Research On Transformer Inrush Current Identification Based On Improved HHT Algorithm

The power transformer is an important equipment in the normal operation of the power grid.It is responsible for the transmission and conversion of electric energy,and its safety and stability are directly related to the safety of the entire power grid.According to the latest statistics[1],the correct action rate of transformer is significantly improved,but it is still low compared with other protection(such as line,bus,generator,etc.).As the main protection of transformer,the correct action of differential protection directly affects the operation safety of transformer.The transformer inrush current(such as inrush current and sympathetic inrush current)is an important factor affecting the malfunction of transformer differential protection.Therefore,how to identify the inrush current and internal fault current accurately and reliably is one of the most important research topics in transformer differential protection.The main work of this thesis is as follows:(1)Firstly,the mechanism of transformer magnetizing inrush current,sympathetic inrush current and internal fault current is analyzed.The simulation of the magnetizing inrush current under different closing angles and different remanence conditions is carried out.and its characteristics are summarized.The parallel and series sympathetic inrush current are studied separately,and the factors affecting them(such as system reactance and resistance,closing angle,transformer remanence,active load,etc)are analyzed,and the characteristics of the change under different factors are summarized.The current of transformer internal faults(such as single-phase faults,inter-phase faults,three-phase faults)is simulated and analyzed,and the differences with transformer inrush current are compared.(2)The EMD algorithm can correctly analyze nonlinear non-stationary signals.However,due to the presence of higher harmonics in the inrush current waveform of the transformer,a cusp wave appears,which is prone to modal aliasing when EMD decomposes it.To solve this problem,this thesis combines the EMD algorithm with the permutation entropy algorithm to form the MEEMD algorithm,and simulation results show that the proposed algorithm can suppress the modal aliasing effectively.It has better effect in the research of transformer inrush current identification.(3)Due to the use of large-capacity transformers,the second harmonic content in the inrush and the dead angle is reduce,and the second harmonic braking and the dead angle principle may cause the protection to malfunction.In this thesis,the MEEMD algorithm can accurately identify the second harmonic even when its content decreases.According to the characteristics of the waveform,the identification criterion of transformer inrush current is proposed and verified by simulation.The specific steps are as follows:Firstly,the MEEMD algorithm is used to decompose the transformer inrush current and the fault current waveform.and then the first IMF signal is extracted for Hilbert transform to obtain the instantaneous frequency.Finally,the inrush current and inner fault current can be identified accurately by the time interval between two adjacent abrupt points of the instantaneous frequency.The correctness and reliability of the proposed criterion are verified.