The objective of this project is to evolve design guidelines based on considerations of deposition and corrosion, for future operation of large stationary combustion turbines for electric power generation. The primary emphasis of the investigation is to establish the operating blade surface temperature, considering all aspects of deposition and corrosion, for a range of fuels including residual oil and coal-derived liquids. The design guidelines that will evolve from this program will provide a data base to determine the trade-offs between metal temperature, performance, maintenance and reliability. There is, at present, some operating and lab experience that indicates the possibility of serious deposition problems as turbine inlet temperature (TlT) increases, and as the differential between TlT and blade surface temperature increases. The concern regarding corrosion effects is centered on two considerations: a) Will metal surface temperature continue to govern the corrosion process as TlT increases, and b) Will present-day additives be effective at higher in-let temperatures and cycle pressures? Evaluation of the initial test results is presented.

Burning residual oil in utility combustion turbines and the consequent deposition on blades and vanes may adversely affect reliability and operation. Corrosion and deposition data for combustion turbine materials have been obtained through dynamic testing in pressurized passages. The deposition produced by the 1900°F (1038°C) combustion gases from a simulated and a real residual oil on cooled Udimet 500 surfaces is described. Higher deposition rates for the doped fuel than for the real residual oil raised questions of whether true simulation with this approach can be achieved. Particles 4–8 μm in diameter predominated in the gas stream, with some fraction in the 0.1–12 μm range. Deposition rates seemed to be influenced by thermophoretic delivery of small molten particles, tentatively identified as magnesium pyro and metavanadates and free vanadium pentoxide, which may act to bond the larger solid particles arriving by inertial impaction to turbine surfaces. Estimated maintenance intervals for current utility turbines operating with washed and treated residual oil agreed well with field experience.

A number of commercially available chemical additives for utilization in modern high temperature gas turbines were evaluated for their corrosion and deposit inhibiting characteristics. The evaluation was made in pressurized passages which simulate conditions of operating gas turbines. Initial screening tests of 10 hr each were made with a contaminant level of 50 ppm vanadium, and promising additives were further tested for times up to 300 hr at the more realistic contaminant level of a No. 3 GT fuel. Characteristics of the deposits were studied with x-ray diffraction and ash fusion tests. Corrosion rates for various alloys tested with treated fuel were measured and compared to test results obtained with untreated fuel. Conclusions were drawn as to the type of elements most effective in reducing corrosion caused by combinations of vanadium and sodium. The role of elements in preventing deposit buildup is also assessed.

Analysis of loss of coolant accident (LOCA) scenarios in nuclear reactor safety evaluation depends on knowledge of many complex phenomena. A primary phenomenon controlling the sequence of events, by determining the residual coolant mass inventory within the primary system, is the critical flow process. Critical flow of a flashing liquid is complicated by marked departure from thermal equilibrium. Several complex models have been proposed to represent the non-equilibrium effects, including six-equation two-fluid models. Amos and Schrock (1983) developed a model based on the premise that the two-phase region is homogeneous and that thermal nonequilibrium is the dominant factor causing the departure from the homogeneous equilibrium idealization. Flashing inception was represented by a modification of the Alamgir-Lienhard (1981) pressure undershoot. Exponential relaxation of the metastable liquid was formulated as suggested by Bauer et al. (1976) and the critical flow criterion used the sound speed formulation of Kroeger (1976). Lee and Schrock (1988) extended the Amos-Schrock work by developing an improved correlation for the pressure undershoot correction factor in terms of Reynolds number and subcooling Jakob number. Improvements were also made in the relaxation constant and in the application of Kroeger’s formulation. In the present paper a new cavity flooding model is used for the evaluation of pressure undershoot at flashing inception. This model is similar to the one developed by Fabic (1964) for the evaluation of liquid superheat required for boiling on a surface subjected to transient heating. The model contains an experimentally deduced factor, which is correlated against stagnation subcooling using the experimental data of Amos and Schrock (1983, 1984), Jeandey et al. (1981), and the Marviken tests (Anon., 1979). The model was then tested against seven additional data sets and shown to be very accurate in predicted mass flux (standard deviation of 10.9 percent for all data). The cavity flooding model is thought to represent the true physics more correctly than does the earlier model, which had its origin in molecular fluctuation theory.

Effects of temperature and contaminant levels on the high-temperature corrosion of superalloys used in gas turbines were investigated using pressurized passages which simulate the operating conditions of present-day gas turbines. The alloys were tested in a cooled configuration realistically simulating the air-cooled vanes and blades of a gas turbine. Conclusions are drawn as to the permissible level of contaminants and the effect of metal cooling on high-temperature corrosion. It is shown that the surface temperature of a blade or vane rather than the gas-stream temperature is the critical factor in determining the amount of attack to be expected at a given contaminant level and the amount of attack is an exponential function of this temperature. Furthermore, in a dynamic-type test no decrease in corrosion rate is noted at higher temperatures. It was concluded that the use of a 5 ppm Na/2 ppm V fuel would result in an excessive amount of attack with a metal surface temperature of 1500 deg F.