Stochastic Multilayer Optimization for an Acrylic Acid Reactor.

In this paper, a multilayer stochastic optimization approach is implemented to solve a dynamic optimization problem under uncertainties for an acrylic acid reactor. The proposed methodology handles different sources of uncertainties (internal, external, process), being a novel approach to obtain more realistic solutions in the context of process optimization. A comparison against deterministic dynamic optimization, single-layer stochastic optimization, and typical PI control loops is carried out. The results show the efficacy of the multilayer stochastic optimization approach for handling different sources of uncertainties, improving the economic profitability of the process while fulfilling the safety constraints in all of the scenarios analyzed.