(392b) Cyber Security of Model Predictive Control Systems for Chemical Processes

Recently, the security of process control systems has become crucially important since control systems are vulnerable to cyber-attacks, which is a series of computer actions to compromise the security of control systems (e.g., integrity, stability and safety) [1]. Among cyber-attacks, targeted attacks are severe threats for control systems because of their specific designs with the aim of damaging the control actions applied to a chemical process (for example, Stuxnet worms aims to modify the data sent to a Programmable Logic Controller) [2]. Additionally, targeted attacks are usually stealthy and difficult to detect using classical detection methods since they are designed based on some known information of control systems (perhaps they make explicit use of the process state measurement). Therefore, designing an advanced detection system [3] and a suitable model predictive control system for nonlinear processes in the presence of targeted cyber attacks is an important open issue.

Motivated by this, in this work, we develop a model predictive control system for nonlinear systems [4] subject to targeted cyber attacks. Specifically, a cyber attack that aims to damage close-loop stability via a sensor tamper is first considered and applied to the closed-loop process, and it is demonstrated that if not suitable action is taken by the model predictive controller the closed-loop system loses its stability. Subsequently, an anomaly-based detection method is developed using a priori knowledge of control system, and a suitable model predictive control method is developed to reduce the impact of cyber attacks and re-stabilize the closed-loop system in finite time. A chemical process example is used to demonstrate the applicability of the proposed approach.

[1] Ye N, Zhang Y, Borror C M. Robustness of the Markov-chain model for cyber-attack detection. IEEE Transactions on Reliability, 2004, 53: 116-123.

[2] Cárdenas A A, Amin S, Lin Z S, et al. Attacks against process control systems: risk assessment, detection, and response. Proceedings of the 6th ACM symposium on information, computer and communications security. ACM, 2011: 355-366.

[3] Ozay M, Esnaola I, Vural F T Y, et al. Machine learning methods for attack detection in the smart grid. IEEE Transactions on Neural Networks and Learning Systems, 2016, 27: 1773-1786.

[4] Mhaskar P, El-Farra N H, Christofides P D. Stabilization of nonlinear systems with state and control constraints using Lyapunov-based predictive control. Systems & Control Letters, 2006, 55: 650-659.