文档介绍:第52卷第2期 航空计算技术 as time domain features
and arc current as frequency domain features. The frequency domain features were analyzed by discrete
fractional Fourier transform. The class divergence was used as an evaluation index to obtain the feature
values in the time frequency domain. Combining the feature importance obtained from the random forest
self- test function to construct a new combination weight. Under the pure resistive, resistive inductive and
resistive capacitive conditions,taking aviation 28 V,270 V DC series and parallel arc as the research ob­
jects ,the fault classification effects of the proposed method and the existing random forest were compared.
The results show that the time domain and frequency domain features used in this paper can effectively
distinguish the degree of arc faults under different voltages and load conditions. The proposed weighted
random forest improves the fault classification accuracy by % compared with the existing random for­
est ,with better diagnostic