Based on authors’ previous work on single column simulation and optimization

this paper focuses on simulation and optimization of a distillation column system composed of a depropanizer and a debutanizer in ethylene production.The distillation column sequence was modeled with open-form equations.It had 4293 variables and 4188 equality equations characterized by its block triangular and highly sparse structure.The operation optimization problem was properly defined for profit improvement.It had relatively small degree of freedom

which could be effectively utilized in RSQP (reduced-space sequential quadratic programming) algorithm.The gradient information was obtained with hybrid automatic differentiation method.Variant implementation of SQP solvers including RSQP(from authors)

SNOPT(from TOMLAB) and FMINCON(from Matlab Optimization Toolbox) were used to solve this problem.The results and efficiency value were compared.It was illustrated in numerical experiments that RSQP outperformed other solvers with the same initial values and convergence tolerance.This work can be useful in the simulation and optimization of the whole process of ethylene production.