摘要
针对水电工程地应力测量中测点较少、量测结果离散及计算的非线性问题,提出了基于人工神经网络(ANN)和遗传算法(GA)的应力测量压力反分析模型。该模型利用ANN预测井底压力,并结合井底量测压力值,建立一个适应度函数,然后利用GA在一个大的搜索空间中找到适应度函数的最优解,即所求的应力参数。最终利用糯扎渡水电工程实测水力压力值对该模型进行验证,结果表明模型识别应力值与实测值的相对误差在5%以内,验证了ANN-GA模型的有效性,为水电工程应力确定提供了新方法。
For the problem of few points,discrete measurement results and calculation nonlinear in stress measurement of hydropower engineering,pressure back analysis model of stress measurement based artificial neural network（ANN）and genetic algorithm（GA）was proposed.In this model,ANN was used to predict bottom hole pressures.By combining measured pressure of bottom hole,a fitness function was established.Then the GA was used to search the optimal solution（i.e.requested stress parameters）in a global space.Finally,the observed water pressure of Nuozhadu hydropower project was used to verify the proposed model.The results show that the relative error between identified and real values is less than 5%.It shows that the proposed ANN-GA model is effective,which provides a new method for stress measurement in hydropower project.
出处
《水电能源科学》
北大核心
2016年第5期137-140,共4页
Water Resources and Power
基金
河南省科技攻关计划项目(152102310318)
河南省高等学校重点科研项目(16A410001)
关键词
应力确定
压力反分析
人工神经网络
遗传算法
determination of stress
pressure back analysis
artificial neural network
genetic algorithm
