摘要
建立马尔科夫—灰色BP神经网络组合模型是为了更加科学合理预测深基坑压顶水平位移,提高预测精度.比较分析BP神经网络模型与串联式灰色BP模型的预测结果,建立马尔科夫链修正的灰色BP组合模型,以汕头市某花园酒店扩建工程的基坑压顶水平位移的实测数据为研究对象,通过比较模型预测结果与实际结果,检验其深基坑预测模型的精度.实例证明,经马尔科夫链改进的灰色BP神经网络组合模型的预测精度优于单一模型,更适合用于样本少、随机波动性大的深基坑变形预测.马尔科夫—灰色BP神经网络组合模型对深基坑压顶水平位移的预测不仅精度高,同时反映出数据序列发展变化的总体趋势和系统之间各状态的规律,为深基坑压顶水平位移预测提供了一种新方法.
To scientifically predict deep foundation pit horizontal displacement, a combined model based on Grey and BP neural network model corrected by Markov chain is established to improve accuracy. On the analysis of BP neural network model and the tandem Grey BP model,a prediction model based on Markov chain and Grey BP neural network model is set up. The foundation pit horizontal displacement of a certain garden hotel in Shantou is used to verify the model and check its precision by comparing the model prediction results and the actual results.The result shows that Grey BP neural network and Markov chain have higher precision. It＇s more suitable for deep foundation pits prediction with less samples and large stochastic volatility. The Markov chain-Grey BP neural network model is not only giving higher prediction but showing the data sequence trends and the internal law between system states. It provides a new method for deep foundation pit deformation prediction.
作者
刘洁
吴鸣
袁继雄
LIU Jie;WU Ming;YUAN Jixiong(Department of Civil Engineering, Shantou University, Shantou 515063, Guangdong, China;Shantou Government Investment Project Construction Management Centre, Shantou 515000, Guangdong, China)
出处
《汕头大学学报（自然科学版）》
2017年第3期53-60,共8页
Journal of Shantou University：Natural Science Edition
基金
住房和城乡建设部科技计划项目(2015-k3-003)
关键词
马尔科夫链
灰色BP神经网络
深基坑
预测
Markov chain
Grey BP neural network
deep foundation pits
prediction
