Study Of High-Speed Train Interior Noise Adaptive Control Technology Based On DSP

With the completion of a large area of high-speed railway, the increasing traffic demands, improving Electric Multiple Units operating speed continuously and the increasing carriers density, railway noise pollution also growing. Compared with ordinary trains, high-speed train noise is more complex, has longer transmission range, causes greater harm to train staff and residents along the railway. Traditional trains mainly optimize the structure, use new materials and install sound barriers, which called passive noise reduction mode. These methods from modeling and simulation analysis to actual application need a long time and large investment. So propose the use of active noise reduction mode which uses electronic equipment in the sound propagation path without changing the train structure.This paper firstly introduces the least mean squared error criterion adaptive algorithm principle (LMS). Research the filter weights coefficient update formula by steepest descent method. Analyze filter order and convergence factor cause effect in convergence rate and steady-state error of adaptive systems. Secondly analyze Filter-XLMS algorithm and delay LMS algorithm in discrete systems. Simulate and model the secondary path adaptive identification and delay response for guidance and reference help in Simulink of MATLAB. Thirdly take TMS320VC5509A type Digital Signal Processor (DSP) as the core to build a high-speed train active noise control hardware system. Introduce in detail clock synchronization, control manner and data transmission between DSP and AIC23 or external memory. Put forward a stable power supply system. Use C/C++ mixed language to program LMS algorithm for simulation and software analysis in Code Composer Studio (CCS), which is used for driving hardware system, the secondary path off-line identification and adaptive noise reduction function. Finally Active Noise Control system is installed in the broad sense comfort EMU experiment platform. Analyze noise reduction in single-frequency noise of 500Hz and 450～650Hz sweep noise noise. Test Active Noise Control system actual noise reduction effect.