| The subject is an important part of national science fund projects “Research on Intelligent Recognition Method of Cutting Dynamic Load of the Super-heavy Rock Roadheader Based on Multi Parameters”(NO: U1510112).The research is aimed at solving the problem of difficulty in recognizing the hardness of cutting rock wall because of the complicated and harsh working environment of roadheader in the coal mine.At present,the driver of the roadheader cannot adjust the swing speed and rotation speed of the cutting head in real time according to the cutting state,which may cause the power of the roadheader to exceed the limit and the picks to be damaged.Therefore,a method for identifying the dynamic load of the super-heavy rock roadheader suitable for underground coal mines is proposed,which provides a basis for the roadheader to adaptively adjust the swing speed and rotation speed of the cutting head according to the cutting load.This research is of great significance.In this paper,through the in-depth analysis of the research status of roadheader and dynamic load identification technology at home and abroad,the existing intelligent level of roadheader is low,and the application of dynamic load identification technology in the cutting load of roadheader is relatively small.Therefore,based on the consideration of the harsh tunneling environment in coal mines,a method for identifying the cutting load of the roadheader based on multisource data fusion is proposed.The specific research contents are as follows:Through the deep understanding of the composition and cutting principle of the cutting mechanism of the roadheader,the dynamic analysis of the cutting mechanism is carried out under different cutting conditions.According to the classical calculation formula of the cutting head load of the roadheader,the main parameters affecting the cutting head of the roadheader include the cutting condition,the cutting rock characteristic,the cutting head depth,the rotation speed,the swing speed or the drilling speed.According to the design parameters of the cutting head of a vertical axis type roadheader,the cutting head load simulation program is written in MATLAB,and the influence of each parameter on the cutting head load is analyzed in depth,which provides the basis for the identification of the cutting head load of the roadheader.Through the analysis of the working principle of the cutting mechanism of the roadheader,the vibration signal of the cutting mechanism,the stator current signal of the cutting motor,the lifting and rotating cylinder pressure signals are used as the monitoring amount to reflect the cutting load of the roadheader.Through the theoretical derivation and simulation verification,the time-frequency relationship between the stator current of the cutting motor and the cutting load is analyzed.The relationship between the propulsion force of the two cylinders and the load of the cutting head is established by the force analysis of the lifting mechanism and the rotating mechanism,which provides a basis for the extraction of the feature quantity of the cutting load.According to the basic requirements of the cutting load identification of the roadheader,a method for identifying the cutting load of the roadheader based on multi-source data fusion is proposed.Firstly,a feature extraction method based on wavelet packet singular value decomposition is proposed.The optimal wavelet parameters of vibration signal,current signal and pressure signal are found with the maximum separability of signal feature vector.The wavelet packet decomposition of each type of signal is performed by using the selected optimal wavelet parameters,and a matrix reflecting the time-frequency information of the signal is constructed.Then the singular value decomposition of the matrix is performed,and separability criterion value based on divergence matrix is used to select singularities sensitive to hardness of different cut rock walls,and the value serves as the feature quantities of various signals.The feature quantity has better distinguishability for different cutting rock characteristics under different working conditions.Then,a probabilistic neural network(PNN)optimization algorithm based on differential evolution algorithm and QR decomposition is proposed,which solves the problem that the network structure of PNN is complex and the smoothing parameters are difficult to determine.The overall scheme of the cutting load identification of the rock roadheader is proposed.In order to reduce the complexity of the hardness recognition network and improve the accuracy,a method of identifying the load after identifying the working condition is adopted.Firstly,the wavelet packet singular value decomposition is used to extract the characteristics of the vibration signal,the cutting motor current signal and the cylinder pressure signal.The extracted pressure characteristics are used to identify the roadheader's working condition information through optimized probabilistic neural network.The LDA dimension reduction algorithm is used to reduce the eigenvectors of the three signal feature quantities to obtain low-dimensional features with better class separability,and form new feature vectors with the cutting parameters.Then the new eigenvectors are used to identify the hardness of the cut rock wall by optimizing the probabilistic neural network under different working conditions.The experiment shows that the method has a high recognition accuracy.The cutting dynamic load identification platform of the Super-heavy Rock Roadheader based on LabVIEW,MATLAB and SQL Server is designed.It mainly includes vibration signal,current signal and pressure signal feature extraction program,load recognition program based on optimized probabilistic neural network,load identification main interface,waveform display interface and data management interface,which realizes visual display of signal feature quantity and load identification information.The storage,query,modification and deletion of the load identification data are completed. |
PDF Full Text Request