(711a) Real-Time Risk-Based Control: Integrating Deterministic Risk Indicators and Probabilistic Safety Constraints for Safe and Efficient Hydrogen Production

Ensuring safe and efficient operation in dynamic process systems remains a key challenge, particularly in safety-critical clean energy applications. This work presents a real-time risk-based control framework that systematically integrates safety into decision-making through a two-stage methodology.

The framework is developed in two sequential steps. First, a deterministic risk-aware control structure is formulated by embedding a dynamic Risk Indicator (RI) within a model predictive control (MPC) scheme. The RI quantifies thermal deviations in real time and enables the controller to respond proactively to emerging risk conditions [1]. Second, the framework incorporates probabilistic safety constraints within an explicit MPC (eMPC) formulation developed using multi-parametric programming [2,3]. This extends to address key challenges in probability-based safety assessment, particularly in translating uncertainty into actionable control laws, and introduces criteria to systematically incorporate probabilistic safety bounds within an explicit control framework. This approach enables real-time control law evaluation without online optimization, while also embedding chance-constrained safety limits. The resulting formulation supports a structured trade-off between performance and safety, allowing the system to operate closer to performance boundaries while maintaining acceptable risk levels.

The proposed methodology is validated on a cyber-physical platform of a proton exchange membrane water electrolyzer (PEMWE), a key technology for green hydrogen production. A linear state-space model is identified from experimental data, and the two-stage controller is implemented to regulate stack current and water flow rate. The deterministic controller ensures thermal stability and safe operation, while the probabilistic extension allows for enhanced flexibility under uncertainty. This work demonstrates a unified and computationally efficient approach to integrating real-time risk assessment into advanced control design, enabling resilient and efficient operation in emerging hydrogen energy systems.

Keywords: Risk assessment, Model Predictive Control, Safety, Multi-Parametric Programming, Optimization, Probabilistic constraints, chance constrained programming

References:

• Ali, M., Cai, X., Khan, F. I., Pistikopoulos, E. N., & Tian, Y. (2023). Dynamic risk-based process design and operational optimization via multi-parametric programming. Digital Chemical Engineering, 7, 100096.
• Akundi, S.S., Braniff, A., Dantas, B., Liu, Y., Tian, Y., Niknezhad, S.S., Khan, F.I. and Pistikopoulos, E.N., 2024. Advanced system control strategies for enhanced safety and efficiency of energy systems. In Methods in Chemical Process Safety(Vol. 8, pp. 243-260). Elsevier.
• Braniff, A., Akundi, S.S., Liu, Y., Khan, F., Pistikopoulos, E.N. and Tian, Y., 2024. A Real-Time Risk-Based Optimization Framework for Safe and Smart Operations. In Computer Aided Chemical Engineering(Vol. 53, pp. 1915-1920). Elsevier.