【作者】 项凤雏；

【导师】 颜伟；

【作者基本信息】 重庆大学 ， 电气工程， 2010， 硕士

【摘要】 随着电力系统的发展,电网结构越来越复杂,为了提高电网运行的安全性和稳定性,对各种辅助监视、分析、决策、控制系统的需求愈加迫切,这也就对状态估计、潮流计算等基础电力系统分析计算的精度提出了更高的要求。而输电网参数的准确性是各种电网分析计算软件的基础。但由于各种原因,电网参数始终存在一定的误差。本文利用模拟的和实际的PMU量测数据,对独立线路参数估计模型和算法进行了深入研究,主要研究内容如下:①高压线路的电阻、电抗和对地电纳在数值上差别很大,常规参数估计方法将其作为独立变量分别估计,导致电阻、电抗和对地电纳的估计值误差不一致,其中小阻抗的误差常常很大。针对上述问题,论文提出了一种基于单位长度参数和长度的线路参数估计新模型。其中,状态变量为线路两端的PMU多时段电压幅值与相角,参数变量为线路的单位长度参数和长度。由于引入了长度作为电阻、电抗和对地电纳的公共变量,突出了线路参数整体对潮流状态的影响,因而新模型可以保证线路参数估计值误差的一致性,避免了常规模型中小阻抗大误差的问题。另外,针对该模型,论文提出了参数估计的混沌量子免疫算法,其具有全局寻优能力强,算法稳定的优点。算例结果验证了论文模型与算法的有效性。②参数估计结果易受量测随机误差的影响,通过多次估计结果均值可以减小其影响,但是现有方法没有提出采用多少样本数合适,样本数过小导致参数估计精度不高,样本数过大导致数据量过大,计算复杂。本文将蒙特卡洛模拟法引入参数估计问题,提出了参数均值估计的蒙特卡洛模拟法。通过多次估计的均值作为参数的近似值,采样的样本数通过方差系数确定。仿真表明,论文提出的方法精度较高,样本数更合理。更多还原

【Abstract】 With the development of the power system, the structure of the network is getting more and more complex. In order to ensure the power grid operating safely and stably, the demand of monitoring, analyzing, determining, controlling is more insistent, which proposes the higher request to precision of state estimate, power flow. The accuracy of the parameter in transmission network is the base of the power system analyzing and computing software. But the parameter has certain errors due to various reasons. Using simulated and actual PMU measurement data, the parameter estimation model and algorithm of independent line is deeply researched in this paper. Major contents are as follows:①Resistance, reactance and susceptance to earth in high voltage lines which have a great difference in value are estimated respectively as independent variables in the traditional parameter estimation methods, which results in the inconsistency of resistance, reactance and susceptance to earth estimation error, and the large error of small impedance. A new line parameter estimation model based on the parameter per unit length and length is proposed concerning above problems. In this model, the parameters per unit length and length of line are given constant, state variables are multi-period voltage amplitude and phase angle of line’s two ends. Because the influence of line parameter on power flow state is highlighted by introducing the length as common variable of resistance, reactance and susceptance to earth, the consistency of line parameter estimation error can be guaranteed, and large estimation error of small impedance which appears in traditional model can be voided in the new model. Besides, the chaos quantum immune algorithm to solve the proposed parameter estimation is presented, which has the advantages of global optimality and stability. The validity of the proposed model and algorithm is verified by the simulation test.②The parameter estimate would be effected by the measure rand errors, which could be reduced by calculate the average value of several estimate results. The present methods don’t propose the appropriate sample number, the small sample number will make the precision low, while the big sample number will make the data large. The Monte Carlo simulation method of parameter average estimate is proposed by introducing the Monte Carlo to the parameter estimate. The parameters’approximate values are ensured by the average value of several estimate, and the sample number is ensured by the variance coefficient. The high precision and reasonable sample number are verified by the simulation. The method in paper can clear the influence of rand measure errors, and has the advantages of high precision in parameter estimate.更多还原