施工技术 2016年7月上 118 CONSTRUCTION TECHNOLOGY 第45卷第13期 D0I:10.7672/5j52016130118 基于MAMPSO-RBFNN的盾构地表变形 时空演化智能预测* 周诚，余群舟 (华中科技大学土木学院工程管理研究所，湖北武汉430074) [摘要]以某地铁越江盾构隧道实际工程为背景，基于Weibull-ARIMA时间过程模型与改进随机介质理论的空间 分布模型，建立了盾构隧道地表任意点变形随盾构掘进的时空演化计算模型，将RBF神经网络与MAMPSO算法进 行有机结合，提出了基于MAMPSO算法的RBF神经网络耦合系统辨识模型及其学习算法，并基于此算法完成盾构 施工地表变形时空演化预测系统的辨识.工程算例结果分析表明，该模型成功应用于某地铁越江盾构隧道右线穿 越长江堤防地表变形时空演化过程的预测，具有较高的计算效率和预测精度，能实现地铁盾构隧道施工过程地表 变形的时空演化预测智能化. [关键词]隧道：盾构；变形；时空演化：RBF神经网络：MAMPSO算法 [中图分类号]U455.43 [文献标识码]A [文章编号]1002-8498(2016)13-0118-07 Intelligent Predictive of Spatio-temporal Evolution of the Surface Deformation Based on MAMPSO-RBFNN Zhou Cheng Yu Qunzhou (School of Ciril Engineering Mechanics Huazhong University of Science Technology Wuhan Hubei 430074 China) Abstract:Based on Weibull-ARIMA time course model and the spatial distribution model sustained by improved stochastic medium theory this paper created a spatio-temporal evolution model for surface deformation in shield tunnel construction.Then it put forward RBF coupling system identification model and its learning algorithm based on MAMPSO algorithm by bining RBF neural network with MAMPSO algorithm to plete the evolution prediction system identification in Yangzi river-crossing shield tunnel construction.Engineering example results show that this model can be applied successfully in evolution prediction of Yangzi river-crossing shield tunnel construction with high putational efficieney and precision of prediction and it can realize th...
基于MAMPSO-RBFNN的盾构地表变形时空演化智能预测.pdf
