Research On Leakage Protection Method Of Downhole Frequency Conversion System Based On PSO-VMD

With the rapid economic development,the frequency converter is used in large quantities in the coal mine.Due to the special structure of the frequency converter,a large number of harmonic components will appear in the feed line,causing misoperation and failure characteristics of the low-voltage leakage protection in the coal mine.It is difficult to extract the problem of weight,which poses a serious threat to the safety production of coal mines and the personal safety of workers.Therefore,it is of great practical significance to study the method of low-voltage leakage protection of coal mine down-conversion system.First of all,the causes of misoperation of leakage protection of underground ungrounded frequency conversion system in coal mines are analyzed to distinguish the zero-sequence currents under normal conditions and fault conditions,and the fault zero-sequence currents are screened out,and the fault line and non-fault of coal mine down-conversion system are analyzed.Fault line characteristics analysis,study of the original fault line selection method in the coal mine downhole frequency conversion system fault line fault in the selection.This paper proposes a Particle Swarm Optimization(PSO)optimization variational mode decomposition(VMD)method combined with an Extreme Learning Machine(ELM)method to implement fault line selection of coal mine underground frequency conversion system.Variational modal decomposition is a new signal processing method.This paper deeply studies the basic principle of the algorithm,and finds that its parameter combination is difficult to set accurately.Therefore,the particle swarm algorithm is used to optimize its parameters,using simulation signals and simulation experiments.It is verified that the variational mode decomposition of particle swarm optimization is more advantageous than the empirical mode decomposition(EMD)in the extraction of leakage faults in coal mine down-conversion systems.On the other hand,the basic principles of ELM and BP neural network are introduced in this paper,and the energy proportion of the integrated modal components of each feeder circuit after the variational modal decomposition of particle swarm optimization is input as a feature into the ELM and BP neural network.Its output identifies faults.Through simulation analysis,in the selection of low-voltage leakage faults in coal mine down-conversion system,the leakage fault line selection method based on PSO-VMD combined with ELM is more time-consuming than PSO-VMD combined with BP neural network.Shorter,more accurate.Finally,the feasibility and correctness of the proposed method are verified by building a 1140 V coal mine down-conversion system model simulation.That is,firstly,the zero-sequence current of the fault and the zero-sequence current in the normal condition are extracted by the variational modal decomposition of the particle swarm optimization,and the fault zero-sequence current is selected.Secondly,the extracted zero-sequence current is used to extract the integrated energy proportion of each feeder circuit,through the ELM for the correct fault line selection.