Coal-Rock Interface Recognition Based on Multiwavelet Packet Energy

The tradition way of recognition of coal-rock interface was usually done by detecting the gamma ray which has several shortcomings such as influence impurities in coal. Moreover, the geological conditions restrictions may have so much influence on the results that may lead this method to out of function. In order to overcome these shortcomings, in this paper the responses of shearer's cutting force was detected to monitor the shearer's cutting state. The response of shearer's cutting force was influenced by multiple factors, such as coal rupture form and working environment. This requires the signal should be processed by using multiple waveforms which can represent multiple factors and finally can find the response of shearer's signal which hide behind the mixed total signal. In order to solve this problem, a multiple scaling functions based multiwavelet algorithm was proposed which can represent the coal-rock interface characteristic signal. A characteristic library was been built by using multiwavelet band energy which can represent the coal-rock response feature of Shearer. By doing numbers of physical simulation tests, it is found that the multiwavelet band energy extract the coal-rock response feature has more advanced than that of the single wavelet analyses. Finally the paper propose a method of detecting the cutting coal-rock state by using Support Vector Machines (SVM) which provide the theoretical basis of the development of simple practical coal-rock Interface Recognition devices.