中文摘要：

针对文献[1]和[2]分别采用模糊聚类争神经网络方法进行煤层注水可行性分析过程中共同存在的煤层适应性单一和计算机运算过程复杂问题，提出建立Fihser判别模型进行煤层注水可行性分析的新方法。选取裂隙发育程度、孔隙率、煤层湿润边角、饱和水分增值、坚固性系数、埋藏深度6个因素，以11组煤层注水实溺标准数据作为样本，通过模糊积分的方法建立Fisher判别函数，并通过实测数据验证该方法准确率为87.5％。将建立的模型应用于2个实际注水工程，与实际情况吻合良好。该方法与传统分类方法相比对各种煤层的适应性强，计算机实现简单，更适合用于指导实际煤层注水工程。

英文摘要：

For fuzzy clustering and neural network used to analysis of coal injection feasibility in literature [ 1 ] and [ 2 ], the common problems are single suitability and complex computing. A new method is proposed for coal injection feasibility analysis with Fihser discriminant model. Factors of cracks development degree, porosity, coal bed wet corners, water saturation value, ruggedness coefficient, and buried deep are Selected,with 11 groups of standard measured data as a sample. Fisher discriminant function is established by fuzzy integral method. The accuracy of the method is 87. 5% by actual measuring verification. The model has been applied to two practical water projects, with good effect. Comparing with the traditional classification methods, this method is adaptable to various coal seams, easy to realize, and more suitable for practical guidance of coal injection.