Research on Spindle Fault Classification Network of Twin Transformer Coded Capsule CNC Machine tool
Spindle is one of the most important mechanical modules of CNC machine tools,timely detection of its faults can ensure the operation efficiency and machining accuracy of the machine tools.Therefore,a twin Transformer coding capsule CNC machine tool spindle fault classification network was proposed.The 2D preprocessing module was applied to obtain more completed vibration data of the spindle bearing of the original CNC machine tool.Then,an improved Transformer code feature extraction module was used to obtain deep-level vibration signal features.The feature mapping was realized through advanced capsule feature transfer network.Finally,the twin Transformer coding capsule classification network was used to complete the classification of CNC machine spindle fault samples.The XK7145 type milling machine was used to complete the fault diagnosis experiments of non-wear tool and wear tool under the condi-tions of healthy bearing,inner and outer ring fault bearing and ball fault.The results show that the average spindle fault diagnosis accura-cy of the proposed method can reach 95.1％,which is 6.9％and 12.3％higher than the methods of ISERAVF-net and VSCPC-net.The visual classification effect of the proposed method is better,and the experimental effect of detecting spindle faults is better.
NETL
NSTL
万方数据
