基于非精确概率的电力设备运行可靠性评估方法研究

发布时间：2018-05-30 02:10

本文选题：电力设备 + 运行可靠性评估　；参考：《山东大学》2017年硕士论文

【摘要】：随着特高压建设、交直流混联、新能源大量并网以及全球能源互联网的建设推进,保障电力系统的安全稳定运行面临新的挑战。电力系统运行安全的重要基础是电力设备的运行安全,架空输电线路、电力变压器作为电网中的关键设备,其健康水平和运行状况直接与整个电网的安全稳定运行水平相关联。对此,提出利用先进、合理的方法对电力设备的运行可靠性水平进行评估,估计架空输电线路的停运概率,对变压器实施故障诊断,对促进电力系统的安全、经济运行具有重要意义。然而,随着光伏、风电等新能源的大规模接入电网和新技术在电力系统中的广泛应用,许多新投用设备缺少甚至没有完备的故障停运统计数据,"统计样本少,统计周期长"是困扰电力设备运行可靠性评估的难点问题。在此条件下,若武断地给出确定性的评估结果,不仅难以具有工程实用价值,甚至会误导调度策略的制定,破坏电力系统运行的经济性或可靠性。非精确概率(Imprecise Probability,IP)是处理非完整样本信息的有效方法,其可应对在设备运行可靠性评估中故障样本缺乏的不利条件,以区间概率的形式代替确定性的精确单值概率评估结果,可为调度决策提供更加完备的概率信息。本文简要介绍非精确概率的表达形式,研究非精确Dirichlet模型(Imprecise Dirichlet Model,IDIM)、信度网络(Credal,Network,CN)、信度网络分类器(Credal Classifier,CC)数学模型的原理及应用,基于非精确概率对电力设备运行可靠性实施评估,包括两部分内容:本文提出一种架空输电线路停运概率的非精确条件估计方法。该方法基于设备停运的历史统计数据和估计目标时段的运行工况数据,构建处理非精确条件概率推断问题的信度网络,并利用多状态随机变量的非精确Dirichlet模型,获得在线路停运样本缺乏条件下信度网络节点的非精确条件概率,从而,估计得到给定运行条件下输电线路停运概率的区间范围。方法体现了暴露型设备停运概率应依其运行工况变化而时变的特点,为解决停运样本缺乏条件下的电力设备运行可靠性评估问题提供了新的思路。同时,本文提出一种基于非精确概率的变压器故障诊断方法,该方法由非精确Dirichlet模型和信度网络分类器构成,可应对变压器故障样本缺失、检测试验数据不完备等客观不利条件,依据方法得出的各故障类后验非精确概率区间的有无重叠,将变压器故障诊断输出为唯一故障模式,或输出包含多个故障类的相似类集合,弥补了传统诊断方法忽略变压器发生其它并发性或潜伏性故障的缺陷,有效提高变压器故障诊断的正判率。
[Abstract]:With the construction of UHV, AC / DC hybrid, a large number of new energy connected to the grid and the construction of the global energy Internet, the security and stability of the power system is facing new challenges. The important foundation of power system operation safety is the operation safety of power equipment, overhead transmission line and power transformer as the key equipment in the power network, its health level and operation condition are directly related to the safe and stable operation level of the whole power network. In view of this, the advanced and reasonable method is proposed to evaluate the reliability level of power equipment, estimate the outage probability of overhead transmission line, implement fault diagnosis of transformer, and promote the safety of power system. Economic operation is of great significance. However, with the extensive application of new energy, such as photovoltaic, wind power and other new energy sources, in power system, many new equipments lack or even do not have complete statistical data of fault outage. Long statistical period is a difficult problem for reliability evaluation of power equipment operation. Under this condition, it is difficult to give deterministic evaluation results arbitrarily, which is not only difficult to be practical in engineering, but also mislead the formulation of dispatching strategy, and destroy the economy or reliability of power system operation. Imprecise probabilistic probability (IPP) is an effective method for dealing with incomplete sample information. It can deal with the disadvantages of fault samples in reliability assessment of equipment operation, and replace the deterministic results of accurate single-valued probabilities with interval probability. It can provide more complete probability information for scheduling decision. This paper briefly introduces the expression form of imprecise probability, studies the principle and application of imprecise Dirichlet model IDIMI, reliability network CredalCast network classifier and reliability network classifier Credal Classifier CCC, and evaluates the reliability of power equipment operation based on imprecise probability. This paper presents an inexact condition estimation method for outage probability of overhead transmission lines. Based on the historical statistical data of equipment outage and the operating condition data of the target period, the reliability network is constructed to deal with the inexact conditional probability inference problem, and the inexact Dirichlet model with multi-state random variables is used. The inexact conditional probability of reliability network nodes is obtained under the condition of lack of line outage samples, and the interval range of transmission line outage probability is estimated under given operation conditions. The method shows that the outage probability of exposed equipment should be time-varying according to its operating conditions, and provides a new way of thinking to solve the problem of reliability evaluation of power equipment under the condition of lack of outage samples. At the same time, this paper presents a transformer fault diagnosis method based on imprecise probability. The method is composed of an imprecise Dirichlet model and a reliability network classifier, which can deal with the transformer fault sample missing. Based on the imprecise probability interval of each fault class obtained by the method, the transformer fault diagnosis is outputted as the unique fault mode. Or the output contains similar classes of multiple fault classes, which makes up for the defect of traditional diagnosis method which ignores other concurrent or latent faults of transformer, and effectively improves the positive judgment rate of transformer fault diagnosis.
【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;TM506

【相似文献】