主动配电网状态估计及鲁棒量测配置研究

发布时间：2018-06-04 21:41

本文选题：主动配电网 + 状态估计　；参考：《山东大学》2017年硕士论文

【摘要】：分布式电源(distributed generations,DG)的大量接入,将逐渐改变传统配电网的形态。主动配电网具有良好的配电自动化水平,可以通过网络重构,实现配电网的灵活、经济、安全运行。主动配电网(active distribution network,ADN)将成为未来配电网的发展趋势。ADN的主动性要求实时感知系统运行状态,配电网状态估计(distribution system state estimation,DSSE)是获得系统运行状态信息的重要途径。然而,当前配电网中实时量测数量不足,需要添加大量伪量测实现系统的可观测性。但伪量测的误差较大,为保证状态估计的精度,需合理地进行量测配置。研究适用于ADN的状态估计及量测配置方法对保证系统安全稳定运行具有重要的理论意义与实用价值。本文围绕ADN的状态估计及量测配置问题展开研究,主要工作如下:(1)提出考虑DG影响的ADN状态估计方法。以高斯混合模型(gaussian mixture model,GMM)模拟未配置实时量测DG出力的概率密度分布(probability density function,pdf),GMM的各个参数由最大期望算法(expectation maximization algorithm,EM)估计得到。在状态估计中,将DG的出力当作伪量测,为表征DG出力的不确定性,其数值为服从GMM的随机变量,对应的权重为GMM方差的倒数。另外,为反映邻近DG间出力的相关性对状态估计的影响,采用多维GMM同时模拟多个DG的出力,并以GMM协方差矩阵的非对角元素作为量测误差协方差矩阵的非对角元素。(2)理论证明量测系统的饱和量测特性,提出饱和量测数量的概念。针对非线性测量方程分析了系统估计精度与量测数量的关系,利用Sherman-Morrison公式和子模特性从理论上推导出量测系统具有饱和量测特性。根据饱和量测特性,定义了饱和量测数量,为量测数量的选择提供了依据,对配电系统量测配置的建设投资给出指导性方案,避免了盲目安装大量量测装置所造成的经济浪费,具有实际意义。(3)以饱和量测数量为约束,建立了兼顾经济性和多种网络结构估计精度的鲁棒量测配置模型。该模型考虑了正常、环网、优化重构及故障重构四种网络结构的估计精度,提高了量测配置的鲁棒性。经济性、估计精度及不同网络结构对量测配置的影响权重由层次分析法(analytic hierarchy process,AHP)确定。在此基础上,采用启发式算法进行配置。仿真结果表明所建模型可兼顾多种网络结构的估计精度,同时能够较好地协调配置经济性与估计精度的关系。
[Abstract]:The large number of distributed generation DGs will gradually change the form of traditional distribution network. The active distribution network has a good level of distribution automation and can be operated flexibly, economically and safely through network reconfiguration. Active distribution Network (ADN) will become the development trend of distribution network in the future. The active requirement of ADN is real-time perception of system running state. Distribution system state estimation is an important way to obtain system running state information. However, the quantity of real-time measurement in distribution network is insufficient, so it is necessary to add a large number of pseudo-measurements to realize the observability of the system. However, the error of pseudo-measurement is large, so to ensure the accuracy of state estimation, the measurement configuration should be reasonably carried out. It is of great theoretical significance and practical value to study the state estimation and measurement configuration method suitable for ADN to ensure the safe and stable operation of the system. In this paper, the state estimation and measurement configuration of ADN are studied. The main work is as follows: 1) A ADN state estimation method considering the influence of DG is proposed. The probability density distribution (PDD) of unconfigured real time DG force measurement is simulated with Gao Si mixed model / GMMs. The parameters of the probability density distribution (PDC) are estimated by the maximum expectation maximization algorithm (EM). In the state estimation, the output force of DG is regarded as a pseudo measure, and to represent the uncertainty of the output of DG, its value is a random variable from GMM, and the corresponding weight is the reciprocal of GMM variance. In addition, in order to reflect the effect of the correlation between the output forces of adjacent DGs on the state estimation, multi-dimensional GMM is used to simulate the output forces of several DGs simultaneously. Using the non-diagonal elements of the GMM covariance matrix as the non-diagonal elements of the measurement error covariance matrix, the saturated measurement characteristics of the measurement system are proved by using the theory of the non-diagonal elements of the measurement covariance matrix, and the concept of saturated measurement quantity is put forward. Based on the nonlinear measurement equation, the relationship between the system estimation accuracy and the measurement quantity is analyzed. By using the Sherman-Morrison formula and the submodel, the saturated measurement characteristics of the measurement system are deduced theoretically. According to the characteristic of saturation measurement, the quantity of saturated measurement is defined, which provides the basis for the selection of measurement quantity, and provides a guiding scheme for the construction investment of measurement configuration of distribution system, thus avoiding the economic waste caused by blindly installing a large number of measuring devices. With the restriction of saturated measurement quantity, a robust measurement configuration model with both economic and network structure estimation accuracy is established. In this model, the estimation accuracy of normal, loop, optimal reconfiguration and fault reconfiguration is considered, and the robustness of measurement configuration is improved. The economic efficiency, estimation accuracy and the influence of different network structure on the measurement configuration are determined by Analytic hierarchy process (AHP). On this basis, heuristic algorithm is used for configuration. The simulation results show that the proposed model can take into account the estimation accuracy of various network structures, and the relationship between configuration economy and estimation accuracy can be well coordinated.
【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】：TM727;TM76

【相似文献】