Stripper modeling for CO₂ removal using monoethanolamine and piperazine solvents
MetadataShow full item record
This dissertation seeks to reduce the energy consumption of steam stripping to regenerate aqueous amine used for CO₂ capture from coal-fired power plants. Rigorous rate-based models in Aspen Plus® were developed, and rate-based simulations were used for packed vapor/liquid separation units. Five main configurations with varying levels of complexity were evaluated with the two solvents. 8 m piperazine (PZ) always performed better than 9 m monoethanolamine (MEA). More complex flowsheets stripped CO₂ with higher efficiency due to the more reversible separation. Multi-stage flash configurations were competitive at their optimal lean loadings, but they had poor efficiency at low lean loading. The most efficient configuration was an interheated column, with more effective and distributed heat exchange. It had a secondary benefit of a cooler overhead temperature, so less water vapor exited with the CO₂. Using a rich loading of 0.40 mol CO₂/mol alkalinity in 8 m PZ, the optimal lean loading was 0.28 and the energy requirement was 30.9 kJ/mol CO₂. Case studies were also performed on cold rich bypass and the use of geothermal heat. When cold rich bypass is used with the 2-stage flash and 8 m PZ, it reduces equivalent work by 11% to 30.7 kJ/mol CO₂. PZ benefited the most from cold rich bypass because it had a higher water concentration in the overhead vapor than with MEA. In an advanced 2-stage flash with 8 m PZ, geothermal heat available from 150 down to 100 °C requires 35.5 kJ work/mol CO₂. The heat duty and equivalent work was higher than other optimized configurations, but it would be a valid option if separating the heat source from the steam cycle of a coal-fired power plant was highly valued. Pilot plant campaigns were simulated with the available thermodynamic models. Two campaigns with 8 m PZ were simulated within small deviation from the measured values. The average absolute errors in these campaigns were 2.5 and 2.7%. A campaign with 9 m MEA in a simple stripper demonstrated that the MEA model did not predict the solvent properties well enough to appropriately represent the pilot plant operation.