Open AccessXLPE cable health assessment based on Relief-F feature weighted FSVM
Author(s) -
Xiaohuan Wu,
Yang Liu,
Wang Li-qun,
Xiaozhen Ren,
Xianjun Tan
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/675/1/012147
Subject(s) - feature (linguistics) , noise (video) , dimension (graph theory) , fuzzy logic , computer science , process (computing) , reliability engineering , feature vector , data mining , statistics , pattern recognition (psychology) , engineering , artificial intelligence , mathematics , philosophy , linguistics , pure mathematics , image (mathematics) , operating system
For the cross-linked polyethylene (XLPE) cable health status assessment process, there are noise data for each feature parameter and the importance of each dimension feature is difficult to determine, which leads to deviations in the cable status assessment results. An XLPE cable health assessment algorithm based on Relief-F feature weighted fuzzy support vector machine is proposed. The Relief-F algorithm is used to calculate the importance of each dimension feature. The feature quantity that is weakly related to the evaluation of the health status of the XLPE cable is deleted. Furthermore, the fuzzy support vector machine is used to evaluate the health status of the XLPE cable, which gives the noise data a smaller degree of membership and reduces the influence of the noise data on the evaluation results. Experimental analysis shows that this method can accurately assess the health status of XLPE cables. It can provide a reference value for relevant departments to carry out cable maintenance and replacement work and save the human resources, and material resources required for maintenance and replacement of cables.