Fault Diagnosis Strategy Design and Simulation for an Automatic Gas Supply System

In this paper, a fault diagnosis probabilistic neural network (PNN) for an automatic gas supply system is proposed. An adaptive fuzzification method, reducing the scale of training data, and a simplified radial Gauss function, accelerating the identification process, are used in the network. The network maximum error of the training data is less than 10
⁻¹⁰
which is much lower than the unimproved network. Based on the fault diagnosis PNN, a joint simulation model combining the modeling and analysis advantages of AMESim and Simulink is built to test the performance of the network. The simulation results show that the improved PNN can identify failure mode efficiently and accurately when faults occur.