Comparison between Thermodynamic Model and Neural Network Model Approach

In this work, two different approaches – an iterative thermodynamic method and a neural network model – are proposed and compared for the modeling of steam turbines. The iterative thermodynamic model can predict steam mass flow, pressure, temperature, enthalpy and power on each turbine drum. The NN model can predict the generated mechanical power. Both models have been trained and validated on a massive dataset created through the internal sizing design tool, which contains the turbine geometrical and mechanical data. Further validation tests have been successfully executed by exploiting field data coming from a solar power plant where a high-pressure and a low-pressure turbine were installed.