

100

~

9o

_m o

8o

;

7o

~

6o

~

5o 4o 3o 2o 10 0

.it. i l l

L...: L :

~

I

0

I

!

2

4 (Theulandl)

Fluid Hnl, o

Mg l e v e l - , -

450

CSD

+

Fe l e v e l s -

450

rng

FIG. 8--[] CSD 450 Mg levels. The peak Mg level at 2219 fluid hours was taken approximately two months after a maximum in water levels was detected on 11/10/86. + Engine 450 Fe levels. The peak Fe level at 1125 fluid hours was taken on 11/25/86. Starter shear was concurrent with the observed Fe increase, starting at approximately 4467 fluid hours.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further

Turbine Fluid Charge Monitoring V i l ~ o s i t y o t 3 7 . 8 " C v l . Fluid Hrs.

50.0 411.0 46.0 A

44.0 42.0 40.0

U

1

A

tt

38.0

0

o

36.0

E =

34.0 32.0 30.0 28.0

I

I

I

2 +

3 2 4 Eng

I

housonda) uid H o u r s o 5 0 0 Eng

4 8 9 Eng

4 A

481

s

Turbine Fluid Charge Monitoring Total Anfloxldont Copaolty vs Fluid Hr~

330

B 320

&.

310

E g I.-

_~

300

0 Q. 0

P m 0 (/1

290

280

270

i

I

0

I:1

324 [ng

I

2 +

4 8 9 Eng

i

4 FTIh o u s a n d l l ) uld H o u r s r 5 0 0 Eng

A

461 Eng

FIG. 9--Legend for both Viscosity and Total Antioxidant Capacity Graphs. [] Engine 324 (new, zerotime test engine). A scheduled engine teardown inspection was performed at 3200fluid hours. The average fluid consumption rate was O.031 L/h. + Engine 469 (zero-time test engine). A scheduled engine teardown inspection was performed at 3700 fluid hours The average fluid consumption rate was 0.025 L/h. 0 Engine 500 (high time test engine, with approximately 8000 h since its last shop visit). A scheduled engine teardown inspection was performed after 3600 h on the new fluid. The average fluid consumption rate was 0.052 L/h. A Engine 461. A foreign object damage (FOD) initiated teardown inspection was performed after 4600 fluid hours. The average fluid consumption rate was 0.042 L/h. Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreemen

PACHUTA ET AL. ON MONITORING PROGRAM

71

and replenishment were being attained. Examples of typical TAC behavior for Engines 324, 469, 500, and 461, as well as their corresponding viscosity curves and average fluid consumption rate data, are given in Fig. 9. Plots of corresponding phosphorus levels for Engines 324 and 461 (excluding the abberational points previously described) shown in Fig. 1 indicate that other major additives reached their "steady state" levels at about the same time as TAC and viscosity parameters. Phosphorus levels also tended to follow an inverse relationship to viscosity data. TAN curves for test engines, CSDs, and APUs normally showed relatively small changes from new (unused) fluid, with maximum typical values below 1.0 mg KOH per gram. Comparisons of the TAN curve for Engine 461, shown in Fig. 7, with the corresponding TAC and viscosity curves for this engine, shown in Fig. 9, demonstrate that "steady states" with respect to these three parameters were attained coincidentally. For this flight test study, this concidental establishment of "steady states" between several key chemical and physical parameters along with the attainment of such "steady state" conditions within expected operational periods were found to be useful for assessing the normalcy of fluid performance prior to obtaining confirmatory engine inspection data. In cases where significant departures from established "steady state" conditions were observed, either fluid contamination or consumption rate changes were subsequently diagnosed as the cause.

Case Ili: Establishment of the Fluid's Relative Condition During "Steady State" Operations The absolute levels of a lubricant's additives remaining during "steady state" operations are related to many variables. Thermooxidative and operational stress factors such as engine temperature regimes, residence times within these regimes, consumption rates, and formulation design are among the major variables influencing additive levels within these "steady state" periods. Throughout flight test applications it is desirable to ascertain the degree of thermooxidative stress imposed and fluid response to various stress factors. Such information can indicate the likely "reserve" of additive protection remaining and potentially signal the need for a precautionary fluid change if the estimated "reserves" of critical additives become severely depleted. Basic viscosity and TAN data cannot supply such compositional information but can be valuable when interpreted along with appropriate analytical data. Historically, viscosity increases of between 25 to 35% at 37.8~ (100~ have been primarily used as "warning limits" indicative of excessive thermooxidative stress and fluid degradation. Filter changes and fluid charge refills have normally been recommended if such limits are exceeded. Test results from this study, however, indicate that viscosity changes at approximately 40~ that are on the order of 35% or more do not necessarily indicate that a serious reduction in additive "reserve" levels has occurred. For example, as shown in Fig. 9A, Engine 461 exhibited "steady state" viscosities that were in excess of a 35% change. Corresponding "steady state" TAN data for this engine (see Fig. 7) indicated a maximum TAN of 0.5 mg KOH per gram, which was well below the currently used "warning limit" of 2.0 mg KOH per gram. A subsequent tear-down inspection of Engine 461 revealed that it did not exhibit significant amounts of sludge or carbonaceous engine deposits. This inspection did reveal that some of the heat shielding surrounding a high-temperature-bearing compartment had been damaged, although the time of this occurrence and its overall significance to engine fluid properties were not known. TAC curves for this engine shown in Fig. 9B indicated that "reserves" of between 50 to 70% above estimated minimum levels were always present. This estimation assumes that fluid degradation mechanisms in aircraft engines are closely simulated by CRC-type oxi-

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No furthe

72

TURBINE OIL MONITORING

dation-corrosion test behavior. It also assumes that TAC decreases linearly with constant thermooxidative stress. The levels of proprietary antioxidants and antiwear agents that were monitored by GC-NPD, GC-FID, and ICPES (total phosphorus) also tended to support TAC data which indicated that sufficient levels of additive protection remained in Engine 461 and all other engine fluid charges during "steady state" operations. SEC-UVD data was also useful in estimating the relative condition of used fluids. These profiles revealed that higher relative levels of in-situ generated oligomeric antioxidants were present in engine fluids displaying higher viscosities. This association is demonstrated by the SEC-UVD profiles shown in Fig. 10 for Engines 500 and 461 when both had approximately 1000 fluid hours and viscosities at 37.8~ (100~ of 32.1 and 38.7 mm2/s, respectively. Such analytical data along with low TAN changes and results from engine inspections supported the interpretation that Engine 461 viscosity increases were predominantly associated with synergistic reactions involving the new fluid's antioxidant package rather than from extensive basestock degradation. Additional engine teardown inspections performed on Engines 324,469, and 500 showed relatively low, acceptable levels of engine sludge and deposits associated with the "steady state" TAC and viscosity levels shown in Fig. 9. The essentially horizontal slope of the TAC curves also appeared to indicate that the service lifetime of this fluid was not approaching a thermooxidative stress related limit. In another recent commercial turbine engine study where analytical methods including PDSC were employed it was reported [21] that it is also possible to monitor relatively small viscosity and TAN changes in modern commercial aircraft engines, yet have a highly stressed fluid generating significant amounts of coke deposits within fluid-wetted areas. The TAC and chromatographic data shown in that work apparently indicated that antioxidant levels remained relatively high. Such discrepancies indicate that much more field experience, including engine inspection work, may be required before DSC- or PDSC-based methods for TAC determinations and results from other analytical techniques can be fully evaluated for relevance to engine cleanliness. Due to the complexity and diversity of engine deposit formation mechanisms and formulation dependence of TAC data, it is unlikely that DSC-based methods alone can ever be used to reliably predict the cleanliness of fluid-wetted hardware. A more probable application of TAC measurements is that like established viscosity and TAN behavior, significant changes in TAC trendlines should be useful in diagnosing certain types of abnormal operational changes developing within an engine. When correlations of TAC data to existing bench test data such as the CRC test are known, TAC monitoring appears to be useful for supplying information about the relative thermooxidative stability of a given fluid.

Case IV." Estimating the Mechanical Integrity of Fluid-Wetted Hardware An inherent problem associated with flight tests of any new turbine fluid involves the lack of a prior service record for characterizing normal hardware behavior in a given system. Especially in the early phases of this study, much dependence was placed upon prior knowledge of fluid performance gained from mechanical-rig and test-stand studies. However, due to the fact that a significant number of test engines were involved, correlations of flight test data to subsequent test engine teardown results soon developed. When applicable, a history of prior engine teardown results was included in these correlation studies. Engine history searches can be invaluable for distinguishing which types of hardware problems may be engine design related from those that may be caused by flight operations or fluid formulation limitations. The development of reliable "hardware condition" tests that can be easily interpreted has

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No furt

PACHUTAET AL. ON MONITORINGPROGRAM

A

73

SEC-UVD Profile Eng. 5 0 0 930h Sample 2

Vis. = 32.1 mm / s IdalorUVActive <-- Additives

Higher M.W.-Ollgomerle Anlloxldant Serles

---~,~

High I d . W .

Exo,-,_,on Limit.

>

~

.

/

/

\

Low M.W. Exclusion

Limit--->1

B

SEC-UVD Profile Eng. 461 s

~

1055h Sample

2

Vis. = 38.7 mm /

.ox.t. / Higher

Id.W.-Ongomerlc /

High Id.W. / Exclusion j Llmlt.-->[/~-~

/

\A

/~

v/

= ~ II

\ \

<--

\.

Uolor UV Active Additives

Low M.W. Excluelon Limit - - - > l

FIG. IO--SEC-UVD profiles of used engine fluids with similar fluid hours and consumption rates. (A) Engine 500 at 930 fluid hours. This sample had a viscosity of 32.1 mm2/s at 37.8~C (100~ (B) Engine 461 at 1055 fluid hours. This sample had a viscosity of 38. 7 mm2/s at 37.8~ (IO0~

historically been and continues to be the most complex and important goal of turbine fluid monitoring. From the "fluid-centered" emphasis of this study, it was assumed that if the new fluid was not performing its primary design functions of cooling, minimizing wear rates due to adhesion, spalling, abrasion, corrosion, etc., that ICPES-based wear metal data or results from other program tests designed to monitor the fluid's condition would detect most forms of fluid-induced mechanical stress. The primary limitation of this program was that only progressive, method-detectable hardware failure modes which would occur over a time period longer than the unit's sampling schedule could be diagnosed in time for corrective actions to be taken. An example of a different type of limitation related to "hardware-condition" monitoring that was encountered during this study involved occasional communication problems associated with sample background. For example, as shown in Fig. 7, Engine 450 exhibited

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreemen

74

TURBINE OIL MONITORING

a sharp, rapid jump in Fe content from 39 to 144 ppm. A resampling from this engine confirmed the unusually high jump and also indicated that an apparent leveling off was taking place. No unusual vibrations or other changes in engine operating parameters were initially reported. Further investigations revealed, however, that the engine's starter had sheared prior to taking the first abnormal sample, which apparently accounted for the Fe behavior. This critical information was not initially reported to appropriate airline and laboratory personnel. Such failures to communicate essential background information can lead to potentially costly and unnecessary fluid changes, engine replacements, and teardown maintenance. This in turn can lead to an undermining of the reliability and importance placed in the overall monitoring program. Therefore, besides instituting immediate resampling when atypical results are obtained, a thorough and persistent inquiry into the background facts surrounding such episodes can be critical to the accurate interpretation of test data and to taking the proper course of remedial action.

Conclusions Several advantages of integrating advanced analytical techniques into monitoring programs designed for commercial flight tests of new turbine fluids were clearly demonstrated. The analytical methods featured were rapid, rugged, easily automated, and required minimal sample preparation. Because of these factors and the strategy employed for their use, both labor and supply costs associated with performing these methods were nominal. Relatively high sample throughput was possible when necessary. Therefore, both economic and work capacity considerations should not limit the incorporation of these analytical techniques into commercial flight test applications of similar scale. These analytical methods should also be useful in routine monitoring programs when suspected problems arise that cannot be solved by standard physical-chemical tests. Included in the abilities of the two-stage analytical monitoring scheme were the detection of turbine fluid intermixing as well as contamination from water and phosphate-based hydraulic fluid. Thermal methods used in conjunction with chromatographic and standard physical-chemical tests were found to be useful in determining the relative thermooxidative stability of used fluid samples and in studies of fluid degradation mechanisms. The information obtained from this flight test generally supported results from another recently reported commercial monitoring study [21] which highlighted some of the limitations of established engine manufacturer programs and possible problems associated with currently recognized "warning limits" for initiating costly turbine fluid changes. Additional basic research, method development work, and correlation studies are required in order to further improve the diagnostic capabilities of turbine fluid monitoring programs. Continued advances in this field are important because existing "on-line" condition sensors (that is, chip detectors) and "off-line" laboratory-based techniques still possess limited abilities with respect to providing reliable, predictive information about the cleanliness and state of fluid-wetted hardware within an aircraft turbine engine.

Appendix Alphabetical Listing of Acronyms Employed APU CSD CRC

Auxiliary power unit Constant speed drive Coordinating Research Council

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No fu

PACHUTA ET AL. ON MONITORING PROGRAM

DSC FID FOD GC GC-FID GC-NPD ICPES JOAP LC NOAP NPD PDSC RID SEC SEC-UVD SEM-XRF TAC UV UVD

75

Differential scanning calorimetry Flame ionization detector Foreign object damage Gas chromatography Gas chromatography with a flame ionization detector Gas chromatography with a nitrogen phosphorus selective detector Inductively coupled plasma emission spectrometry Joint oil analysis program Liquid chromatography Naval oil analysis program Nitrogen phosphorus selective detector Pressurized differential scanning calorimetry Refractive index detector Size exclusion chromatography Size exclusion chromatography with an ultraviolet absorbance detector Scanning electron microscopy with an X-ray fluorescence detector Total antioxidant capacity Ultraviolet Ultraviolet at~sorbance detector

References Bishop, G. J., Journal of Synthetic Lubrication, 1987, Vol. 4, No. 1, p. 25. Jensen, R. K., et al., Journal of Synthetic Lubrication, 1984, Vol. 1, No. 2, p. 91. Zeman, A., et al., Journal of Synthetic Lubrication, 1985, Vol. 2, No. 2, p. 2. Zeman, A., et al., Journal of Synthetic Lubrication, 1986, Vol. 3, No. 4, p. 309 Sniegoski, P. J., Lubrication Engineering, Vol. 41, No. 1, 1985, p. 11. Maier, K., Journal of Synthetic Lubrication, Vol. 3, No. 3, 1986, p. 163. Zeman, A., Schmiertechnik and Tribologie, Vol. 29, 1982, p. 55. Bard, P., et al., Frezenius Zeitschrift Analytische Chemie, Vol. 314, 1983, p. 25. Zeman, A. and Bartl, P., Frezenius Zeitschrift Analytische Chemie, Vol. 316, 1983, p. 43. [10] Zeman, A., et al., Thermochimica Acta, Vol. 80, No. 1, 1984. [11] Centers, P. W. and Smith, H. A., Journal of Synthetic Lubrication, Vol. 1, No. 3, 1984, p. 176. [12] Coates, J. P., "Structural and Chemical Characterization of Neopentyl Polyol Esters by Infrared Spectroscopy," ASLE Transactions, Vol. 9., No. 2, 1985, p. 185. [13] Jones, W. R. Jr. and Morales, W., Lubrication Engineering, Vot. 41, No. 1, 1985, p. 22. [14] Pachuta, D. G., et al., "Modern Methods of Turbine Fluid Analysis: Comparative Study of Several Commercial MiI-L-23699Products," presented at the Aerospace Technology Conference, Society of Automotive Engineers, Long Beach, CA, October 1985. [15] Kauffman, R. E., et al., "Assessment of Remaining Lubricant Lifetime," November 1986, UDRITR-86-14, AFWAL-TR-86-2024, Dayton Ohio Research Institute, Dayton, OH. [16] Pachuta, D. G., et al., "Determination of Total Antioxidant Capacity in MiI-L-23699 Aviation Turbine Oils by a Dynamic Differential Scanning Calorimetric Method," presented at the 26th Eastern Analytical Symposium, New York City, NY, Paper No. 254, September 1987. [17] Eisentraut, K. J., et al., Analytical Chemistry, Vol. 56, No. 9, 1984, p. 1087A. [18] Stecki, J. S. and Kuhnell, B. T., Lubrication Engineering, 1985, Vol. 41, No. 8, p. 485. [19] Rippen, M. E., et al., "Laboratory Spectrometer for Wear Analysis of Engine Lubricants," P & W/ED FR-18951, AFWAL-TR-85-2101United Technologies Corp., Pratt & Whitney Engineering Div., April 1986. [20} Pachuta, D. G., et al., "Experiences with a Sequential ICPES System Applied to a Commercial Flight Evaluation Study," presented at the Pratt & Whitney Spectroscopic Oil Analysis Conference, West Palm Beach Gardens, FL, March 1987. [21] Pragnell, J. W. A. and Von Tilberg, J. E., "Gas Turbine Oil Monitoring Programme," Paper No. 871014, Society of Automotive Engineers, Warrendate, PA, April 1987. [22] Roberton, R. S., Lubrication Engineering, 1986, Vol. 42, No. 8, p. 466. [1] [2] [3] [4] [5] [6] [7] [8] [9]

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep

76

[23] [24] [25]

TURBINE OIL MONITORING

Kight, R., "The History of NOAP in the Military," presented at the Pratt & Whitney Spectroscopic Oil Analysis Conference, West Palm Beach Gardens, FL, March 1987. Suess, R. R., "Spectrometric Oil Analysis--A Current Overview," presented at the Pratt & Whitney Spectroscopic Oil Analysis Conference, West Palm Beach Gardens, FL, March 1987. Humphrey, G. P., "The Development of Water Analysis as a NOAP Diagnostic Tool," presented at the Pratt & Whitney Spectroscopic Oil Analysis Conference, West Palm Beach Gardens, FL, March 1987.

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further reproduct

Copyright by ASTM Int'l (all rights reserved); Sat Sep 5 19:40:48 EDT 2015 Downloaded/printed by Pontificia Universidad Catolica del Peru (Pontificia Universidad Catolica del Peru) pursuant to License Agreement. No further rep