隧道施工过程的三角模糊数与神经网络安全评估

Safety assessment for subway construction process with triangular fuzzy number and neural network

为动态地对地铁隧道施工过程进行安全评估,针对地铁超浅埋暗挖段隧道,构建三角模糊数与神经网络相结合的地铁隧道施工过程安全评估模型.根据监测和数值仿真数据建立安全指标体系;划分地铁隧道施工过程评估等级,并将评估等级语言变量转换为三角模糊数;用混合遗传算法优化的BP神经网络,找出综合指标数据与评估等级的非线性映射关系.研究结果表明:大连地铁超浅埋段隧道施工过程处于安全等级;通过与既有方法比较分析,证明该评估方法与实际情况具有较强的贴近度.

In order to realize dynamic safety assessment of subway tunnel construction process, safety evaluation model of subway tunnel construction process based on combined triangular fuzzy number and neural network was established in the super shallow buried tunnel for Dalian subway. Firstly, according to the monitoring and numerical simulation data, the safety index system was established; then, assessment level of subway tunnel construction process was divided and converted to triangular fuzzy numbers; finally, the nonlinear mapping relationship between comprehensive index data and assessment level was found with BP neural network optimized by hybrid genetic algorithm. Results of study show that construction process of the super shallow buried tunnel in Dalian subway is in the safety level; Compared with those assessment results given by other methods, the assessment results given by this method is more consistent with actual situation.