Research on Early Warning of Security Risk of Hazardous Chemicals Storage Based on BP-PSO

Abstract

The storage and transportation of hazardous chemicals has always been a serious problem. In this paper a scheme for the safety monitoring and management of hazardous chemicals storage is proposed. Multiple sensors are used for real-time detection of external and internal environmental, and the data is transferred to the cloud. Further more the data processing in the cloud. Even if the corresponding early warning is made, it will help the storage security management of hazardous chemicals. For data in the cloud, BP neural network, and multi-particle swarm algorithm (PSO) are used to optimize the weights and thresholds of the network, and to avoid the shortcomings of using BP network alone, converge faster, and obtain the global optimal solution. Through the simulation study of hydrogen storage tank data, the results show that the results predicted by the BP-PSO neural network algorithm are more consistent with the actual situation, and are significantly better than the BP network results. This model can accurately make early warnings of hazardous chemicals and promptly help operators analyze existing security risks, which has a high reference value.