Steam methane reforming with carbon capture and storage (SMR-CCS) is a vital route for producing "blue" hydrogen. In 2018, a baseline SMR-CCS process design (BPD) was modeled by DOE/NETL to produce blue hydrogen via steam natural gas reforming. The BPD included a pressure swing adsorption (PSA) unit for H2 purification and employed two chemical solvents for the CO2 capture. The objective of this work is to modify and model the BPD in Aspen Plus V. 12. 1 to improve its performance and flexibility and reduce costs. The modified process design (MPD) included a hydrogen hollow fiber membrane (HFM) and a PSA unit and employed thirty-six physical solvents for CO2 capture. The Capital Expenditure (CAPEX), Operating Expenditure (OPEX), Fuel Expenditure (FEX), Levelized Cost of CO2 Capture (LCOC), and Levelized Cost of H2 Production (LCOH) of the MPD were calculated, and techno-economic analysis (TEA) of the process was performed. The calculated values were used to develop an Artificial Neural Network (ANN) in MATLAB to efficiently predict the TEA of the MPD. The Aspen Plus modeling proved that physical solvents were effective for CO2 capture; a comparison between the BPD and MPD revealed that the MPD had superior performance metrics, including higher H2/CH4 and H2/CO2 selectivities, lower CAPEX, and lower LCOH. The 1AB-DECAM physical solvent was the most economically viable one used in the MPD due to its lowest LCOH (0. 93/kg H2). Also, using ANN as a machine-learning (ML) tool offers significant time savings in screening physical solvents and predicting TEA of the MPD for blue hydrogen production.
Wang et al. (Thu,) studied this question.