Document G5Yag4GZeop1z8g8KRJ5zkQ0q

GAS TURBINE ON TEST STAND (I to r) dc generators, reducing gear, olr Wet, compressor, combustors, gas turbine, I 1350-F 2000-Hp Gas Turbine Runs | On Test Stand for Over 1000 Hours ^ A ykar aco Westinghouse announced an experimental 2000-hp land gaa tur bine, (Powen, Jan 1947, pp 98-9,102). Testa run since, while incomplete, have gone far enough to confirm the hopes of its originators. Subsequent experi ence gives substantial evidence of eventual success (or industrial and transportation use. In more than 1000 hr of tests, per. formsnee has been essentially in accord with design predictions. Operation under all types of loading and up to the design temperature has caused no objectionable distortion and no serious creepage. Difficulties have not been fundamental and can be readily cor rected in new designs. On the whole, experimental evidence points to the soundness of the general design. . Tati Results. Hie turbine, on test since August 1, 1946, has operated about 1000 hr, of which more then 850 hr have accumulated aince July 9,1947. Evaluating performance of the unit and its components took 300 hr. Remolnder of the time ws used in simulating the more aevere operating cyclea expected in actual service. Accurate over-all performance has been established by reliable measure ments of power output, fuel Bow, speed, air-inlet temperature and atmospheric pressure. Over-all fuel rate at full load of 0.78 lb per bhphr corresponds to s Experimental unit generates full load at 16.7% thermal eflV clency with 19,500-Bto-per-lb fuel oil. Maximum output dol veloped equalled 2220 hp with 48-F inlet air to the compressor By T i PUTX, Manager lotonetlrv end Gat-Tutbini Engineering Section Wett/ngbeuM ffoctric Corporation i thermal efficiency of 16.7%, baaed on the fuel having a heating value of 19,500 Btu per lb. Maximum output developed has been 2220 hp when operating with an air-inlet temperature of 48 P. Test Procedure. Evaluation of com ponent performance has been difficult. Many changes in instrumentation and laborious heat-balance calculations were made to achieve the desired results. Compressor performance was estab lished by measuring air flow, inlet and discharge pressures and tempereturo rise. AdiabsUc compression efficiency varied from 80 to 86% over the entire speed and load range. To determine turbine efficiency In a complete plant without a dynamometer between compressor and turbine re quires accurate measurement of inlet end exhaust temperatures and pres sures. Of these, temperature of the combustion gea as it enters the torbiss Is particularly difficult to tfieasoi* Three methods were used; (l) dired measurement, using specially designed,' shleldcd-type temperature probes (1). calculation, taking combustion effid' ency os 95%, measured air and fud' flow, measured combustion inlet te\ peraiure and neglecting all losses (31. calculation, using the measured totblMj exhaust pressure and temperature^ measured Inlet pressure and tarblas work being determined by beai-balan<a ; calculation. Turbine efficiency, obtained by uslal the temperature recorded by direct measurement, gave least reliable resold while the second and third method* were quite consistent and fn close agre* menu Turbine efficiency varied fro*1 84 to 86% over the operating range-.1 This is about 2 points lower than tbs* obtained with earlier test result* 80 (150] POWER More* b erased by the increased radial Up fU.Kwrg found necessary for rapid (fciiiei in loading. Cenbastlon efficiency, using specially fcjjjped air-atomising nozzles, found tyhest-bstance calculations, varied bointo 94 to 96%. These values agree doth with those of combustion tests n ecr research laboratories. loading Is continuous. This corresponds to rapid temperature changes of 600 to 700 F on turbine and combustor, tur bine-inlet temperature being 1350 F at full load and 600 to 750 F at no load. To simulate locomotive operation, the unit ran at full load 30 min, then was Immediately unloaded and run 30 mht at no load, whereupon load was re applied in 10 to 20 sec and the cycle repeated. This cycle was then changed to limit the loaded and unloaded time to 10 min instead of 30. To accelerate the program, a further change to 5 min was made when testa established that this time was sufficient to heat and cool parts subjected to rapid tempera, ture variation. Chart at left shows typical load cyclo run. Verting. The nnU starts easily, one generator being used os a motor. Time required is a function of starting power available. When this power is limited to a maximum of 35 kw, the unit coo start In about 2Vh <nln. With a maxi mum of 80-kw starting power, it runs In 1 min. With 20 kw the time U 8 min. When the rotor reaches 15% peed the acetylene igniters are turned on and at 2S% speed the fuel is re jected. Starting power is shut of! at the Operating Cycles. The unit has been started from a cold standatill condition 550 times and has undergone several thousand rapid load cycle changes from end of 1% min end the unit reaches a stable self-sustaining speed In about 2>b min. A gas turbine o! this type can operate at full capacity 10 min from no load to full load. Loading and un the time otarting Is initiated, or even loading cycle testa have proved its loadresponse characteristics. Probably the most severe load oyele will bo in loco motive service where loading and un less if necessary. Troubles end Remedies. During the early teat weeks the eompreaser Inlet and exhaust duets were equipped with March 1948 (151| 81