高心墙土石坝坝料流变模型参数的智能识别研究

通过在遗传算法中引入位爬山法，有效地改善了遗传算法的局部优化能力，提高了其优化速度与精度。采用这一改进的遗传算法对糯扎渡水电站心墙土石坝坝料流变模型参数进行了大范围识别，实现了参数空间内全局意义上的快速搜索。对取自不同坑号的各种坝料进行了流变参数识别，识别结果较准确地反映了不同应力水平下坝料的流变特征，拟合精度较高。结果表明：不同坑号的坝料在力学特性上各有差异，为了全面了解各种坝料的流变特性，对不同料场的坝料进行试验研究是有必要的；同时，不同应力状态下坝料的流变力学参数有差异，说明不同围压对坝料结构的改变也不同，从而对其流变特征的影响也会有差异。

Rheological parameters identification for materials of high rockfill dam with core wall intelligent method

Bit climbing method is introduced into genetic algorithm (GA) to improve generalization ability, calculating speed and precision of GA. Then this improved GA is used to identify the rheological parameters for materials of high rockfill dam with core wall of Nuozhandu Hydropower Station. Rheological parameters of various dam materials in different holes are identified, and the calculating results using the identified parameters perfectly agree with the tested data. Results also show that rheological parameters of different dam materials greatly differ from each other; so systematic experiments for all the dam materials are necessary. Also, there are some differences among rheological parameters of different dam materials, which may mean that various confining stresses influence the structure of dam material differently.