Application of distance discriminant analysis method to predict the fault water inrush

In order to analyze the phenomenon of water inrush from faults theoretically, it is essential to make simplification and takes only the most significant factors related to the problem into the theoretic model, such as existence of water source, water pressure, thickness of block, mining failure thickness of floor, and fault displacement etc. Fifteen representative engineering cases of water inrush from coal floor were taken as training samples. A correspondent forecast model of water inrush quantity was established based on the principle of Mahalanobis distance discriminant analysis theory. The forecast model was applied in five faults of Zhangcun mine. The results show that the carboniferous limestone aquifer causes fault water inrush more easily than the Cambrian limestone aquifer does. And the water inrush level is the maximum when the fault lies in the carboniferous limestone aquifer. Otherwise the bigger the fault displacement, the higher the water inrush quantity is. The prediction results also consistent to the real ones. So the method to predict the fault water inrush proposed in the article is feasible.