https://doi.org/10.1016/j.seppur.2021.118959
”
Power plants and most of the CO2 emitting industrial sectors, e.g. cement and steelwork plants, do not use strong acidic (or basic) solutions within their manufacturing process. Therefore, the advanced configurations for the control of NH3 emissions aim at avoiding the requirement of acid washing thus the handling and consumption of aqueous H2SO4 solutions. As a result, three different advanced configurations for the control of NH3 emissions have been proposed, as shown in Fig. 2. All of them modify the FG post-conditioning section of the reference configuration (process section A in Fig. 1) by introducing operating and/or design modifications into the FG-WW column, which enable the removal of the acid washing. Further details are provided below for each advanced configuration of the FG-WW column.
Configuration A1 with the SNA, which still considers the FG-WW column configuration of the reference CAP but allows for the removal of the acid-wash column by tuning the operating conditions of the SNA. The SNA consists of two sections: (i) a lower section whose goal is the removal of the bulk amount of NH3 contained in the CO2-depleted FG coming from the CO2 absorber, and (ii) an upper section in which the NH3 concentration in the FG is further decreased to values in the order of magnitude of tens to hundreds of ppmv. While the upper section of the NH3-absorber in the reference CAP configuration decreases the NH3 content in the FG, yNH3FG−WW,out in Fig. 2, to values around 200 ppmv, the removal of the acid-wash column downstream requires to reach concentrations below 10 ppmv instead, in order to meet the NH3 emissions limits at the stack. The NH3-lean solution regenerated in the NH3 desorber is chilled before entering the top of the NH3 absorber. While flowing down along the column, the aqueous NH3 solution captures NH3 and CO2 from the FG flowing upwards in counter-current, increasing both its NH3 and its CO2 concentration. The NH3-rich stream obtained at the bottom of the NH3 absorber is sent back to the NH3 desorber, where NH3 and CO2 are stripped off from the liquid solution, regenerating the NH3-lean stream to be used again in the NH3 absorber. A fraction of the NH3-rich solution leaving the bottom of the NH3 absorber is recycled to an intermediate stage of the column after being chilled. The goals of this pumparound stream are: (i) to approach the NH3 concentration in the FG leaving the lower section of the NH3 absorber to the NH3 partial pressure in equilibrium with the pumparound stream using high enough a liquid-to-gas flowrate ratio, and (ii) to maximize the NH3 removal in the lower section of the NH3 absorber by decreasing the temperature of the liquid solution.
Fig. 2. Flow diagram of the FG post-conditioning section of the CAP without acid-wash column using different configurations of the FG-WW column with respect to the reference process. The operating variables that govern the performance of the configuration, i.e. the decision variables of the optimization problem, the process specifications and constraints, and the input variables have been colour-coded in
,
and
, respectively. (For interpretation of the references to colour in this figure caption, the reader is referred to the web version of this article.)
In order to assess the performance of the process configuration when removing the acid-wash column using the SNA, a multi-objective optimization problem of the FG-WW section has been solved for each cˆNH3 case (thus for different NH3 concentration values in the CO2-depleted FG entering the column, yNH3FG−WW,in). As a result, the optimal set of operating conditions and the SNA design are obtained, aiming at minimizing the specific equivalent work associated with the FG-WW section, ωFG−WW, and at maximizing the productivity of the NH3 absorber, PrFG−WW. In practice, such multi-objective optimization problem is run as a series of single-objective rigorous optimization problems based on Successive Quadratic Programming (SQP) whereby the minimum ωFG−WW is found when varying the FG-WW column height, HFG−WW [m]. The decision variables of the optimization problem, which are shown in Fig. 2, are the FG-WW column decision variables introduced in Section 2, i.e. cˆNH3FG−WW, TleanFG−WW, (L∕G)FG−WW, fsFG−WW and TpaFG−WW, along with the relative height at which the pumparound stream enters the packing of the NH3 absorber, hpaFG−WW [-], defined as:(5)hpaFG−WW=HpaFG−WWHFG−WWwhere HpaFG−WW [m] is the height of the packing of the FG-WW column at which the pumparound stream is fed. Additionally, the liquid-to-gas flowrate ratio for the bottom section of the SNA, (Lbot∕G)FG−WW [kg kg−1], can be computed from fsFG−WW and (L∕G)FG−WW in order to provide a parameter that facilitates the understanding of the operation within the SNA, as follows:(6)(Lbot∕G)FG−WW=(L∕G)FG−WW1−fsFG−WW
Besides the specification of yNH3FG−WW,out, constraints to the optimization problem are that: (i) the NH3 content in the stream purged from the FG post-conditioning section (see Fig. 2), wNH3purge, is below 150 ppmwt, which effectively constrains the value of cˆNH3FG−WW unless there is no purge stream, i.e. FpurgeFG−WW=0kgs−1, and (ii) TleanFG−WW and TpaFG−WW are, on the one hand, lower than the temperature of the streams entering their corresponding chillers but, on the other, higher than 1.5 °C in order to enable the use of simple industrial refrigeration systems.
Configuration A2 with the NAW. The NAW consists of three sections: (i) a lower section where the bulk of NH3 contained in the CO2-depleted FG is removed from the gas, as for the SNA, (ii) an intermediate section in which the NH3 concentration in the FG is further decreased to values of a few tens of ppmv, and (iii) an upper section where the NH3 concentration is decreased to values below 10 ppmv. With this aim, in this configuration the acid-wash packing section included in the FG post-conditioning section of the reference configuration is exchanged for a water-wash packing section, which is added to the top of the NH3 absorber. A water make-up stream at ambient temperature enters the top of the column and captures NH3 from the FG flowing upwards in counter-current. As a consequence of the increase in NH3 concentration, the liquid is expected to capture also some CO2 from the FG. The polluted water exiting the bottom of the water-wash section of the FG-WW column is then fed to the top of the intermediate packing section, together with the NH3-lean stream. As for the rest of the column, the NAW is similar to the SNA.
In order to assess the performance of the NAW, a multi-objective optimization of the FG-WW section has been carried out following the same strategy and approach detailed above for the SNA. As far as the decision variables are concerned, we consider eight in total, which are shown in Fig. 2; in addition to those considered for the optimization of the SNA, we have included the following operating variables of the NAW: (i) the liquid-to-gas flowrate ratio for the make-up water stream entering at the top of the water-wash section and the CO2-depleted FG entering at the bottom of the FG-WW column, (L∕G)WW [kg kg−1], and (ii) the relative height at which the NH3-lean stream enters the packing of the FG-WW column, hleanFG−WW [-], defined as:(7)hleanFG−WW=HleanFG−WWHFG−WWwhere HleanFG−WW [m] is the height of the packing of the FG-WW column at which the NH3-lean stream enters the packing of the NAW. Regarding constraints to the optimization problem, wNH3purge is limited to values below 150 ppmwt in all cases for this process configuration, since a purge stream is always required. This effectively limits cˆNH3FG−WW to values below 0.01 molNH3kgH2O−1.
Configuration A3 with the ANA, which is a slight variation of the NAW. The FG-WW column consists of the same three sections, but the polluted water leaving the bottom of the water-wash section is withdrawn from the column instead of being mixed with the NH3-lean stream entering the NH3 absorber. Instead, the polluted water stream is partly recycled to the top of the column, partly purged from the process to avoid the accumulation of NH3 and CO2 in the liquid flowing through the water-wash section of the ANA. Therefore, a water make-up stream is required at the top of the FG-WW column to maintain the liquid-to-gas flowrate ratio in the water-wash section. While in the case of the NAW polluted water is mixed with the NH3-lean stream thus cleaned in the NH3 desorber, the ANA aims at avoiding the additional steam required by the NAW for water cleaning. Consequently, the amount of NH3 captured from the gas in the water-wash section of the ANA is limited by the maximum NH3 concentration allowed in liquid streams purged from the process, i.e. below 150 ppmwt. In practice, the ANA consists of two different sections that operate independently, i.e. a SNA (Configuration A1) followed by a water-wash section, which improves its operability and controllability with respect to the NAW configuration.
As far as the performance assessment of the configuration including the ANA is concerned, a multi-objective optimization of the FG-WW column has been carried out following the same strategy and approach as for the SNA and for the NAW. In this case, seven decision variables have been considered in the optimization problem, as shown in Fig. 2. These are the same decision variables used for the optimization of the configuration with the NAW, except for (L∕G)WW. While increasing (L∕G)WW is penalized in the NAW configuration by a greater reboiler duty in the NH3 desorber, it does not have any direct effect on the specific equivalent work of the ANA configuration. Therefore, including (L∕G)WW as decision variable for the optimization of the ANA configuration would lead to values enabling the capture of all gaseous NH3 in the water-wash packing section of the ANA. As a consequence, no energy consumption would be required in the NH3 desorber, at the cost of uncontrolled water make-up requirements. On the contrary, if the flowrate of water make-up is constrained, the process optimization would lead to the maximum allowable (L∕G)WW value. Therefore, (L∕G)WW has been considered as a parameter instead of as a decision variable in the case of the ANA optimization. This implies that the multi-objective optimization of the process configuration including the ANA has been carried out for different values of (L∕G)WW, covering the process water consumption typical of amine-based post-combustion CO2 capture processes, i.e. ≈500kgtCO2captured−1 for aqueous MEA applied to the same inlet FG composition considered in this study [39]. As for the process constraints, these are the same defined for the SNA configuration, along with the NH3 concentration limit in the purge stream from the water-wash section of the FG-WW column.
“