【作者】 张亚洲；

【导师】 刘宪林；

【作者基本信息】 郑州大学 ， 电力系统及其自动化， 2012， 硕士

【摘要】 电力系统数学模型是电力系统规划、运行及控制的基础。当前电网的快速发展对模型参数的准确性要求越来越高。WAMS的出现,将电力系统运行中数据实时采集并上传到调度中心,其中各种扰动下的暂态数据可用来对模型参数的辨识。论文首先对WAMS及其在系统辨识中的应用做了介绍,然后,针对发电机、励磁系统、原动机及其调速系统分别建立状态空间模型,并详细分析了用非线性最小二乘法辨识的可行性,最后结合电科院CEPRI-36节点算例仿真,模拟广域测量环境下采集的数据,对同步发电机组模型参数辨识进行研究。取得的主要成果如下：在前人研究的基础上,着重分析发电机参数在不同扰动类型下的辨识结果,包括励磁参考电压阶跃扰动、调速系统功率给定扰动和负荷阶跃扰动。其中前两种属于人为扰动,而负荷扰动为电力系统自然扰动。论文用非线性最小二乘法对发电机五阶模型参数进行辨识,结果显示,在以上人为扰动和系统自然扰动情况下都能成功辨识。传统方法对励磁系统和调速系统参数的辨识,一般并不涉及对调差系数辨识。论文尝试将调差系数与模型其他参数一起进行辨识。结果显示用非线性最小二乘法也能够在上述三种扰动情况下对包括调差系数在内的参数进行辨识,这将会使在电力调度中心端确定励磁系统调差率和调速系统调差特性成为可能,对电网运行具有重要意义。针对WAMS无选择地采集大量数据的特点,大部分时间的采样数据波动不大而难以进行辨识,论文尝试用某些电气量指标衡量其波动情况,确定动态数据具有足够的扰动强度能够用于辨识,论文从采样数据的采样周期、总时间跨度和扰动电气量的选择三个方面进行分析,分别定量给出了辨识发电机、励磁系统、原动机和调速系统所用的动态数据满足的扰动强度指标。更多还原

【Abstract】 Power system mathematical model is the basis of power system planning, operation and control. With the rapid development of the current grid, we need the model parameters to be more and more accurate and precise. The advent of WAMS real time data acquisition in power system operation can be real-time collection and upload it to the dispatching center, including the transient data in a variety of disturbances, these data may be used for identification of model parameters. The paper first introduces the WAMS and its application in system identification. Then, for generator model, excitation system, prime mover and governor system, the state space model is established respectively and analysis in detail the feasibility of identification using the nonlinear least square method. Finally, combined with the CEPRI-36buses system example, data collected in the simulation of wide area measurement environment to study the power system model parameter identification, the main results are summarized as follows:Based on the previous results, the identification results of the generating under different disturbance types are emphatically analyzed. The disturbances include the excitation reference voltage step disturbance, the power given disturbance and load step disturbance. The former two disturbances are artificial imposed disturbances, and the last one is system natural disturbance. The fifth-order generator model parameters are identified by nonlinear least squares method. The results show that the generator model parameters can be identified successfully under these three kinds of disturbances.The traditional methods for excitation system and governor identification system are usually not related to the identification of the difference coefficients. The paper attempts to identify the difference coefficients with other parameters. The results show that the parameters can also be identified using the nonlinear least squares method under the three disturbance parameter identification. Then, it is possible that the voltage adjustment and governor system difference coefficient can be got at the power dispatch center sides, which have great significance to the power grid.As the characteristics of the power system WAMS have acquired large amounts of data which has no choice, most of the data fluctuations are small and difficult to identify, the paper attempts to use indicators to measure the fluctuations of some electrical quantities to determine the dynamic data using to identify with sufficient fluctuations in intensity. It is analyzed from three aspects of the data sampling period, the total time and the choice of the electrical quantities, for generator model excitation system, prime mover and governor system, the dynamic data using identifying model parameters have to meet the disturbance intensity index respectively and quantitatively.更多还原