Process synthesis and controllability assessment of CO2 capture plants in a parallel environment
The objective of this work is to develop, implement and evaluate a parallel computational framework for the simultaneous process synthesis and controllability assessment of absorption/desorption processes for postcombustion CO2 capture. The framework employs a stochastic optimisation algorithm which is able to handle efficiently discrete design variables, pertaining to process flowsheet structural features represented through a generic superstructure. The discrete design parameters are introduced iteratively into a deterministic optimisation algorithm which is efficient for continuous design variables and operates internally within the stochastic algorithm. Every solution obtained by the continuous algorithm is transferred into a controllability assessment stage, implemented in the form of a non-linear sensitivity analysis approach which evaluates the effect of disturbances within an optimum control scheme. This layout is realized within a synchronous, parallel realization of a Simulated Annealing algorithm, where the primal-dual interior-point optimisation algorithm, as implemented by the Interior Point Optimizer (IPOPT) software, is used for steady-state process design and the predictor-corrector homotopy-continuation algorithm, using the PITCON software, for controllability assessment. The obtained results show that the parallelisation scheme is computationally very efficient and the obtained solution is 52 % better in terms of overall performance than a corresponding, conventional sequential process design and control approach.