中文摘要：

分析并比较用于永磁直驱风电系统的4种典型无速度传感器辨识方法，即基于反电势的锁相环（backEMFbasedphase—lockedLoop，EPLL）方法、基于端电压的锁相环（voltagebasedphase—lockedloop，VPLL）方法、磁链积分方法（fluxintegrationalgorithm，FIA）和简化卡尔曼观测器（simpli—fledKalmanobserver，SKO）方法。由于系统的非线性特性，首先分析了不同辨识算法下系统的小信号稳定性和稳态特性，以比较不同算法稳态特性的差异；其次，通过Matlab／SIMULINK时域仿真和硬件实验，一方面验证了文中理论分析的结果，另一方面进一步揭示了不同控制算法的动态性能、算法复杂度、参数依赖性以及对系统运行效果的影响。结果表明：EPLL是适用于兆瓦级永磁直驱风电系统的辨识策略；FIA与EPLL相比算法较为复杂；VPLL需要更多的参数且控制性能比EPLL或FIA差：SK0方法实现最为简单，但控制效果差于其他3种方法。

英文摘要：

Four typical identification schemes are analyzed and compared for sensodess control of the per- manent magnetic synchronous generator （PMSG） based direct-driven wind energy conversion system （WECS）. The schemes are back electromotive force （EMF） based phase-locked loop （EPLL） method, voltage based PLL method （VPLL）, flux integration algorithm （FIA） and simplified Kalman observer （SKO）. Firstly, the small signal model was derived to analyze the nonlinear system, based on which the system stabilities were analyzed to compare the differences of the steady-state characteristics of the WECS controlled with the four schemes. Then, time-domain Matlab/SIMULINK simulations and hardware exper- iments were implemented. The results verify the theoretical analysis and reveal the dynamic characteris- tics, algorithm complexities, parameter dependences and influences of different schemes on the system performance. Based on the analysis, the following conclusions can be obtained： 1 ）EPLL is suitable for application of PMSG based WECS; 2）FIA is more complex than EPLL; 3）VPLL needs more parameters and has worse performances than FIA and EPLL; 4）SKO is the simplest but it performs worst.