【作者】 刘玉娇；

【导师】 蒋传文；

【作者基本信息】 上海交通大学 ， 电力系统及其自动化， 2013， 博士

【摘要】 短期运行是可再生能源发电运行的最关键环节之一，但是可再生能源发电的出力随机性使其短期运行决策伴随着一定的风险，可再生能源发电的决策者必须对这种风险进行控制才能制定出合理的短期运行计划。不同运营环境下上述风险的承受者和决策者是不同的，因此可再生能源发电的短期决策需要依据不同的风险承受者分别建模和求解。另外近年来可控负荷（如储能，电动汽车，智能家居等）发展迅速，其能够为可再生能源发电的优化提供更多的可利用资源，因此本文将在考虑可控负荷因素的前提下对不同运营环境下的可再生能源发电的短期优化及其风险管理展开研究。可再生能源发电优先入网的环境下风险承受者和管理者是系统管理人员，非市场环境和可再生能源优先市场环境都属于这个范畴，本文首先针对非市场环境展开研究。这种环境下可控负荷的投资主要源于系统运行侧，因此此时考虑可控负荷因素的可再生能源发电的短期优化就变成了含可再生能源发电与可控负荷电力系统的短期优化。本部分研究内容从系统运行角度出发，将包含储能、电动汽车、可控消费负荷等在内的可控负荷作为控制量融入传统的经济运行模型，以机会备用约束的置信度作为系统风险度量和管理的手段，提出了基于机会备用约束的含可再生能源发电与可控负荷电力系统的短期优化运行模型并求解，研究成果可为含可再生能源发电与可控负荷系统提供短期优化及风险管理的工具。在可再生能源优先的电力市场环境下风险承受者仍然是由整个系统承担。由于能量市场的决策模型与非市场环境类似，因此本文第二部分研究内容落点于市场环境下含可再生能源系统的旋转备用优化决策及其风险管理上。首先以旋转备用容量成本、弹性消费负荷的响应成本和传统负荷的停电损失为综合成本，以综合成本期望和条件价值风险为经济性度量指标，建立了考虑弹性消费负荷的含可再生能源发电系统的旋转备用优化和风险管理模型并采用改进的多目标克隆免疫算法和模糊风险决策方法进行求解。在此基础上为了辨识可再生能源发电所需备用占系统总备用的比例以便于能源管理部门制定宏观决策，本文提出采用“备用需求贡献”概念进行备用需求分摊的方法，该方法能够有效的区分市场每个参与方包括各可再生能源发电的备用需求。此外，无论机会备用约束还是基于成本效益的备用决策模型都会使系统面临一定的小概率但有可能是巨额损失的尾部风险，为了进一步减少系统面临的风险，本文还提出了基于保险理论的尾部风险转移的方法，该方法能够给系统提供通过支付保费来减少所面临的小概率厚尾风险的选择。为了鼓励可再生能源自我进步，许多国家在给可再生能源发电一定补贴的情况下让其自由参与电力市场竞争，这种情况下可再生能源发电不确定性造成的风险将由可再生能源发电设备的拥有者自行承担。考虑可再生能源发电设备的拥有者主要为以下两种：一是大规模的可再生能源发电商，二是含小规模可再生能源发电设备的微网。本文的第三部分研究以二者为研究对象对完全市场环境下可再生能源发电的短期优化运行展开研究。首先提出了市场环境下含可再生能源和可控负荷的微网的优化运行模型，所建模型同时考虑了储能设备的负荷转移功能和能量备用功能并采用多目标优化和模糊决策方法求解。然后针对含储能设备的大规模可再生能源发电商建立了基于条件价值风险或基于收益乐观值为风险度量的最优竞标模型并求解，之后为了满足具有主观风险态度决策者的决策需求，本部分研究首创的将累积前景理论用于可再生能源发电商的风险决策中，丰富了当前可再生能源发电商的短期优化决策模型。更多还原

【Abstract】 Short operation is one of the most critical factors for optimizing the utilization ofintermittent renewable energy generations (RGs). But the uncontrollable characteristic ofRGs adds risks on their short-term operation, the decision-makers must take those risksinto account in order to make a good plan. Because the risk takers and managers are not thesame under different operating environments, the optimization of short-term operation ofRGs should be modeled and solved depending on different risk takers. Furthermore, therapid development of controllable loads in recent years (such as energy storage, electricvehicles, smart home, etc.) offers more available resources for optimizing RGs, so thisthesis carries researches on short-term optimization and its risk management of RGs fordifferent operating environments with the consideration of controllable loads.Risk takers of RGs are the whole system and the risk manager is the operators ofpower grid when governments force the power system to meet generating demand of thoseRGs. Both non-market and RG priority market are belonged to this category and the firstpart of the thesis focuses on the non-markets environment. Since the investment ofcontrollable loads in this environment are mainly due to the power grid, such as the batteryswitch stations of electric vehicles, intelligent community and energy storage stations inour country are most invested by the State Grid. So short-term optimization of RGs in thisenvironment is actually the short-term optimization of power system with RGs andcontrollable loads. From the system’s view, it proposes a chance reserve constrainedoptimization model for short-term operation of those power systems by adding the storages,electric vehicles and other controllable loads into the traditional economic operation modeland utilizing the confidence level of chance reserve constraint as the risk measure of theschedule plan. These researches can offer an optimization and risk management tool forpower system with RGs and controllable loads.The risk taker is still the whole power system in RG priority markets. The decision-making of energy markets is similar with the non-market environment, so thesecond part of this thesis pays attentions on making decisions of spinning reserve markets.It builds a multi-objective optimization model with two objectives: expectation andconditional value at risk of tatal cost, which consists of costs of reserve capacity, outagelosses and the cost of elastic demands. The proposed model is solved with the improvedmulti-objective immune algorithm and a risk decision-making method based on fuzzytheory. In order to identify the reserve demand of renewable generations from the totaldemand of power system, this section carries researches on the reserve allocation of powersystem with renewable generations. It proposes a new method to allocate the total reserveby proposing a concept of reserve demand contribution. Furthermore, the reserve plansdecided by risk management tools will face a problem of tail risk which is with smallprobability but large losses, so the third part of this section proposes a method to solve thetail risk by covering insurances which can offer a choice for power system to change theuncertainty losses into a fixed losses by paying insurance fees.In order to promote technological advance of RGs, many countries give them somegenerating allowance and put renewable generations into the competitive markets. Theowners of RGs will bear the risks caused by their uncertainties in this situation. So thethird section of the thesis focuses on the optimization of RGs under a competitive market.It chooses two types of RG owners as the research objective: one is the microgrid withRGs and the other is the large scale renewable energy power producers. It first proposes amulti-objective short-term optimization model for those microgrids with the considerationof risk aversion by energy storage and solves it with a improved multi-objective immunealgorithm and fuzzy decision theory. Then it proposes an optimal bidding model for largescale RGs with the conditional value at risk or the optimal value of profits as risk measures.And in order to satisfy the demand of decision makers who want make decisions with theirown risk experiences, the last part of this section puts forward a new decision model basedon the prospect theory which enriches the short-term optimization model of large scaleRGs.更多还原