Assessment Method for Equivalent Models of Wind Farms Based on Direct-Driven Wind Generators Considering Randomness

Recently, the research on the sub-synchronous oscillation (SSO) caused by direct-driven wind generators (DDWG) became a new hot field. However, no existing literatures have discussed the equivalent model of the DDWG-based wind farm for the SSO research. This paper proposed a methodology to evaluate the applicability of different equivalent model considering the randomness of the wind power. First, taking the average Euclid distance between the corresponding eigenvalues of the detailed system and the equivalent system as the criterion, a second-order polynomial was adopted to approximate the nonlinear relationship between the distance and the injected wind power. Then, the distribution of wind power was modeled as a Gaussian mixture model (GMM), and the analytical calculation method of the distribution function of the distance was derived with the help of GMM's properties. By calculating the mathematical expectation and the variance of the distribution, evaluation can be conducted to different equivalent model. The equivalent model with smaller expectation and variance can represent the dynamics of the actual wind farm more accurately. At last, the structure-reserving aggregated model and the output multiplication based equivalent model were taken as examples to explain the evaluation process of the proposed method. © 2018 Chin. Soc. for Elec. Eng.