Demand-Response PEV Charging Control in Smart Grids for Renewable Supply Utilization and Hedging and Transfer of Forecast Risk

目前有两个主要的发展塑造电力市场。在供应方面，人们越来越多地推动使用可再生能源，但这与能源供应率的高度变化有关。在需求方面，插电式（混合动力）电动汽车（PEV/PHEV）的部署越来越多，这不仅会显著增加平均用电量，还会产生非常突发的需求模式。幸运的是，PEV在能源消耗率和时间表方面也提供了很大的灵活性，我们相信，供应方的变化可以通过灵活性部分“吸收”。这构成了该项目的广泛目标，旨在建立有效的PEV需求调度机制，以有效利用可变费率的可再生能源，并对冲电力公司（聚合商）的预测风险。这些机制将允许将与能源生产/供应的可变性相关的风险部分转移-从电力供应商和聚合商（进而是发电商）转移到消费者（PEV用户），以换取降低的充电价格。该项目针对PEV充电环境产生的解决方案也适用于其他弹性负载，这些负载与一段时间内部分可预测的能源消耗量相关。为了应对PEV的大量部署，并尝试大规模整合可再生能源发电，本项目考虑的问题对于电网的稳定性和盈利性至关重要。具体而言，本项目研究了智能电网中PEV充电的两种需求响应控制策略。在第一种方法中，公用事业公司（或聚合器）为PEV充电设置时间和位置相关的价格，同时让PEV（或与PEV充电相关联的智能仪表）基于PEV所有者的个体最佳利益和约束来主要进行调度。在第二种方法中，公用事业公司（在市场上）提供硬合同和软合同，以便在指定的时间间隔内将PEV充电到特定水平，PEV根据其需求和偏好购买该时间间隔;然后由公用事业公司本身完成PEV充电的调度，以遵守其已出售给PEV的合同条款。在此背景下，将探讨两个广泛的相互关联的研究问题。第一个问题涉及这些合同的设计，以及相关的“合同组合优化”问题--这涉及到决定公用事业公司应该提供多少硬合同和软合同，以达到其期望的风险回报权衡点。第二个涉及通过拍卖和其他“激励相容”的市场清算机制（即，从PEV调用真实的响应），并导致所得到的能量分配的总体“社会效率”。预测，需求响应控制在这两个定价/调度框架的PEV充电带来了挑战性的问题，可以使用随机博弈论，拍卖机制设计，随机微积分和风险管理中的工具和技术建模，但必须适当地调整和扩展，以考虑模型，约束条件和要求的PEV充电和可再生能源供应过程。

There are currently two major developments shaping the electricity market. On the supply side, there is increasing push towards the use of renewable energy sources, which are however associated with a highly variable rate of energy supply. On the demand side, there is increasing deployment of plug-in (hybrid) electric vehicles (PEVs/PHEVs), which will not only increase the average electricity consumption significantly, but also generate very bursty demand patterns. Fortunately, PEVs also provide significant flexibility in terms of their energy consumption rates and schedules, and we believe that variability in the supply side can be partly "absorbed" using the flexibility. This constitutes the broad goal of this project, which aims at efficient PEV demand scheduling mechanisms for effective utilization of variable-rate renewable energy sources, and hedging of forecast risks by electric utilities (aggregators). These mechanisms would allow the partial transfer of risk associated with variability of energy generation/supply -- from the electricity suppliers and aggregators (and in turn the generators) to the consumers (PEV users), in exchange of a reduced price of charging. The solutions resulting from this project for the PEV charging context would also be applicable to other elastic loads that are associated with a partly predictable amount of energy consumption over a time span. The issues considered in this project are critical for the stability and profitable operation of the electric grid as it copes up with the increased deployments of PEVs, and attempts to integrate renewable energy generation at a large scale.More specifically, this project investigates two demand-response control strategies for charging PEVs in the smart grid. In the first method, the utility (or aggregator) sets time- and location-dependent prices for PEV charging, while letting the PEVs (or the smart meters associated with PEV charging) mostly do the scheduling, based on the individual best interests and constraints of the PEV owners. In the second method, the utility offers (in the market) hard and soft contracts to charge PEVs up to certain levels within a specified interval, which the PEVs purchase based on their needs and preferences; the scheduling of the PEV charging is then done by the utility itself such that adheres to the terms of the contracts that it has sold to the PEVs. Two broad inter-related research issues will be explored in this context. The first involves the design of these contracts, as well as the associated "contract portfolio optimization" question - which involves deciding on the how many hard and soft contracts the utility should offer to attain its desired risk-return tradeoff point. The second involves the process of offering these contracts through auctions and other market clearing mechanisms that are "incentive-compatible" (i.e., invoke truthful response from the PEVs), and result in overall "social efficiency" of the resulting energy allocation. Predictive, demand-response control within these two pricing/scheduling frameworks for PEV charging brings about challenging issues that can be modeled using tools and techniques in stochastic game theory, auction mechanism design, stochastic calculus, and risk management, but must be suitably adapted and extended to consider the models, constraints and requirements of the PEV charging and renewable energy supply processes.