Mathematical Model for an Electrode-Type Electric Boiler Based on GSA-optimized Neural Network

Taking an electrode-type electric boiler as the object, the factors influencing the thermal load of the electric boiler are analyzed. A prediction model for the thermal load is established with neural network, and the model parameters are optimized by Gravitational Search Algorithm (GSA)to improve the convergence performance of its training process. The results show that, the neural network model can fit the nonlinear characteristics among input and output variables, and GSA significantly improves the training speed of the neural network model. The model can predict the electric boiler thermal load accurately under various conditions. It can also participate in peak regulation and frequency regulation of power grid combining with thermal power units, which can improve the operational flexibility of thermal power unit.