摘要
在边坡最危险滑动面的搜索过程中,为避免目标函数建立时因公式化简带来的误差及传统优化算法易陷入局部极小值的问题,提出了基于BP-遗传算法的排土场稳定性分析方法。以瑞典条分法为基础,通过BP神经网络建立潜在滑动面的圆心、半径与其安全系数间的非线性关系式,利用遗传算法,搜索边坡的最危险滑动面,并计算其相应的安全系数。弓长岭排土场实例分析结果表明:与传统稳定性分析方法相比,BP-遗传算法的稳定性分析结果更加精确、可靠,具有较好的应用前景。
In the process of searching the most dangerous slip surface, the method for analyzing the stability of the waste dump was proposed based on the BP neural network and ge- netic algorithm, Therefore, it avoided the error of the objec- tive function due to formula simplification and the problem that the traditional optimization algorithm was easy to fall into local minimum value. Based on the Swedish slice meth- od, the nonlinear correlation among the center and radius of the latent slip surface with its corresponding safety factor was established. And according to the genetic algorithm, the most dangerous slip surface and the corresponding safetyfactor were determined. The application results in the waste dump of Gongchangling Open-pit Mine showed that, com- pared with the traditional stability analysis method, the re- sult of the BP neural network and genetic algorithm was more precise and reliable, with a good implementation pros- pect.
出处
《矿业研究与开发》
CAS
北大核心
2016年第8期71-74,共4页
Mining Research and Development
基金
国家自然科学基金项目(51274053)
辽宁省教育厅科研基金项目(L2011040)
关键词
BP神经网络
稳定性分析
遗传算法
排土场
BP neural network, Stability analysis, Genetic al-gorithm, Waste dump