Prediction of the Hydrogen Leak Concentration based on Improved PSO-BP Neural Network

摘要

The change of ambient temperature and humidity will affect the measurement accuracy of hydrogen leakage sensor, resulting in false alarm or missed alarm. Based on the analysis of the influence of environmental temperature and humidity changes on the sensor measurement results, this paper proposes a nonlinear compensation scheme based on improved pso-bp neural network (PSO-BP). We combine PSO algorithm and BP neural network, and optimize pso-bp neural network by adopting the strategy of adjusting relevant parameters according to the changing trend of individual fitness value, aiming at the deficiency of precocious convergence caused by the fixed related parameters of PSO algorithm during the training process of the network. The experiment shows that the improved pso-bp neural network can effectively reduce the influence of environmental temperature and humidity on the sensor, and has a good compensation.