| Steel wire rope is a spiral steel wire bundle made of steel wire,which meets the requirements of mechanical properties and geometric dimensions according to certain rules,and plays the role of traction and bearing.It has the characteristics of high strength,not easy deformation and enough flexibility.It is listed as a national standard industrial product due to its many advantages.It is widely used in tourism,construction,logistics,chemical industry,heavy machinery,robots,marine ships and other social and economic fields.Wire rope in the use process will still appear wear,corrosion,deformation and even broken phenomenon,its bad state directly threatens the safety of people and equipment.In this paper,the intelligent recognition system of wire rope defects is taken as the research object.Based on the mechanism of wire rope magnetic flux leakage,ICEEMD(Improved Complete Ensemble Empirical Mode Decomposition)and the back propagation neural network,the following aspects of design and research are carried out around the problems of wire rope magnetic flux leakage signal detection and recognition :(1)On the basis of summarizing the research on nondestructive testing of steel wire ropes in China and abroad,the principle of magnetic flux leakage testing of steel wire rope defects was analyzed,and the types of steel wire rope defects were summarized.A 360-degree circumferential annular uniform arrangement of magnetic flux leakage detection sensing system of steel wire rope was designed,which provided a reliable basis for the subsequent signal acquisition,analysis and detection of steel wire rope defects.(2)Aiming at the problems of complex operation and incomplete defect signal acquisition of traditional wire rope magnetic flux leakage signal acquisition device,a data acquisition system for wire rope defect magnetic flux leakage detection is designed.The system enlarges and filters the weak signal output by the sensor,collects all 16 channel signals synchronously with equal distance and small interval to STM32 microprocessor,and realizes the functions of data storage,data processing and analysis of wire rope magnetic flux leakage signal.It meets the requirements of complete acquisition signal and high accuracy of acquisition signal in magnetic flux leakage detection of wire rope defects.(3)Aiming at the problem of wire rope defect signal processing,an ICEEMD-WTF-WF multi-stage noise reduction method is proposed.Firstly,the wire rope defect signal is analyzed.According to the signal characteristics,the defect signal is adaptively decomposed based on the mathematical characteristics of the signal by using the ICEEMD method,and the IMF(Intrinsic Mode Function)series are obtained.According to the data relationship of energy ratio,permutation entropy and correlation coefficient of each IMF component,the IMF stock wave noise component is taken out,and the WTF(Wavelet Threshold Filtering)is used to remove the stock wave noise in the frequency concentration,so as to screen the useful IMF component.The reconstructed signal is smoothed by WF(Wiener Filtering)method to eliminate random environmental noise.The effective removal of stock wave noise and random noise and the full retention of defect signal characteristics are realized,which lays a good foundation for the accurate identification of small defects in wire rope.(4)Aiming at the problem of classification and identification of wire rope defects,a precise identification model of wire rope defects based on DE-BP neural network is constructed.On the basis of expounding BP neural network and differential evolution algorithm,the BP neural network is optimized and improved by introducing DE(Differential Evolution)algorithm,so as to improve the overall effect of wire rope defect recognition system and realize the accurate recognition of wire rope defects.After the comparison and verification of the model,the results verify that the defect judgment rate of the DE-BP algorithm model is higher than that of the traditional BP algorithm model. |
PDF Full Text Request