Performance and ANN-based optimization of an advanced process for wet CO2-to-Methanol using a catalytic fluidized bed reactor integrated with separators

•An advanced process for wet CO2-to-MeOH using a turbulent fluidized bed reactor was developed.•The advanced process was better than the conventional one in techno-economic analysis.•Performance analysis and artificial neural network-based optimization were performed.•Optimum operational domains were suggested for producing MeOH at 98.5–99.9 wt% purity.•An optimal design method was proposed for large-scale production of green methanol. Large-scale green processes to produce methanol from massive industrial sources of CO2 with a certain amount of water are increasingly attractive for carbon neutralization. In this study, an advanced CO2-to-Methanol (CTM) process, consisting of a turbulent fluidized bed (TFB) reactor, separators, and other units, was developed, analyzed from start-up to steady state, and optimized for high-purity methanol production (98.5–99.9 wt%) from a wet CO2 source (1.3 MtCO2/year with 1.35 mol% H2O). A two-stage flash drum with direct feeding of wet CO2 into a flash drum was applied to the developed CTM process to remove water and enhance unconverted gas recovery. At a methanol purity of 99.9 wt%, the advanced CTM process increased the methanol yield (0.15–0.31%), production capacity (5700–10100 tMeOH/year), and net CO2 reduction (0.0009–0.0033 kgCO2/kgMeOH) but decreased production cost (3.7–7.4 $/tMeOH) compared to the conventional CTM process under various of reactor temperatures (478–538 K) and pressures (40–90 bar). During the start-up process, the CTM process indicated the peaks in the methanol molar fraction and flow rate as the TFB reactor performed the stable profile of solid temperature (548–553.75 K) and volume fraction (∼0.249). In the performance assessment, the main units (compressor, TFB reactor, and distillation column) accounted for 96, 95, and 82.5% of capital expenditure (CAPEX), operating expenditure (OPEX), and energy consumption, respectively. Meanwhile, lowering the price of the annual CO2 and H2 increased CAPEX and OPEX contributions from 10.92 (present) to 28.64% (in 2050). To optimize the process using an artificial neural network (ANN) model, three main objectives and four main operating variables were selected as features through data analysis. The ANN model precisely predicted process performances (>99% accuracy) could reduce the computational cost by 900-fold compared to the process simulation. For methanol purity of 98.5–99.9 wt%, the CTM process should be operated in the optimum domain with a reactor pre