Spatial-Temporal Evolution Prediction of Gas Distribution based on PSO-Elman Neural Network

The rapid development of the industrial economy has increased the use of hazardous chemicals in chemical plants. Real-time monitoring of dangerous gas leakage has become an important approach to ensure the safety of life and property. When a gas leak occurs, it is difficult to estimate the distribution of dangerous gases in space due to the limited detection range of monitoring points, and it is impossible to understand how the gas will evolve in space in the future. In this study, we propose a prediction algorithm for the temporal and spatial evolution of gas concentration distribution. At first, this study uses Elman neural network (ENN) to predict the trend of concentration distribution, and it is optimized by particle swarm optimization (PSO). Then, Inverse distance weighting (IDW) was used to spatially interpolate the gas concentration data predicted by the PSO-Elman to obtain the predicted concentration values of undistributed monitoring points in the experimental area. Simulation validations are conducted on the basis of public data set from Mobile Robotics and Olfaction lab in Orebro University, and the results show that the model established by the PSO-Elman can accurately predicted the trend of gas concentration distribution for a long time. The proposed method can be applied to important fields such as environmental monitoring, chemical leak detection, inspection of landfills and so on.