Análise transiente e avaliação de um sistema de controle para uma planta de captura de CO2 via contatores de membrana

Abstract

Hollow fiber membrane contactors (HFMCs) are efficient in separating CO2 from postcombustion gases, given their high gas-liquid (solvent) contact surface area. It combines membrane and absorption technologies, where the level of technological maturity is expanding and knowledge of the process dynamics is crucial for scaling up. Therefore, accurate models of HFMCs for carbon capture are necessary to understand and optimize their operation, but most of those available are steady-state and are limited to the analysis of design parameters. Furthermore, as membrane gas absorption processes become more consolidated and their applications grow, there is a need to evaluate control strategies to better operate them. In this work, two rigorous dynamic models for an HFMC were evaluated, validated with experimental data, an isothermal model and a non-isothermal model. Therefore, the HFMC was inserted into a CO2 capture plant (plant equipment: absorption tower, integrated heat exchanger and buffer/solvent tank) replacing the main equipment of the system, the absorber tower. Dynamic simulations were performed and demonstrated that the nonlinear model reacted well to disturbances, reaching steady states in line with practical stability limits. Conventional control strategies were inserted into the process, which include a ratio controller for the CO2 capture rate-𝜃(%) (CC) and two feedback controllers, the buffer tank temperature controller-TBT (TC01) and the rich solvent outlet temperature controller - TC2 (TC-02). Furthermore, to overcome the multivariable problem (3 x 3), strong nonlinearities of the distributed parametric model (set of partial differential equations-PDEs) and operational constraints of the system, an adaptive predictive controller based on quadratic programming (AQDMC) was inserted. Thus, according to the results presented, it was verified that the CC, TC-01 and TC-02 controllers were able to reject disturbances and, after correction with the PID control action, maintain the variables at their setpoints. However, it was noticed that, for some disturbances inserted in the process, the CC controller was not able to reject the inverse response dynamics of and/or limit the variable 𝜃(%) to its maximum value. Finally, it was noted that the AQDMC, compared to CC, TC-01 and TC-02, allows smaller fluctuations for variables 𝜃(%), 𝑇 𝐶2 , 𝑇𝐵𝑇, smaller under/overshoot percentages and meets the barrier restrictions imposed on the system, demonstrating to be more efficient and a potential control system for the carbon capture plant under study.