Author: Charles A. Kang

Publisher:

Published: 2015

Total Pages:

ISBN-13:

Climate change mitigation will require large reductions in CO2 emissions from electricity production. Some of these cuts will come from increased use of renewable energy resources, but it is likely that thermal power plants will be used for an extended period of time to maintain grid stability and accommodate seasonal variability in renewable generation. Therefore, thermal power plants with CO2 capture and storage (CCS) capability may coexist with renewable generation to provide reliable low-carbon electricity. Moreover, CCS-enabled facilities designed for constant operations are not necessarily optimal under the conditions that are likely to occur with increased renewable penetration. There is therefore a need to devise optimal designs and operating plans for flexible thermal power stations equipped with CCS. In this work, computational optimization is used to determine the design and operating plan of a coal-natural gas power station with CO2 capture, under a CO2 emission performance standard. The facility consists of a coal power plant undergoing a retrofit with solvent-based post-combustion CO2 capture. The heat for CO2 capture solvent regeneration is provided by a combined cycle gas turbine (CCGT) designed for combined-heat-and-power service. Variable facility operations are represented by discrete operating modes dispatched using the electricity price-duration curve. Two problem formulations are considered. In the `simplified-capture' problem formulation, the CO2 capture system is represented using a single variable for capacity, while heat integration (including a detailed treatment of the heat recovery steam generator component of the CCGT) is optimized jointly with variable operations. In the `full-system' problem formulation, the detailed design of the CO2 capture system is optimized alongside a full treatment of heat integration and variable operations. To accomplish this, a computationally efficient proxy model of the CO2 capture system is developed that reproduces the behavior of a full-physics Aspen Plus model. Both problem formulations are incorporated in a bi-objective mixed-integer nonlinear program in which total capital requirement (TCR) is minimized and net present value (NPV) is maximized. Pareto frontiers are generated for six scenarios constructed from recent historical data from West Texas, the United Kingdom, and India. All six scenarios are considered using the simplified-capture problem formulation. The West Texas base scenario and the India scenario, which differ greatly from each other, are considered using the full-system problem formulation as well. Results between the two formulations are quite consistent and show that hourly electricity price variability and the choice of objective function can have a large effect on optimal design and planned operations. In the West Texas base scenario, which has high price variability, the maximum NPV facility in the full-system formulation (NPV of $201 million, TCR of $510 million) has a time-varying operating plan in which the CO2 capture system has a utilization factor of 66% (out of a maximum of 85%). In this scenario the minimum TCR facility (NPV of $101 million, TCR of $333 million) has a constant operating profile. In contrast, low price variability in the India scenario results in constant operations regardless of objective. Two advanced CO2 capture processes -- the mixed salt and piperazine processes -- are considered using the simplified-capture formulation for the West Texas base scenario. The advanced processes are shown to outperform the standard monoethanolamine (MEA) process, with the mixed salt process outperforming the MEA process by 16% for maximum NPV and 14% for minimum TCR. The full-system formulation using the MEA process provides generally similar results to those from the simplified-capture formulation in both the India and West Texas base scenarios. However, the inclusion of the detailed design of the CO2 capture process in the full-system problem formulation provides valuable design information, such as the effect of the integer nature of the number of CO2 capture trains. Taken in total, the results of this study highlight the value of applying computational optimization to consider integrated plant design and variable operations together.