Analysis Of Noise On Coal Mine Working Surface And Research Of Noise Reduction Algorithm

In the coal mine, the equipment noise seriously affects the normal voice communication between miners and between the staff and ground console. But till now, the problem of coal mine voice communication has not been effectively solved. To achieve normal and high-quality voice communication in mine working environment, a deep analysis about noise environment in the coal mine, noise and voice features is needed. Under the circumstance of high quality of voice communication, reducing environmental noise effectively to ensure normal communication between the staff. This not only improves the working efficiency but also reduces coal mine accidents.In this thesis, an analysis on characteristics of mine noise source and voice is made with MATLAB tool and three effective approaches for reducing mine noise are put forward. A simulated analysis and performance comparison between classic LMS adaptive algorithm and two kinds of variable step sizes adaptive algorithm is made. Based on the real mine noise environment, noise scenes about single noise source reflection, multiple noise sources and so on, are simulated. With collected coal mine noise, including coal winning machine noise, water pump noise, the trams noise, transmission belt mechanical noise, the three algorithms for complex mine noise filtering are tested on MATLAB simulation platform. Simulation results show LMS adaptive filtering algorithm has a large difference on noise processing effect when filter orders are different. In the real environment, because of the distance between the sound source and caller having a random variation, LMS filter algorithm whose filter order is fixed has some shortcomings in noise cancellation. Then a corresponding improvement is made on the basis of analysis on LMS adaptive filtering algorithm. Taking advantage of the characteristics of the filter weight vector, to estimate real-time distance between the source and the caller quickly and accurately, then real-time update filter order according to distance change, to improve the process effect of filter. Finally, the noise reduction algorithm is realized based on DSP.