利用励磁电压扰动后的同步发电机参数辨识

基于Matlab/Simulink建立同步发电机系统在励磁电压发生扰动时的电气模型,并通过Park变换,快速Fourier分解（FFT）得到同步发电机实用参数辨识所需d轴和q轴电压、电流的基频分量和励磁电压等电气数据。然后根据汽轮同步发电机六阶电气模型以及对应的电气微分方程进行辨识,微分方程的求解采用改进Euler法。在寻优辨识中先计算稳态参数,从而减少待求参数的维数,再利用改进的粒子群算法寻优辨识暂态参数。仿真算例的计算结果验证了该方法的正确性。     

基于实测系统扰动的同步发电机参数辨识与验证

为提高发电机参数的有效性,提出了一种基于实测系统扰动数据的同步发电机参数辨识与验证方法。首先,根据扰动前后不同参数的特性被激发程度不同的特点,对发电机参数进行分步辨识,即利用稳态量测辨识稳态参数,利用扰动量测辨识暂态/次暂态参数。进一步根据发电机功角可测时d、q轴的解耦特性,对d轴、q轴的暂态/次暂态参数进行解耦辨识。利用分步、解耦的辨识策略进行参数辨识,每一步待辨识的参数不超过4个,从而提高辨识结果的鲁棒性。然后,提出一种基于外部时序受控源的混合动态仿真验证策略,实现发电机参数的灵敏验证。最后,利用某实际电网的实测扰动数据对实际发电机的参数进行了辨识和验证,结果证实了上述方法的有效性。     

同步发电机实用参数分时段辨识研究

同步发电机在动态过程的不同时段起主导的实用参数不同。对比研究同步发电机实用参数在励磁扰动故障下的轨迹灵敏度特性,给出不同参数在不同时段内的灵敏度大小的不同。提出一种分时段辨识策略,即先利用励磁扰动前、后的稳态数据辨识同步发电机的稳态参数,然后分别根据次暂态和暂态时段数据辨识次暂态和暂态参数,最后以单机无穷大系统为例验证此方法的有效性。     

同步发电机参数辨识的蚁群算法及扰动分析

以发电机5阶实用模型为研究对象,利用励磁电压参考值的变化产生动态响应,采用蚁群算法辨识同步发电机参数,分析干扰的大小和类型对发电机参数辨识的影响.在RTDS仿真机组上验证了蚁群算法的可行性,结果表明,从扰动的大小上看,扰动越大,参数辨识的精度越高;从扰动的类型上看,白噪声激励下辨识精度较好,该激励更有利于激发同步发电机的次暂态过程;无论是哪种扰动方式,稳态参数的精度没有大的变化,改善的主要是次暂态参数的辨识精度.     

直驱永磁同步风力发电机的电气参数辨识

文中首先运用机组空载和多组稳态运行点数据,研究直驱永磁同步风力发电机电气模型参数的可辨识性问题,证明机组的电气参数可唯一辨识;然后,计及测量数据的误差效应,给出优选初值—微变搜索辨识算法的具体化实用步骤表达式;最后,运用实测数据,辨识了直驱永磁同步风力发电机的特性参数,证明了理论分析的正确性和辨识算法的有效性。     

同步发电机参数的一次性任意抛载测试方法

为简化同步发电机的抛载试验方法,在同步发电机功角可以测量的基础上,提出了一种用于同步发电机参数辨识的一次性任意抛载测试方法。对于同步发电机5阶模型,当同步发电机励磁电压不变,调速器不动作时,详细推导了任意抛载情况下的发电机端口电压包络线方程、励磁电流变化方程。论证结果表明,利用一次性任意抛载过程中的动态以及抛载前的稳态,以发电机励磁电流、定子电压、转速作为量测量,可以同时辨识全部d、q轴电气参数,而不需分d、q轴两次抛载试验,简化了试验过程。仿真计算验证了上述方法的有效性。     

基于蚁群算法的同步发电机参数改进频域辨识方法

同步发电机参数频域辨识方法一般假设小扰动条件下励磁电压保持不变,但实际上,试验过程中励磁电压不断变化,扰动大小的划分也带有一定经验性。为此,提出了基于蚁群算法的同步发电机参数改进频域辨识方法,该方法需要实时测量扰动后的发电机功角、定子电压、电流以及励磁电压等变量,并应用快速傅里叶变换及功率谱算法将各电气量转换为频域信号。然后,在计及励磁电压变化的条件下,应用蚁群优化算法拟合发电机电压q轴分量的频域特性,从而辨识d轴最优参数。q轴运算电抗不受励磁电压影响,故可按照传统方法辨识q轴参数。对某电站水轮发电机的参数辨识验证了改进频域方法的有效性,与传统方法相比,改进频域辨识方法可适用于较大扰动的情况,且d轴参数辨识精度较高,收敛性更好,具有较好的实用性。     

风电场等值模型可辨识性研究

研究了风电场等值模型参数的可辨识性,对风电场模型参数的辨识具有重要的理论意义和实用价值。结果表明,利用风电场扰动前、后稳态条件和动态过程,证明了风电场等值模型参数是可唯一辨识的。     

基于饱和模型的同步发电机参数辨识研究

由于饱和效应对同步发电机的参数辨识有重要影响,因此在考虑饱和效应对d轴、q轴参数辨识影响的基础上,提出同步发电机参数辨识方法。对这一方法的辨识过程进行了介绍,并通过仿真算例确认,基于饱和模型的同步发电机参数辨识方法具有较高的精度,同时可以减小辨识过程中的计算量。     

基于多新息最小二乘法的同步发电机一次性抛载试验参数辨识

针对同步发电机d、q轴抛载试验的不足,提出了任意工况下一次性抛载试验参数辨识方法.通过分析同步发电机抛载试验动态过程,建立了任意工况下一次性抛载试验d、q轴传递函数模型.考虑到传递函数输入/输出均可测量,通过积分变换将微分方程转化为积分方程,将同步发电机任意工况抛载试验传递函数模型转化为线性回归模型,在此基础上通过分析参数可辨识性和引入多新息最小二乘辨识算法,解决了同步发电机物理参数辨识问题.算例结果验证了该方法的有效性.     

The Squirrel-Cage Induction Motor Model and Its Parameter Identification Via Steady and Dynamic Tests

One of the most important bases for designing robust closed-loop controllers applied to induction motor with high performance is establishing its mathematical model and state observers, as well as the parameter identification with high accuracy. In this paper, a step-by-step mathematical model of the squirrel-cage induction motor is described at αβ coordinate frame where the parameters are defined in detailed form; the rotor flux linkages and load torque are estimated via an asymptotic observer; the induction motor parameter identification is performed via a data acquisition board, applying dynamic and steady-state tests. Inductances of the induction motor model are calculated using the proposed relationships between the magnetically coupled circuit and equivalent circuit model. The mathematical model, state observers, and parameter identification procedure of squirrel-cage induction motor are validated via comparison of simulation signals with their corresponding real-time signals. This validation is made experimentally by a steady-state test, where load conditions are changed via a dynamometer which is belt coupled with the squirrel-cage induction motor.     

基于RBF-Volterra的非线性电路故障诊断

故障诊断的重要前提是故障特征可比性原则,而可比性原则的条件是模型结构的不变性,这就要求同一参数模型起码能稳定地拟合同一电路的所有故障状态。针对故障状态下非线性强弱变化造成的参数估计数值稳定性问题,提出一种基于RBF-Volterra级数的非线性电路故障诊断方法。在平移联动约束下,由基函数多项式(BFP)导出RBF-Volterra级数,分析了RBF-Volterra级数的截尾不截维特性和对非线性强弱变化的适应能力.RBF-Volterra级数不仅数值稳定性好,而且结构紧凑能有效避免维数灾难问题。最后,通过实例仿真说明该方法的有效性。     

基于参数辨识的消弧线圈接地电网单相接地故障测距方法

为解决消弧线圈接地电网单相接地故障测距困难的问题,提出一种基于最小二乘参数辨识的消弧线圈接地系统单相接地故障测距方法。以消弧线圈接地电网故障零序全波等效电路作为辨识模型,将故障距离、过渡电阻、对地电容作为模型的未知参数,利用该模型所反映的单相接地故障电压、电流时域关系构造参数辨识目标函数,用最小二乘辨识方法进行最优参数求解,得到故障位置到母线端的距离。对于带分支线路的情况,在参数辨识测距方法的基础上建立了推理判断逻辑,提高了测距方法的适用性。所提出的测距算法可集成在故障选线装置中,无需增加额外设备即可实现故障选线、测距一体化功能。仿真和实验验证了所提测距方法的有效性与准确性。     

Estimation of induction motor parameters using shuffled frog-leaping algorithm

This paper introduces a shuffled frog-leaping algorithm based method to approximate the equivalent circuit parameters of induction machines from the manufacturer data, such as nameplate data and motor performance characteristics. The steady-state equivalent circuit is applied for the simulations. The circuit parameters are found as the result for the error minimization function between the estimated and maker data. The suggested algorithm solves the parameter estimation problem and surpasses the solutions reached by differential evolution, particle swarm optimization and genetic algorithms. Therefore, this algorithm can be employed in motor energy management system for bettering the overall energy savings in industry.     

Incipient fault location formulation: A time-domain system model and parameter estimation approach

Power systems faults are unavoidable events which affect distribution networks reliability. The fault process in underground cables is gradual and characterized by a series of sub-cycle incipient faults associated with an arc voltage. These events often are unnoticed and eventually result in permanent faults. This paper presents an incipient fault location formulation for distribution networks with underground cables. Presented formulation is composed by a time-domain system model and parameter estimation strategy. System model derivation considers distribution networks inherent features as unbalanced operation and underground distribution cables capacitive effect. Further, incipient fault characteristics as fault arc voltage are considered. The proposed system model is an overdetermined linear system of equations in which the fault location is estimated through a parameter estimation approach. Parameter estimation is made through a Non-Negative Weighted Least Square Estimator (NNWLSE). Smoothing and curve-fitting procedures are applied to input data aiming to decrease the noise effect. A load current compensation strategy is proposed to reduce its effect in the fault current estimation and a back substitution method is proposed for estimation refinement. Validation is performed using real-life distribution network with underground cable data simulated on ATP/EMTP. Test results are encouraging and demonstrate the method’s potential for real life applications. An average error of 1.95% is obtained when compared with 6.48% derived using the state-of-art.     

Estimating parameters of synchronous generators using square-root unscented Kalman filter

A new method to estimate the parameters of a synchronous generator using the square-root unscented Kalman filter (SRUKF) is presented in the paper. A third-order model for the parameter estimation of both round rotor and salient generators is developed first and then the SRUKF method is applied to the third-order model to perform the joint estimation of state variables and unknown generator parameters. The simulation results on a test system demonstrated the effectiveness of the proposed method in parameter recognition of a synchronous generator. The estimation processes of generator parameters steadily converge to the estimated values whereas the estimation processes of state variables are consistent with the dynamic responses in the numerical simulations.     

一种基于PMU的线路自适应故障测距算法

提出一种新的基于相量测量单元(PMU)的输电线路故障测距的自适应算法。该算法利用PMU装置获得高压线路两端的电压和电流相量,在线计算线路参数,解决了线路实际参数与电力局所提供参数的不同、线路参数在运行过程中的不确定性等问题。采用前置带通滤波器与全波傅氏算法相结合的滤波算法,提取相当精确的突变量基频分量,用于输电线路故障测距。大量的EMTP仿真计算结果和实际系统参数验证结果表明,该测距算法不受系统的运行方式、故障点过渡电阻、故障类型、故障距离等因素的影响,具有很高的测距精度。     

Estimation of synchronous machine parameters from performance relations

The parameters of synchronous machines are required in system analysis, fault calculations, sizing of cables, circuit breaker selection, power factor correction, and voltage drop calculations. One of the major stumbling blocks is the general lack of design and application data. The design information is generally restricted and not commonly available to all users. Furthermore, the parameters are difficult to obtain through field measurements. In this paper, a mathematical model is presented to estimate the parameters of a synchronous machine from name-plate details. The data used in the estimation include output kVA, voltage, current, speed, rated power factor, field voltage, field current, and short-circuit ratio. The equations are derived from a well-known performance model and solved using a spreadsheet program available on any personal computer. Two case studies are presented in this paper. There is reasonable agreement between the estimated and available parameter values. Finally, the usefulness of this procedure and some of the limitations are outlined.     

电力系统多台发电机参数的整体辨识

提出对多台发电机参数进行整体辨识的思路,即根据广域测量系统（WAMS）采集到的发电机的功角轨迹,同时辨识多台发电机的参数。文中针对不同阶次的Park模型,采用基于轨迹灵敏度的可辨识性分析方法,研究了多台发电机参数的可辨识性,结果表明大部分参数能够区分辨识而一部分参数不能够区分辨识。提出了多台发电机参数的辨识策略,以及基...     

基于PMU数据和时间序列模型的发电机参数识别

电力系统动态仿真的精度往往由同步发电机等元件模型及参数的精度所决定。在对同步发电机模型进行时间序列分析的基础上,提出了一套依靠同步相量测量单元对惯性时间常数、发电机转速调节常数以及涡轮调速器的时间常数等发电机动态参数进行识别的方法。该方法首先将发电机的经典模型和摇摆方程统一转换为一个时间序列模型,并利用线形参数辨识方法对此时间序列模型的参数予以辨识;最后通过求解发电机模型参数与时间序列模型参数的等式关系得到发电机的模型参数。对不同功率变化波形进行了仿真试验,按照本文方法计算得出的发电机参数与实际参数一致,验证了本文提出方法的有效性。