基于多元逐步线性回归的煤矿瓦斯涌出量影响因素研究

Research on Influencing Factors of Coal Mine Gas Emissions Based on Multiple Stepwise Linear Regression

针对煤矿瓦斯涌出影响因素较多,难以准确分析各影响因素的相关性,以及在实际生产作业中瓦斯涌出量难以进行准确预测的问题,采用双变量相关性分析与多元逐步线性回归相结合的方法对此进行了研究。首先,对可能影响瓦斯涌出量的影响因素进行相关性假定,并利用双变量相关性分析确定主要影响因素。其次,将满足条件的影响因素利用多元逐步线性回归方法建立了瓦斯涌出量预测模型,模型的模拟精度为0.954 7,对比实际值和预测值计算得到均方根误差RMSE=0.056。最后,利用canoco5分析了瓦斯涌出量与各个影响因素之间夹角关系,对瓦斯涌出量与各因素之间的正负相关性进行了验证,为煤矿瓦斯涌出量研究提供了相关依据。

Aiming at the problems that there are many influencing factors of coal mine gas emission, it is difficult to accurately analyze the correlation of various influencing factors, and it is difficult to accurately predict the gas emission in the actual production operation, studies this problem by using the combination of bivariate correlation analysis and multiple stepwise linear regression. Firstly, the correlation assumption is made for the influencing factors that may affect the gas emission, and the main influencing factors are determined by bivariate correlation analysis. Secondly, the prediction model of gas emission is established by using multiple stepwise linear regression method. The simulation accuracy of the model is 0.954 7. Compared with the actual value and the predicted value,the root mean square error RMSE is 0.056. Finally, the angle relationship between gas emission and various influencing factors is analyzed by using canoco5, and the positive and negative correlation between gas emission and various factors is verified, which provides a relevant basis for the study of coal mine gas emission.