استراتژی های عملیاتی و مشارکت در نیروگاه های با ضبط کربن

For the post-combustion carbon capture (PCC) process with MEA solvent, most of the relevant literature discussed the optimal operation under a cost-minimum target. A power plant integrated with the PCC process, however, prefer to maximize its lifetime cumulative profits. To maximize the profits, we may identify the following issues. For the power plant operation, trade-offs should be made between electricity output and the energy-intensive carbon capture process; for the CO2 allowance bidding, a fossil-fuel power plant should bid and win adequate allowances from the market to balance its demand of CO2 emission. We apply the Sarsa temporal difference (TD) learning algorithm to search for a strategy that maximizes profits during the power plant lifetime with the above issues. This strategy includes both the operation and the bidding of the power plant. Our results show it is better than the independently-designed bidding and operation strategy with a fixed CO2 capture level. In addition, the Sarsa TD algorithm can find a better strategy than Sarsa(λ) if training data can be generated cheaply.