【作者】 胡军峰；

【导师】 闫庆友；

【作者基本信息】 华北电力大学 ， 技术经济及管理， 2014， 博士

【摘要】 全球气候变化是全世界都面临的重大挑战,其中人类活动所排放的C02数量急剧增加是最主要的原因,其对环境的影响也越来越重要。在世界C02排放总量中,中国已经成为全世界C02排放量最多的国家,占了20%以上的比重,而且比重还有进一步加大的趋势。因此在未来全世界控制C02排放的努力中中国必将承担越来越大的压力和责任。其中电力行业C02排放占了中国排放总量的40%左右,是C02排放总量最多的行业,因此电力行业C02强度下降对于中国2020年C02减排目标的实现至关重要。大规模发展可再生能源对电力行业C02减排具有积极作用,中国可再生能源也取得了超常规发展。可再生能源包括风力发电、太阳能发电、潮汐能发电、地热发电、生物质能发电等,目前中国具备大规模发展条件的主要是风力发电,所以本论文主要研究风力发电的发展。由于风力发电具有波动性、不确定性以及不可调峰性等特征,在接入电网发电时需常规电源提供调峰和备用等辅助服务,才‘可以满足电网调度的要求。随着风力发电的大规模发展,其对常规电源辅助服务的需求也将越来越大,常规电源是否可以提供足够的辅助服务对于风力发电的发展也至关重要。原电监会的统计显示,2010年1-6月份,风电弃风导致损失电量27.76亿千瓦时,占同期风电上网电量的12.47%,其最主要的原因就在于电力系统能够为风电接入提供辅助服务的能力不足。因此,提高电力系统为风电接入提供辅助服务的能力,对于风电真正入网发电,具有非常重要的意义。但是,提高电力系统辅助服务能力需要较大投入,单纯依靠电力系统自身很难承担,同时风电大规模接入电网所带来的电力系统辅助服务成本提高由电力系统承担也有失公平,需要在此基础上进一步研究风电接入所带来的辅助服务成本提高的分摊和补偿机制,激励电力系统为风电接入提供辅助服务。只有这样,电力系统辅助服务能力才能真正提高,风电发电入网也才能够真正得到保障。因此,为研究上述问题,本论文分以下内容进行分析。首先对风电接入导致的辅助服务成本进行分析,得到风电接入导致的辅助服务成本的特性及其影响因素,然后利用静态合作博弈方法分析辅助服务成本如何在风电场之间进行分摊,并对各种静态合作博弈分析结果进行对比讨论,在静态合作博弈分析基础上,进一步研究利用动态合作博弈方法分析辅助服务成本如何在风电场之间动态分摊,最后在辅助服务成本分摊静态合作和动态合作博弈模型的基础上,得到适合中国的风电接入导致的辅助服务成本分摊机制。大规模风电接入电网主要影响的辅助服务有调峰、备用和无功。随着风电发电技术的发展,尤其是双馈感应异步风力发电机的出现,其自身可以提供无功补偿,对电力系统无功辅助服务的需求很小。因此风电接入对电力系统无功辅助服务的影响可以忽略。本论文将重点研究风电接入对调峰和备用辅助服务的影响。风电接入对辅助服务影响程度的分析主要采用“有无对比法”,即通过对比风电是否接入电网,两者之间的差值即风电接入所带来的影响。风电接入对不同的辅助服务影响的原因也不相同。调峰辅助服务主要是由于风电出力波动性所带来的,因此风电接入导致的调峰辅助服务应该采用无风电接入负荷持续曲线与风电接入净负荷持续曲线(负荷与风电综合)所导致的调峰辅助服务的差值进行计算。备用辅助服务主要是由于风电出力预测偏差所带来的,因此风电接入导致的备用辅助服务应该采用无风电接入预测偏差与风电接入综合预测偏差所导致的备用辅助服务的差值进行计算。同时本论文进一步发现只要两个风电场出力的相关系数不为1,风电场联合接入电网导致的调峰辅助服务成本和备用辅助服务成本就比风电场单独接入电网导致的辅助服务成本之和要小,即风电联合入网会对调峰辅助服务成本和备用辅助服务成本产生“平滑效应”。因此需要进一步分析风电场联合接入电网导致的辅助服务成本如何在风电场之间进行分摊。本文进一步构建了风电接入电网导致辅助服务成本分摊静态合作博弈模型和动态合作博弈模型。静态合作博弈模型包括按照EANS方法分摊模型,Shapley值方法分摊模型,Owen值方法分摊模型,核仁方法分摊模型,动态合作博弈模型包括开环纳什均衡动态合作分摊模型、反馈纳什均衡动态合作分摊模型以及动态Shapley值合作分摊模型。模型对比分析结果表明,对于风电接入导致的调峰辅助服务成本而言,在数据不充分的情况下,按照电量分摊是一个比较可行的方案,在数据不断完善的情况下,则可以逐渐的过渡到按照核仁方法或者Shapley值方法进行分摊,而对于风电接入导致的备用辅助服务成本而言,按照电量分摊则完全不可行,初期可以考虑按照各自引致比例进行分摊,然后逐渐过渡到按照核仁方法或者Shapley值方法进行分摊。目前的辅助服务补偿机制采用发电厂按照上网电量(或者电费)的比例承担辅助服务成本,这种机制用来补偿风电引发的辅助服务成本不是特别理想,如果风电场也按照这种方式和火电厂一样承担辅助服务成本,则有失公平。因此辅助服务补偿机制需要为适应补偿风电辅助服务成本作进一步改进。对于风电接入导致的调峰辅助服务成本,需要事先知道非常详细的负荷每分钟波动状况以及风电每分钟出力状况以及风电出力之间每分钟的平滑效应,这就需要非常庞大的数据监测和数据积累,相应需要付出的成本也比较高。所以要想确定风电接入导致的调峰辅助服务成本在数据不完善的情况下很难确定出来,而且风电出力波动主要是由风能资源决定的,这一因素也主要是客观原因,不是风电场努力能够改善的。因此风电出力波动性所导致的调峰成本也不是风电企业自身努力可以改善降低的。所以初期可以考虑风电接入导致的调峰辅助服务成本不由风电场承担,而是改由消费者来承担。在有了充分的数据积累和良好的数据监测能力后再考虑由风电场承担各自所导致的调峰成本,初期可以考虑按照电量分摊,未来在逐渐过渡到更合理的按照核仁方法或者Shapley值方法进行分摊的方式。对于风电接入导致的备用辅助服务成本,其主要是由风电场的不确定性所导致的,这一不确定性可以通过提高风电预测水平进行改善,所以备用成本由风电场自身承担和补偿也可以激励风电场改进自身出力的预测精度,这也可以降低风电接入导致的备用辅助服务成本。备用成本补偿和分摊初期可以考虑按照风电场各自引致比例进行分摊,然后逐渐过渡到按照核仁方法或者Shapley值方法进行分摊。由于改革难度不同,因此本论文建议先易后难,实现风电场对其导致的辅助服务成本进行补偿和分摊。第一阶段可以考虑在不改变现有辅助服务考核和补偿办法的情况下增加对风电预测精度进行考核的内容,基于风电出力预测精度考核结果由风电场对风电导致的备用辅助服务成本进行补偿,风电导致的调峰辅助服务成本则通过上调火电标杆上网电价,进而通过销售电价进行补偿。第二阶段则可以考虑改变目前辅助服务考核和补偿办法的资金来源,打破成本补偿资金在发电系统内循环的状况,改为由电网公司承担考核后成本补偿资金的不足部分,并计入购电成本通过销售电价进行补偿,而不再间接通过上调上网电价来补偿辅助服务成本,同时风电承担自身引发的备用成本。由于通过辅助服务考核方式确定的辅助服务成本可能不太经济,没有达到成本最低的状况,第三阶段可以考虑建立电力现货市场,通过发电企业报价方式确定调峰辅助服务成本,在明确风电接入导致的调峰辅助服务成本状况的基础上,由风电按照其电量进行补偿和分摊,同时建立电力备用辅助服务市场,通过发电企业报价方式确定备用辅助服务成本,消费者和风电进行分别对其引发的备用成本进行补偿。当然对于风电接入导致的辅助服务成本如何进行补偿和分摊还有很多值得深入研究的内容,包括：1、风电导致的无功等其他辅助服务成本及如何补偿；2、辅助服务提供如何在不同技术之间进行优化；3、省间甚至是区域间如何对风电接入导致的辅助服务成本进行补偿；4、在考虑风电正外部性的情况下如何分摊风电接入导致的辅助服务成本；5、电网公司如何参与辅助服务成本补偿；6、在电力市场交易机制设计时如何避免市场力的影响；7、如何设计合理的辅助服务交易机制让风电场以合适的角色参与。更多还原

【Abstract】 Global climate change, resulting mainly from rapid increase of CO2emitted by human activities, is a considerable challenge facing the whole world and exerts increasingly significant impacts on environment. China is the world’s biggest CO2emitter, taking up20%of total amount. It is estimated that the proportion will continue increasing. China has an increasingly heavy responsibility to CO2mitigation in the world. In China, power sector is the biggest CO2emitter, accounting for about40%of the total emission in China. In order to realize the2020target for CO2emission in China, reduced CO2intensity in power sector is crucial.Large-scale wind power development plays an active role in CO2reduction in power sector. For this purpose, China’s wind power has witnessed rapid development. However, wind power is intermittent, uncertain, and cannot be moved down. In order to meet the requirements of dispatching, ancillary services such as peak-shaving and operational reserve from conventional power are required for integration. With the large-scale wind power development, demand for ancillary services provided by conventional power is increasing. The capacity of conventional power’s provision of ancillary services is crucial to wind power development. Former SERC statistics show that in the period between January and June in2010,2.776billion kWh of electricity was lost due to wind curtailment, taking up12.47%of the total integrated electricity in the same period. The major reason for the loss is the lack of ancillary services. Therefore, increased capacity of ancillary service provision in power system has very important significance towards wind power integration.Improved capacity of ancillary service provision in power system requires huge investment, which can hardly be afforded by power system itself. Also, it would be unfair to have the power system afford the cost. Hence, in order to incentivize ancillary service provision for wind power integration, further research is necessary on distribution and compensation mechanisms for the ancillary service cost. Only in this way can capacity of ancillary service provision in power system be enhanced, and wind power integration be guaranteed.Therefore, in order to study the above problems, this paper will conduct the following analysis. Firstly, the author will examine the ancillary services costs resulting from wind power integration to summarize its characteris and main influencing factors. Then the author will adopt static cooperation game theory to allocate the wind power incurred intergration ancillary service cost, with thorough comparison and disucssion on the anlaysis results. Base on the anlysis resutls from the staic cooperation game theory, the author will further use a dynamic cooperation game theory to allocate the ancillary service cost between wind farms in a dynamic modelling. Lastly, by combining the analysis results from both static and dynamic cooperation game theories, the thesis will come up with a wind power integration incurred ancillary services cost distribution mechanism that fits China’s context.In China, the scope of ancillary services covers primarily the peak-shaving, operational reserve and reactive services. Most of them share identical concepts with international ones. Along with the development of wind power generation technoglogy, particularly the invention of double-fed induction asynchronous wind turbines which can suffice its own reactive service, the power system requires little reactive ancillary service. Thus this paper will focus on the wind power integration’s impact on peak-shaving and operational reserve services.The main analyzing tool to the wind power integration’s degree of impact on the ancillary services is called "With&Without Antitheses", that is by comparing the difference of the power grid between wind power connected and disconnected situations to determine the wind power intergration impact. Furthermore, different types of ancilary services have different impacts on wind power integration incurred ancilary services cost. Peak-shaving ancillary service was mainly due to the volatility of wind power output, so that the peak-shaving ancillary cost brought by wind power integration should be the difference of the ancillary service cost between non-wind power load duration curve and wind power integrated load duration curve (load and wind power integrated). Operational reserve ancillary service was mainly caused by wind power output prediction error, so that the extra ancillary service cost brought by wind power integration should be the difference between operational reserve ancillary service cost of the non-wind power output prediction error and the wind power integrated output prediction error.Meanwhile, the paper found that as long as the correlation coefficient of two wind farms’ power output is not1, the total peak-shaving and operational reserve ancillary service costs of two wind farms’joint intergration is less than respective integration, which means the combined integration of wind power would have a "smoothing effect" on peak-shaving&operational reserve ancillary services. Therefore, it is necessary to further analyze the allocation of combined wind intergration ancillery service cost among wind farms. The paper further build wind power integration incurred ancillary service cost distribution in a static cooperative game model and a dynamic cooperative game model. Static cooperative game model consists of static methods in accordance with EANS allocation model, Shapley value allocation model, Owen value apportion model and nucleolus apportion model. While dynamic cooperative game model includes open-loop dynamic cooperative Nash equilibrium allocation model, dynamic cooperative feedback Nash equilibrium allocation model and Shapley value cooperation dynamic allocation model. The comparative analysis of the model results show that wind power integration peak-shaving inccured ancillary services cost, in the case of insufficient data, distribution in accordance with the electricity output is a more viable option; while in the case of improving data, gradual transition to the nucleolus method or Shapley value allocation method is feasible. On the other hand, for wind integration incurred operational reserve ancillary service cost, distribution in accordance with the electricity output is not feasible. It is suggested that at the initial stage, cost can be allocated in accordance with their power proportion integrated, and then gradually transfer to nucleolus allocation methods or Shapley value method.The current compensation distribution mechanism for ancillary service which in accordance with on-grid electricity output (or electricity tariff) proportion isn’t suitable for wind power integration inccured ancillary service cost distribution. It is unfair for the wind farm and thermal power plants to bear the cost following current distribution method. Therefore, the ancillary service compensation distribution mechanism needs to be adjusted and improved to fit the wind energy.To calculate the wind power peak-shaving ancillary services cost, it requires a costly large data monitroing and accumulaion which includes:every minute fluctuation of the power load, wind power output per minute and the wind power output per minute smoothing effect. Thus it is difficult to calculate the wind integration incurred peak-shaving ancillary service cost without sufficient data. However, wind power output fluctuations are mainly determined by the characteristic of wind energy resource that can not be overcomed by the wind farm. Therefore, at early stage, the wind power integration inccured peak-shaving ancillary cost is suggested to be borne by the consumer instead of the wind farms. After establishing a sufficient data accumulation and data monitoring capacity, the peak-shaving cost can be transferred back to the wind farm respectively. The distribution methods can start from distributing in accordance with the electricity output, and gradually transfer to a more rational nucleolus method or Shapley value apportion approach. Considering the difficultis at different stages of reform, this paper suggests to start the reform from the easier sector, and to achieve a rational distribution and compensation mechanism for the wind power integration incurred ancillary service cost. At the first stage, while keeping the current ancillary service assessment and compensation measures, it is suggested to add a wind power prediction accuracy assessment, based on which compensating wind farm’s operational reserve ancillary service cost, whereas the wind power inccured peak-shaving ancillary service cost can be compensated by raising the feed-in tariff that transmitted to the sales tariff. At the second stage, it is suggested to change the internal loop of fund rasing and compensating for the ancillaery service assesment and compensation within the power generation sector. And the grid companies will bear the shortfall of the compensation after assessment which can be reimbursed in the sales tariff, instad of feed-in tariff. And wind farm bears its self-inccured spare ancillary cost. However, the assessmet determined ancillery service cost is not economical, hence it won’t reach the lowest cost. At the third stage, with the establishment of electricity spot trade market, the peak-shaving ancillery cost will be determined by power generation companies offer. With sufficient data and knowledge of wind integration incurred peak-shaving ancillary service cost, the wind power compensation and distirubtion will be based on wind electricity output proportion. On the other hand, while building a power operational reserve ancillary service market, the generation company’s offer can determine the ancillary service cost which the consumer and the wind power will compensate their inccured ancillary costs respectivly.Further study on the compensation and distribution of wind power integration inccured ancillary service cost is required, including:1. The compensation to wind power inccured reacive and other ancillary service cost;2, Technologies to optimize ancillary service provision;3, The compensation to wind power integration inccured ancillary service cost among provinces and even among regions;4, The distribution of wind power integration inccured ancillary cost considering wind power’s positive externalities;5, grid companies invovlement in ancillary service cost compensation;6, The design of the electricity market trading mechanism to avoid the impact of market power;7, The design of a reasonable ancillary service mechanism that promote wind power’s appropriate participation.更多还原