基于RBF神经网络的地铁隧道施工坍塌事故应急车辆需求预测  

Demand prediction of emergency vehicles for subway tunnel collapse accidents based on RBF neural network

作　　者：邓林[1] 陈玉斌 刘湘慧 陈赟[2] DENG Lin;CHEN Yubin;LIU Xianghui;CHEN Yun(School of Architectural Engineering,Hunan Communication Polytechnic,Changsha 410132,China;School of Traffic&Transportation Engineering,Changsha University of Science and Technology,Changsha 410114,China;CRCC Harbour&Channel Engineering Bureau Group Survey&Design Institute Co.,Ltd.,Guangzhou 511442,China)

机构地区：[1]湖南交通职业技术学院建筑工程学院,湖南长沙410132 [2]长沙理工大学交通运输工程学院,湖南长沙410114 [3]中铁建港航局集团勘察设计院有限公司,广东广州511442

出　　处：《铁道科学与工程学报》2022年第7期2100-2106,共7页Journal of Railway Science and Engineering

基　　金：2018年湖南省应急管理厅安全生产科技研究及推广项目(201801)。

摘　　要：现有生产安全事故应急救援体系未能充分利用历史经验辅助应急决策,并多为资源总体上的预测和调度。为高效、准确地应对地铁隧道施工坍塌事故,制定应急救援方案,提高应急救援效率,以确定应急车辆需求量为例,提出一种基于RBF神经网络的应急车辆需求预测模型。在确定预测指标的基础上,整理分析历史事故案例并提取关键指标值,通过训练确定扩展速度及隐含层神经元个数,分别构建以坍塌位置、次(衍)生事故、坍塌面积和被困人数为输入,救护车及消防车需求量为输出的RBF神经网络预测模型。在预测结果的基础上,确定所需应急车辆类型及数量,推算医护人员和消防人员的配备数量,并结合事故实际特征对预测结果进行修正。以某实际地铁隧道施工坍塌事故数据进行案例分析,预测该事故所需的应急车辆、相关人员及设备等应急资源的需求量,并将其与实际数据对比分析,验证了预测模型的可行性和有效性。该预测模型可为地铁隧道施工坍塌事故应急救援方案的制定提供新的思路和方法。The existing emergency rescue system for production safety accidents fails to make full use of historical experiences to assist emergency decision-making,and mostly uses the overall prediction and scheduling of resources.In order to efficiently and accurately deal with subway tunnel construction collapse accidents,emergency rescue schemes were formulated and emergency rescue efficiency was improved by focusing on the determination of the demand for emergency vehicles as an example.A model for forecasting emergency vehicle demands was proposed based on the RBF neural network.On the basis of determining the prediction index,the historical accident cases were sorted out and analyzed,and the key index values were extracted.Through training,the expansion speed and the number of neurons in the hidden layer were determined,and the collapse position,secondary(derivative)accidents,collapse area and the number of trapped people were taken as inputs.The output of the RBF neural network prediction model included ambulance and fire truck demands.Based on the prediction results,the types and quantities of emergency vehicles needed were determined,the number of medical staff and firefighters was calculated,and the prediction results were modified according to the actual characteristics of the accidents.The collapse accident data of an actual subway tunnel construction were introduced for case study,the demand of emergency vehicles was predicted,the personnel,equipment and other emergency resources required by the accident were related,and the results were compared against the actual data to verify the feasibility and effectiveness of the prediction model.The prediction model can provide a new frame work and method reference for the formulation of emergency rescue plans for subway tunnel collapse accidents.

关 键 词：地铁隧道施工 坍塌事故 应急车辆需求 RBF神经网络 

分 类 号：U455.1[建筑科学—桥梁与隧道工程]