KELM-TE方法在石油储运系统安全信息风险传播中的应用

Application of Kernel Extreme Learning Machine and Transfer Entropy Method in the Risk Propagation of Security Information in Oil Storage and Transportation System

为控制生产系统内安全信息传播带来的风险,明确安全信息风险传播的路径和方向,提出一种基于核极限学习机(KELM)和传递熵(TE)结合的生产系统安全信息风险传播路径分析方法(即KELM-TE方法)。该方法首先针对安全信息意外释放,分析其在生产系统内的传播发展过程,并根据TE的原理建立安全信息风险传播推演模型,确定了安全信息在过程变量中的传播方向;然后基于KELM的预测方法,通过信息计算变量间的关联系数、扰动变化率和影响因数,推算出生产系统内安全信息风险传播的可能路径,并通过生产系统安全信息风险传播方向和路径的预测,形成了完整的生产系统安全信息风险传播模型;最后将该方法应用于某储运系统的憋压泄漏事故过程中,验证方法的有效性。结果表明:KELM-TE方法可预测生产系统内可能的安全信息风险传播路径,能为管理人员提供及时的风险控制措施,保障系统的安全运行。

In order to control the risks brought by the spread of security information in the production system,to clarify the path and direction of risk propogation of security information,this paper proposes a risk propagation analysis method for security information in production system based on Kernel Extreme Learning Machine(KELM)and Transfer Entropy(TE).First,the propagation and development process of security information in the system is analyzed.Based on the principle of TE,the risk propagation reasoning model of security information is established,and the transmission direction of security information in the process variables is realized.With the prediction method of KELM,the possible path of security information risk propagation in the system is determined by calculating the correlation coefficient between variables,the rate of change of disturbance and the influence factor through the information.Next,through the prediction of the propagation direction and path,a complete risk propagation model for security information in production system is proposed.Finally,the method is applied to the pressure leakage accident process of a storage tank system to verify the validity of the method.The results show that the KELM-TE method can predict the possible risk propagation path of security information in the system,and can provide timely risk control measures for the managers and ensure the safe operation of the system.