基于遗传规划的岩石流变模型辨识

将遗传规划方法用于岩体流变模型辨识中,该方法不需要事先假定流变模型的形式,可以一次性地同时确定流变模型的结构形式及参数,避免了模型假定所带来的误差.采用Matlab语言编制的遗传规划程序对红板岩的蠕变实验进行了研究,得出了常应力水平下,应变与时间的非线性关系,并将其结果与基于最小二乘的西原体蠕变模型参数识别的结果进行了对比,结果表明:遗传规划用于岩石的流变模型辨识是可行的.

Identification of Rheological Constitutive Model of Rock Based on Genetic Programming

Genetic programming was adopted to identify the rheological constitutive model of rock,genetic programming's such merits as simultaneous determining the structure and parameters of rheological constitutive model without presumption of rheological constitutive model can avoid the error from the assumption of model efficiently.The creep test of red slate rock materials were studied by means of genetic programming in Matlab language,the nonlinear relation between strain and time under constant stress level were obtained,whose prediction results were compared with that of traditional method for the identification of creep model parameters,the results show that it is feasible to apply the genetic programming to the identification of rheological constitutive model of rock.