Research Of Security State Prediction Of Coal Production Logistics System Based On Improved Particle Swarm Optimization-Support Vector Machine

Coal resource is an important strategic energy resource, which is the lifeblood of economic development and social progress in the next few decades in China. With the increasing demand, the safety of coal mine production is facing tremendous pressure, which is related to the sustainable development of our coal industry and national energy security. The safety of coal mine production logistics system is guarantee and base of the coal production’s normal operation and economic efficiency. Therefore, it is significant to enhance the safety management of coal mine production logistics system, especially to strengthen the effective prediction of the state of mine safety.This paper is aimed to study the mine production logistics system which is characterized by complex nonlinear, real-time dynamic and vague uncertainty. This paper focuses on the view of security state prediction, and proposes a complete index system of the safety status of the coal mine production logistics system, and also proposes a series of corresponding prediction methods. The main contents are as follows.(1) This paper combines the complex nature of mine production logistics system, stands on the view of mid-control and dynamic analysis, is classified into seven categories, i.e. human factors, operational factors, mechanical, electrical factors, transport factors, factors ventilation, drainage and gas factor, so as to analyze the safety of coal production logistics system comprehensively, and explore the mechanism between coal production logistics.system and each factor.(2) This paper proposes a method of building index system of the security status of coal production logistics based on gray theory. Firstly, it establishes the initial index system by the analysis of influencing factors; then it reduces the correlation index by gray cluster analysis; Finally, it filters the optimization index system by gray correlation analysis. Such method can not only avoids the subjectivity of human interference, but also streamlines the index system.(3) This paper aims at the characteristic of complex and data acquisition difficulty of coal mine production logistics system, propose a forecasting model of the security status of the coal production logistics system based on improved particle swarm optimization and support vector machine (IPSO-SVM). Firstly, improve the convergence speed and search accuracy of particle swarm optimization (PSO) by introducing the linear adjustment method of inertia weight and learning weights; then combine with the IPSO and support vector machine (SVM), in order to solve the problem of SVM parameters; finally, test the validity of the model through the case experiment.The results of this paper would not only expand the field of coal production logistics system, enriches the research of coal mine safety management, but also can help to improve the safety situation of coal mine production, and help enterprises improve the safety level of coal production management. So it has a positive meaning for the sustainable development of Chinese coal industry.