基于张量场的复杂系统事故风险动态预测模型研究

A Dynamic Prediction Model of Complex System Accident Risk Based on Tensor Field

复杂的安全相关系统在运行过程中面临的事故风险受到系统中不同组件的共同影响。在系统运行过程中,要想在确定时刻对事故发生的概率和严重程度做出准确的分析和预测,需要通过构建一系列基于事件的模型并通过描述极其复杂的因果关系来实现。为了克服这一难点,本文基于事故风险是一种张量场函数的观点,构建了基于张量场的系统事故风险模型。该模型通过对事故风险的观测坐标系进行转换,从事故发生概率和严重程度组成的参数坐标系转换到其他可观测的安全相关物理量所组成的参数坐标系,实现对事故风险的监测。实际数据分析结果验证了模型的合理性。

The accident risk faced by complex safety-related systems during operation is influenced by different components in the system.In the process of system operation,in order to make an accurate analysis and prediction of the probability and severity of accidents at a certain time,it is necessary to build a series of event-based models and describe extremely complex causal relationships to achieve this.In order to overcome this difficulty,based on the view that accident risk is a tensor field function,this paper constructs a system accident risk model based on the tensor field.The model realizes the monitoring of accident risk by transforming the observation coordinate system of accident risk from the parameter coordinate system composed of accident probability and severity to the parameter coordinate system composed of other observable safety-related physical quantities.The rationality of the model is verified by the results of actual data analysis.