Globalization of distributed parameter self-optimizing control

Numerous nonlinear distributed parameter systems (DPSs) operate within an extensive range due to process uncertainties. Their spatial distribution characteristic, combined with nonlinearity and uncertainty, poses challenges in optimal operation under two-step real-time optimization (RTO) and economic model predictive control (EMPC). Both methods necessitate substantial computational power for prompt online reoptimization. Recent local distributed parameter self-optimizing control (DPSOC) achieves optimality without repetitive optimization. However, its effectiveness is confined to a narrow range around a nominal operation. Here, globalized DPSOC is introduced to surmount the limitation of the local DPSOC. A global loss functional concerning controlled variables (CVs) is formulated using linear operators and Fubini's theorem. Minimizing the loss with a numerical optimization procedure yields CVs exhibiting global optimality. Maintaining these CVs at constants ensures such a process has a minimal average loss in a large operating space. The effectiveness of the proposed method is substantiated through a transport reaction simulation.