A hybrid method and its applications to analyse the low frequency oscillations in the interconnected power system

In this study, a hybrid method is proposed to calculate oscillatory parameters and the approximate mode shapes (AMSs) of low frequency oscillations based on the measured signals from wide area measurement system. This method not only overcomes the limitations of the single algorithm, but also integrates the advantages of the used techniques. First of all, two-level decomposition is presented to minish the scale-mixing influence in the empirical mode decomposition. Then, the Gibbs phenomenon of traditional Hilbert transform (HT) is reduced by using the normalised HT. Next, the relative phase calculation algorithm is introduced to realise the generator grouping and mode shape identification. Finally, the AMSs of dominant oscillation at different time ranges and frequency distributions are determined. The EPRI-36 bus simulation model and actual measured signals are used to evaluate the performances and validities of the proposed hybrid method.