Optimal Design of Pressure Swing Adsorption Systems

Abstract

Pressure swing adsorption PSA is an effective technique for gas separation and can be used for several applications such as carbon dioxide capture, air separation, and hydrogen purification. The algorithm developed in this project seeks the design parameters that minimize the cost of the unit over its life time while addressing process and product constraints. The objective function is to minimize the net present value of arisen costs (NPVC) considering all process constraints during a life period of 30 years for a PSA plant. The NPVC function consisted of 2 types of costs; the Total Capital Investment, and the Bed Replacement Cost. The function was then transformed into a nonlinear model, and the design factors chosen included both integer variables as well as continuous variables. By means of the assigned constraints and the remaining cost correlations of the objective function, the latterly established model was then integrated, forming a Mixed Integer Non-Linear Programming (MINLP) problem. The model was later reduced into various NLPs, which were subsequently solved using an SQP method that ensures local optimality. MatlabTM language has been used to apply the algorithm and the fmincon function was utilized to optimize the NLPs. The above-mentioned algorithm was then implemented on an industrial case to prove its efficiency. The NPVC calculated by the proposed optimization algorithm in this project is less than the previous optimization approach by $1.75M. Moreover, the time taken for computation of the algorithm was 112 seconds, which is 50 times lower than the former optimization methodology.

Date of Award	Sep 2019
Original language	American English
Supervisor	Yasser AlWahedi (Supervisor)