Analysis of coal gangue recognition capability based on vibration characteristics of the tail beam and experimental study on coal gangue recognition in fully mechanized top coal caving

Abstract

With the progress of control and communication technology, coal mining is gradually transformed and upgraded from traditional mechanism to automation, the concept of intelligent mining in fully mechanized caving is also put forward in the mining technology research of thick and extra-thick coal seams. As a key technology and necessary condition of intelligent mining in fully mechanized caving, intelligent recognition of coal and gangue is a technical bottleneck to be broken through, which has seriously affected the coal mining efficiency and recovery rate. Based on this, this paper carries out a systematic study on coal gangue recognition in fully mechanized caving from the perspective of tail beam vibration. Firstly, the impact-contact dynamic theoretical model between coal gangue and the hydraulic support tail beam and its rigid-flexible coupling dynamic simulation model was constructed respectively. The theoretical and numerical simulation was combined to study the vibration characteristics induced by the contact between coal gangue and the tail beam, and the vibration characteristics caused by the difference of coal and gangue properties were compared and analyzed. Meanwhile, coal gangue recognition capability is summarized, which provides theoretical support for the vibration characteristics-based coal gangue recognition. Secondly, through analysis of fully mechanized caving environment, a single coal drawing experimental platform and vibration information acquisition system are designed and built in the laboratory environment. Then, a research on signal eigenvalue recognition sensitivity based on “coal gangue binary classification” recognition was carried out through the combination of signal acquisition and features extraction as well as machine learning, the optimal recognition eigenvector based on recognition effect analysis was obtained. On that premise, recognition research of “coal gangue multi-category recognition” based on the single point and multi-point vibration signals were proposed and conducted, which successfully realized coal gangue mixture ratio recognition under the condition of multi-category. The research provided the theoretical basis and technical support for accurate and real-time coal gangue mixture ratio recognition in top coal caving.