Off-Design Performance of 9F Gas Turbine Based on gPROMs and BP Neural Network Model

Gas turbines are increasingly and widely used, whose research and production reflect a country’s industrial capacity and level. Due to the changeable working environment, gas turbines usually work under the condition of simultaneous changes of ambient temperature, load and fuel. However, the current researches mainly focus on the change in single condition, and do not fully consider the simultaneous change in different conditions. On the basis of single condition, this paper further studies the dual off-design performance of gas turbines under three conditions: temperature-load, fuel-load and fuel-temperature. Firstly, the whole machine model of a gas turbine is established, in which the compressor model has the greatest impact on the performance of gas turbines. Therefore, this paper obtains a more accurate compressor model by combining the engineering modeling advantages of gPROMs and the powerful mathematical calculation ability of MATLAB neural network. Then, according to the established gas turbine model, the dual off-design performance is studied, which is mainly based on the parameter of output and efficiency. The result shows that the efficiency and power output of gas turbines will decrease with the increase of ambient temperature. With the decrease of fuel calorific value, power output and efficiency will increase. As the load decreases, the efficiency of the gas turbines will decrease, and these changes are consistent with the single off-design performance. However, when the fuel and temperature change simultaneously, only adjusting the IGV angle cannot avoid the surge when the temperature is above 30°C. At this time, it is necessary to adjust the extraction rate in order to ensure the safe and stable operation of gas turbines. Therefore, the research on dual off-design performance of gas turbines has an important significance for the peak shaving operation of gas turbines.