A Mixed Model- and Data-Driven Approach for Subsynchronous Oscillation Analysis of VSC-HVDC Grid-Connected Offshore Wind Farm

The previous research on the subsynchronous oscillation (SSO) caused by Voltage source converter-based high voltage direct current (VSC-HVDC) grid-connected offshore wind farm is based on model-driven method, but the model-driven research method has much inapplicability in multiple scenarios affected by external factors. The data-driven method has the merits of mining the characteristics from the external factors, so as to improve the applicability of the model-driven method in multiple scenarios. This paper proposes a mixed model- and data-driven approach for SSO analysis of VSC-HVDC grid-connected offshore wind farm. In model-driven stage, the small signal model of the system is established. SSO modes and the strongly correlated parameters are identified and their influence on the damping characteristics of the system is analyzed. In data-driven stage, data is processed based on the model-driven results. The Gaussian Mixture Model (GMM) algorithm is applied to cluster Wind speed & current data, and the relationship between clusters and system stability is obtained.