Fuel Problems With Marine Gas Turbine Engines

Size: px

Start display at page:

Download "Fuel Problems With Marine Gas Turbine Engines"

Mary Willis
5 years ago
Views:

r 67-GT-5 2 The Society shall not be responsible for statements or opinions advanced in papers or in discussion at meetings of the Society or of its Divisions or Sections, or printed in its

1 r 67-GT-5 2 The Society shall not be responsible for statements or opinions advanced in papers or in discussion at meetings of the Society or of its Divisions or Sections, or printed in its publications. $1.50 PER COPY 750 TO ASME MEMBERS Discussion is printed only if the paper is published in an ASME journal or Proceedings. Released for general publication upon presentation Copyright 1967 by ASME Fuel Problems With Marine Gas Turbine Engines J. P. ATTIANI Department of the Navy, Naval Ship Engineering Center, Philadelphia, Pa. This paper describes a test program undertaken by the Navy to determine the causes of two problems in the fuel-oil systems of gas turbine engines. The first problem concerns short filter life; the second, copper deposits that cause clogging of fuel-oil nozzles. Results are given for operating experience with va-ious filter units from both laboratory tests and fleet experience. The effect of copper deposits on thermal stability is discussed. The author concludes with a review of remedial actions being taken to solve these problems. Contributed by the Gas Turbine Division for presentation at the Gas Turbine Conference and Products Show, Houston, Tex., March 5-9, 1967, of The American Society of Mechanical Engineers. Manuscript received at ASME Headquarters, January 20, Copies will be available until January 1, THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS, UNITED ENGINEERING CENTER, 345 EAST 47th STREET, NEW YORK, N.Y

2 Fuel Problems With Marine Gas Turbine Engines J. P. ATTIANI INTRODUCTION The Naval Ship Engineering Center, Philadelphia Division, is actively engaged in the study, development, test, and evaluation of ship propulsion systems, including boilers, steam turbines, gas turbines, and diesel engines. These assignments encompass new design and modification as well as operational problems encountered with these systems. In addition to and because of our laboratory experience, we have an additional function of, being the first line of recourse for technical assistance to our ships afloat. Through this second responsibility we come in direct contact with operational problems encountered by the Navy wherever our ships are located. This enables us to have an appreciation of conditions peculiar to propulsion systems in shipboard environments. In both our "in-house" test programs and in our Fleet experience we have encountered fuel oil problems. One problem area, that has been given some detailed study recently, has been insufficient filter life. With both coalescent type filters and with standard particulate filters, short life is a problem. A survey program was undertaken to determine the cause of the problem and laboratory studies were made of used filters and fuel oil samples. A problem of a different nature that occurred was one of fuel oil nozzle clogging aboard destroyer escort type ships using gas-turbine-driven emergency generators. These deposits were found to be predominately copper. TYPICAL OF SEPARATOR \ ELEMENT MOUNTING ASS YS \, ''---'% TOTAL OF DOUR N \: T N TYPICAL OF MALESCER ELEMENT MOUNTING ASS'YS TOTAL OF SEVEN FLOAT CONTROL VALVE ASSEVRLY VORTEX RAFFLE FUEL OIL SYSTEMS COMPONENTS In-House Experience The Division was assigned the task of evaluating a Pratt & Whitney FT4A gas turbine engine for marine use and, at that time, to evaluate all necessary shipboard auxiliary equipment and systems. The fuel oil system chosen includes a Bowser-Briggs type coalescent filter, an Andale Strainer, and a Garsite floating suction as part of this auxiliary equipment. The Division has accumulated a total of 3100 hr of operating time with this fuel oil system. Coalescent Filters. The Bowser-Briggs coa- Fig. 1 Bowser Briggs coalescent filter lescent filter, Fig.1, is designed to direct fuel oil through a coalescent element and then through a paper filter separator to remove water and dirt. Particulate matter as small as 0.25 microns in size is separated by this unit. During the first 1083 hr of operation accrued on this system, the coalescent elements were changed a total of 11 times, with intervals between changes ranging from 15 to 265 hr. The fre- 1

3 Location Denmark Sample Source TABLE 1 DATA FROM HMDS PEDER S/CRAM Ship's Main Supply Ship's Day Tank Fuel Oil Azores Auxiliary Supply Coalescent Filters Before Azores Large Coalescer After Azores Large Coalescer Particle Size Distribution ( Microns) 1-5, 5-15, 15-25, 25-50, quent changes were due to the fact that the element would reach the maximum allowable 15-psi pressure-drop limit. It was found also that the coalescent elements were rupturing at a point lower than this limit. At this time the manufacturer was contacted, and his analysis of the used elements revealed that fine particulate matter, suspended in the fuel oil, was being trapped by the outer wraps of the elements, causing undue strain at that point and resulting in rupture. As a result of this experience, the manufacturer has increased the number of spot welds on the outer retaining mesh of the elements. Subsequent experience with the elements has shown that, in order to avoid casing rupture, the pressure drop should still not exceed 15 psi. During this time, replacement elements from another manufacturer were used for evaluation, and although they did not rupture they exceeded their allowable pressure drop of 20 psi at intervals ranging from only 15 to 38 hr. All subsequent operating experience was with Bowser-Briggs replacement units and was essentially unchanged up to the final 664 hr of testing. During this last testing period, the elements Were changed a total of six times. The first four changes were typical, ranging from 49 to 291 hr; however, the fifth replacement ran for a 660-hr period before dirty, water-entrained fuel, caused by low storage barge levels, increased the filter pressure drop to the maximum allowable. The significant increase of life on the 660 hr elements is difficult to specifically explain at this time. There were, however, two items which unquestionably affected the life of the elements. A floating suction was installed in the fuel oil day tank, and the tank was cleaned in conjunction with this installation. During the same time period in which the experience noted in the foregoing was being gained, it was found that the Danish Navy was having similar problems with coalescent filters installed on their gas turbine ship, HMDS Feder Skram. The coalescent filters on this ship are supplied by another manufacturer, but the experience was similar. They have been averaging 55 hr on each filter replacement, with some intervals as short as 12 hr. Used elements were obtained from the Danes, as well as samples of their fuel oil. These items were analyzed and it was found that the coalescent elements contained (approximately) 95 percent 0-5 micron particles and the fuel oil samples contained the same high percentage of fine particles, Table 1. Strainers. In the system, an Andale Strainer is used to remove particulate matter from the fuel oil prior to its entering the coalescent filter. This strainer is a duplex basket-type, rated at 300 gpm. The baskets consist of a 450 x 450 mesh screen with a coarser mesh protective overlay. Testing of this strainer has shown that it will remove 95 percent of all particles above 37 microns. This testing was performed with glass beads, and when their smooth surface is compared with actual field contamination, which is irregular, it can be assumed that its field catch efficiency will exceed the 95 percent of 37-micron particles. This strainer has been used for 3600 hr of in-house service and has performed without problems. The baskets are designed for outside to inside flow, and Division operators developed a bonus feature in that, by connecting the bottom drain to a purifier, the baskets can be backflushed and cleaned effectively without removal of the basket from the casing. Floating Suction. A floating suction system manufactured by the Garsite Company was installed in the fuel oil day tank prior to installing the fifth set of coalescent elements. Following the installation of this system, the improvement in the coalescent element life was noted. The floating suction consisting of a float, a bellmouth suction, and a swing joint arrangement together provide for a continuous low velocity suction near the fuel oil surface. A baffle, placed across the bellmouth inlet, provides for a horizontal flow at the suction. A stop leg limits the height of suction to 18 in. from the tank bottom. Using a floating suction will minimize contamination by water and large particles; however, it is expected that particulate matter in the 1-5 micron size range would continue to be picked up by the suction. Particulate matter in this size range appears to be characteristic of marine diesel fuels,,and settles at an extremely slow rate owing to the viscosity of the marine diesel fuel. The settling rate of the most common suspended particles, 5 microns in size, ranges from 0.69 to 1.43 in. per hour, depending on their density. 2

4 TABLE 2 PARTICULATE DISTRIBUTION - FUEL OIL FILTERS (LOSE) Pa cle Size Distribution Location Sample Source 1-50 (Microns) 5-15,, 15-25p 25-50p Boat A - L.P. Filter Boat A - L.P. Filter Location Source Tank Farm LCSR Boat Fuel Tank MARINE TABLE 3 DIESEL FUEL DATA (LCSR) Particle Size Distribution ( Microns) 1-5p 5-15p 15-25, y Wt. Particulate Matter Mgs/100ML Boat A - M.P. Filter Boat 8 - H.P. Filter Boat C - L.P. Filter LCSR Before LCSR After Boat C - L.P. Filter Boat C - H.P. Filter LCSR Boat Fuel Tank Boat D - H.P. Filter LCSR Before Boat D - H.P. Filter LCSR After As noted in the foregoing, the increase in life of the coalescent filters did coincide with the installation of the floating suction. However, if the author's suspicions are correct that the 0-5 micron particles are the limiting factor in coalescent element life, then the floating suction should not have been a significant factor in the increased life. It was noted, however, that the frequency of Andale Strainer cleanings did decrease after installing the floating suction. This is a reasonable result, since floating suctions should be very effective against the larger particles. Fleet Experience Landing Craft Swimmer Reconnaisance (LCSR) Boats. A recent problem of short fuel-oil-filter life (1 to 3 hr) has been experienced aboard gasturbine-powered boats of the LCSR type. Although the complaints came predominantly from the West Coast, there were indications that it was not a local problem. Used low and high-pressure filters and fuel A- samples were obtained from several boats and from several fuel sources on both the East and s. The filters from these boats were analyzed to determine the size of the particulate matter retained. A summary of these analyses in Table 2 shows that particles below 5 microns constitute percent of all the material present. These filters were rated at 5 micron, nominal; so basically the problem is one in which the filter is performing more efficiently than their nominal rating would indicate they should. Analyses of the fuel oil samples, Table 3, indicates essentially the same particulate distribution as that found on the fuel oil filters. These data also tend to substantiate the opinion that particles in the 1-5 micron range are an inherent property of marine diesel fuel. To alleviate this situation of short filter life, recommendations have been made for (a) the installation of filters with greater surface area, TK Farm A TK Farm B (Before Separator) TK Farm B (After Separator) TK Farm C (b) increasing the filter rating from 5 micron nominal to 10 micron nominal, and (c) installation of a duplex filtering system that would provide for continued engine operation when one unit becomes contaminated. FUEL OIL - COPPER CONTAMINATION Fleet Experience Experience With DE-Type Ships. Another fuel oil problem occurring in the Fleet concerns the gas-turbine-driven emergency generator sets installed on the DE 1006 and DE 1021-class vessels. These units have a continuing problem with fuel oil nozzle clogging, which in turn affects their starting and operating characteristics. Samples of the nozzle clogging deposits from these ships were analyzed and found to contain a high percentage of copper (30 to 50 percent). The source of the copper is not definitely known; however the fuel oil transfer lines are composed of copper-nickel alloys, both 70/30 and 90/10. The effects of copper on thermal stability of jet and marine diesel fuel are well known. A report issued several years ago by Pratt & Whitney Aircraft Company l indicated, in the absence of a metal deactivator, that as little as 125 ppb of copper would affect the thermal stability of marine diesel and JP-5 fuels. In order to more fully under- 1 G. F. Jonke, "Effect of Trace Copper and Lead Contaminants on Fuel Thermal Stability," Pratt & Whitney Aircraft Report FL-65-8, March 29,

5 TABLE 4 EFFECT OF COPPER ON THERMAL STABILITY ence of even trace amounts of copper can act as Fuel Oil Source a catalyst and accelerate the rate of gum and Pressure Drop ** Tube (Hg in.) Rating sediment formation by initiating oxidation and JP-5 Fuel polymerization reactions. Destroyer Escort (400 ppb Cu) Marine Diesel Fuel. Samples of marine diesel Destroyer Escort. 200 ppb Cu * fuel, currently being used in the in-house gas Destroyer Escort 250 ppb Cu * turbine studies, were tested in a manner similar Destroyer Escort ppb Cu* to that used with the JP-5 fuel. These fuels were Destroyer Escort ppb Cu * found to contain metal deactivators. The effect of the metal deactivator is demonstrated in the Marine Diesel Fuel data given in Table 4. Thermal stability is main- Source A NAVSEC Storage Tk (25 ppb Cu) tained even at excessively high levels of copper. Source A NAVSEC Storage Tk ppb Cu * The effect on tube rating may be due to the naph- Source B NAVSEC Storage Tk ppb Cu * thanate present in the synthetic copper additive. Source B NAVSEC Storage Tk ppb Cu * Source A NAVSEC Storage Tk ppb Cu * Source A NAVSEC Storage Tk ppb Cu * Source A NAVSEC Storage Tk ppb Cu * 2/5 2 Source A NAVSEC Storage Tk ppb Cu * 0.9 Source A NAVSEC Storage Tk ppb Cu * Source A NAVSEC Storage Tk * ppb Cu * ** Test Terminated at tip = 25 in. Hg * As Copper Naphthanate stand this problem, a laboratory program was carried out to define (a) The copper threshold of significant thermal stability deterioration in fuel oil currently supplied. (b) The amount of copper that could be expected to be picked up when in contact with copper-bearing piping materials. IN-HOUSE EXPERIENCE Effect of Copper on Thermal Stability JP-5 Fuel Oil. A sample of fuel oil (JP-5) was taken from a destroyer escort encountering nozzle clogging. Thermal stability determinations, as measured by the ASTM procedure D-1660, were performed on the fuel oil as received, as well as after the addition of varying amount of copper in the form of copper naphthanate. The results of these tests are presented in Table 4 and indicate the degree to which thermal stability is affected by the presence of copper. Thermal stability, as defined by the test procedure used, indicates the tendency of a gas turbine fuel oil to form decomposition products in the fuel oil system components. By this test, any gum or sediment formation will cause a pressure drop across a sintered steel filter. The pressure drop is measured in inches of mercury, with a LID of 12 in. being considered the maximum allowable. The pres- 4 4 Copper Pickup in Fuel Transfer Lines The presence of metal deactivators may offset the gum-forming role of copper by complexing the copper with an organic compound. However, there remains the problem of nozzle deposits. The presence of high amounts of copper at this point may be the result of decomposition of the coppermetal deactivator complex. A program is underway to provide quantitative data concerning the rate of and extent of copper pickup from fuel oil transfer lines. By this program, the susceptibility of JP-5 and marine diesel fuel to copper contamination from transfer lines of various copper alloys will be determined. SUMMARY The effectiveness of the floating suction in the in-house fuel oil system is directly dependent on the size of the particles and settling time of the fuel oil. Following the installation of the floating suction in the gas turbine facility, the time period between Andale Strainer back-flushings had increased. Since the Andale Strainer separates essentially all particulates 40 microns and larger from the fuel oil, this suggests that larger particles are being eliminated from the fuel oil system by the floating suction. As previously mentioned, a characteristic of marine diesel fuel is that it contains particulate matter of the 1-5 micron size. We would not expect the floating suction or Andale Strainer to significantly affect this size range of particles. Therefore, the coalescent filter will still be exposed to these smaller particles. The reason for the long life of the coalescent filter during this final 1000-hr test can be attributed to a reduction in the quantity of small particulate matter in the fuel oil. It should be noted, however, that the day tank had been cleaned just prior to experiencing the long period of coa-

6 lescer operation. Thus, a combination of initially clean fuel oil, a rust and scale-free tank, and the addition of a floating suction may have been contributing factors in the increase in filter life. However, in view of our finding regarding the fine particles in marine diesel fuel and their effect on filter life, it would be desirable to have a coalescent filter which would pass 0-5 micron particles. It is recognized that this requires a modification to the present military specification; however the effect of such a change on its water-separating ability is not known. As previously mentioned, the filters aboard the LCSR's were all being plagued with fine contamination and, as Table 2 shows, 95 percent of the particles removed were in the 1-5 micron size range. For this particular installation, the gas turbine manufacturer has stated that the engine required protection from all particles above 10 micron in dia. Based on this, a 10-micron (absolute) filter will be evaluated which would significantly increase filter life aboard those boats. Pending completion of a fuel oil circulation study through piping of various copper alloys, a program is underway to provide immediate remedial action for fuel oil nozzle clogging aboard DEtype ships, which includes a maintenance procedure to insure clean fuel oil systems, a nozzle spray and flow check, and a method for cleaning nozzles for reuse. 5

Hydraulic Starting of Marine Gas Turbine Engines

Hydraulic Starting of Marine Gas Turbine Engines 68-GT-29 The Society shall not be responsible for statements or opinions advanced in papers or in discussion at meetings of the Society or of its Divisions or Sections, or printed in its publications.