Demonstrating full-scale post-combustion CO₂ capture for coal-fired power plants through dynamic modelling and simulation

Abstract

This study aims to provide insights into the design and operation of full-scale post-combustion CO₂ capture for a 500 MWe sub-critical power plant through dynamic modelling and simulation. The development and validation of the dynamic models of the power plant and CO₂ capture plant are described. In addition, the scale-up of the CO₂ capture plant from pilot plant scale (where it was validated) to full scale is discussed. Subsequently the manner in which the two plant models were linked is discussed. A floating IP/LP crossover pressure configuration is used. A throttling valve is included between the LP turbine and draw-off point to prevent pressures at the crossover from dropping below required levels in the reboiler for solvent regeneration. The flue gas from the power plant is treated before it is sent to the CO₂ capture plant. Four case studies are considered. The first investigates the effect of increasing solvent concentration on the performance of the power plant with the capture plant. The second investigates which absorber packing height offers a good balance between capital and operating costs. The two dynamic case studies show that the CO₂ capture plant has a slower response than the power plant. They also reveal an interaction of CO₂ capture level and power plant output control loops making it difficult to achieve steady power output levels quickly.

Introduction

Chemical absorption of CO₂ using MEA solvent has been recommended as a suitable technology for post-combustion power plants especially for retrofit purposes. Although the process technology has been applied in natural gas sweetening, the scale of process required to achieve up to 90% CO₂ capture in fossil fuel-fired power plants is typically several times larger than what is commercially available at present. For instance, Mitsubishi Heavy Industries (MHI) has built some of the largest plants that process up to 450 tonnes of CO₂ per day from natural gas fired boilers [1]. A 500 MWe supercritical coal-fired power plant operating at 46% efficiency (LHV basis) [2] would release over 8000 tonnes of CO₂ per day. To separate CO₂ from flue gas at that scale, a relatively large chemical absorption facility is required which would be a significant capital investment to the operator.

The operation of a power plant integrated with a chemical absorption process would present additional challenges. The efficiency of such plants would drop significantly because steam that would have been used to generate electricity is drawn off to regenerate the solvent. Several studies have demonstrated that significant energy penalties would be incurred with the inclusion of carbon capture technology [3], [4], [5]. In addition, resultant reduced steam flows to the low pressure (LP) turbines would ultimately result in reduced pressures upstream of this point in the turbine. This and the possible process modifications required are described in the study. Another concern is whether such power plants could continue to play their role in meeting electricity demand. Coal-fired power plants currently operate flexibly in meeting varying electricity demand. Having power generation processes that could operate flexibly would be increasingly important with the growth of renewable power generation. Though renewable sources are virtually carbon neutral, they suffer the drawback of being intermittent in nature. Such variations in power generation increase the requirement for flexible operation in other plants on the same electricity grid. Hence, both the coal-fired power plant itself and the downstream CO₂ chemical absorption process would have to be capable of flexible operation.

Design and operational studies are typically carried out with pilot and larger scale demonstration plants. Current pilot plant studies worldwide are on a much smaller scale than would be required commercially. Even at such scales (typically less than 5 MWe), the cost of construction for these facilities typically runs up to several million dollars [6]. Once built, these plants are limited in the range of studies that could be carried out. Full scale demonstration projects are estimated to cost over a billion dollars [6]. A lot of useful insights could be derived from accurate dynamic models of the post-combustion capture process at a much lower cost.

Most process models developed to study cost and performance implications of carbon capture and storage (CCS) have been steady state models which cannot account for the various transients associated in the power generation process. Transients occur during plant start-up and shutdown operations. Load-following operations are common in coal-fired power plants. Operational problems could be further compounded if there are tight restrictions on CO₂ emissions. The downstream absorption plant may have to closely follow load changes. Finally, to improve overall efficiency, increased process integration of the power generation process and capture process would be required. This would likely further complicate the operation of the integrated facility. Insights regarding the integrated plant operation could also be provided through studies using dynamic modelling and simulation.

A number of studies have been carried out on dynamic model development of the chemical absorption plant. Kvamsdal et al. [7] and Lawal et al. [8] present the dynamic model development and simulation of the absorber only. Lawal et al. [8] also discusses different types of models used for modelling reactive absorption and the developments made in this regard. Rate-based models are shown to be more accurate than equilibrium-based ones. Ziaii et al. [9] describes the dynamic model development and simulation of the regenerator only while Lawal et al. [10] extends this to the two stand-alone absorber and regenerator column models. Analysis on stand-alone columns may be inaccurate due to the inevitable coupling of the two columns linked with a recycle loop. Lawal et al. [11] describes the dynamic model development of the chemical absorption process (absorber and regenerator linked by recycling the solvent). This model, however, was developed and validated at pilot plant scale, three orders of magnitude smaller than what is required for processing flue gas from a coal-fired power plant generating 500 MWe. Lawal et al. [12] investigated the performance and dynamic response of the chemical absorption process downstream an enhanced oxygen coal power plant and demonstrated how there was room for further improvement of performance by addressing the increased absorber temperatures.

A number of studies explored the dynamic response of power plants. Lu [13] combined a dynamic model with steady state correlations to simulate power plant component dynamics in MATLAB/SIMULINK. The components could be linked to form a power plant. The dynamic model developed was not validated though some simulation results for a whole plant were shown. Åström and Bell [14] derived a model from first principles to describe the drum, downcomer, and riser components of a natural circulation drum-boiler. Bhambare et al. [15] modelled from first principles a 250 MW coal-fired natural circulation boiler. The boiler system was divided into seven submodels: downcomer, riser, waterwall, drum, superheater and reheater, attemperator, and furnace. Chaibakhsh et al. [16] developed a dynamic model of a sub-critical once-through Benson type boiler based on the experimental data obtained from a complete set of field experiments. Genetic algorithm (GA) was executed to estimate the model parameters and fit the models response on the real system dynamics. Stephenson et al. [17] developed dynamic models of a sub-critical coal-fired power plant in gPROMS and validated the results with plant data. Power plant models described by these authors do not include integration with a CO₂ capture plant.

Steady state models of the CO₂ capture plant integrated with the power plant have been developed. A steady state, in-house model was used by Aroonwilas and Veawab [18] to compare the performance of different process configurations and different solvents (including solvent blends) for the chemical absorption plant integrated with a supercritical power plant. Sanpasertparnich et al. [19] carried out a study of the integration of the two plants using a steady state model. The model also considered CO₂ compression. Another steady state integration study was carried out by Cifre et al. [20]. In this study, the impact of CO₂ compression was considered as well and efficiency penalties estimated were up to 16%.

This study was conducted with a modelling and simulation approach. The gPROMS (Process Systems Enterprise Ltd.) advanced process modelling environment has been used to implement the proposed work. Dynamic models of the CO₂ chemical absorption plant and a 500 MWe sub-critical coal-fired power plant were developed. A sub-critical plant was used rather than a supercritical one because there was plant data available for dynamic validation obtained from one of the plants RWE npower operates. The two dynamic process models were linked together to carry out a unique study of their integrated operation.