The 4th International Symposium on Mine Safety Science and Engineering (ISMSSE2018)

200 / 2018-07-03 15:17:22

Prediction of maximal water bursting discharge from coal seam floor based on multiple nonlinear regression analysis

water inrush from coal seam floor; prediction of maximal water bursting discharge; multiple nonlinear regression modified model; bivariant multiple regression equation; Matlab; SPSS

Mine Dynamics Disaster > Water inrush

Aimed at the influences of various complicated factors for water inrush from floor during coal seam mining, the prediction of the maximal water bursting discharge from coal seam floor is regarded as a problem of pattern identification with nonlinear, high dimensions and finite samples. In this paper, based on the information fusion theory, regarding the prediction of maximal water bursting discharge from coal seam floor as a process of multi-source information fusion and state estimation, establish an evaluation system of main controlling factors for predicting the maximal water bursting discharge from coal seam floor, and finally an evaluation model for forecasting the maximal water bursting discharge from coal seam floor was structured according to multiple nonlinear regression theory. Taking the major coalmines of Xinwen Coalfield as the research background, the paper selects six factors as the basic discriminant factors, including aquifer thickness, unit water inflow, water pressure, aquiclude thickness, fault influencing factor and depth of destroyed coal seam floor caused by rock pressure. Deng's grey relational theory was used to analyze the correlation degrees between each main controlling factor and the maximal water bursting discharge from coal seam floor. Structural equation model was adopted to reveal the internal connecting links among the key influencing factors, determining the weights of each main controlling factor for predicting the maximal water bursting discharge from coal seam floor. Then, the SPSS scatter plot and Matlab functional programming were applied to fitting the correlativity curves between each main controlling factor and the maximal water bursting discharge from coal seam floor in Xinwen Mining Area, and the optimal unitary nonlinear regression models between the maximal water bursting discharge from coal seam floor and each main control factor were established. Ultimately, the multiple nonlinear regression modified model for predicting the maximal water bursting discharge from coal seam floor was acquired by using the multiple nonlinear regression analysis with the combined weights of each main control factor. Compared with the bivariant multiple regression equation and the measured data, the results showed that the average values of forecasting errors predicted by multiple nonlinear regression modified model and bivariant multiple regression equation are respectively 10.53%, 16.76%, indicating that the multiple nonlinear regression modified model and the bivariant multiple regression equation have higher prediction accuracy and relatively smaller error ranges, with comparatively better application value in prediction of maximal water bursting discharge from coal seam floor.