煤层注水效果分析的仿真研究

Simulation of Coal Injection Effect Analysis

研究煤层注水标准优化问题,由于煤层因素复杂,需进行煤层注水效果分析,为确定注水模型参数提供依据。传统煤层注水效果分析模型,由于煤层适应性单一,造成分级误差大的问题。为建立正确模型,提出采用Fisher判别模型进行注水效果分析的新方法。模型选取裂隙发育程度、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组煤层注水实测标准数据作为样本,通过模糊积分的方法建立Fisher判别模型,并对结果处于可注性临界状态的煤层进行BP网络二次修正。经过一组实测数据的仿真,结果表明模型分级准确率为93.8%,并将改进方法应用于2个实际注水工程与实际情况吻合良好,表明提出的模型对各种煤层都有很强的适应性,为煤层注水技术提供了科学依据。

Scientifically analysis of the coal injection effect is an important basis of determining water process parameters,and it is also a core to develop coal mine water policy.Against the common problems of coal seam＇s single adaptation and large classification error in coal injection effect analysis model,the paper proposed a new method of Fisher discriminant model for coal injection effect analysis.It selected six factors of cracks development degree,porosity,coal bed wet corners,water saturation value,ruggedness coefficient,and buried deep,and with 11 groups of standard measured data as a sample,it established Fisher discriminant model by fuzzy integral method.Secondly,it amended the results in critical state with BP network.Through simulation analysis of a group of actual measuring data,the accuracy of the method reached by 93.8%.The model has been applied in two practical water projects,and it has shown good effect.Compared with the traditional classification methods,this method is adaptable to various coal seams,and more suitable for practical guidance of coal injection.