A Comparative Experimental Study Between LMS & FXLMS Algorithm on Industrial HVAC System
With the recent technology enhancement in industrial Heating, Ventilation, and Air Conditioning (HVAC) systems, there has been a significant increase in noise pollution. Active noise reduction (ANR) was introduced to help in reducing the noise from the industrial HVAC system. This is a better alternative when compared to passive noise reduction (PNR) which is costly and unable to reduce lower frequency ranged noises. The ANR system in this paper employed LMS and FXLMS algorithm with different step-sizes, testing approach and placements of microphone and speakers to perform noise reduction. The system was tested using simulation testing, singleton testing, multi-tone testing and industrial testing. Each testing was done using different step-sizes to observe the convergence rate and steady-state performance in different positions of microphone and speakers. The simulation testing shows the convergence rate to steady state of LMS algorithm was higher compared to FXLMS algorithm with the step-size 0.00001. Conversely, the FXLMS algorithm performed better than LMS algorithm during the singleton and multi-tone with 0.001 step-size. The similar trend was also observed in industrial testing with 0.00001 step-size. To conclude, the use of different step-sizes and position of microphone and speakers affect the noise reduction when using LMS and FXLMS algorithm for the ANR system.