2017 IEEE 3rd Information Technology and Mechatronics Engineering Conference (ITOEC 2017)

41 / 2017-06-25 15:10:58

Prediction Control of Working Process of Biomass Combustion Boiler based on Multilayer Perceptron Neural Network Model

prediction control; multilayer perceptron; neural network;working process; biomass combustion boiler

The structure of biomass direct fired boiler differs greatly from that of common fuel powder boiler, so the difference of operation process is great, which will inevitably lead to the difference of operation regulation law. Therefore, it is very important to analyze its technological process and combustion process in detail.All data were analyzed by SPSS17.0 for statistical analysis, P =0.05 was considered statistically significant. We used the IBM SPSS Modeler 14.1data software to carry out modeling and prediction. The results show that there are 100 neurons in hidden layer and the area under the curve of ROC is 0.912. The model accuracy, sensitivity and specific is 91.96%, 81.22% and 93.77%. Through validation data set validating, the model accuracy, sensitivity and specific is 92.15%, 80.32% and 94.01%. Therefore working process of biomass combustion boiler could accurately predict by MLP neural network model based on characteristics as the input layer variables of prediction model.