基于BP神经网络的城市燃气管网系统脆弱性评价

发布时间：2019-01-01 13:52

【摘要】：目前我们对系统安全性的研究主要是从系统风险的角度进行的，而非从如何提高系统本质安全方面加以考虑。但长期的安全工作实践证明，单纯从风险的角度评价系统安全状况，做出安全管理决策略显不足，为研究城市燃气管网系统的本质安全性，文章借鉴城市燃气管网风险的评价指标体系及其评价方法，提出了适用于城市燃气管网系统的脆弱性概念及其脆弱性评价的研究思路。本文根据城市燃气管网自身的行业特点与实地调研结果，界定了适合于城市燃气管网系统的脆弱性概念。从物理、结构和社会功能三个角度出发，建立了一个由四个层次组成的城市燃气管网脆弱性指标体系，且第二层包括“敏感性程度”和“应对能力”两个指标，并通过借鉴相关方法和专家建议设定了5个等级，确定了各等级的范围及输出值。然后通过MATLAB R2010a，应用BP神经网络的研究方法分别训练“敏感性程度”神经网络和“应对能力”神经网络，并构建脆弱性函数模型。最后将城市燃气管网脆弱性等级划分为5个，用于评价城市燃气管网的脆弱性程度。选取重庆市某城镇的城市燃气管网系统进行实例应用，得出研究区域的燃气管网脆弱性等级值为0.81，被评定为“不脆弱”。通过模糊综合评价法得出的脆弱性评价结果、管网近10年的运行状况与前文得到的结果进行对比验证，得出三者的结果吻合，说明本文所建立的城市燃气管网系统脆弱性指标体系和脆弱性评价模型具有一定得可靠性和适用性。基于BP神经网络的评价模型能够将脆弱性样本数据进行一系列的科学处理和计算，脆弱性评价结果自动输出，具有较强的泛化能力，能够较好的用于评价未知样本，，为城市燃气管网脆弱性评价方法提供了一种新思路。
[Abstract]:At present, our research on system security is mainly from the point of view of system risk, not from how to improve the essential security of system. However, the long-term practice of safety work has proved that it is insufficient to evaluate the safety situation of the system from the risk perspective and make the safety management decision, in order to study the essential safety of the urban gas pipeline network system. Based on the evaluation index system of urban gas network risk and its evaluation method, this paper puts forward the concept of vulnerability and the research ideas of vulnerability evaluation for urban gas pipeline network system. According to the industry characteristics and field investigation results of urban gas pipeline network, this paper defines the concept of vulnerability suitable for urban gas pipeline network system. From the three aspects of physics, structure and social function, a vulnerability index system of urban gas pipeline network is established, which consists of four layers, and the second layer includes two indexes: "sensitivity" and "coping ability". By drawing lessons from relevant methods and experts' suggestions, five grades are set up, and the range and output value of each grade are determined. Then, through MATLAB R2010a, the method of BP neural network is applied to train the sensitivity degree neural network and the coping ability neural network, and the vulnerability function model is constructed. Finally, the vulnerability grade of urban gas pipeline network is divided into 5 grades, which is used to evaluate the vulnerability degree of urban gas pipeline network. The urban gas pipeline network system of a town in Chongqing is selected for an example, and the vulnerability grade of the gas pipeline network in the study area is 0.81, which is evaluated as "not fragile". Through the result of vulnerability evaluation obtained by fuzzy comprehensive evaluation method, the operation status of the pipe network in the past 10 years is compared with the results obtained in the previous paper, and the results are in agreement with each other. It shows that the vulnerability index system and vulnerability evaluation model of urban gas network system established in this paper have certain reliability and applicability. The evaluation model based on BP neural network can process and calculate the data of vulnerability samples scientifically and automatically output the results of vulnerability evaluation. It has a strong generalization ability and can be used to evaluate unknown samples. It provides a new idea for the vulnerability evaluation method of urban gas pipeline network.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TU996;TP18

【参考文献】