Novel knowledge-driven active management and control scheme of smart coal mining face with digital twin

In the previous digital twin system of smart coal mining face,the knowledge chain embedded in 'environment-equipment-material' twins is not fully excavated, and the experiences of operators for determining and controlling the mining operation are not exploited. Especially,a rational information interaction mode between physical space and virtual space is necessary. However,a signal passive information interaction method is now employed. All above weaknesses lead to the inefficient adaptability of operation control and safety decision-making for the uncertainty of mining process,and the poor robustness of comprehensive control from virtual space to physical space,even the low utilization of cooperative physical entities. To address the above issues,a new active control and management mode for virtual spaces is proposed to deeply mining and utilizing knowledge from smart intelligent face with digital twin. The five-dimensional digital twin framework of smart mining face is constructed. The specific meaning of each dimension is illustrated in detail. Following that,the knowledge is classified in terms of their attributes along the whole process of smart mining face. More especially,three kinds of prior knowledge are described by rules and models. After analyzing coupling relationship among multi-energy flows of 'environment-machine-thing',the dynamic evolution model of twin knowledge is built. Finally, a knowledge-guided active control mechanism for virtual and real spatial information is developed. It is solved by machine learning and optimization methods,which provides a new paradigm of extracting,transferring and utilizing knowledge in the digital twin smart mining face. The proposed new active control mode can cope with the uncertainty of geological conditions,the spatio-temporal dynamics of environment and the diversity of equipment during the mining process. This provides a technical support for the efficient low-carbon intelligent mining,and the high-fidelity mapping framework and decision-making control mechanism at the knowledge level to promote "information visible,track traceable and state traceable"in parallel mine.