Research On Intelligent Identification Technology Of Cutting Dynamic Load Of The Roadheader

Aiming at the problem that the dynamic load of roadheaders can not be obtained directly through measurement,which makes it impossible to accurately judge the state of cutting load and cause serious loss of pick.Therefore,developing intelligent load identification system for rock roadway driving is of great practical significance.In this paper,dynamic analysis of the dynamic load of the cutting head of a rock tunnel roadheader has been analyzed,and a new intelligent fusion recognition method based on data mining has been determined in combination with a series of dynamic response signals that can reflect the dynamic load.The contents of this paper are listed as follows:(1)In order to ensure the reliability of the data collected in the operation of roadheaders and the accuracy of dynamic load identification,the influence of the motion parameters of the cutting head on the dynamic load identification has been analyzed.The influence of the motion parameters of the cutting head on the dynamic load has been obtained by simulation analysis,and the motion parameters of the cutting head have been adjusted.(2)The mechanism of the roadheaders and the force analysis of the cutting head are discussed,and the sensitive signals that can reflect the driving load are analyzed,which are the cutting head vibration signal,the pressure signal of the lifting hydraulic cylinder and the current signal of the cutting motor,and the improved algorithm has been introduced into the research and application of the feature extraction.(3)According to the requirements of the driving load identification,a driving load identification scheme has been designed,including the data acquisition scheme,the feature extraction scheme combining the wavelet decomposition with the Hilbert transform,and the data fusion recognition scheme.The recognition effect of the typical traditional RBF neural network and topology structure for the single hidden layer feedforward neural network limit learning machine algorithm form a contrast and the recognition effect of feature level fusion and decision level fusion are compared respectively.Compared with the feature level fusion,the decision level fusion has shown outstanding advantages in fault tolerance,while the single hidden layer feed-forward neural network Extreme Learning Machine algorithm applied to the fusion algorithm not only ensure good recognition effect,but also improve the real-time performance of the overall scheme.On the MATLAB simulation platform,the software development of the driving load identification has been completed,and the signal feature extraction program and the data fusion intelligent recognition program such as the limit learning machine have been designed respectively.