Dynamic MAWQA: Towards efficient real-time optimization of slow dynamic processes

This paper deals with the real-time steady-state optimization of slow dynamic processes under plant-model mismatch. A novel scheme that iteratively adapts the process set-points (or operating parameters) to attain an economic optimum is proposed. Based on computing the next steady state from the process transient response to the current set-point change, this scheme can significantly reduce the time to reach the optimum steady state of slow dynamic processes. The key idea here is to integrate the modifier adaptation with quadratic approximation optimization scheme with a steady-state estimation algorithm. Simulation studies with two benchmark processes illustrate the performance of the proposed scheme.