Practical optimization of complex chemical processes with tight constraints

Stochastic optimization algorithms are introduced for complex chemical processes—using semicontinuous distillation with reaction in a middle vessel as a case study. The processes are dynamic, consisting of several operating modes, and have tight operating constraints to prevent weeping or flooding in the distillation column of the case study. Often intensification leads to fewer processing units, operated in many modes over a cyclic campaign, introducing tight operating constraints. The optimization algorithms examined include a modified univariate technique for N-dimensions, the particle swarm method, the simplex method, and others. These methods are used jointly in a bi-level algorithm, and are tested for speed and quality of result.