复杂工程地质体地应力场智能反演

Intelligent inversion method of in-situ stress field for a complicated engineering geological body

摘要:

地应力是地质构造和自重共同作用在地质体内形成的原始应力，是影响工程稳定性和灾变失稳的重要因素.由于受漫长的地质构造作用和地质演化，地应力场随时间和空间变化，由此使准确反演地应力场造成困难.以金川矿区为工程背景，借助地应力测量结果，开展工程地质体的地应力场反演研究.首先，建立矿区工程地质体三维数值模型，并采用正交数值分析和遗传规划算法，建立地应力与岩体参数和侧压系数的函数关系；然后，根据实测的地应力值与计算的地应力值之差平方和最小为优化目标，建立工程地质体的地应力场反演优化模型.采用遗传算法求解，获得矿区岩体参数和侧压系数；最后，将其代入数值模型进行正分析由此获得初始地应力场.通过4个测点地应力测量值与反演值对比分析可知，反演地应力的最大误差为16%，最小误差仅为0.62%.研究结果表明，地应力智能反演方法可用于复杂工程地质体的地应力场反演，且获得的地应力场满足地质工程分析所需要的精度.

Abstract:

The in-situ stress is the original stress in the geological body forming from the geological structure and gravity. It is an important factor which influences the stability and instability of the geological engineering. Due to the long geological tectonic setting and geological evolution, the in-situ stress field is a function of time and space, thus it is very difficulty for us to make accurate inversion the in-situ stress field. Taking Jinchuan mine as the engineering background and with the aid of in-situ stress measurement in Jinchuan mine, the in-situ stress field was inverted. First the 3D numerical mode of geological engineering body in Jinchuan mine was established, and the relation between in-situ stress and rock mass parameters and coefficients of horizontal pressure was obtained by orthogonal numerical analysis and genetic programming(GP). Then the optimization mode was established which takes the sum of squares of differences between the measured initial stress and calculation initial stress up to the minimum value as the objective function. The rock masses parameters and coefficients of horizontal pressure would been obtained by solving the optimization model using genetic algorithm. Finally, the initial stress field can be obtained when the 3D numerical analysis is carried out again by inputting the parameters of rock masses and coefficients of horizontal pressure. Based on comparing inverting in-situ stress with the measured values for the 4 gauging points, the maximum and minimum error of the principal stress is 16% and 0.62% respectively. The results show that the method of intelligent inversion of in-situ stress field can be applied for complicated geological engineering body to simulate in-situ stress field and the accuracy meets the engineering demand.