Results from an initial solvent testing programme at a post-combustion carbon capture pilot plant at Niederaussem, Germany, will help to bring the technology to full-scale commercialization.

Process diagram of the Niederaussem pilot plant and online measurements for the detailed investigation of the process

The lignite fired 1000 MW Niederaussem power plant was the first of its kind in Germany to be fitted with carbon capture equipment. It was pre-assembled and then commissioned at the plant in July 2009 as a development corporation between RWE Power, BASF and Linde, with supporting funding from the German Ministry of Economics and Technology.

The pilot plant comprises optimized components such as a direct contact pre-scrubbing cooler unit for the flue gas as well as a lean liquid cooler and interstage cooler that both allow the solvent to be returned at different absorber heights. The carbon capture takes place downstream of the plant’s flue gas desulphurization (FGD) plants and is typically able to scrub 7.2 tonnes of CO2 per day from the exiting slipstream. There are two options for feeding flue gas to the CO2 scrubbing pilot plant, either from a conventional FGD from the 1000 MW unit, called BoA1, or from a high-performance FGD pilot plant, called REAplus.

The flue gas properties were also influenced by dry lignite co-combustion in the 1000 MW power plant (up to 30 per cent of the furnace thermal rating), where dry lignite is supplied by a prototype fluidized bed dryer with internal waste heat utilization (WTA technology). This means that the flue gas conditions are expected to be close to those of future pre-dried lignite fired power plants.

This initial testing stage lasted for six months, during which time the operators sought to discover the optimum configuration for post-combustion carbon capture at the plant using monoethanolamine (MEA) as benchmark solvent, while also assessing over an additional six-month period the performance of a new amine-based solvent, called GUSTAV200, that has been developed by BASF.

Currently available solvents for removing CO2 from flue gases are unable to meet all the demands of post-combustion capture technology. At present, the state of the art is MEA. This makes development of solvents better than MEA one of the most critical steps in making CCS technologies cost-effective.

Solvent selection

BASF chose its solvent after characterizing and evaluating around 180 others in experiments at its laboratories. After an in-depth evaluation of the lab screening results the most promising solvent candidates were selected for evaluation runs at the company’s small test plant, called mini-plant. This was followed by the development of a standard experimental programme to characterize one solvent regarding its performance and optimal operating conditions.

One of the mini-plant’s most promising experiments assessed the properties of the novel solvent GUSTAV200, which would later be tested in the Niederaussem pilot plant in comparison to MEA. Under mini-plant conditions, it requires 25 per cent less energy for regeneration and has a lower circulation rate than MEA.

In the mini-plant the tested the stability of the solvent in the presence of oxygen was tested for two weeks, during which time the concentration was analyzed daily using gas chromatography. This showed no change in the solvent composition and performance, which supports the findings of oxygen stability tests executed in the laboratory.

These promising results showed that GUSTAV200 had the potential, at least in theory, to fulfill the demands for post-combustion carbon capture while performing much better than MEA. Nevertheless, since the results were based on the mini-plant conditions, it was necessary to validate them in a full-scale pilot plant under real gas conditions and over a much longer period of time.

Real world conditions

The Niederaussem pilot plant is playing a vital role in developing carbon capture technology that could be commercially applied to full-scale coal or lignite fired plants, particularly in assessing the accuracy of the current process modeling predictions used to design a plant.

The pilot plant had some 250 measurement and online analysis systems installed to validate the results of detailed operating parameter variation tests, as well as an extensive sampling and measurement programme to allow the operational performance and the energy demand for CO2 capture to be precisely determined.

During each test run, the optimal operating parameters for a selected solvent were first identified and then maintained over an extended time. This long-term test assesses the solvent’s stability and performance over time. Emissions were measured after the capture process had been running for long enough for degradation products and trace elements introduced into the solvent by the flue gas to accumulate. This approach was replicated in the MEA phase and with GUSTAV200.

The testing programme using MEA officially started on 28 July 2009, although the plant had also been using it during the initial start-up and commissioning period the month before.

The main objective of the MEA test phase was to gain some hands-on experience with the new pilot plant process using a benchmark solvent. Comparatively good data on the physical-chemical properties of MEA allows a distinction to be made between progress through process optimization and progress through better solvent performance.

Extensive operating parameter variation studies using MEA (30 per cent weight, no inhibitors or additives) were launched at the end of July 2009 to determine the optimal operating parameters in terms of energy consumption. The long-term behaviour of the solvent under these optimized operational settings was then tested. After some 5000 hours of operation, the plant switched to the GUSTAV200 solvent and this comprehensive testing process continued for a further 4000 operational hours, starting from January 2010 and ending in July the same year.

The pilot plant was connected to the conventional FGD absorber for the first half of the GUSTAV200 assessment, and later switched to use flue gas from the high-performance FGD pilot plant. The test is of interest because different pre-treatment steps could result in different concentrations of contaminants in the flue gas that then affect performance.

A comparatively undisturbed, continuous operation of the pilot plant during the MEA and GUSTAV200 test phases reached an outstanding high availability of 97 per cent. Step-like changes observed in the solvent flow rates reflected systematic parameter studies for both solvents. During the test campaigns, the standard analyses of the solvent as amine and water concentration, CO2 load, foaming behaviour and organic acid concentration were performed at the laboratory on site, while more detailed analyses regarding organic and inorganic trace elements have been conducted by specialized laboratories at BASF Ludwigshafen. The test programme was also accompanied by a comprehensive material testing programme and trace element measuring campaigns regarding the gaseous and liquid inlet and outlet streams of the capture plant.

Temperature control

The flue gas inlet flow at the Niederaussem pilot plant is cooled down to typically 40 °C by a direct contact cooler at the absorber inlet. When operating without the high-performance FGD, acidic reacting trace components can be removed from the flue gas by adding sodium hydroxide to the recirculating water flow of the cooler (pre-scrubber).

The absorber is equipped with several beds and offers the option of by-passing beds to investigate the effect of different packing heights. There is also the option of interstage cooling by using different side draws. A water wash section at the top of the absorber consisting of a packed bed and a cooled water cycle is installed to reduce possible emissions due to volatile compounds and entrainment.

In the pilot plant, the treated flue gas and the captured CO2 are fed back to the existing flue gas duct of the power plant and vented to the atmosphere. Some characteristic design parameters of the pilot plant and typical operating parameters obtained during the test campaigns with MEA and GUSTAV200 are listed in Table 1.
An electric steam generator is used to heat up and evaporate part of the solvent. A slipstream of the solvent cycle is equipped with a mechanical and an activated carbon filter to remove possible particles, such as fly ash or gypsum from the FGD, or degradation products from the solvent flow. No reclaiming unit is currently installed, because the qualities of the solvents’ degradation products are not yet known, while it was also important to observe impact any contaminants had on performance and emissions.

Table 1: Design parameters of the Niederaussem pilot plant and typical flue gas parameters obtained in the test phases

The pilot plant in Niederaussem is equipped with around 250 measuring points that provide information for evaluating the experiments. For each set of conditions, a time frame of at least one hour, but normally two, of steady-state operation was selected for analysis of the experimental values. To guarantee the quality of the results, a maximum deviation of 5 per cent was selected for the global mass and energy balances.

Some tests conducted at the beginning of each solvent test phase confirmed the reproducibility and plausibility of the experimental values. For constant flue gas conditions, CO2 content, flow rate and also for a given pressure at the desorber, the circulation rate was varied until the optimal operating point was found in terms of energy demand for regeneration.

The results of these hysteresis tests show that, at less than 2 per cent deviation, the reproducibility of the energy values is better than the average deviation in the plant’s energy balance. Additionally, the typical shape for the energy versus circulation rate curve can be clearly recognised and the effect of the pressure on the energy demand is as expected for MEA. It can be concluded that the experimental results of the pilot plant in Niederaussem are highly reliable and that the pilot plant offers excellent options for studying CO2 removal from flue gases using amine-based solvents.

MEA and Gustav200

For the comparison of two different solvents, many experiments have to be performed in order to find the optimal operating conditions in terms of energy demand for each of them. Only upon completion of these experiments may the results of two different solvents be compared.

Variations in desorber pressure, lean solvent temperature and feed position, interstage cooler position and temperature, flue gas temperature, removal rate etc. were systematically tested under real flue gas conditions. At the same time, the solvent circulation rate was varied until the minimum energy requirement for a set of given conditions was found.

The results of variations in pressure at the top of the desorber were studied from 1.5 bara to 2.0 bara. For each pressure, the solvent circulation rate was varied to find the minimum energy requirement at a removal rate of 90 per cent. Under these conditions, GUSTAV200 had an energy demand of about 2800 MJ/t CO2, which is 20 per cent below that of MEA. Moreover, the circulation rate of GUSTAV200 is lower than that of MEA.

The study shows no dependence of the energy demand for GUSTAV200 on regeneration pressure. For MEA, a difference of about 4 per cent was measured between 1.5 bara and 1.75 bara, but no significant difference was found between 1.75 bara and 1.90 bara.

The energy demand for regeneration can be divided into four main contributions: strip steam, heating of the condensate reflux and of the solvent, and finally the CO2 heat of absorption/desorption. In general with increasing the regeneration pressure the strip steam demand decreases, while the CO2 heat of absorption/desorption increases with growing pressure because of the increasing temperature.

These two opposite effects could explain the fact that the dependence of GUSTAV200’s energy demand on the regeneration pressure was not measurable. To confirm this thesis the absorption enthalpy for GUSTAV200 was estimated based on an energy balance around the desorber and compared with the MEA values, which were calculated using BASF’s equilibrium model for MEA. This comparison shows a higher dependence of the heat of absorption/desorption for GUSTAV200 than for MEA, which corroborates the postulated thesis.

Based on these results, regeneration of GUSTAV200 appears possible at a lower temperature than MEA, without raising reboiler duty. The evaluation of these results must take account of the integration of the CO2 capture process into the power plant. On the one hand, a lower regeneration temperature is an advantage since lower-quality steam can be used; on the other, power consumption for CO2 compression will increase.

Interstage cooling is a well-known technique used in acid gas removal. The pilot plant at Niederaussem is able to test the benefits offered by interstage cooling for process performance and the optimization of the cooling temperature and cooling position in the column.

The experiments were carried out at a constant flue gas temperature and flow, and a removal rate of 90 per cent. The circulation rate was varied until the optimal operating point for energy regeneration was found. The difference in energy demand between the process configurations with and without interstage cooling for a given flue gas temperature and interstage cooling with standard cooling conditions is about 4 per cent for GUSTAV200 and about 3 per cent for MEA. An additional temperature reduction in the interstage cooler of about 10 K did not show any significant effect for MEA and GUSTAV200.

Based on these results it can be said that the interstage cooler is a cost-effective way to reduce the energy demand of the process.

The Niederaussem pilot plant has been an ideal choice to assess solvent and process performance, due to a design allowing highly flexible and extensive process variation studies, comprehensive measurement equipment and high operational availability. It has allowed a profound understanding of capture process configurations and solvent behaviour to be achieved.

The test data and simulation results for the MEA benchmark solvent and the GUSTAV200 are in agreement. The high reproducibility under real power plant operating conditions provides a reliable basis for sound process optimization and the design of large-scale capture plants. The much lower energy demand for the regeneration of GUSTAV200 means a huge step forward in post-combustion capture technology.

The article is based on a paper ‘The post-combustion capture pilot plant Niederaussem – Results of the first half of the testing programme’ published in Energy Procedia, which was co-authored by Peter Moser and Sandra Schmidt of RWE Power AG, Germany, Georg Sieder and Hugo Garcia of BASF SE, Germany, and Veselin Stamatovc of Linde-KCA Dresden GmbH, Germany. The author would like to thank his co-authors for their invaluable contribution.

More Power Engineering International Issue Articles
Power Engineering International Archives
View Power Generation Articles on