信息安全威胁下油气智慧管道系统失效新型致灾机理与早期预警

我国油气管道已步入智慧管道建设阶段，管道实体和数字孪生高度融合，具有“信息-物理-社会”系统属性。信息空间安全问题已加速向设施物理空间和人员作业社会空间交互与渗透，呈现“跨域”迭代破坏现象；且攻击模式也呈动态、多样及智能化趋势。目前国内外研究焦点多为网络层的入侵检测，然而从网络入侵至最终的生产安全失控期间（风控“时间窗”），如何快速、精准识别、跟踪和预警，成为难点。其原因为不同攻击模式下管道系统安全状态失稳判据与主控因素等规律不明并缺乏定量认知；不同尺度下受影响节点的故障动态特征如何表征与测度尚无科学系统的理论。本项目通过信息安全威胁下智慧管道系统失效新型致灾机理的研究，揭示攻击行为与管道系统漏洞的交互关系、故障跨域传播动力学等新规律；创建短时/长时攻击和变工况下的稳定跟踪检测以及跨尺度早期预警方法，为保护智慧管道系统免受信息安全攻击破坏提供最后一道关键的屏障。

With the developments of new generation information technologies, a smart manufacturing of oil and gas pipeline era is emerging. It is clear that achieving cyber physical and social fusion is a crucial step for the implementation of smart oil and gas pipeline system. With respect to the digital twin, for each physical pipeline object, there exists a virtual mirror model with abilities of analyzing, evaluating, optimizing and predicting, etc. The two parts interact with each other and remain synchronous in forming a closed loop. In addition, data from both the physical and virtual sides can be obtained and fused to generate more comprehensive information. Therefore the security problem in information space has accelerated its interaction and penetration into the physical space of facilities and the social space of personnel operations, which presents the phenomenon of "cross-domain" iterative destruction. Meanwhile the way of security attack shows a more dynamic, diverse and intelligent trend. The occurrence of cascading effect of cyber-attack accidents has made enough attention, however the current research methods mostly focus on the intrusion detection at the network layer. How to quickly and accurately identify, track, and provide early-warning within the risk control “time window” which lies between the attack invasion and further damage to the physical system, becomes a difficult task. The reason is that the coupling relationship between information security threats and facility failures, loss of functions, and main control factors are unknown and insufficiency of quantitative cognition. The scientific and systematic theories are relatively scarce on how to characterize and measure the fault dynamic characteristics of affected nodes at different scales. The new failure and accident mechanism of smart pipeline system under the threat of information security is studied in this project. New mechanisms such as the interaction between cyber-attack behaviors and pipeline system vulnerabilities, and the dynamics of cross-domain propagation of faults will be revealed. Fast, accurate and stable monitoring and detection methods for short-term and long-term attacks respectively will be developed, especially to have the ability to cope with the variable pipeline operating conditions. In this way, a new theoretical system involving trans-scale early-warning according to micro observable information can be established, which plays an important role as the last safety barrier for protecting smart pipeline systems from information security attacks.