Spatio-temporal modeling of wind energy and applications to optimal operation of battery storage

TransAlta公司是一家总部位于卡尔加里的发电公司。他们拥有并经营太阳能、风能、水力、天然气和煤炭发电设施，其业务遍布加拿大，总发电量为8775兆瓦。其大部分发电机位于艾伯塔省，他们是艾伯塔省电力市场的参与者。在这个项目中，我们将重点关注与TransAlta在阿尔伯塔省电力市场上的运营有关的决策过程。TransAlta每小时面临的主要决策问题是确定各种电源的价格和数量对。另一个并行的决策问题是电池存储。因为TransAlta可以选择购买电力并将其储存在电池单元中，然后在需求较高或供应较低的时候出售。我们在这个项目中的目标之一是使TransAlta能够将风力发电的时空方面整合到总风力发电的预测中。从长远来看，这将有助于TransAlta更好地量化与预测相关的错误，并更有效地对冲与预测错误相关的风险。此外，我们的目标是推导出电池存储决策问题作为马尔可夫决策问题（MDP）的公式，该问题可以使用现代优化软件来解决。

TransAlta Corporation is an electricity power generation company with headquarters in Calgary. They own and operate solar, wind, hydro, gas and coal power generation facilities, and their operations stretch across Canada with a total power generation capacity of 8775 MW. A large portion of its generators are situated in Alberta and they are participants in the Alberta electricity market.In this project, we will focus on the decision processes pertaining to TransAlta's operations in the day ahead electricity market of Alberta. The main decision problem that TransAlta faces on an hourly basis is determining the price quantity pairs for various power sources. Another decision problem that runs in parallel is battery storage. Because, TransAlta has the option to buy electricity and store it in battery units, and sell it at a later time when the demand is higher and/or the supply is low. One of our goals in this project is to enable TransAlta to integrate the spatio-temporal aspect of wind power into the forecasting of the aggregate wind power. In the long term, this will help TransAlta to better quantify the errors associated with the forecasts and hedge more effectively against the risks associated with the forecast errors. Also, we aim to derive a formulation of the battery storage decision problem as a Markov Decision Problem (MDP) which can be solved using modern optimization software.