In order to reduce the high-decibel low-frequency noise generated by power transformer

it is necessary to analyze the active control of the transformer noise system. According to the time-domain and frequency-domain characteristics of the transformer noise and the experiment condition

we selected an appropriate type active noise control system. On the basis of simulation and theory analysis

we investigated the effects of step side

the filter order

sampling rate on the stability

convergence

and noise reduction of LMS adaptive filter algorithm for transformer power noise reduction. Then we designed the transformer noise on-line monitoring and active noise control software system

and chose the hardware devices based on the noise characteristics. The results show that

in the transformer noise reduction exp

eriment

the low frequency part of the noise will be reduced effectively

and the system has achieved 5

1

0 d B noise reduction and 68.38%

9

0% decline in average sound energy density

which verify the feasibility of LMS algorithm and the reliability of on-line monitoring system with software platform and hardware devices.