计及PMU的鲁棒电力系统预测辅助状态估计

发布时间：2019-02-13 08:07

【摘要】：电力系统状态估计在现代能量管理系统(EMS)中扮演着至关重要的作用,是调度人员进行正确决策的基础。但量测数据常因量测装置内在误差、传输噪声等原因受到污染,干扰状态估计结果,误导调度人员。因此,提高状态估计算法的鲁棒性以及抑制不良数据的能力,对保证电力系统稳定运行有重要意义。随着量测装置技术的发展,同步相量量测装置(PMU)在电力系统中推广应用,为状态估计提供高精度、高同步的量测。同时,PMU量测与SCADA量测来源不同,相互独立,从而互为备用,可以有效抑制SCADA量测中的不良数据,进一步提高算法的鲁棒性。因此,本文主要是提出一种鲁棒性更好的预测辅助状态估计算法,并探究PMU量测对该算法估计精度和鲁棒性的影响。本文主要内容如下:1.简单介绍了几种应用在状态估计中的算法,包括加权最小二乘法、卡尔曼滤波和扩展卡尔曼滤波。2.基于SCADA量测,在扩展卡尔曼滤波(EKF)的基础上改进,提出了广义最大似然类型一扩展卡尔曼滤波算法(GM-EKF)。基本思路:首先,利用EKF的状态方程与量测方程构建线性回归框架。然后,利用投影统计算法(PS)辨识异常值并构建等价权函数。接着,评价函数选择Huber函数,构建类似WLS形式的目标函数并利用IRLS求解。为验证算法的有效性和鲁棒性,将GM-EKF算法在IEEE标准测试系统中仿真,并与相关算法进行结果比较。3.基于SCADA/PMU混合量测,探究PMU量测对于GM-EKF算法估计精度和鲁棒性的影响。基本思路:针对PMU量测与SCADA量测不同融合方式,一种是状态变量为极坐标,直接添加PMU量测,形成非线性的鲁棒预测辅助状态估计算法。另一种是首先处理收集到的SCADA量测,将处理的状态估计值与PMU量测作为新的量测,在直角坐标系下,形成线性的鲁棒预测辅助状态估计算法。将算法在IEEE标准测试系统中仿真,分析仿真结果。
[Abstract]:Power system state estimation plays an important role in modern energy management system (EMS) and is the basis for dispatcher to make correct decision. However, the measurement data are often contaminated by the inherent errors of the measuring device and transmission noise, and the estimation results of the interference state are misled by the dispatcher. Therefore, it is important to improve the robustness of the state estimation algorithm and the ability to suppress bad data to ensure the stable operation of power system. With the development of measuring device technology, synchronous phasor device (PMU) is widely used in power system, which provides high precision and high synchronization measurement for state estimation. At the same time, the sources of PMU measurement and SCADA measurement are different and independent, which can effectively suppress the bad data in SCADA measurement and further improve the robustness of the algorithm. Therefore, this paper mainly proposes a more robust predictor-aided state estimation algorithm, and explores the effect of PMU measurements on the estimation accuracy and robustness of the algorithm. The main contents of this paper are as follows: 1. This paper briefly introduces several algorithms used in state estimation, including weighted least square method, Kalman filter and extended Kalman filter. Based on SCADA measurement and the improvement of extended Kalman filter (EKF), an extended Kalman filter algorithm (GM-EKF), a generalized maximum likelihood type, is proposed. Basic ideas: firstly, the linear regression framework is constructed by using EKF's equation of state and measurement equation. Then, the outliers are identified by projection statistic algorithm (PS) and the equivalent weight function is constructed. Then, the evaluation function selects the Huber function, constructs the objective function similar to WLS and solves it by IRLS. In order to verify the effectiveness and robustness of the algorithm, the GM-EKF algorithm is simulated in the IEEE standard test system, and the results are compared with the related algorithms. 3. Based on SCADA/PMU mixed measurement, the effect of PMU measurement on estimation accuracy and robustness of GM-EKF algorithm is investigated. The basic idea: according to the different fusion methods of PMU measurement and SCADA measurement, one is that the state variable is polar coordinate and PMU measurement is added directly to form a nonlinear robust predictive auxiliary state estimation algorithm. The other is to process the collected SCADA measurements first and take the state estimators and PMU measurements as new measurements to form a linear robust predictive auxiliary state estimation algorithm in rectangular coordinates. The algorithm is simulated in IEEE standard test system and the simulation results are analyzed.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM73

【参考文献】