【作者】 张进；

【导师】 贺仁睦；

【作者基本信息】 华北电力大学（北京） ， 电力系统及其自动化， 2005， 博士

【摘要】 电力系统动态仿真是人们认识电力系统的基础性工具,也是指导电力系统规划、设计和运行的根本依据,具有极高的权威性。电力系统动态仿真的准确度直接影响到系统安全稳定策略的效果。1996年美国西部大停电事故后仿真发现,使用现有的仿真数据库不能重现实际系统失稳,随后研究表明仿真的准确性存在很多问题。我国在全国联网工程的仿真计算中也发现不少难以理解的现象。这引起了人们对电力系统动念仿真有效性的极大忧虑。如何正确认识电力系统动念仿真的有效性,进而提高其有效性,确保仿真能够正确地反映电力系统的动态特性成为目前迫切需要解决的问题。 论文深入分析了电力系统动态仿真与其他系统仿真理论的异同,首次提出电力系统动态仿真的应用目的——先验仿真无法验证,只能评估其可信度。而先验仿真的可信度靠后验仿真的准确度保证,模型、参数有效性则是二者之间的桥梁。 论文通过分析电力系统动态仿真的特点,电力系统动态仿真评估必须综合考虑元件模型、参数准确度度评估和全网动念仿真输出准确度评估。针对电力系统变量数量多、层次复杂的特点,按照分层原则首次提出主客观相结合的基于模糊层次分析法的电力系统动态仿真评估框架,并使用IEEE9节点系统验证了该方法的有效性。 论文详细分析了电力系统动态变量的特点,提出以prony算法作为电力系统动态变量特征提取工具,按照分时原则将信号分时,并针对各个时段的特点提出第一摆误差、频率相似度、阻尼相似度、误差能量比等仿真误差计算指标。 针对电力系统模型维数高、参数多,各部分耦合紧密的特点,按照分块原则首次提出基于PMU实测数据的分块仿真误差溯源算法。给出网络分割和PMU安装选点的原则,将实测数掘引入仿真来实现仿真解耦。电力系统精确等值,将仿真模型、参数误差源范围缩小到一个较小区域。 论文使用基于灵敏度分析的参数辨识方法进一步寻找仿真模犁参数误差源。提出针对仿真误差计算指标的摄动灵敏度计算方法,该方法具有灵活,准确的特点。并以负荷模型为例,验证了该方法减少待辨识的参数,简化参数辨识的效果。 介绍了国内某大区电网大扰动试验。给出该次大扰动试验中实测负荷建模的方法;对比、分析了实测数掘与仿真输出的差别,给出模:型、参数调整的方法,使仿真与实际系统扰动的第一摆基本符合,并讨论了其中的机理;讨论了后续摇摆的拟合。讨论了分时、分块、分层仿真验证原则在实际电网仿真验证中的应用。更多还原

【Abstract】 Power system dynamic simulation is a basic tool to understand power system. It also is essential guidance to layout, design and running of power system. So it possesses very high authority. The veracity of power system dynamic simulation has direct influence on the strategy of power system security. The post simulation of west United States blackout in 1996 found that real power system collapse can’t be recurred using existing simulation library. The research indicates the veracity of simulation is not good. Many unreadable phenomena are found in countrywide power grid connecting. Those problems make people worry about the creditability of power system dynamic simulation. How to rightly understand the validity of simulation, how to improve the validity of simulation and how to insure the simulation rightly describe the dynamic characteristic of power system are problems crying for solve.This paper analysis the characteristics of model validation and power system dynamic simulation. Content and key factors of model validation of power system dynamic simulation are introduced. Model validation of Component and model validation of entire power grid should be entire. And they supplement each other. A model validation frame is purposed that based on power system dynamic Fuzzy analytic hierarchy process.This paper researches the characteristic of dynamic error between real system measured data and simulation output. A tool to distill dynamic characteristic based on prony algorithm is purposed. Power system signal is divided according to time. Physical character of power system dynamic variable is the basis of it. First swing error, frequency comparability, damp comparability and energy error are purposed as error index. Those indexes measure the error of power system dynamic simulation. And it provides important reference to simulation error trace.All component of power system is tight coupling. And the dimension of power system model is very high. A network partition method is purposed based on spur track density. Measured data are imported into simulation calculation. Subarea simulation are used to find model and parameter errors that make simulation error. Parameter identification method based on sensitivity analysis is used to ulteriorly find errors of simulation model and parameter. All kinds of sensitivity analysis methods are compared. And a load model parameter sensitivity analysis method considering simulation error index is purposed. The aim of it is reducing the parameters that need to更多还原