Research On Key Technology Of Coal And Rock Cutting State Identification

90% of China’s coal is mined by underground mining,and the intelligent degree of mining equipment is low,which leads to many coal mining disasters,weak adaptability of coal machinery,high failure rate,and low efficiency.Improving the intelligent level of coal machinery equipment is one of the main tasks of coal mine intelligent development.Shearer is the core equipment of a fully mechanized mining face.The accurate identification of its cutting state is not only the key to realizing the intelligent and efficient cutting of shearer,but also the necessary basic guarantee of intelligent unmanned mining in a fully mechanized miningface.This paper combined the field test sampling,basic theory of CPS perception analysis,construction technology of complex coal seam,virtual prototype technology,bidirectional coupling technology,data processing theory,image fusion method,deep learning theory,and the basic principle of Wireless Sensor Networks.It focuses on the key technologies of the coal and rock cutting state identification,such as "the construction of high-precision three-dimensional simulation model of complex coal seam that can update and replace particle set,the acquisition and processing of the coal and rock cutting state data information set,sample expansion,identification and classification,and the design of information transmission network with high timeliness and low loss".The main research work and results are as follows:The overall architecture of coal and rock cutting state recognition system was designed according to the characteristics and requirements of coal and rock cutting state recognition system,integrating the concepts of the sensor,actuator,controller,and network in the traditional CPS sensing analysis structure,and integrated the virtual prototype technology.The specific implementation of the coal and rock cutting state recognition system was also analyzed.The typical complex coal seam containing gangue,inclusion,and minor fault in Yangcun mine of Yanzhou mining area was taken as the engineering object.Based on the test results of coal and rock physical and mechanical properties,the plug-in application of the complex coal seam was compiled using the EDEM application program interface technology.Also,the high-precision three-dimensional simulation model of the complex coal seam that can update and replace particles was constructed.The rigid-flexible coupling pattern of shearer cutting section was constructed based on virtual prototype technology.The EDEM and Recur Dyn were used to build the high-precision three-dimensional simulation model of complex coal seam and the bidirectional coupling model between shearer cutting parts,and the simulation group was set based on the actual cutting state of the shearer.Based on the simulation results of Discrete Element Method-Multi Flexible Body Dynamics(DEM-MFBD),the one-dimensional original vibration acceleration signals of the key components of the shearer cutting part were determined,including spiral drum,rocker arm shell,and square head.After transforming one-dimensional original signal data into the time-frequency images by STFT,morphological wavelet image fusion technology was used to realize the effective fusion of characteristic information of spiral drum,rocker arm shell,and square head under different working conditions.The basic database for coal and rock cutting state recognition was also constructed to provided original samples with rich and precise characteristic information for the expansion of recognition system data.Based on the deep learning theory,a DCGAN-RFCNN recognition network model with strong ability to perceive the cutting state information of coal and rock and rapid recognition and classification was constructed.By taking the characteristics of the vibration time-frequency sample image of the coal and rock cutting state as the perception basis,an improved DCGAN image expansion model was designed.By optimizing the structure of the generator model and discriminator model of the network,its ability to synthesize high-quality sample images was improved,so as to enrich the data base information.The improved RFCNN algorithm combining CNN convolutional neural network and random forest recognition classifier was used to train the recognition model.The training set and test set of the network were composed of the original data samples obtained by simulation and the data samples synthesized by the augmented model.Through the comparative experimental analysis of the RFCNN network model with different recognition classification layer models,different recognition network models,and different synthetic sample numbers in the recognition network,the effectiveness of the recognition network model was verified.The results show that: When synthetic samples are not included in each working condition in the RFCNN model,the average recognition rate is 90.641%.With the continuous increase of synthetic image samples in the network,the evaluation index of the coal and rock cutting state recognition accuracy changes rapidly.When 5000 synthetic image samples are integrated into each working condition,the average value of recognition accuracy is the highest,the variance value and average deviation value are the smallest,the recognition effect of the model reaches the best state,and the generalization ability is also enhanced.Therefore,using the designed improved DCGAN network to enrich the image sample database of the coal and rock cutting state has significant advantages for the efficient perceptual classification of recognition network.Also,the RFCNN outperformed the other variants: CNN-K,CNN-Softmax,and CNN-SVW,in terms of recognition accuracy and inference speed.Further,it obtained higher recognition accuracy by 25.805%,12.958% and 9.326%,respectively,over SVW,CNN,and Alex Net.Also,the experimental platform of shearer cutting coal and rock was built,where the coal and rock cutting state recognition network was trained and tested based on the migration learning theory.Through the statistical test results,the accuracy of coal and rock cutting state recognition is 98.64%,which realizes the accurate recognition of coal and rock cutting state.We also designed the topology of the WSN system for transmitting coal and rock cutting state data information in a coal mine,where the routing protocol of the WSN system was improved by combining fuzzy C-means(FMC)with particle swarm optimization(PSO).Combined with Matlab/Simulink and Ptolemy II platform,an information transmission network system(UCM-WSN)with a high prescription and low loss in the cutting state of structural coal and rock was designed.The results show that the data transmission accuracy of the system reached 98.94%,which is outperforming performance compared with the traditional mesh structure WSN system.It indicates that the timeliness of data transmission of coal and rock cutting state information was improved,and network energy loss was reduced.This paper has 164 figures,26 tables,207 references.