基于弹球损失支持向量机和端口能量函数的暂态稳定评估方法研究

发布时间：2019-01-05 18:59

【摘要】：电力系统暂态稳定评估—直是关系到电力系统安全稳定运行的重要问题。随着可再生能源及大量电力电子装备的接入,区域电网互联规模不断扩大,使电力系统的调度运行方式和安全稳定控制面临严峻考验,需研究满足当前运行状态且有利于复杂电网调度的暂态稳定评估在线分析方法。本文主要从弹球损失支持向量机和端口能量函数法暂态稳定评估两方面进行分析,主要研究内容如下:(1)针对传统支持向量机在进行电力系统暂态稳定评估时,易受系统临界稳定干扰样本的影响问题。提出一种基于弹球损失支持向量机的暂态稳定评估模型。首先,采用系统指标和投影能量函数指标构建暂态稳定指标的原始特征集,通过最大相关最小冗余特征选择方法对暂态指标集进行特征压缩,寻找对电网暂态变化敏感度高的特征子集;然后,基于弹球损失支持向量机思想将特征子集映射到高维空间,实现非线性暂态稳定评估问题的线性转化,进而引入分位数改变系统稳定类与不稳定类之间的最近点位置,将暂态稳定分类问题转化为在弹球损失支持向量机中寻找最优分位数距离问题,以减小系统临界稳定干扰样本的影响,提高电力系统暂态评估方法的评估准确率和稳定性。(2)针对传统能量函数构造易受系统模型影响,当网络拓扑改变时需要重新构造能量函数问题。提出一种基于端口能量函数法的暂态稳定评估模型。将端口能量描述为端口状态变量和控制变量的函数,进而借助端口能量表征VSC-HVDC系统的暂态能量聚集效应,以避免VSC-HVDC内部复杂控制过程的建模,降低能量函数的构造难度,且可扩展到网络中任意含电力电子器件的端口;进一步,构建基于端口能量的全系统能量函数,通过迭代势能边界面法(iteration potential energy boundary surface,IPEBS)进行暂态稳定评估,借助点积判据确定系统失稳时间,在系统失稳的转子运动轨迹上搜索势能最大值作为临界能量,确定极限切除时间,实现暂态稳定的快速判别。
[Abstract]:Power system transient stability assessment is an important problem related to the safe and stable operation of power system. With the access of renewable energy and a large number of power electronic equipment, the scale of regional power grid interconnection is expanding, which makes the dispatching operation mode and safety and stability control of power system face a severe test. It is necessary to study the on-line transient stability assessment method which satisfies the current operation state and is beneficial to complex power grid dispatching. In this paper, the loss support vector machine and the port energy function method for transient stability evaluation are analyzed. The main contents are as follows: (1) traditional support vector machine is used to evaluate the transient stability of power system. The problem of vulnerability to critical stability disturbance samples of the system. A transient stability evaluation model based on loss support vector machine (SVM) is proposed. Firstly, the original feature set of transient stability index is constructed by system index and projection energy function index, and the feature compression of transient index set is carried out by the method of maximum correlation and minimum redundancy feature selection. Looking for the feature subset with high sensitivity to the transient change of power network; Then, the feature subset is mapped to high dimensional space based on the idea of loss support vector machine (SVM). The linear transformation of nonlinear transient stability assessment problem is realized, and the quantile is introduced to change the position of the nearest point between the system stability class and the unstable class. In order to reduce the influence of critical stability disturbance samples, the transient stability classification problem is transformed into finding the optimal quantile distance in the ball-loss support vector machine. To improve the accuracy and stability of power system transient evaluation methods. (2) for the traditional energy function construction is vulnerable to the system model, when the network topology changes need to reconstruct the energy function problem. A transient stability evaluation model based on port energy function method is proposed. The port energy is described as a function of the port state variable and the control variable, and then the transient energy aggregation effect of the VSC-HVDC system is represented by the port energy to avoid the modeling of the complex control process in the VSC-HVDC. It reduces the difficulty of constructing the energy function and can be extended to any port containing power electronic devices in the network. Furthermore, the energy function of the whole system based on port energy is constructed. The transient stability is evaluated by iterative potential energy boundary surface method (iteration potential energy boundary surface,IPEBS). The point product criterion is used to determine the instability time of the system. The maximum potential energy as the critical energy is searched on the rotor trajectory of the unstable system, and the limit cutting time is determined to realize the fast identification of transient stability.
【学位授予单位】：东北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM712

【参考文献】