Assessment of Moving Bed Temperature Swing Adsorption Process for Post-Combustion CO2 capture

Abstract

In the context of post-combustion CO2 capture, adsorption-based processes are considered a promising alternative to absorption technologies thanks to their lower environmental impact, absence of corrosion problems, and potentially lower energy requirements. In conventional Temperature Swing Adsorption (TSA) processes, the adsorbent is packed in a series of fixed bed columns that cyclically alternate between the adsorption and regeneration steps to separate the CO2 from the rest of the flue gas components. One drawback of TSA systems for post-combustion capture is the large temperature swings usually required to achieve high product specifications in terms of both CO2 purity and recovery. These large temperature swings do not only imply high energy penalties and parasitic losses, but also lead to long cycle times, large system footprints, and low process productivity due to the poor heat transfer within the packed beds.

One way to overcome the aforementioned limitations is by performing the adsorption-desorption cycle in a Moving Bed Temperature Swing Adsorption (MBTSA) system, whereby the adsorbent is circulated through different sections that correspond to each cycle step. The main feature distinguishing moving and fixed bed processes is that the former can be operated at steady state. This is beneficial because it renders complex cycle schedules unnecessary and eliminates the parasitic energy losses associated with intermittent heating/cooling of the heat exchanger walls. The possibility to operate in steady state is also an advantage in terms of internal heat recovery and process integration, which is particularly relevant within post combustion applications where the power cycle can be adapted to supply the heat required by the capture process in an efficient way. In view of its potential advantages, and comparatively lower maturity level, the objective of this thesis is to investigate the application of the MBTSA technology in the context of post-combustion CO2 capture.

The design of an MBTSA system is a complex task involving a large number of inter-related process parameters such as the choice of adsorbent material, process configuration, operating conditions, and the size and geometry of the components. In this context, process modeling and simulation stands as an essential tool for the conceptualization and analysis of new MBTSA systems. In order to study the performance of the MBTSA technology for different applications, a one-dimensional model of the process was developed. The model was obtained by applying the mass, momentum and energy balances to the different sections of the MBTSA system, and it was implemented in the gPROMS environment. One distinguishing feature of the MBTSA model is that, instead of imposing a fixed wall temperature in the heating and cooling sections, it includes additional energy balances for the heating/cooling fluid and heat exchanger walls. In addition, the model accounts for the internal heat recovery achieved when coupling the preheating and precooling sections, which is necessary to reduce the external energy duty required by the process.

The performance of the heat exchangers employed to provide and remove heat from the sorbent depends largely on the sorbent-side heat transfer coefficient, as it is the dominating thermal resistance between the gas/solid phases and the heating/cooling fluid. The correct estimation of this parameter is therefore crucial for the design of MBTSA systems. In order to assess the technology with realistic parameters, the heat transfer coefficient on the sorbent side of a lab-scale MBTSA apparatus was measured at different flow rates and temperatures. The heat transfer coefficient increased with the flow rate of adsorbent particles, while no dependence was observed on sorbent temperature. The heat transfer coefficients obtained (70 – 120 W/m2K) were significantly higher than those typically encountered in fixed bed configurations (10 – 50 W/m2K). This confirmed that the moving bed configuration has the potential to address one of the main limitations of the fixed bed TSA technology for CO2 capture, namely, the low process productivity due to the slow heating and cooling of the adsorbent.

The results of the experimental campaign were used to develop a correlation for the sorbent-side Nusselt number as a function of the Péclet number. This correlation was incorporated into the MBTSA computational model, which was then used to design and analyse an MBTSA process for a waste-to-energy plant with a power output of 16.8 MWel, a thermal output of 64.6 MWth, and an exhaust flue gas flow rate of 56 kg/s with 11%vol CO2 concentration. Despite the low selectivity of the activated carbon adsorbent considered, the proposed MBTSA process reached high CO2 purity (97.2%) and capture rate (90.8%). These product specifications were achieved at the expense of adopting a high regeneration temperature (187°C) and solid-to-gas ratio (11.6 kg of adsorbent per kg of flue gas). Nevertheless, the designed MBTSA system was able to attain high process productivity (181 kgCO2/tadsh), which can be attributed to the short cycle time associated with the fast heating and cooling of the adsorbent. The results of the waste-to-energy case study indicate that the MBTSA technology is suited to capture CO2 at high purity and recovery, while achieving higher process productivity than fixed bed TSA processes. In addition, it is believed that the thermal energy required by the proposed MBTSA system may be significantly reduced by replacing the activated carbon material by other adsorbents having higher capacity and selectivity towards CO2, such as zeolites or metal organic frameworks (MOF).

This thesis also evaluated the suitability of the MBTSA technology in the context of power generation from natural gas. The case study considered was an 800 MW Natural Gas Combined Cycle (NGCC) power plant with an exhaust gas flow rate of 916 kg/s containing 5.15%vol of CO2. Two different MBTSA capture processes were proposed: one using a commercial zeolite 13X and other using a novel CPO-27-Ni MOF adsorbent. The two systems were able to meet the target specifications in terms of CO2 purity (>95%) and capture rate (>90%), while achieving higher productivity than conventional fixed bed TSA processes. Even if the separation performance of both processes was similar, the distinct physical properties of the adsorbents led to different system dimensions and operating conditions, demonstrating the flexibility of the MBTSA technology.

In addition, the influence of the capture system and its auxiliaries on the performance of the power plant was analyzed by integrating the MBTSA model with a process model of the NGCC power plant. The simulations showed that the process using CPO-27-Ni required more thermal energy for sorbent regeneration (125.6 vs 100.7 MWth). However, the energy penalty associated with the steam extraction used as thermal input was lower (25.8 vs 29.1 MWel) because the steam extraction was performed at lower temperature and pressure. This advantage was partially offset by the higher pressure drop in the adsorption section of the MOF process, which led to a higher power consumption in the flue-gas booster fans (17.3 vs 10.1 MWel). Despite the distribution of the energy penalties associated with the CPO-27-Ni and zeolite 13X processes was different, the net electric efficiency of the NGCC power plant was very similar in both cases. In particular, both MBTSA capture processes led to a reduction of about 7 percentage points with respect to the reference plant without CO2 capture.

Moreover, the proposed MBTSA processes were benchmarked against a state-of-the-art absorption process using monoethanolamine (MEA) as solvent. One of the main differences between the amine-based process and the MBTSA systems is that the latter require an additional energy input to dry the flue gas because the adsorbent materials considered (i.e., zeolite 13X and CPO-27-Ni) are incompatible with water. This drying process accounted for the 17% of the energy penalty associated with the MBTSA capture systems. By contrast, the thermal energy required, and hence the power penalty associated with steam extraction, was significantly higher for the MEA process. Despite the breakdown of energy penalties between the MBTSA systems and the MEA process was different, no significant difference was observed in terms of overall power plant performance. In particular, the net electric efficiency of the reference power plant was 63.1%, while the efficiency of the power plant with CO2 capture was 54.7% for the case of MEA, 56.1% for the MBTSA using CPO-27-Ni, and 56.2% for the MBTSA using zeolite 13X. These results suggest that the MBTSA process applied to NGCC power plants is suitable for capturing CO2 at high purity and high capture rate, while being competitive with the state-of-the-art MEA capture process in terms of energy penalty. Considering the much earlier stage of development of this technology with respect to the MEA process, the MBTSA seems to offer a large potential for process improvement and should be considered for further development.