Detection of Cyber-Attacks and Resilient Operation of Nonlinear Processes Under Economic Model Predictive Control

In recent years, the cyber-security of chemical plants has become increasingly important as more communication networks are replaced or complemented by wireless networks in addition to point-to-point communications. Stable and secure operation of chemical plants require accurate information and reliable communication technologies. In the last few years, increasing research efforts have been dedicated to developing system and control designs to address cyber-attacks using machine-learning tools for cyber-attack detection [1,2,3]. Despite this progress, the integration of artificial intelligence and advanced control algorithms to address the problem of cyber-security is an emerging field of research interest with many open issues.

This work presents a resilient operating scheme for nonlinear processes subject to standard types of cyber-attacks, and proposes detection and handling strategies of these cyber-attacks. In particular, considering a general class of nonlinear systems, a modified Lyapunov-based Economic Model Predictive Controller (LEMPC) using combined closed-loop and open-loop control action implementation schemes is proposed with the control objective of optimizing economic benefits while maintaining closed-loop process stability. Process stability and resiliency against particular types of targeted destabilizing attacks will be ensured under the proposed controller modification and operation strategy. Moreover, machine-learning methods are used to develop sensor-data-based cyber-attack detectors, which are periodically activated online during process operation. Simulation results based on a continuously stirred tank reactor example demonstrate the effectiveness of the resilient LEMPC control strategy in the presence of various cyber-attacks, as well as the capability of the detection algorithm in reporting and differentiating the different types of cyber-attacks.

[1] Cárdenas, A., Amin, S., Lin, Z., Huang, Y., Huang, C., Sastry, S., 2011. Attacks against process control systems: risk assessment, detection, and response, in: Proceedings of the 6th ACM Symposium on Information, Computer and Communications Security, pp. 355–366.

[2] Chen, S., Wu, Z., Christofides, P.D., 2020. A cyber-secure control-detector architecture for nonlinear processes. AIChE Journal 66, e16907.

[3] Durand, H., 2018. A nonlinear systems framework for cyberattack prevention for chemical process control systems. Mathematics 6, 169.