回采工作面瓦斯涌出量的聚类分析及归类预测模型的建立

Cluster Analysis of the Gas Emission from Coal Face and Establishment of Its Categorization Prediction Model

为了预防瓦斯灾害的发生,以实现回采工作面瓦斯涌出量的准确预测,在对回采工作面瓦斯涌出量实测数据统计的基础上,应用模糊C均值聚类算法进行回采工作面瓦斯涌出量的聚类分析,得出各数据的聚类中心和样本数据对于分类的隶属度;建立回采工作面瓦斯涌出量的归类预测模型,对待测瓦斯涌出样本进行预测,实现了瓦斯涌出量的归类预测;最后,用实例论证了该方法的可行性和有效性.研究结果表明:隶属度表征样本属于各个类别的程度,同时也证明了瓦斯涌出量与各个影响因素之间的关系是非线性的;采用归类预测模型对待测样本进行归类预测,通过比较样本与聚类中心的关联度大小,判定样本的归属类别,避免了人为的主观性和盲目性;该方法可行、实用,是一种有效的瓦斯涌出量归类预测方法.

Based on the statistical data of gas emission from coal faces, the gas emission data are analyzed by using the fuzzy C-mean cluster algorithm, and the cluster centers and the membership degrees of samples are obtained. The categorization prediction model for gas emission from the coal face is established, and is applied to the prediction of the gas emission. The categorization prediction of gas emission is achieved based on the combination of the fuzzy C-mean cluster algorithm with the categorization prediction model. Finally, the method is applied to the practical example. Practical application demonstrates that the relations of the gas emission and its influence factors are nonlinear; when the gas emission samples are predicted by using categorization prediction model, the adscription of each sample is determined by comparing the related degrees. As a result, the artificial subjectivity and blindness is avoided; the method has a good feasibility and a preferable practicability, thus it is an efficient categorization prediction method for gas emission.