Integrating dynamic economic optimization and encrypted control for cyber-resilient operation of nonlinear processes

Abstract

This article proposes a two-layer framework to maximize economic performance through dynamic process economics optimization while addressing fluctuating real-world economics and enhancing cyberattack resilience via encryption in the feedback control layer for nonlinear processes. The upper layer employs a Lyapunov-based economic model predictive control scheme, receiving updated economic information for each operating period, while the lower layer utilizes an encrypted linear feedback control system. Encrypted state information is decrypted in the upper layer to determine the economically optimal dynamic operating trajectory through nonlinear optimization. Conversely, the lower layer securely tracks this trajectory in an encrypted space without decryption. To mitigate the cyber vulnerability of the upper layer, we integrate a cyberattack detector that utilizes sensor-derived data for attack detection. We quantify the errors stemming from quantization, disturbances, and sample-and-hold controller implementation. Simulation results of a nonlinear chemical process highlight the robustness and economic benefits of this new control architecture.