基于模糊模式识别的煤与瓦斯突出区域预测

Regional forecast of coal and gas burst based on fuzzy pattern recognition

借助于成熟的模糊理论与技术可以较客观地实现突出预测中不精确信息与不精确关系的正确表达与处理,综合考虑突出影响因素,提出了煤与瓦斯突出区域预测的模糊模式识别新方法.首先采用模糊聚类分析对煤与瓦斯突出的样本集合进行分类,建立不同程度的模糊模式.然后对待预测样本进行模糊模式识别,根据择近原则,哪一种模式与预测样本最接近,待预测样本即属于哪一种突出类型.此方法克服了模糊聚类单一分析方法的不确定性,实现了多指标定量化的预测,提高了预测效果的准确程度.通过实例验证,证明了预测的可靠性.

Recurring to well-rounded fuzzy theory and technology can realize exact expression and manage between imprecise information and imprecise relation on outburst prediction. Advanced a new method based on fuzzy pattern recognition to forecast coal and gas burst by considering outburst-factors comprehensively. Firstly, the simple integration was classified for the coal and gas outburst by adopting clustering analysis method. The fuzzy model was set up for different outburst degree. Afterwards, the danger degree of the undetermined forecasting sample was predicted by applying fuzzy pattern recognition. In principle of choosing nearest, which model was most similar to predict sample, so the predict sample was that outburst model. This method can overcome the indeterminateness and realize quantitative forecast including many indexes. The accuracy of forecast was improved. Through actual validation, the reliability of the prediction was tested and verified.