Analysis And Research On Fault Location Method Of Hybrid Three-terminal HVDC Transmission Line

Hybrid HVDC transmission technology has a broad prospect and has become the mainstream of HVDC transmission engineering because it has the ability to be independent of active and reactive power adjustment and can avoid commutation failure and transmit power over long distances with large capacity.As a new type of transmission method,the LCC-MMC hybrid three-terminal DC transmission system has problems such as strong structural complexity and difficulty in locating line faults.Therefore,it is necessary to study the fault location method of the hybrid three-terminal DC transmission system.The rapid recovery of faulty lines and the improvement of power supply reliability have certain value and significance.The research status of HVDC technology and its fault location algorithms are illustrated.Then,the model is built and verified in PSCAD/EMTDC by analyzing the topology structure,operation principle,mathematical model,control method and modulation mode of hybrid three-terminal HVDC system.In order to make up for the deficiency of traditional location method in precision,a fault location method based on the combination of wavelet packet and Elman neural network is designed for hybrid three-terminal HVDC system.Firstly,the voltage and current fault components of the measured points after the fault are collected.Secondly,the transient current fault component method is used to select the lines.Then,the spectrum energy of the voltage traveling wave with rich fault information is extracted by the wavelet packet algorithm,which is used as the input of Elman.By setting different transition resistors in the case of failure,MATLAB simulation verifies that this mechcsd has a strong ability to withstand transition resistors.A hybrid three-terminal HVDC fault location method based on the combination of wavelet packet and BP neural network is designed.Firstly,the grounding fault is set in the PSCAD model,and the fault voltage and current data are collected at the measurement points.Secondly,the fault section is identified by line selection method,which extracts the energy ratio of low frequency and low frequency of the fault components of the current by wavelet packet.Then,the energy obtains from the fault voltage is decomposed and reconstructed by the three-layer wavelet packet,which is used for the training of BP neural network.The final output is the specific location of the fault.By setting the transition resistance and noise interference,the simulation results show that this method is suitable for fault location of hybrid three-terminal HVDC transmission lines with high accuracy.By comparing and analyzing the two fault location methods proposed,it can be seen that both of them have high ability of resistance to transition resistance,and can also be resistant to a certain degree of noise interference,and the location is accurate.Because the accuracy and speed of the two methods are different in fault location,the methods can be selected according to different positioning requirements and actual situations.