Potential machine-grid interactions caused by large-scale wind farms have drawn much attention in recent years. Previous work has been done by analyzing the small-signal modeling of doubly-fed induction generators (DFIGs) to obtain the oscillation modes. This paper, by making use of the metered power data of wind generating sets, studies the correlation between oscillation modes of the DFIG system and influence factors which includes wind speed and grid voltage. After the metered data is segmented, the Prony algorithm is used to analyze the oscillation modes contained in the active power. Then, the relevant oscillation modes are extracted in accordance with the small-signal analysis results. Meanwhile, data segments are clustered according to wind speed and grid voltage. The Apriori algorithm is finally used to discuss the association rules. By training the mass of data of wind generating sets, the inevitable association rules between oscillation modes and influence factors can be mined. Therefore, the prediction of oscillation modes can be achieved based on the rules. The results show that the clustering number quite affects the association rules. When the optimal cluster number is adopted, part of the wind speed/voltage clusters can analyze the certain oscillation modes. The predicted results are quite consistent with the practical data.Energies 2018, 11, 2370 2 of 21 (doubly-fed) induction motor wind turbine, providing a lot of flexibility for wind turbines and their controllers. In [7], a small-signal model of the direct-drive permanent magnet synchronous generator is established to study the stability of the grid-connected wind generating sets after the small disturbance, and the parameters of the controller are designed effectively. Reference [8] aimed at the suppression of torsional vibrations caused by small electrical disturbances from the grid side in a DFIG-based wind turbine system. The Prony algorithm is a complementary method used in the time-domain model system. It decomposes time-domain signals into damped sinusoids with four parameters per mode: frequency, damping, amplitude and phase. Compared with small-signal analysis, the Prony analysis shows an advantage when the system turns complicated because the former has difficulty solving high-order matrix [9,10]. All the papers above investigated the wind power oscillation mechanism and suppression method by modeling methods. However, the operation of wind generating sets is affected by factors which are more complex than the modeling simulation. A more direct and convenient way is to directly use the metered data.With the advent of the big data era, the vast amounts of data collected from wind farms conceals vast information [11]. Data mining and machine learning theory including correlation analysis, cluster analysis, classification analysis, have brought new ideas to power research [12]. Among them, the Apriori algorithm, one of the correlation algorithms, provides an effective solution for exploring the association relationship between large dat...