Variable convergence train noise active control algorithm based on feedforward reference

• Published in: 机车电传动 pp. 25 - 30
• Main Authors: LI Tao, HE Yuyao, WANG Ning, LI Zhuang, SHEN Zhaoyuan, XIAO Gang
• Format: Journal Article
• Language: Chinese
• Published: Editorial Department of Electric Drive for Locomotives 01.01.2022
• Subjects: active noise control; feed-forward signal reference; low-frequency train noise; simulation; variable convergence coefficient
• ISSN: 1000-128X

An active control algorithm for train noise based on feed-forward reference with variable convergence was studied. By constructing a variable convergence function based on the feed-forward signal reference, the non-zero offset problem of the system was improved. Then it was applied to the low-frequency noise reduction of trains. Simulation results show that the algorithm can improve the balance between convergence speed, tracking speed and steady-state error. And the experimental results show that the algorithm achieves an average noise reduction of 12.0 dB for low-frequency train noise which below 1 000 Hz in the target area. Compared with the fixed-convergence FxLMS, variable-convergence NLMS, and SVSLMS control algorithms, it has the advantage of noise reduction up to about 3.6 dB.