Sound quality control method of 110 kV transformer based on distributed variable stiffness dynamic vibration absorber.

Using the A-weighted sound pressure level (L) as the only evaluation index for transformer noise is unreasonable. To measure the actual impact of transformer noise on humans more effectively, this paper proposes to introduce the sound quality evaluation method into the evaluation of transformer noise. First, the field noise of a 110 kV transformer was measured, and the objective parameters and subjective scores of the noise samples were analyzed, and a subjective prediction model of sound quality was established using the multiple linear regression method. Subsequently, a sound quality control methodology was proposed, which targets noise attenuation at the 100 Hz fundamental frequency and its first six harmonic components, while excluding non-peak frequency bands. To attain this goal, band-stop filters were employed to attenuate the noise energy at the specified frequencies. Following that, subjective prediction models were applied to analyze the filtered noise data, thereby validating the effectiveness of the proposed approach. Additionally, a distributed variable stiffness dynamic vibration absorber (DVA) was designed to mitigate structural-borne noise and improve sound quality. Laboratory experiments quantified its vibration and noise reduction efficacy: with a single-DVA configuration, vibration reduction rate was over 30%, while the noise level decreased by ≥ 1 dB. Finally, a field noise measurement was carried out. The results demonstrate that the linear sound press level decreased from 82.2 dB to 76.2 dB, the loudness decreased from 24.0 sone to 18.3 sone, and the subjective score improved from 4.7 to 6.1, meaning that subjective feelings were improved by two levels.