Research on fault diagnosis of gas steam boilers based on deep neural networks

Abstract Gas steam boilers are special equipment subjected to high temperature and pressure, it is particularly important detect deal with operation faults in time. Against the 4 types of common fault gas boilers, diagnosis model was established based on deep neural networks (DNN), parameters were optimized verified by experiments. The results show that performs best as Tanh activation function, 100 Batch_Size, Adam Optimizer, 10 -5 learning rate 0.2 Dropout since output accuracy above 92% begins converge iterative times 12. identification reaches 100% verification experiments can still effectively recognize dataset from early stage which suggests has provides a reliable solution for safety supervision equipments.