基于多源数据融合的油气联合站风险预警研究

Risk Assessment of Oil and Gas United Stations Based on Multi-source Data Fusion

针对联合站蒸汽云爆炸事故场景,确定了相关评价因子及风险等级阈值区间,构建了基于云模型的基本概率分布(BPA)函数,通过可信度权重与信息量权重对初始BPA进行修正,建立了一种改进证据理论的融合模型,并针对某联合站的油罐区进行了模型应用。结果表明:提出的BPA构建方法可以有效利用数据;与以单传感器数据为依据的风险预警结果相比,多源数据融合获得的结果更准确;在证据存在高度冲突时,提出的方法可以对存在高度冲突的数据进行合理处理,使最终结果更加可靠。

For the vapor cloud explosion(VCE)accident scenario of the united station,the relevant evaluation factors and the threshold range of risk level are determined.The basic probability distribution(BPA)function based on cloud model is constructed.The initial BPA is corrected by the weight of credibility and information.A fusion model of improved evidence theory is established.The model is applied to the oil tank farm of a united station.The results show that:(1)The proposed BPA construction method can effectively use data.(2)The risk warning results based on single sensor data are unstable,and the results of multi-source data fusion are more representative.(3)When there is a high degree of conflict between evidences,the proposed method can reasonably process the data with a high degree of conflict,so that the final result is more reliable.