【作者】 王勇；

【导师】 韩学山；

【作者基本信息】 山东大学 ， 电力系统及其自动化， 2012， 博士

【摘要】 科技的进步、人们生活水平的提高及电力系统本身的特点,使得电网规模日益扩大,电网事故的后果也越来越严重,导致电力系统的可靠性和风险问题受到了越来越多的关注。用户负荷需求的波动和系统中设备的随机故障,使得电力系统的运行具有较强的随机特性,而电力市场的引入和可再生能源发电容量的快速提高,带来了更多的随机性影响因素,仅靠确定性的方法已无法满足现代电力系统运行决策的需求。进行某项决策时,必须要考虑尽可能多的运行情况,综合各个事件发生的概率及其后果,在保证整个电力系统处于可接受的风险水平下,审慎做出决断。经过数十年的研究与实践,通过将可靠性工程原理与电力系统的实际问题及电力系统数学模型、计算方法相结合,在基于事件发生概率的可靠性概念和评价指标的定义、计算模型及其参数的累计和统计等方面的研究与应用已经相当成熟。但是,受计算方法和系统复杂度的影响,目前的电力系统可靠性分析和风险评估研究在模拟的精确度和速度上还存在冲突,在可靠性灵敏度的概念和工程实用方法上还有待于进一步完善,尤其是在运行环境下,需要一种快速电力系统可靠性分析与评估方法,使得在有限的计算时间内,可以足够的精度获得电力系统当前的可靠性状态,并预测未来一段时间内的发展变化。在此背景下,论文对运行环境下电力系统可靠性分析与风险评价的模型和方法展开研究,其主要工作和创新成果体现在如下几个方面：(1)针对厂站主接线可靠性分析中,尚未有机考虑元件故障切除及故障恢复的问题,以非序贯Monte Carlo模拟为基础,提出将元件故障的切除和恢复有机衔接的厂站接线可靠性分析方法。该方法基于拓扑分析理论,首先获得距离故障元件最近的断路器割集,模拟故障切除的操作,并在此过程中考虑保护失灵和开关拒动的影响；然后,闭合所有非故障开关,并利用距离故障元件最近的开关割集,来模拟故障恢复的操作,分别统计两个过程中无法满足的功率需求,可以获得更合理的故障后果。该方法的特点在于提高可靠性分析精度前提下,既可避开模型复杂度的处理,又能使主接线可靠性分析贴近实际。通过算例的分析与比较,验证了方法的正确性,以及研究的必要性和算法的有效性。(2)提出广义电网(将输电线路、变压器等有阻抗元件和断路器、隔离开关等无阻抗元件一体化考虑)可靠性分析的蒙特卡罗模拟方法。首先通过厂站拓扑分析,得到等值节点及与有阻抗元件的连通性,然后,再进行网络拓扑分析,获得等值电网,由此计算并统计系统可靠性测度指标。在广义电网模拟过程中,依据邻接矩阵传递闭包阵不变的思想,采取快速局部拓扑技术,以减少计算量。算例分析表明,该方法获得的结果更加符合实际。(3)针对电力系统可靠性分析中蒙特卡罗模拟法需要大量计算资源和收敛速度较慢的缺点,从运行角度出发,在设备具备故障率指数分布特性前提下,基于马尔科夫链,建立可快速对电力系统运行可靠性解析分析的模型。其核心体现在：首先,提出便于工程实现的电力系统运行状态划分为3类的准则,及其对应的状态空间转移表达；然后,基于电力系统运行历史数据样本(或模拟数据样本)得到电力系统运行等值的马尔科夫状态转移概率矩阵。在此基础上,利用状态转移概率矩阵和电力系统当前的运行状态,可快速解析未来一段时间内电力系统运行状态变化概率、平稳状态概率及首次故障平均时间等可靠性指标,实现电力系统运行可靠性的快速评估。相关算例验证了电力系统状态转移的马尔科夫特性,表明了本研究的有效性。(4)对电力系统进行可靠性灵敏度分析可以辨识系统中的薄弱环节,对改善系统可靠性具有重要意义。针对传统可靠性灵敏度方法在理论与实践中的不足之处,通过分析可靠性灵敏度的图形表达和电力系统工程实际需求,提出了潜在提高灵敏度的改进分析方法,使之更加符合实际需要的同时,有效降低了可靠性指标计算误差对元件灵敏度评价结果的影响。同时,该方法与广义电网可靠性分析方法有机配合,可直接评价和比较电力系统中常用的一次设备的灵敏度,可更好地满足检修决策等工程实际决策的需要。更多还原

【Abstract】 With the development of science and technology, improvement of the living standards of the people and the characteristics of the power system, power network has been expanding. This enlarges the range of the power system accidents, which has been catching more and more spotlights about the reliability of the power system.The operation of power system has random characteristics for the volatile of user’s load demand and random failures of the element in the system, and power industry deregulation and fast-growing renewable generating capacity results in more highly stressed and unpredictable operating conditions. Deterministic approach can not satisfy the needs of modern power system. To make a decision must be taken into account operation scenes as much as possible, integrated all of the probability of occurrence of each scene and its consequences, and in ensure the risk of whole power system can be managed to an acceptable level.After decades of research and practice, it is already quite mature that the application and study of the reliability concepts and evaluation indices based on probabilities, reliability computational model, accumulation and statistic of component reliability parameters, through combination of basic reliability engineering principles and practical situation, mathematical model of power system. Because of calculation methods and complexity of power system, there have some conflict between the calculation speed and accuracy of the Power System Reliability Analysis and Risk Assessment, and there are many matters in the concept and engineering practice of element’s importance, it must be more develop and completed. Especially in operation environment, a quick approach for power system reliability analysis and risk assessment was required, to obtain the current reliability state and forecast the developing trend of the power system with enough accuracy, in a finite time. In this thesis, a theoretical study of power system reliability analysis and risk assessment in operation environment is carried out. The main works and innovative achievements of the thesis are as follows:(1) In traditional substation scheme reliability analysis, the effects of fault clearance and fault restoration are not elaborately investigated. In this thesis, a novel methodology for the substation scheme reliability analysis is proposed. In which, the effects of both fault clearance and fault restoration are considered. The proposed method is on the basis of non-sequential Monte Carlo simulation and topology analysis. First, cut set of breakers surrounding the failed elements is found to simulate fault clearance, meanwhile the influence of protection and breaker failure is considered. Then, all non-failed switchgears are closed and cut set of switchgears surrounding the failed elements is found to simulate fault restoration. Unsatisfied load demand of two processes is calculated respectively, which can get more reasonable result. Compared with the traditional model considered before, the computation complexity is reduced and the accuracy is enhanced in the proposed method. The proposed method is applied for a study case and the necessity and effectiveness is demonstrated.(2) This thesis presents a Monte Carlo simulation method for reliability analysis of generalized power network which integrates impedance components, such as transmission lines and transformers, and non-impedance components, such as breakers and switches. First, topology analysis for sub station was carried out to get equivalent nodes and the connectedness between these nodes and impedance components. Second, system topology analysis was carried out to get the equivalent network which is used to calculate the measure indicatrix on reliability of power network. During the simulation procedure for generalized power network, a quickly local topology analysis technology based on same transitive closure of incidence matrix was proposed to save computation resources. Test system shows that the proposed method was more practical.(3) The Monte Carlo Simulation Approach used in reliability analysis of the power system requires vast computing resource due to low convergence speed. Aiming to solve this problem, a fast power system reliability analyzing model based on Markov chain is founded from view of operation, by assuming that equipments’fault rate has exponential distribution characteristics. first, a criterion to sort the power system running state into3states which can be easy implemented, as well as corresponding state space transition expression, are proposed; second, based on the historical running data or simulation data, the Markov transition probability matrix of the power system is formed. By using the transition probability matrix and initial distribution, power system reliability indices including state transition probability, steady state probability and the mean time to first failure can be fast analyzed, thus quick reliability assessment of power system can be realized. The example testified the Markov property of the power system state transition and the effectiveness of this research.(4) Weak parts of a power system can be recognized using sensitivity analysis, which has been an important technique for system reliability assessment and enhancement. Aiming at inadequacy of conventional sensitivity analysis, an improved algorithm of improvement potential component importance analysis is founded to make it be fit for actual demands better, and efficiently avoid calculating error of reliability indices affecting results of component importance analysis. Using this algorithm in the generalized power network reliability analysis, this thesis evaluates and compares importance of primary equipments directly to satisfy the need for maintenance strategy better.更多还原