Thermal runaway avoidance using Hamilton–Jacobi reachability and model predictive control

Abstract

This work investigates application of Hamilton–Jacobi (HJ) reachability and model predictive control (MPC) to exothermic batch processes with the aim of preventing thermal runaway while also improving process efficiency. First the set of unsafe states is identified and the avoid set, the set of states that inevitably become unsafe, is computed using HJ reachability. A novel MPC based scheme incorporating the avoid set as a constraint is then presented. Simulations demonstrate that it results in improved safety and process efficiency when compared to MPC schemes without the avoid set. The results indicate that HJ reachability and MPC can be used to achieve safe process intensification of exothermic batch processes.