Study On Coal And Rock Character Recognition Method In Fully Mechanized Caving Faces

Over50%in quantity of coal production is exploited via fully mechanized caving mining. Topcoal caving is under manual control by coal workers through the observation of coal and rock traitswith naked eyes or ears, which results in "less caving" or "over caving" phenomenon easily becauseof many restrictions in fully mechanized caving face. The author analyzed the technical bottlenecksof the traits identification of coal and rock. The top coal caving experiments were carried out atfully mechanized faces underground, and the on-spot data in quantity were acquired. The vibrationdata from different measuring points under different conditions were analyzed in time domain andwavelet Packet analysis, found that (1) the measurement point at the rear beam of the sublevelcaving hydraulic support were the better point,(2) the kurtosis index is more sensitive to theworking conditions,(3)the distribution of the frequency band energy is different under theconditions of the caving of top coal and the caving of top rock. The acoustic data at the rear beamof the sublevel caving hydraulic support under different conditions were analyzed in time domainand wavelet Packet analysis, found that (1) time domain indexes are changed under differentworking conditions, and the variance's variation is most obvious,(2) the frequency band energydistribution is different under different working conditions. The gray level histograms of the imagesignal at the coal falling outlet were studied under the conditions of the caving of top coal and thecaving of top rock and the mean values were calculated, found that (1) the gray level of the imagemainly distributes in the range of10-100under the condition of coal falling, while in range of90-220under the condition of rock falling,(2) The mean values of the gray level are different obviously, about66under the condition of coal falling while about130under the condition of rockfalling. At last, the method on coal and rock character recognition were investigated based onvibrating signal, acoustic signal and image signal respectively. This Paper riches the theory andmethod on the traits recognition of coal and rock in fully mechanized caving mining, and providessome theoretical basis and technical premise for the automation and intellegence offull-mechanized caving mining.