Dynamic Update Method of Working Face Geological Model Driven by Multi-Source Data

In order to build a high-precision dynamic geological model to serve the intelligent mining, 11 working face is explored step by step through the comprehensive prospecting technology. A multi-source 12 data fusion method was applied to realize mutual verification, supplement, fusion and interpretation of 13 non-uniform heterogeneous geological data to obtain a high-precision geological data volume. Also, the 14 dynamic update model method was proposed to update 3D geological model of working face quickly so 15 that the accuracy of the geological model can be improved effectively. Furthermore, cutting path 16 planning technology was developed based on the dynamic geological model. The field test showed that 17 the cutting path planning based on the high-precision dynamic geological model can improve the coal 18 mining efficiency and improve the fusion efficiency between geology and coal mining systems. Dynamic 19 update of multi-attribute geological information should be studied and developed to improve the 20 automatic level of mining driven by geological data.


Introduction 26
Intelligent mining is the key of achieving the goal of safe, efficient and green coal mining, and the Before the excavation period, ground prospecting such as 3D seismic surveys, electromagnetic 82 prospecting, and ground drilling is implemented. Before mining period, further prospecting of inside and 83 near the working-face such as in-seam seismic survey, drilling logging, electromagnetic prospecting etc. 84 is implemented. During mining period, detailed prospecting in the mining area such as seismic surveys, 85 resistivity monitoring, geological cataloguing, dynamic image identification, etc.is implemented. 86 3 High-precision geological modeling method for working face 87

Multi-source data fusion 88
The multi-source heterogeneous data is featured with different sources and structures. In order to 89 construct a high-precision geological model, the multi-source heterogeneous data must be effectively 90 fused and unified into a same spatial coordinate system. Since the single geological exploring data source 91 is insufficient in prospecting accuracy and interpretation reliability, cross-validation method of 92 geological exploration data based on the spatial fusion of multi-source heterogeneous geophysical data 93 is proposed to improve exploration's overall accuracy and reliability. Liu et al. (2020) proposed the 94 spatial location of coal seams and the internal geological structure can be predicted by mutual cross-95 validation method of geophysical survey data, drilling data and mining exposing data. When seismic data 96 is available, the multi-source data fusion based on seismic dynamic interpretation can be applied. The 97 dynamic interpretation of geological and seismic data is used to achieve the goal of predicting the 98 geology structures and the coal seam surface of the working face. In static interpretation stage, layers are 99 labeled. Also, coal seam floor and structures are interpreted according to high-quality seismic 100 superposition or migration data set. Meanwhile, coal seam thickness and lithology of roof and floor are 101 predicted, according to wave impedance data volume, Quasi-natural gamma body, etc. The geological 4 information revealed by excavation roadway, including roof and floor, coal seam thickness and structures, 103 is collected using measurement technology. Then, the spatial form and structure of the coal seam floor 104 are updated with the constraints of the new geological information. The primary task of dynamic 105 interpretation is to determine the basic geological condition of the working face, build the basic 106 geological framework of transparent working face, detect the main geological anomalies, and provide a 107 high-precision data volume for the geological modeling of the working face.

Model interpolation algorithm 123
Constrained interpolation is applied with non-uniform discrete geological data points to construct a For constraints for interpolation of the working face C , consider the roughness as main constraints, 133 neglecting the soft constraints. According to the definition of smooth discrete interpolation, the global 134 roughness is calculated as follows: 135 In Equation (1) roughness that is calculated as follows: is vector function of the 141 known points. 142 After obtaining the global and local roughness, the implicit function in the working face can be built. 143 While neglecting soft constraints, the whole constrains function shows as follows: 144 is the soft constrains. 147 Then find the first-order partial derivative of the constraint equation with respect to  a  and let the 148 derivation function equals to 0.  a  is calculated as follows: Finally, the objective function is solved by equation (5) The optimal planning cutting curve of the mining working face in a specific range is calculated based 175 on coal mining requirements. Intelligent cutting is implemented based on the unified data integration 176 platform, using the data integration process between the central control system and the cutting curves

Multi-source geological data fusion analysis 217
Multi-source heterogeneous data were fused and unified into a same spatial coordinate system due to 218 the different sources and structures of data. For the prospecting data in time domain, time-depth 219 conversion is essential in spatial fusion. For the targeted intelligent working face, the geological anomaly 220 was predicted to be a coal erosion zone by cross-verification between sandstone zone revealed by 221 roadway and in-seam seismic survey. The coal erosion zone was verified by in-seam borehole logging. 222 In addition, the cross points between in-seam boreholes and coal seam surface were determined 223 comprehensively according to roof and floor lithology revealed by roadways, surface drilling, in-seam 224 seismic survey, and logging. 225

High-precision dynamic geological modeling 226
The multi-source heterogeneous geological data were processed using space-time fusion technology. 227 Then, discrete smooth interpolation (DSI) algorithm was applied to construct high-precision geological 228 model of the mining working face. The geological data collected before mining, including surface drilling 229 data, roadway refinement measurement, in-seam boreholes logging, were used to build static model.

Conclusions 256
(1) Before the mining period, the high-precision geological prospecting method could be applied to 257 obtain the geological conditions accurately. Also, according to alignment, cross-validation, and fusion 258 analysis of the geological data using multi-source data fusion method, the high-precision geological 259 model of the working face could be constructed. 260 (2) During the mining period, the high-precision dynamic geological model for intelligent mining 261 could be refined using dynamic update method on the basis of geological data revealed in the working 262 face. 263 (3) The field test shows that the cutting curve planned by high-precision dynamic geological model in 264 the mining period can be sent to the shearer through the central control center for automatic mining. 265 At present, the intelligent and automatic coal mining can be applied based on the high-precision 266 geological model. In future research, dynamic update of multi-attribute geological information, mining 267 disturbance effect analysis and automatic prospecting should be studied and developed to improve the 268 automatic level of mining driven by geological data. 269

Declarations 270
There are no available data and materials. The authors declared that they have no conflicts of interest 271 to this work. We declare that we do not have any commercial or associative interest that represents a 272 The authors would like to thank the editors and reviewers for comments and suggestions.