A competitive Markov decision process model for the energy-water-climate change nexus

作者：Nanduri Vishnu^*; Saavedra Antolinez Ivan

来源：Applied Energy, 2013, 111: 186-198.

DOI：10.1016/j.apenergy.2013.04.033

• CMDP
• Carbon and water taxes
• Reinforcement learning
• Climate change
• Bidding strategies

Drought-like conditions in some parts of the US and around the world are causing water shortages that lead to power failures, becoming a source of concern to independent system operators. Water shortages can cause significant challenges in electricity production and thereby a direct socioeconomic impact on the surrounding region. Our paper presents a new, comprehensive quantitative model that examines the electricity-water-climate change nexus. We investigate the impact of a joint water and carbon tax proposal on the operation of a transmission-constrained power network operating in a wholesale power market setting. We develop a competitive Markov decision process (CMDP) model for the dynamic competition in wholesale electricity markets, and solve the model using reinforcement learning. Several cases, including the impact of different tax schemes, integration of stochastic wind energy resources, and capacity disruptions due to droughts are investigated. Results from the analysis on the sample power network show that electricity prices increased with the adoption of water and carbon taxes compared with locational marginal prices without taxes. As expected, wind energy integration reduced both CO2 emissions and water usage. Capacity disruptions also caused locational marginal prices to increase. Other detailed analyses and results obtained using a 30-bus IEEE network are discussed in detail.

• 出版日期2013-11