https://doi.org/10.1016/j.seppur.2021.118959
”
Fig. 3(a) shows the optimization results considering only the performance of the FG-WW column and of the NH3 desorber, in the plane ωFG−WW–PrFG−WW: Greater NH3 concentration in the solvent circulating within the CO2 capture section, thus higher NH3 concentration in the CO2-depleted FG entering the NH3 absorber, increases the requirements of the FG post-conditioning section, in terms of both energy consumption and size of the FG-WW column.
Fig. 3. Resulting pareto fronts of the optimization of the FG-WW column belonging to the state-of-the-art benchmark configuration for different apparent NH3 concentration values in the solvent circulating in the CO2 absorber-CO2 desorber loop, i.e. low cˆNH3 (light blue), mid cˆNH3 (blue) and high cˆNH3 (dark blue), in the plane: (a) specific equivalent work associated with the FG post-conditioning section and the NH3 desorber vs. productivity of the NH3 absorber; (b) total specific equivalent work vs. total productivity; and, (c) as of (a) but referenced to the NH3 captured in the FG-WW column. The performance at the operating conditions of the FG-WW column provided in Table 3 (hereinafter referred to as “Base” operating conditions for the reference configuration) is shown by means of the circles filled in grey at low productivity values. The optimal performances of the FG-WW column in the state-of-the-art benchmark configuration for each solvent concentration are shown as filled circles. Such optimal performances have been selected qualitatively as the last pareto point where the process productivity is still significantly increased with a minor increase of the specific energy consumption. (For interpretation of the references to colour in this figure caption, the reader is referred to the web version of this article.)
The results provided in Fig. 3(b) show that the performance of the FG post-conditioning section (and of the NH3 desorber) has a significant impact on the performance of the full CO2 capture process: While Pr can be increased by more than 50% with respect to the base operating conditions of the FG-WW column, this would be achieved at the cost of increasing ω up to 10%. Nevertheless, Pr can still be significantly increased with respect to the base operating conditions of the FG-WW column at a low energy cost, regardless of the cˆNH3 level. As far as the comparison among different cˆNH3 cases is concerned, the optimization of the reference configuration of the FG post-conditioning section improved the most its performance for higher NH3 concentration levels in the process with respect to the base set of operating conditions. As a result, while the mid cˆNH3 case performed the best at the base set of operating conditions of the FG-WW column in terms of specific equivalent work and productivity, the high cˆNH3 case with optimized FG-WW column conditions outperformed the rest in terms of Pr, approaching the PrCO2abs value reported in Table 3.
On the contrary, the high cˆNH3 case outperforms the lower NH3 concentration cases in terms of both productivity and specific equivalent work associated with the FG post-conditioning, if referred to the amount of NH3 captured in the FG-WW column, as shown in Fig. 3(c). That implies that the removal of NH3 in the SNA is still efficient when the NH3 concentration in the CO2-depleted FG is as high as 8,000 ppmv and hence, that cˆNH3 values above 9 molNH3kgH2O−1 might further increase Pr with an acceptable increase of ω if the FG-WW column of the reference CAP configuration is optimized as in this work—our previous optimization work showed that increasing cˆNH3 improves the productivity of the CO2 absorber with a small increment of energy demand [14].
In addition, the optimal values for each FG-WW column decision variable along the pareto fronts shown in Fig. 3 are plotted in the Supplementary Material. Although the effect of the FG-WW column decision variables on the process performance is subject to complex interdependencies, qualitative trends of the optimal conditions can be inferred:
- •
-
Each decision variable shows similar trends when comparing among different cˆNH3 cases.
- •
-
Decision variables reach their optimal values within the studied ranges, except for (L∕G)FG−WW, which stays at the lower bound for lower productivity values; values of (L∕G)FG−WW below 0.1 kg kg−1 have not been considered in order to avoid operational issues at industrial scale associated with liquid maldistribution.
- •
-
In general, we can move towards higher productivity values while minimizing the increase in energy requirements by simultaneously decreasing cˆNH3FG−WW, increasing (L∕G)FG−WW, decreasing TleanFG−WW, increasing (Lbot∕G)FG−WW and decreasing TpaFG−WW. These trends increase the capacity of the solvent to absorb NH3 from the FG. On the other hand, the relative height at which the pumparound stream enters the FG-WW column, hpaFG−WW, reaches a maximum and then decreases for increasing Pr values.
- •
-
However, the former trends suffer from discontinuities at intermediate Pr values, which stem from the change of operation regime of the FG-WW column. As mentioned above, higher Pr values can only be achieved by decreasing the pumparound and lean stream temperatures and by increasing the liquid-to-gas flowrate ratio along the NH3 absorber, which decreases the temperature of the FG exiting the column. Therefore, a purge stream to avoid water accumulating in the FG-WW column-NH3 desorber loop is required at high Pr values, which constraints the NH3 concentration in the NH3-lean stream entering the top of the SNA. On the contrary, no purge stream to control the water balance is required at low Pr values, which allows for the selection of the optimal cˆNH3FG−WW value; as soon as the constraint on the NH3 concentration of the NH3-lean stream becomes active, a step or discontinuity is noticed in the trends of the optimal values of the remaining decision variables.
“