A wavelet de-nosing algorithm of XLPE cable partial discharge signals based on chaotic simulated annealing

For the poor adaptive capability and de-nosing effect of wavelet de-nosing for partial discharge detection in XLPE cable, a wavelet de-nosing algorithm based on Chaotic simulated annealing was proposed to improve adaptive soft threshold value. And then we verify its validity. The algorithm combined Global Chaos rough search with local simulated annealing fine search. This algorithm employs generalized cross-validation criteria for selecting adaptive threshold and obtains optimal wavelet soft threshold after processing. Then the collected raw signalde-noised. Finally, we met the requirement of retaining useful information and effectively reducing noise. The results show that SNR increase of at least 13.3% by using this algorithm compared with other traditional threshold methods. It can be applied to the detection signal de-noising on site, which is beneficial for the assessment of XLPE cables running.