摘要
为了从管道大数据中挖掘有效信息,高效准确地识别管道存在的风险,建立城市管网风险预测体系,结合深度置信网络,提出了基于深度置信网络的管道风险预测方法。以某燃气公司12段管网为例,其预测结果与实际情况相符,进一步选取未评级管道,利用多层次模糊风险分析法与所建立模型同时进行风险评估,二者结果一致。研究结果表明,该模型稳定性强、诊断速度快、识别准确率高,可较好地实现城市燃气管道的风险模式识别。
In order to mine effective information from pipeline big data, identify the risk of pipeline efficiently and accurately, a city pipeline network risk prediction system was established. Combining with deep trust network, a pipeline risk prediction method based on deep confidence network was proposed. Taking a 12-segment pipe network of a gas company as an example, the prediction results were consistent with the actual situation, and the unrated pipeline was further selected to use the multi-level fuzzy risk analysis method to conduct risk assessment simultaneously with the established model. The results were consistent. The research results showed that the model has strong stability, high diagnosis speed and good discrimination,which can have better recognition effects about the risk pattern of urban gas pipelines.
作者
王新颖
张惠然
张瑞程
赵斌
陈海群
WANG Xin-ying;ZHANG Hui-ran;ZHANG Rui-cheng;ZHAO Bin;CHEN Hai-qun(School of Environmental & Safety Engineering, Changzhou University Jiangsu Changzhou 213164, China;School of Petrochemical Engineering, Changzhou University, Jiangsu Chang zhou 213164, China)
出处
《消防科学与技术》
CAS
北大核心
2019年第6期902-905,共4页
Fire Science and Technology
基金
国家安监总局安全生产重大事故防治关键技术科技项目“天然气管网泄漏检测及风险管理系统关键技术研究”(安监总厅科技[2013]140号)
常州市科技项目“城市地下燃气管网信息化管理与应急决策支持系统”(CZ20170017)
关键词
深度学习
大数据
燃气管道
风险评估
deep confidence network
big data
gas pipeline
risk assessment
