No description available.

Results are reported from a series of experiments simulating the combustion and expansion processes of a heavy-duty combustion turbine engine burning a heavy residual fuel oil. The tests were carried out in a turbine simulator device, consisting of a combustion chamber and a turbine first-stage nozzle cascade sector. Both film, air-cooled and closed-circuit, water-cooled nozzle sectors were tested. These sectors were four-vane, three-throat sections with throat cross-sectional areas of approximately 50 (10⁻⁴) m². The test fuel was simulated by adding the appropriate contaminants to no. 2 fuel oil. A series of seven full-length tests were performed, ranging in length from 22.5 to 88.2 hours. Four of the tests involved the watercooled nozzle sector and the remaining three used the air-cooled nozzle. The principle objectives of the tests were to assess the rate at which ash accumulates in the turbine nozzle and the relative difficulty in removing these deposits. The variable used to evaluate the extent of the ash deposit on the nozzle was the effective throat area, determined using the calculated gas flow rates, turbine nozzle inlet temperature, and the measured combustion chamber pressure. The parameters varied in the test program, other than the nozzle sectors, were the gas temperature and the gas pressure. The gas pressure variations served to vary the gas path surface temperatures at constant gas temperature. The test conditions were nominal turbine firing (nozzle exit) temperatures of 1283 and 1394 K and combustor pressures of 3 and 6 atmospheres. A 2-to-l pressure ratio was maintained across the nozzle to insure sonic conditions at the throat sections. With the exception of one test, the data show that the deposit rates in the water-cooled turbine nozzle were lower than in the air-cooled nozzle. The effect of increasing the gas temperature was to dramatically increase the ash deposition rates. Decreased gas pressures (and hence surface temperatures) resulted in reduced deposition rates. Ash cleanability was enhanced by water-cooling. Heat transfer data were analyzed from the water-cooled tests and gave significant insight into the ash deposit formation and removal phenomena. One of the more significant conclusions drawn from these data was that the major portion of the effective area decrease observed in a turbine nozzle because of ash deposits is due to the pressure face deposits. A computer simulation of a combustion turbine engine was developed to aid in the evaluation of the turbine simulator test data. Results from field tests of full-sized production engines burning residual oil were used in the simulation to determine the relationship between the extent of ash deposition (throat area reduction) and turbine efficiency. This result was then combined with data from the turbine simulator tests to produce a real-time computer simulation of full-sized combustion turbine engines having air- and water-cooled first-stage turbine nozzles. It was found that water-cooling of the turbine nozzle would result in an increase in engine availability of 27 per cent when operating on heavy residual fuel oil. / Ph. D.

M.Ing. / This dissertation deals with the structural integrity analysis of a low pressure (LP) turbine with transverse cracking. Cracks are initiated in the centre ring keyway of the shaft by a fretting mechanism and propagate during barring operation. The cracks arrest during normal operation when they are still relatively shallow. The aim of the calculations is to predict the transition where cracks start propagating by high cycle fatigue during normal operation of the shaft. Most influencing factors, including continuum mechanics, mathematical modelling, material behaviour, service loads and industry experience, are studied in detail as a precursor to the case study. The calculated results show that the case study is an example where the application of fracture mechanics on the crack-free stress field leads to erroneous results. There is a significant redistribution of stress in the presence of the crack so that the actual stress ratio, as calculated from three dimensional cracked models, varies significantly compared to the crack-free model. Calculated results, together with carefully researched material properties, confirm the postulated crack growth during barring operation and predict limiting crack sizes where high cycle fatigue would ensue during normal operation. The case study concludes that the shaft has a significant remaining life and that it can be returned to service with periodic non-destructive examinations.

Silicon carbide and silicon nitride turbine engine components are susceptible to hot corrosion by molten sodium sulfate salts which are formed from impurities in the engine's fuel and air intake. Several oxide materials were identified which may be able to protect these components from corrosion and preserve their structural properties. Ta20, coatings were identified as one of the most promising candidates. Thermochemical calculations showed that the chemical vapor deposition(CVD) of tantalum oxide from O2 and TaCI5 precursors is thermodynamically feasible over a range of pressures, temperatures, and reactant concentrations. The deposition of Ta205, as a single phase is predicted in regions of excess oxygen, where the reaction is predicted to yield nearly 100% efficiency. CVD experiments were carried out to deposit tantalum oxide films onto SiC substrates. Depending on the deposition conditions, a variety of coating morphologies have been produced, and conditions have been identified which produce dense, continuous Ta205 deposits. Preliminary corrosion tests on these coatings showed no apparent degradation of the CVD deposited tantalum oxide coatings. The feasibility of depositing ZrTi04 as a coating material was also investigated based on thermochemical considerations. Since no data were available for this material, thermodynamic values were estimated. Thermochemical calculations indicated the chemical vapor deposition of zirconium titanate from O2, ZrCl4, and TiCl4 occurs over a range of temperatures in a very narrow region of the phase diagram. Deviations from the single phase region predicted the codeposition of either Zr02 or Ti02 with ZrTi04. These results suggested that the chemical vapor deposition of ZrTi04 may be difficult from a process handling perspective. Additionally, the process is predicted to be very inefficient, leaving substantial amounts of unreacted chlorides in the reactor exhaust. / Master of Science