SONG Renjie, ZHAO Meng, GUAN Xiaozhuo, et al. Assessment of Gas Insulation State in GIS Equipment Based on Relevance Vector Machine[J]. High voltageapparatus, 2016, 52(12): 237-243.
SONG Renjie, ZHAO Meng, GUAN Xiaozhuo, et al. Assessment of Gas Insulation State in GIS Equipment Based on Relevance Vector Machine[J]. High voltageapparatus, 2016, 52(12): 237-243. DOI: 10.13296/j.1001-1609.hva.2016.12.039.
Assessment of Gas Insulation State in GIS Equipment Based on Relevance Vector Machine
A multi-parameter assessment method of the gas insulation state in GIS equipment is proposed. Factors affecting the insulation conditions of GIS equipment are analyzed in detail
and the most influencing factors determined. Assessment of insulation using a parameter set of gas composition
temperature
humidity
pressure and purity is carried out. Finally
a one-on-one(ORO)classification method is used to establish a GIS insulation state model based on a relevance vector machine(RVM). The GIS gas insulation condition can be divided into four categories corresponding to "normal"
"attention"
"abnormal"and "malfunction". The model is trained with training samples before it is used for the evaluation of gas insulated equipment state. A use case analysis shows that the method has better accuracy in the classification of insulation state of the GIS equipment.