作者: Ziang Li , Zhengtao Ding , Meihong Wang
DOI: 10.1016/J.IFACOL.2017.08.454
关键词:
摘要: Abstract For the post-combustion carbon capture (PCC) process with MEA solvent, most of relevant literature discussed optimal operation under a cost-minimum target. A power plant integrated PCC process, however, prefer to maximize its lifetime cumulative profits. To profits, we may identify following issues. operation, trade-offs should be made between electricity output and energy-intensive process; for CO2 allowance bidding, fossil-fuel bid win adequate allowances from market balance demand emission. We apply Sarsa temporal difference (TD) learning algorithm search strategy that maximizes profits during above This includes both bidding plant. Our results show it is better than independently-designed fixed level. In addition, TD can find Sarsa(λ) if training data generated cheaply.