基于改进Monte Carlo法的发输电系统可靠性评估

发布时间：2018-05-04 05:11

本文选题：发输电系统 + 可靠性评估　；参考：《山东大学》2017年硕士论文

【摘要】：电力系统的可靠性评估方法主要包括确定性方法和概率性方法,概率性方法由于计及了系统的随机行为,能够更真实地反映系统的风险度,因此获得了广泛的应用。概率性评估方法又可分为解析法和蒙特卡洛法两种,对于含有较多元件且需要考虑各种随机因素的发输电系统而言,其状态空间巨大,使用解析法逐个枚举故障状态是件极其困难的事情,而蒙特卡洛法不受系统规模限制的特点,使其更适用于此类系统的可靠性评估。但是蒙特卡洛法的不足之处在于:评估时间和评估精度密切相关,换言之,获得高精度的评估结果意味着消耗大量的评估时间。目前,各类电力系统可靠性评估软件大多采用减小抽样方差的方法来加快蒙特卡洛法的收敛速度,减少方差的方法主要有重要抽样法、控制变量法、对偶变数法和等分散抽样法等。上述方法各有特点,且都能够在一定程度上加快蒙特卡洛法的收敛速度,但是它们有个共同的缺点,对系统的适应性较差。为解决这一问题,本文将等分散抽样法和分裂与赌法相结合,提出了一种改进的电力系统可靠性评估方法——等分散与赌法。该方法按照发电系统和输电系统对系统整体可靠性指标贡献度的不同将电力系统分为重点区域和非重点区域,对于非重点区域,利用“赌”的方式对其选择性地抽样;对于重点区域,利用等分散技巧对其抽样,从而在减小抽样方差的同时节约了随机数的生成,提高了评估效率;该方法不但具有等分散抽样法能够有效减小抽样方差的特点,同时具备分裂与赌法对不同系统的良好适应性,从而能够结合系统自身的特点更高效地进行模拟。基于等分散与赌法,本文分别对IEEE-RTS79系统及其修改后的MRTS系统进行了可靠性分析,通过与常规蒙特卡洛法和分裂与赌法的评估结果进行比较,验证了本文所提方法的可行性、高效性以及良好的系统适应性。本文使用自主研发的可靠性评估软件对包含83台发电机,71台变压器,549条母线,1173条线路,523个负荷的某省级实际电网进行了可靠性建模,分别用常规蒙特卡洛法、分裂与赌法以及本文所提等分散与赌法对该电网进行了可靠性评估,通过分析比较评估结果,验证了等分散与赌法在实际工程中的适用性。
[Abstract]:The reliability evaluation methods of power system mainly include deterministic method and probabilistic method. The probabilistic method is widely used because it can reflect the risk degree of the system more truthfully because it takes into account the stochastic behavior of the system. The probabilistic evaluation methods can be divided into analytical method and Monte Carlo method. For a transmission system with more elements and a variety of random factors, the state space is huge. It is very difficult to enumerate the fault states one by one by using the analytic method, but the Monte Carlo method is not limited by the system size, which makes it more suitable for the reliability evaluation of this kind of systems. But the shortcoming of Monte Carlo method is that the evaluation time is closely related to the evaluation accuracy. In other words, obtaining the high accuracy evaluation results means consuming a lot of evaluation time. At present, most of power system reliability evaluation software adopt the method of reducing sampling variance to accelerate the convergence speed of Monte Carlo method. The main methods of reducing variance are important sampling method and control variable method. Dual variable method and equal dispersion sampling method etc. These methods have their own characteristics and can speed up the convergence of Monte Carlo method to a certain extent, but they have a common shortcoming and poor adaptability to the system. In order to solve this problem, an improved power system reliability evaluation method, equal dispersion and gambling, is proposed by combining equal dispersion sampling method with split and gambling method. According to the contribution degree of power generation system and transmission system to the whole system reliability index, the power system is divided into the key area and the non-key area. For the non-key area, the method of "gambling" is used to selectively sample the power system. For the key areas, the equal-dispersion technique is used to sample the samples, thus reducing the sampling variance and saving the generation of random numbers, thus improving the evaluation efficiency, this method not only has the characteristics that the equal-dispersion sampling method can effectively reduce the sampling variance, At the same time, it has good adaptability to different systems by splitting and gambling, so it can be simulated more efficiently according to the characteristics of the system itself. Based on the method of equal dispersion and gambling, the reliability of IEEE-RTS79 system and its modified MRTS system are analyzed, and the feasibility of the proposed method is verified by comparing the results with those of conventional Monte Carlo method and split and gambling method. High efficiency and good system adaptability. In this paper, the reliability model of a provincial power network with 1173 lines and 523 loads has been established by using the self-developed reliability evaluation software, which consists of 83 generators, 71 transformers and 549 busbars, and the conventional Monte Carlo method is used respectively. The split and gambling method and the equal-dispersion and gambling method proposed in this paper are used to evaluate the reliability of the power network. Through the analysis and comparison of the results, the applicability of the equal-dispersion and gambling method in practical engineering is verified.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM732

【相似文献】