摘要
以某石化公司典型装置为例,运用贝叶斯网络设计了泄漏风险预警模型。针对贝叶斯网络的叶节点,采用业界通用的同类设备失效概率作为先验概率,并从气候条件、设备的复杂性、工艺参数等多个方面对网络参数进行优化调整。针对网络结构,采用合理划分风险区域、风险单元的方法避免由于节点增多而造成的计算效率问题。最后,采用建模工具对预测推理场景和诊断推理场景进行模型的运行效果展示。
Taking typical petrochemical equipment as example, the leakage risk prediction model with Bayesian Network was designed. As for node parameters, the universally adopted homogeneous failure possibility was referred as the prior possibility of the equipment failure. And the network factors were regulated according to the collected climate variation, equipment complexity, technological parameters, etc. For the purpose of improving network structure design, the reasonable partition of threat establishment as well as threat unit were introduced to avoid dramatic performance decline when network node quantity increased rapidly. In conclusion, leakage prediction and leakage source diagnose were demonstrated with the modelling tools to show the model effectiveness.
作者
王学岐
赵瑞彬
孙秉才
曹淑霞
WANG Xueqi;ZHAO Ruibin;SUN Bingcai;CAO Shuxia(China Petroleum Safety and Environmental Protection Technology Research Institute Co.,Ltd.,Dalian,Liaoning 116000,China;Beijing Huachuang Zhongshi Technology Development Co.,Ltd.Beijing 100029,China)
出处
《自动化与仪器仪表》
2023年第10期68-72,78,共6页
Automation & Instrumentation
基金
中国石油天然气股份有限公司科学研究与技术开发项目:油气类危险化学品泄漏监测风险预警系统与应用装备研发(2020D-4626)。
关键词
化工装置
贝叶斯网络
泄漏风险自动预警
先验概率
petrochemical equipment
bayesian network
automatic early warning of leakage risk
prior possibility
