摘要
结合隧道施工的特点,提出了基于径向量基函数神经网络修正的模糊期望值决策法。首先采用三角模糊数的形式给出评价指标的取值以及评价者的主观经验值,基于模糊期望值决策法得到隧道施工安全的评价期望值;然后构造适用于高维输入的径向量基函数神经网络算法,建立网络自组织调整隐节点优化规则,采用RBF神经网络修正模糊决策得到期望值,从而建立了RBF神经网络修正模糊期望值的安全评价方法。从安全管理、环境条件等8个方面建立了隧道工程安全评价指标体系。结合工程案例,运用该方法对隧道工程的施工安全进行评价。结果表明,总体上该方法与模糊评价法结果一致,但更具合理性和准确性。
The paper is aimed at presenting our research results on the application of the improved fuzzy expectancy decision-making method to the tunnel excavation risk assessment. So far as we can see, there exist two main problems with this kind of safety assessment. That is, on the one hand, it is lack in precise index systemwith a lot of subjective factors and experienced factors, which are hard to quantify and standardize. On the other hand, traditional fuzzy comprehensive eva-l uation methods prove to be complicated and mainly predictable in nature. In order to solve the problems involved, we have done necessary research on the fuzzy expected value decision-making method （FDM） and established a fuzzy expected value decision-making method modified by the radial basis function（RBF） neural network. To be exact, we have done the following innovations： Firstly, in order to establish a scientifically-based model of safety assessment of tunnel construction, we have established a safety evaluation index system for the tunnel construction according to the risk level of factors and special technologies in tunnel construction, which can be divided into eightmajor subsystems： the securitymanagement, the blasting equipment with blasting operations, the environmental conditions, the tunnel muck hauling and transportation installations and vehicles, workmen-protecting facilities, construction ventilation, construction equipment and facilities and, finally, the power supply system. The index value and engineers’subjective experience factors are all illustrated in the form of triangle fuzzy expertise. The safety evaluation expectancy value can also be achieved by usingthe fuzzy expected value decision-making method. Secondly, this paper hasworked out a RBF neural network algorithm suitable for mult-i dimensional input and set up the optimized rules for adjusting hidden nodes automatically. The safety evaluation of the tunnel construction can thus be evaluated in a comprehensive way by integrating the safety evaluation e
出处
《安全与环境学报》
CAS
CSCD
北大核心
2013年第3期248-251,共4页
Journal of Safety and Environment
基金
云南省自然科学基金项目(2011FZ048)
关键词
安全工程
安全评价
隧道施工
径向量基函数
神经网络
高维输入
safety engineering
safety evaluation
tunnel construction
radial basis function（RBF）
neural network
highdimension inputs
