基于粒子滤波的电力系统机电暂态状态估计研究

发布时间：2019-04-09 20:14

【摘要】：同步相量测量单元(phasor measurement unit,PMU)作为一种信息测量装置,已经广泛应用于电力系统运行的各个环节当中,当电力系统处于机电暂态过程中时,PMU能够直接测量系统运行状态的相量信息。然而,由于这些信息是利用传感器进行测量的,且需要通过一定的方式进行信息传输,所以最终使用的数据不可避免存在随机误差和坏数据。在电力系统安全监控方面,为了得到更准确的控制方案或者结果,在应用前需要对实际量测数据进行滤波处理。文中提出了基于粒子滤波(Particle filtering,PF)算法的电力系统机电暂态状态估计方法,主要内容包括以下几个方面:首先,对粒子滤波算法进行了深入研究,以基本PF算法为基础,提出了基于序贯重要性重采样(sequential importance resampling,SIR)的PF算法,为了验证本文提出算法的优越性,同时研究了传统解决非线性状态估计问题的扩展卡尔曼滤波算法(extended Kalman filter,EKF),从理论上对两种算法进行了对比研究分析。其次,将提出的基于SIR的粒子滤波算法应用于电力系统实际状态估计中。首先对发电机在机电暂态过程中的运行状态进行了状态估计,建立了相应的发电机四阶状态空间模型,包括系统方程和观测方程;在发电机四阶模型的基础上,对暂态过程中状态方程的噪声误差进行了分析;为了能够科学合理的定量评价估计的效果,提出了基于Copula理论的观测路径相关性评价指标;最后将提出的方法应用于CEPRI7节点系统的机电暂态状态估计当中,从多个角度,定性定量的对估计结果进行了评价。结果表明基于PF的估计结果与实际结果相关性较高、与真实值的均方根误差小,估计效果优于EKF的估计结果,有效减小了误差数据的影响。最后,提出了对全系统进行机电暂态状态估计的方法。在进行发电机暂态状态估计的基础上,提出了机网接口的直接解法,将对发电机节点的机电暂态状态估计结果用全系统节点电压相量误差方差表示;建立了考虑发电机暂态过程状态估计的全系统动态状态估计模型,通过引入发电机状态估计约束提高对全系统暂态状态估计的精度。通过对仿真算例的计算分析,可以得出本文提出的电力系统暂态过程全系统状态估计方法,能够有效的滤除实际PMU测量过程中可能出现的随机误差,获得更加准确的节点电压相量值。
[Abstract]:As a kind of information measuring device, synchronous phasor measurement unit (phasor measurement unit,PMU) has been widely used in every link of power system operation, when the power system is in the electromechanical transient process, PMU can measure the phasor information of the running state of the system directly. However, because this information is measured by sensors and needs to be transmitted in a certain way, the end-used data will inevitably have random errors and bad data. In the aspect of power system security monitoring, in order to obtain more accurate control scheme or result, it is necessary to filter the actual measurement data before application. This paper presents a power system electromechanical transient state estimation method based on particle filter (Particle filtering,PF) algorithm. The main contents are as follows: firstly, the particle filter algorithm is deeply studied, which is based on the basic PF algorithm. In this paper, a PF algorithm based on sequential importance resampling (sequential importance resampling,SIR) is proposed. In order to verify the superiority of the proposed algorithm, the traditional extended Kalman filter (extended Kalman filter,EKF) algorithm, which is used to solve the nonlinear state estimation problem, is studied. In this paper, the comparison and analysis of the two algorithms are carried out theoretically. Secondly, the proposed particle filter algorithm based on SIR is applied to the actual state estimation of power system. Firstly, the operating state of the generator in the electromechanical transient process is estimated, and the fourth-order state space model of the generator is established, including the system equation and the observation equation. On the basis of the fourth-order model of generator, the noise error of state equation in transient process is analyzed, in order to evaluate the effect of quantitative estimation scientifically and reasonably, the correlation evaluation index of observation path based on Copula theory is put forward. Finally, the proposed method is applied to the electromechanical transient state estimation of the CEPRI7 node system, and the results are evaluated qualitatively and quantitatively from several angles. The results show that the estimation results based on PF have a high correlation with the actual results, and the root mean square error between the real values and the estimation results is small. The estimation effect is better than that of EKF, and the influence of the error data is reduced effectively. Finally, a method of electromechanical transient state estimation for the whole system is proposed. Based on the estimation of generator transient state, a direct solution of machine-network interface is proposed. The result of electromechanical transient state estimation of generator node is expressed by the variance of voltage phasor error of the whole system node. The whole system dynamic state estimation model considering generator transient process state estimation is established, and the accuracy of the whole system transient state estimation is improved by introducing the constraint of generator state estimation. Through the calculation and analysis of the simulation example, we can get the whole system state estimation method of power system transient process proposed in this paper, which can effectively filter out the random errors that may occur in the actual PMU measurement process. More accurate node voltage phasor values are obtained.
【学位授予单位】：东北电力大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM732

【参考文献】