A New Strategy and Algorithm of Minlp Based on Module for Synthesis of Chemical Processes

This article presents the formulation and the solution of a class of Mixed-integer Nonlinear Programming (MINLP) problems, applied to optimal synthesis of process systems in a modular simulator environment. The proposed algorithm makes use of module, reduces the work of calculation on model equations, makes the size of MINLP smaller. The algorithm consists of solving alternate sequences of MILP master problem and NLP subproblem whose sizes have been reduced. The NLP subproblem provides the linearization information of the nonlinear constraints that relate the output variables to input variables specified explicitly. At the same time Data-Oriented Process Simulation system (DOPS) technique is used in the modular simulator environment. The HDA process is taken as a case study the capabilities of the process synthesizer. It proves that the proposed solution deals with complex MINLP problems satisfactorily.