Multi-objective prediction of coal-fired boiler with a deep hybrid neural networks

Abstract

Our works frequently examine the emission of pollutants and the prediction of the thermal efficiency of boilers from power plants. Power plant systems are strongly coupled. Thus, multi-objective modelling and prediction is always a difficult problem. Artificial neural network (ANN) modelling is one of the methods used to meet this challenge. With the increasing requirements of environmental protection, the classical shallow neural network can no longer meet the needs of high precision. In recent years, deep neural networks have gradually demonstrated their powerful capabilities. However, can deep neural networks be used to improve model prediction performance? After many experiments, we successfully construct a sophisticated and stable deep hybrid neural network model to achieve this requirement. The experimental results show that the performance of the hybrid model is superior to that of the classical model; we diagram the detailed structure of the model and provide the corresponding parameter settings in this paper.