(419d) Lyapunov-Based Economic Model Predictive Control with Cyberattack Detection for Process Actuators

Recent years have seen industries working to more tightly integrate physical processes on process shop floors with communication and control systems ^[1]. This cyber-physical systems (CPS) integration is a consequence of advantages like increased production efficiency and improved transparency into the steps involved in production for better quantity control. However, this integration also opens the integrated processes and components to cyberattacks which can cause the degradation of process safety resulting in effects ranging from loss in profits to endangering human lives. Various aspects of a physical system can be implemented by an attack, but these different channels lead to actuators taking actions they would not have taken in the absence of an attack. However, there is a difference between having these actuators directly targeted via an attack (through, for example, improper control signals being sent to them), compared to having the sensors manipulated so that the actuators are receiving the correct control action for the state measurement that the controller received. Much of the prior work in our group on developing a control-theoretic framework for analyzing detection and control policies as they relate to cyberattacks on nonlinear systems has focused on attacks on the process sensors in the sense that the sensor measurements received by the controller are the part of the control loop that becomes in accurate.^[3,4] These detection and control policies have focused on the use of an advanced optimization-based control framework known as Lyapunov-based economic model predictive control (LEMPC) ^[2] in the detection of attacks when the actuators are assumed to function correctly. However, when they function incorrectly, it is necessary to consider how these strategies might be redesigned.

This work discusses our advances in integrated cyberattack detection and control policies when actuators are impacted.^[5,6] We first discuss how three integrated cyberattack detection and control policies from ^[4] can be modified to for the case that actuators are attacked instead of sensors. The first of these methods involves constantly probing for cyberattacks by developing new steady-states throughout state-space over time where the closed-loop state measurements must show that the state is approaching the new steady-state during operation at all times. The second method involves using predictions of the state under the expected control action and comparing this with the state measurements observed. The third involves using redundant state estimators to catch when the closed-loop state estimates are inconsistent. We demonstrate closed-loop stability under the first two of these strategies in the actuator attack case, and discuss challenges with implementing the third. We also discuss how these strategies relate to profits and to what aspects of a system would need to be fully secured to gain detection guarantees. The implementation strategy of the constant probing strategy is demonstrated through a continuous stirred tank reactor example. This example serves to showcase that part of the implementation strategy must involve how to ensure that the alternative steady-states developed over time are steady-states within the input bounds.

References:

[1] Lezzi, M., Lazoi, M., Corallo, A. Cybersecurity for industry 4.0 in the current literature: A reference framework. Computers in Industry (103), 97-110 (2018)

[2] Heidarinejad, M., Liu, J. and Christofides, P.D. Economic model predictive control of nonlinear process systems using Lyapunov techniques. AIChE Journal, 58(3), 855-870 (2012).

[3] Rangan, K. K., Oyama, H., & Durand, H. Integrated Cyberattack Detection and Handling for Nonlinear Systems with Evolving Process Dynamics under Lyapunov-based Economic Model Predictive Control. Chemical Engineering Research and Design (2021).

[4] Oyama, H. and H. Durand. Integrated Cyberattack Detection and Resilient Control Strategies Using Lyapunov-Based Economic Model Predictive Control, AIChE Journal (66), ee17084 (2020).

[5] Rangan, K. K., Oyama, H., & Durand, H. Actuator Cyberattack Handling Using Lyapunov-based Economic Model Predictive Control. 13th IFAC Symposium on Dynamics and Control of Process Systems, including Biosystems (DYCOPS 2022).

[6] Oyama, H., Rangan, K. K., & Durand, H. Lyapunov-Based Economic Model Predictive Control for Detecting and Handling Actuator and Simultaneous Sensor/Actuator Cyberattacks on Process Control Systems. Frontiers in Chemical Engineering, 11.