A Nonlinear Risk Evaluation Method for Water Inrush Through the Seam Floor

A risk evaluation method based on principal component analysis, fuzzy mathematics, particle swarm optimization, and support vector classification, is proposed due to the limitations of the water inrush coefficient method. The many factors that influence coal floor water inrush were categorized as structural, hydrogeological, coal seam occurrence and mining condition principal components, according to principal component analysis, among which the structural and mining condition components were the most important. The principal component factors were then fuzzified. Finally, an evaluation method was established, with the parameters of the supporting vector classification optimized through the particle swarm algorithm. The validity of the method was demonstrated by comparing its predictions with actual conditions at ten coal seam mining faces.