Subway Crowded and Stampede Neural Networks Safety Assessment Basing on SHEL Model

Abstract

Abnormal crowded subway is extremely easy to cause the stampede accidents. This paper made typical stampede accident analysis and statistics in recent years to conclude the subway crowded stampede influencing factors. Then introducing the SHEL model, centered on human factors it is researched to study the relationship between personnel, hardware, software and environment causing the subway crowded stampede. In view of the characteristics of fuzzy, incomplete and uncertain of analysis object data, it introduces the BP artificial neural network. And an instance subway station is chosen as an analysis example to quantitative study training as well as fitting of crowded stampede process. The analysis results show that there is risk in Liveware–Liveware(L–L) and Liveware–software(L–S), which has been verified in the actual operation of the instance subway station. The crowded shoving, collision, quarrels in the instance subway often occur. But the management measures of the subway station traffic grooming are limited and the grooming effectiveness is little. Therefore, using BP neural network evaluation method to construct the comprehensive assessment system, based on the SHEL model, can quantitatively assess the crowded subway stampede accident risk, so as to promote the safety of metro operation to provide basis for decision making.