基于不确定测度的电力系统抗差状态估计

发布时间：2018-10-05 19:39

【摘要】：随着电力信息物理系统的快速发展,对电力系统的精确预测、精当决策和精准控制成为必然要求,而这依赖于精确的数据。作为数据滤波器的状态估计模块是能量管理系统中的基础和核心,传统的加权最小二乘状态估计方法及基于残差的不良数据辨识方法对强相关的多不良数据的辨识能力有限,已不能完全满足日益增长的生产应用和安全的要求,电力系统状态估计的抗差性能和计算效率急需提高。为此,本文在不确定理论体系下对状态估计进行了深入研究,论文的主要工作如下:(1)基于不确定测度重新定义了正常测点、异常测点、测点正常率,进而提出兼顾测点正常率和量测量偏离真值程度的新的状态估计结果评价新指标,可以应用在真值已知或未知的情况。(2)基于兼顾测点正常率和偏离度的评价指标定义正常测点参与度函数及状态估计准则函数,在此基础上提出了最大正常率最小偏离度估计方法(MNLD)。此法是以测点正常率最大和正常测点偏离度最小为目标的多目标规划模型,从理论上保证了其估计结果具有较高的合理性。(3)为使MNLD模型适用于实际求解,作了3点等效转化。1)通过目标规划法将SE的多目标规划转化为单目标规划问题;2)采用双曲正切型矩形脉冲(RPF)替代正常测点参与度函数;3)采用改进凝聚函数逼近max(7)(8)型函数(无穷范数型函数);并采用内点法进行求解。(4)通过对MNLD方法与已有的状态估计方法进行收敛性试验、抗差性能、计算效率等方面的仿真实验。大量算例表明,MNLD方法在抗差性能和计算效率方面具有良好表现。
[Abstract]:With the rapid development of power information physical system, accurate prediction of power system, precise decision-making and accurate control become necessary requirements, which depends on accurate data. As a data filter, the state estimation module is the foundation and core of the energy management system. The traditional weighted least square state estimation method and the residual based bad data identification method have limited ability to identify the strong correlation multi-bad data. It can not fully meet the requirements of increasing production application and safety. The robust performance and computational efficiency of power system state estimation need to be improved urgently. The main work of this paper is as follows: (1) based on the uncertainty measure, the normal point, the abnormal point, the normal rate of measurement point are redefined. Furthermore, a new evaluation index of state estimation results, which takes into account the normal rate of measurement points and the degree of deviation from the true value, is proposed. It can be applied to the situation where the true value is known or unknown. (2) based on the evaluation index which takes into account the normal rate and deviation of the measured point, the participation function and the state estimation criterion function of the normal measuring point are defined. On this basis, a method for estimating the maximum normal rate and minimum deviation degree, (MNLD)., is proposed. This method is a multiobjective programming model with the maximum normal rate and minimum deviation of normal measuring points as the objective, which theoretically ensures the high rationality of the estimated results. (3) in order to make the MNLD model suitable for practical solution, In this paper, the equivalent transformation of 3 points is made. 1) the multiobjective programming of SE is transformed into a single objective programming problem by the goal programming method. 2) the hyperbolic tangent rectangular pulse (RPF) is used to replace the participation function of normal measuring points. 3) the improved condensate function is used to approximate max (7) (8). The inner point method is used to solve the problem. (4) the convergence test of MNLD method and the existing state estimation method is carried out. Simulation experiments on robust performance, computational efficiency and so on. A large number of examples show that the MNLD method has good performance in robustness and computational efficiency.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM732;TM711

【参考文献】