基于环套原理的ANN型矿井小构造预测方法与应用

应用“多重环套理论”和人工神经网络（ANN）技术，提出了预测预报矿井小构造的一套完整理论体系和工作方法。以淄博岭子煤矿为例，优选确定煤层倾角、厚度、瓦斯聚集量和涌水量变化是主要控制岭子煤矿回采工作面前方小构造的四大因子，建立了1号井的回采工作面前方小构造预测的非线性模型。现场检验结果表明，在煤层回采过程中，只要精细掌控主控因子的大量信息变化，就能准确预测预报工作面回采前方的小构造展布特征。基于环套原理的ANN型矿井小构造预测方法解决了诱发矿井重大灾害事故小构造的预测预报难题。

Methodology and application on size-limited structure predictions with ANN based on “Loop Overlapping Theory”

Abstract: Based on “Loop Overlapping Theory” and the artificial neural networks (ANN), a set of complete theory system and work method were put forward to forecast the size-limited structure. Take the Lingzi coal mine in Zibo for example, the four master factors, that is the coal seam dip angle, thickness, the gas accumulation quantity and the gushing water volume change, were selected. These factors control the size-limited structures of the front working face in Lingzi coal mine. Then, the nonlinearity model was established to predict the size-limited structures in the front coal mine working face in the first well. The field test result indicates that if the massive information of the master factors can be mastered in coal exploitation, the distributing feature of the size-limited structure can be predicted accurately in the front face mining. The technique of ANN based on “Loop Overlapping Theory” solves the difficult problem in predicting the size-limited structure which induced to great mine accident. That has the important theory guiding sense and the practical value to the mine safety production.