Active interior noise control for rail vehicle using a variable step-size median-LMS algorithm. (1st September 2018)

• Record Type: Journal Article
• Authors: Guo, H.; Wang, Y.S.; Liu, N.N.; Yu, R.P.; Chen, H.; Liu, X.T.
• Is Part Of: Mechanical systems and signal processing. Volume 109(2018)
• Page Start: 15
• Page End: 26
• Publication Date: 2018-09-01
• Subjects: Rail vehicle interior noise -- Active noise control (ANC) -- Median least-mean-square (MLMS) algorithm -- Variable step-size MLMS algorithm; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2018.02.040
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: A novel variable step-size median-LMS (VS-MLMS) algorithm is proposed. Parameters in VS-MLMS for active noise control of rail vehicle are determined. The VS-MLMS algorithm is compared with other 2 improved LMS algorithms. The VS-MLMS shows advantages for suppressing interior noise of a rail vehicle. The VS-MLMS can be used for ANC of other noises with impact components. Abstract: A variable step-size median least-mean-square (VS-MLMS) algorithm, for active noise control (ANC) of rail vehicle interior noises is presented in this paper. Firstly, the interior noises of an urban rail vehicle under normal operating conditions are measured and analysed. Results show that the interior noise includes obvious impact noise components, and the sound energy of this noise is mainly concentrated in a low-frequency range below 1000 Hz. Utilising the measured interior noises, the adaptive filter order, median window length, and the optimal step size of the MLMS algorithm are determined. According to the step-size scope, the VS-MLMS algorithm is developed using a Sine step-size adjusting function. The MLMS, VS-MLMS and other algorithms are applied to the measured interior noise signals. Results suggest that the proposed VS-MLMS, which can balance the convergence speed and the steady-state error well, is more suitable for active control of a low-frequency noise with impact noise components. The VS-MLMS algorithm can be directly used in an ANC system of the sample urban rail vehicle, which is helpful to improve the ride comfort of the passengers.