Inversion Of Stress Field In Fault Area Based On PSO-SVR Algorithm

In situ stress exists in rock mass,which is an important basis for geotechnical engineering design,construction and disaster prevention.With the development and change of engineering construction,the construction of Geotechnical Engineering presents the trend of deepening and complexity.The influence of in-situ stress on engineering stability is more and more obvious,which has become the key factor of engineering safety construction and protection.At the same time,the existence of fault structure area obviously increases the difficulty of obtaining in-situ stress.It is necessary to propose a more perfect and targeted in-situ stress inversion method to obtain accurate and reliable regional in-situ stress data.Based on the regression prediction function of support vector regression algorithm and the global optimization ability of particle swarm optimization algorithm,a nonlinear in-situ stress inversion method based on particle swarm optimization and support vector regression is proposed in this paper.The inversion problem of in-situ stress field in fault structure area is studied,and the feasibility of the method is verified.The main research work of this paper is as follows:(1)Considering the nonlinear correlation between the distribution of in-situ stress field and its influencing factors in structural area,an in-situ stress inversion method based on support vector regression(SVR)algorithm is proposed.The advantages and disadvantages of this method are analyzed.In view of the shortcomings,the global optimization algorithm particle swarm optimization algorithm combined with inversion method is introduced to iteratively optimize the value of model parameters,and an initial in-situ stress field inversion method based on particle swarm optimization algorithm combined with support vector regression algorithm(PSO-SVR)is established.(2)The learning ability of inversion methods is compared and analyzed,and different methods are used to compare and verify the learning ability of the nonlinear mapping relationship between in-situ stress and its influencing factors.Through model sample learning,the effectiveness and superiority of the inversion method are verified,and the learning ability and regression prediction accuracy are obviously better than other methods.(3)Based on the engineering geological data and the field measured in-situ stress value,according to the principle of "inversion forward calculation",the in-situ stress in the fault structure area of coal mine is inversed and calculated by PSO-SVR algorithm.The results show that the error between the in-situ stress inversion value of each measuring point and the field measured value is small,the inversion accuracy is high,the regional stress field distribution is reasonable,and the inversion results are effective and reliable.The in-situ stress distribution in the fault structure area is mainly affected by the formation lithology and tectonism.The in-situ stress values around the fault change significantly,and the in-situ stress difference inside and outside the fault is large.(4)An in-situ stress field inversion method based on particle swarm optimization and support vector regression is proposed and applied to practical working conditions.This method can provide reasonable and reliable in-situ stress data and effective theoretical basis for in-situ stress field inversion and structural area engineering design and construction,and play a certain guiding role.