Performance evaluation of reaction and impulse turbines using ANN, ANFIS, and ANN-PSO models in MATLAB

Turbomachines are categorized as either reaction or impulse machines based on the proportion of dynamic and static heads involved in energy transfer. Reaction turbines are characterized by the passage of the working gas through a set of stationary blades, referred to as stator blades. These stator blades serve to control the uid's movement and enhance its velocity. On the other hand, impulse turbines feature a set of xed nozzles through which the working uid ows. These nozzles convert the pressure energy of the working medium into kinetic energy. Reaction turbines operate at high temperatures and pressures, constituting oil-free exhaust systems. They are typically employed in applications where such conditions are essential. Impulse turbines, on the other hand, are designed for extracting energy from swiftly moving uids, making them particularly well-suited for hydroelectric power generation. Both reaction and impulse turbines have undergone extensive experimental research. A dataset comprising one hundred and ninety-seven (197) data points was generated from these turbine machines. Of these, one hundred and twenty (120) were derived from reaction turbines, while seventy-seven (77) were obtained from impulse turbines. The datasets were employed to develop models using ANN, ANFIS, and ANN-PSO within the MATLAB environment. For the reaction turbine, the ANN-PSO model achieved the highest regression test value (R²) of 0.9972. On the other hand, for the impulse turbine, the ANN model attained the highest regression value, with an R² of 0.99973.