【作者】 王春义；

【导师】 刘玉田；

【作者基本信息】 山东大学 ， 电力系统及其自动化， 2010， 博士

【摘要】 随着电网互联和远距离交直流输电技术的发展,大容量机组、超(特)高压设备、电力电子装置以及分布式发电技术等在电力系统中的大量引入,系统规模日益扩大,动态行为愈加复杂,由局部故障引发电力灾难的风险大大增加。2003年下半年至今,在北美和加拿大、英国伦敦、瑞典和丹麦、意大利、希腊、科威特、俄罗斯首都莫斯科、西欧8国以及巴西等国家和地区都先后发生过大停电事故。我国近20年来在各大电网发生的停电事故也有100余起。国内外电力系统的运行实际充分说明,新设备和新技术固然能为系统的安全稳定运行提供有力保障,但无法从根本上避免大停电事故的发生。由于现代社会对电力供应的依存度越来越高,大停电事故造成的后果日趋严重,一旦发生停电事故,必须采取措施安全和快速地恢复系统供电。由于大停电事故发生的概率极小,调度员缺乏实战经验。事故发生后,为使其在时间紧、任务重和压力大的情况下,能够给出正确的恢复决策,保证系统的安全快速恢复,调度部门迫切需要切实可行的电力系统恢复优化决策方法,以及能为调度员提供辅助决策和恢复培训的软件系统。因此,应当分析系统恢复问题的特点,完善系统恢复策略,研究能够综合协调系统安全和恢复速度的系统恢复优化决策方法,以解决现有方法在处理大规模电力系统恢复问题时灵活性和综合协调能力方面的不足；设计开发具有统一数据平台,且可与EMS／DTS系统接口的电力系统恢复辅助决策与培训仿真系统,以弥补现有电力系统恢复辅助决策系统在数据一致性、仿真分析工具的完备性、图形和数据维护工作量大等方面的不足,拓展DTS系统在电力系统恢复培训方面的理论和功能。上述工作的开展对于丰富电力系统的恢复控制理论,完善电力系统的应急指挥系统,提高电力系统应对停电事故的能力和减少停电损失都具有重要的理论意义和应用价值。在广泛阅读电力系统恢复与辅助决策方面相关文献的基础上,论文针对以上几个方面开展了深入的研究,主要研究工作和创新成果如下：1)针对电力部门对电力系统恢复辅助决策支持系统的实际需求,提出了基于数据仓库的输电网架恢复群体智能决策支持系统框架。该系统包括综合数据平台、智能控制子系统和决策支持子系统,采用三段式输电网架恢复策略和多属性效用理论综合考虑各属性评价值并做出最终恢复决策；可以综合考虑网架恢复过程中的暂态过电压、工频过电压、自励磁、负荷恢复引起的频率波动、机组／负荷特性及其重要程度、恢复时间、线路可用传输容量、设备故障概率等约束条件和影响因素。该系统的综合数据平台能够解决恢复过程中各安全校验和分析评估算法所需数据和模型不一致的问题。采用的分布式计算技术将各计算任务交给不同的计算机节点处理,能够加快网架恢复方案的生成速度。智能控制子系统和决策支持子系统能够保障网架恢复过程的安全性与快速性。该系统既可用于制定网架恢复方案和调度员培训,以便应对停电事故后的系统恢复,又可在实际恢复过程中为省级或地区级调度部门的决策者提供在线支持。2)针对网架恢复的优化决策问题,结合分层分区的输电网架恢复策略及各分区内网架恢复方案的制定流程,提出了基于多属性决策理论的输电网架恢复决策方法。该方法在决策前进行方案筛选,减少了计算量；利用几何平均法对多个决策者的方案评价值进行聚合,使其支持群体决策；结合最小加权法和信息熵算法,使选取的组合权重能够综合考虑决策者的主观意志和属性评价值的客观信息；利用欧氏距离和非线性规划,对方案排序结果进行了敏感性分析,给出了最优方案的绝对敏感指数及确定权重时需重点关注的属性。该方法可用于系统恢复的在线优化决策,它能够综合协调系统安全和恢复速度,给出每一恢复步骤的优化决策,提高恢复决策的可行性和灵活性。山东电网的仿真结果表明了本文方法的有效性。3)针对黑启动方案的自动生成、校验和评估问题,以及对调度员进行黑启动培训的应用需求,提出了在DTS系统中扩展黑启动方案制定与培训仿真子系统的设计方案,以及黑启动操作方案的生成算法,完成了与该子系统相关的开发工作,实现了黑启动方案制定与培训仿真子系统与DTS系统的跨平台集成。该子系统可与DTS系统共享数据库、图形库、人机交互界面,以及通讯平台等功能模块,避免了独立型黑启动辅助决策支持系统庞大的数据和图形维护工作量。扩展后的DTS系统可根据停电后的电网运行状态自动生成黑启动操作方案,并对其进行校验、调整和评估,克服了黑启动预案更新速度慢和灵活性差的缺点；可利用黑启动方案或教案对调度员进行黑启动培训,提高其应对停电事故的能力；可对工频过电压、暂态过电压、自励磁、大型辅机启动引起的电压和频率波动进行仿真,可根据黑启动方案自动验证黑启动过程中保护定值和安全自动装置的设置是否合适,保证黑启动方案的可行性。山东电网的仿真结果表明该系统具有较好的应用前景。4)针对抽水蓄能电站提供黑启动服务时的最优预留水量问题,建立了能够综合考虑黑启动效益、静态效益和边际运行位置约束的抽水蓄能电站黑启动最优预留水量计算模型；提出了抽水蓄能电站的黑启动效益、静态效益、黑启动所需水量以及边际运行位置的计算方法。利用山东电网的黑启动现场试验数据,计算获取了泰安抽水蓄能电站#1机组在超低负荷工况下的发电效率,并以试验数据为基础计算了泰安抽水蓄能电站提供黑启动服务时的黑启动效益、静态效益和黑启动耗水量,分析了影响黑启动效益、静态效益和黑启动耗水量的主要参数,可为泰安抽水蓄能电站的运行调度提供参考。该方法将黑启动服务引入抽水蓄能电站的运行调度,完善了抽水蓄能电站的运行调度模型。该方法也可应用于其他以抽水蓄能电站作为黑启动电源的系统。更多还原

【Abstract】 With the development of power grid interconnection and long-distance AC and DC transmission technology, high-capacity units, super (ultra) high voltage equipments, power electronic devices and distributed generation technology are largely introduced into power system. Power system becomes larger and its dynamic behavior becomes more complex than before. The risk of blackout caused by partial failure has increased significantly. Since 2003, several large blackouts have occurred in North America and Canada, England, Sweden and Denmark, Italy, Greece, Kuwait, Moscow, Western Europe, Brazil and other countries. In recent 20 years, there were also more than 100 blackouts occurred in China. The actual operation of power system at home and abroad shows that new equipments and technologies can make power system operate more safely and stably, but the occurance of blackout is unavoidable. As modern society becomes increasingly dependent on electricity supply, consequences caused by blackouts become more and more serious. Once blackout occurred, measures must be taken to restore the power supply as soon as possible.As the probability of blackout is extremely small, dispatchers are commonly lack of restoration experience. After blackout, in order to make right decisions and guarantee power system restored safely and quickly, utilities urgently needs a feasible power system restoration decision-making method and the software which can provide power system restoration decision support and training simulation for dispatchers. Therefore, it is necessary to analyze the characteristics of power system restoration problem and improve restoration strategy. The decision-making method which can comprehensively coordinate system security and restoration speed should be studied to improve the existing methods on flexibility and coordination ability in dealing with large scale power system restoration. The power system restoration decision support and training simulation system, which has unified data platform and can integrate with EMS/DTS, should be designed and developed to improve the existing system in data consistency, simulation tools completeness and graphic and data maintenance. These studies have important theoretical significance and application value in expanding the theory and function of DTS system on power system restoration. They can enrich power system restoration control theory, enhance power system emergency command system and reduce blackout losses. Based on literature review of researches on power system restoration and its decision support, thorough studies on skeleton restoration decision-making method, power system restoration decision support and training simulation system and the optimal reserved water of pumped storage power station have been made in this dissertation. The main research work and innovative fruits are as follows.1) According to the actual needs of utilities on power system restoration decision support system, a group intelligent decision support system framework for power system skeleton restoration based on data warehouse is proposed. This system is composed of three parts:they are comprehensive data platform, intelligent control subsystem and group decision support subsystem. A three-stage skeleton restoration strategy is proposed to control restoration process. According to evaluation values of every benefit-oriented attribute, the multi-attribute utility theory is employed to dynamically determine the target and plan of next restoration step. The system can comprehensively consider the constrains such as transient overvoltage, sustained power frequency overvoltage, self-excitation, frequency fluctuation, units and loads characteristics, restoration duration, available transfer capability of transmission line and equipment failure probability. The comprehensive data platform can guarantee the data needed by every security verification and assessment algorithm are consistent. By using distributed computing technology assign computing tasks to different computers, restoration plan generation rate is accelerated. Intelligent control and decision support subsystem can ensure the safety and rapidness of the restoration process. Simulation results of a practical power system demonstrate that the system not only can generate feasible cases for case-based reasoning, but also can provide online supports in actual restoration process.2) Combining the subsystem paralleling restoration strategy with the restoration plan development processes, a novel group multi-attribute decision-making method for power system skeleton restoration is presented. The proposed method improves the technique for order preference by similarity to an ideal solution (TOPSIS) in supporting group decision-making. Subsequently, the combined weights can comprehensively take into account the subjective preference of decesion maker and the objective information of attribute values by combining the weighted least-square method with Shannon entropy theory. Furthermore, the ranking results sensitivity of the improved TOPSIS is analyzed and the absolute sensitive index of the optimal plan was presented by using Euclidean distance and non-linear programming method. This method can comprehensively consider the safety margin and restoration rate of power system, and can be used to support the restoration decision-making of a subsystem. Simulation results of Shandong Power Grid in China show the proposed method is feasible.3) A black start plan generation and training simulation subsystem is designed on the data and graph platform of DTS, to which the tables, fields and metafiles are added, according to the actual requirements of utilities on black start plan generation, checking, evaluation and dispatcher training. A new black start operation plan generation algorithm is proposed. This subsystem has been developed and integrated with the cross-platform DTS system. It can share database, graphics library, human-machine interface, communications platform and other modules with DTS system. In this way, the large maintenance workload of data and graphics in independent black start decision support system is saved. The extended DTS generates black start plans according to the power system conditions after blackout. Then, these plans are checked, adjusted and evaluated. The settings of protections and automatic devices are automatically verified. The system has the training simulation functions of power frequency overvoltage, transient overvoltage, self-excitation, the voltage and frequency fluctuation caused by load or large motor restoration. Simulation results of Shandong power grid show its good application prospect.4) In the power grid with a larger proportion of thermal generator set, the pumped storage power station needs not only to provide black-start ancillary service but also to meet the demands of peak load shifting, frequency regulating and so on. However, the reserve water which is used to meet the needs of black start has a direct impact on the benefits of peak load shifting. In order to obtain the largest benefits of pumped storage power station, the calculation methods for black-start benefits, static benefits, marginal running position and black-start consumed water are proposed. Considering the black-start benefits of pumped storage power station, static benefits and the constraint of marginal running position, the mathematical model of optimal reserve water is constructed. Using the data acquired from the black start field test in Shandong power grid, the generating efficiency of Taian pumped storage unit #1 in ultra low load condition is calculated, and the black-start benefits, static benefits and black-start consumed water of it is computed. Then, the main influence factors on black-start benefits, static benefits and black-start consumed water are analyzed. These results can provide reference to the operation of Taian pumped storage power station. It is the first time to introduce black start service into the operation scheduling of pumped storage power station, which improves the operation scheduling model of pumped storage power station. This method can be used in other power grids which use pumped storage power station as black start power.更多还原