以某水电站地下厂房区软岩部位试验洞监测资料为基础，运用基于均匀设计-遗传算法－神经网络的人工智能反演分析方法，在考虑了仪器埋设滞后的位移损失的情况下，对岩体在宏观结构条件下的流变参数进行三维粘弹塑性位移反分析，得到了岩体的流变参数并进行了位移预测，与监测成果进行了对比分析。结果表明：预测围岩流变位移、速率与监测结果比较接近，规律性较好，说明辨识效果较好。

Parameters identification of rheological constitutive model is an important content of geological rheology. On the basis of the in-situ measuring displacement information of the test cavity surrounding rock of some hydroelectric station underground powerhouse, rockmass rheological parameters are identified through 3D viscoelastic-plastic numerical analysis and intelligence method that include uniform design (UD) ，BP artificial neural network (BP-ANN) and genetic algorithm (GA). And then the surrounding rock mass rheological displacements as a result of the excavation of test cavity are forecasted. In comparison with the monitoring displacements, the forecasting displacements and velocities are approximate to the measured displacements and velocities, which prove the rightness and reliability of the above method.