Online Discrimination System for Mine Water Inrush Source Based on PCA and BP Neural Network
Water inrush is one of mine geological disasters to threaten safety mining in china. Water inrush sources recognition is an effective method to predict mine water inrush disaster in time. Compared with overlong processing time using conventional Hydro-chemistry methodology, the paper proposed in-situ mine water sources discrimination model using principal component analysis (PCA) and Back Propagation (BP) neural network based on in-situ water monitoring system. The system is constructed by sensor nodes, an information collector and a ground monitoring center. The measured data were collected from in-situ sensors such as fluorescence spectra, PH value, conductivity, Ca2+，Na+，HCO3- and Cl- ions transducers in different water layers of LiJiaZui Coal Mine in Huainan. PCA is utilized to eliminate correlation and BP neural network is used to recognize mine water sources. The results show that the proposed model achieves 91% accuracy to recognize water sources in mine. Thus, the proposed Model is a rapid and an effective way to recognize mine water sources and further water inrush disaster prediction.