Combined inversion algorithm of coal seam stress field based on wavelet neural network and gradual loading algorithm

摘要

Joints extension change the physical and mechanical properties of coal due to in-site stress, then the redistribution of coal seam stress and deformation is caused, so the calculation method of coal element joints extension is given, and relationship between joints extension and stress is analyzed. Appling wavelet neural network and gradual loading algorithm, a combined inversion algorithm of coal seam stress field is presented. According to the element transition principle, the computer element program is corresponding developed, and the distribution regularity of in-site field is studied for JinCheng region in QinShui basin using combined inversion algorithm. The results show that the maximum principal stress increases from northwest to southeast in the region, and it is among 8–13Mpa; the minimum principal stress is located in Southern JinCheng, and the value is about 5–9MPa. Compared combined inversion algorithm results with measured values for the key points, the principal stress relative error is only 3%, and azimuth error is less than 1.3°. So combined inversion algorithm has higher calculation accuracy and can satisfy the engineering demand, also it has a series of advantages, such as simple model, fast calculating speed, and so on.