Safety Risk Early-warning System Based On Data Mining

The thesis is based on the project of"Qinxin safety risk early-warning management system", which is part of great project of the safety management and culture construction of Qinxin Group in Shanxi. Some support technologies related risk early-warning system are introduced, including safety risk early-warning management theory and data mining technologies. The early-warning rules are derived from the associated rules of data mining. System model based on the theory of safety risk early-warning management and association rule of data mining are established. And further more, the Qinxin safety early-warning management system software are developed according to this model.Considering the real situation of the coal industry, the construction principles, system architecture, interfaces, databases and program modules of the risk early-warning system are designed based on the theories of above mentioned. The combined early-warning system of job-sheet and period on the basis of association rules of data mining are emphasized. Job-sheet warning can uncover and alert the hidden dangerous factors in the production process, and thus can correct or prevent unsafe behaviors timely and effectively. Therefore, the accidents caused by miss-operation and unreasonable management are avoided.The risk early-warning system of Qinxin is supported by two basic databases, risk factor database and hazard database. It's also subdivided six professional areas: coal mines, coal preparation, coke, alumina, calcium carbide. On the other hands, the hazards in Qinxin Group are grouped into 10 types, they are: heavy machinery, boilers, pipes, tailings, hazardous substance reservoirs, power distribution equipments, upgrading system, electromechanical equipments, transportation systems and high temperature and pressure equipments.The system has the functions of analyzing and evaluating the risk in the six professional areas, and ability of early control to the 10 kinds of hazards. It's been proved that the system is effective in the management of accident, predicting and disaster control.