基于OED-GRA评价模型的采场岩体参数敏感性分析

Sensitivity analysis of rockmass parameters in stope based on the OED-GRA evaluation model

针对岩体参数对采场稳定性的敏感性问题,提出一种正交—灰色关联评价模型(OED-GRA),可分析相互影响的多种岩体参数与采场稳定性之间的关联度。选取Hoek-Brown准则岩体参数中的地质强度指标GSI、岩石单轴抗压强度σci、岩石软硬程度常数mi、岩体扰动系数D等4个因素,采用正交试验方法和数值模拟方法进行16次典型的试验。取顶板沉降位移、顶板水平应力、矿柱水平位移、矿柱竖直应力4个指标进行指标满意度的敏感性分析。基于灰色关联理论,求得各评价指标与岩体性质参数的灰色关联系数,并获得Hoek-Brown准则岩体性质参数与Mohr-Coulumb准则岩体力学参数之间的关联系数。研究结果表明,GSI对采场稳定的敏感性最大,D的敏感度明显最小,σci、mi参数敏感性较GSI次之。采用正交灰色关联评价模型,避免了单因素分析敏感性的局限,同时可分析不同量纲的多因素关联度,结论更为客观、可靠,为多因素关联的敏感性分析提供一种新的思路。

Abstract. In allusion to the sensitivity analysis of rock mass parameters to stope stability, the evaluation model of Orthogonal Experimental Design and Gray Relational Analysis （OED-GRA） is put forward, which can analyze correlation between multiple parameters of rock mass influence each other and stope stability. The 16 typical testing schemes are investigated with the methods of orthogonal experiment and numerical simulation considering four factors, i. e the geological strength index GSI, the rock uniaxial compressive strength σd, the constant mi reflected degree of rock soft and hard, the degree of disturbance to rock mass D in Hoek-Brown strength criterion. The sensibility analysis based on indicators of satisfaction are carried out with four indexes selected, i. e the roof displacement, the roof horizontal stress, the pillar horizontal displacement, the pillar vertical stress. The correlation coefficient between the evaluation index and property parameter of rock mass, and the correlation coefficient between parameters and mechanics parameters are obtained. The results show that the maximal influential degree of rock mass parameters to stope stabilization is the GSI while the minimum is the D, and the GSI more sensitive comparing to the σd and mi. The OED-GRA evaluation model is adopted to avoid the limitations of the single factor analysis sensitivity and multiple factors correlation of different dimensional can be analyzed at the same time, the conclusion is more objective and reliable, which provides a new train of thought for the sensitivity analysis of multiple factors associated.