| Bearing cylindrical rollers are the core parts of cylindrical roller bearings,and the deviation of their diameters has an important effect on the bearing capacity and rotation accuracy of the bearing.At present,bearing manufacturers mainly use manual inspection methods for inspection,which have disadvantages such as high false detection rate,poor stability,and low efficiency.The same research team researched a bearing cylindrical roller diameter measurement system,which can achieve high-speed dynamic measurement of cylindrical rollers,but the dynamic error has a greater impact on the accuracy of the measurement system.The purpose of this paper is to improve the measurement accuracy of the measurement system,taking the measurement system as the research object,to analyze the dynamic error source of the measurement system,to simulate the dynamic error and dynamic characteristics of the measurement system,and to compensate the dynamic error of the measurement system through the neural network algorithm.Analysis of the dynamic error source of the measurement system.Firstly,analyze the structural composition and measurement principle of the measurement system;secondly,analyze the dynamic error source of the measurement system,analyze the influence of probe pre-elongation,measurement speed,rod deflection deformation and probe vibration on the measurement accuracy,and get the functional relationship between the measuring speed and the pre-extension of the probe when measuring the cylindrical rollers of different sizes;finally,through the beat analysis,the measurement speed of the measurement system when the working beat is satisfied is determined,which provides a theoretical basis for the subsequent simulation analysis.Simulation analysis of the dynamic error and dynamic characteristics of the measurement system.Taking three cylindrical rollers with diameters of 28 mm,32mm,and 34 mm as the research objects,the measurement process of cylindrical rollers is simulated and analyzed by ADAMS simulation software,and the displacement,velocity,and contact force simulation curves of the probe are obtained,which verifies the correctness of the theoretical analysis.Through ANSYS Workbench software to analyze the dynamic characteristics of the measurement system,including the modal analysis and harmonic response analysis,analyze its anti-vibration performance,and get the excitation frequency that affects the accuracy of the measurement system.Research on dynamic error compensation of measurement system.Three sizes of cylindrical rollers are selected for the experiment,60 of each cylindrical rollers are selected as the experimental sample,and the measurement value are obtained by the measurement system,and compared with the true value of the cylindrical roller diameter.Through MATLAB software,the measured value of the measurement system is used as the input of the neural network,and the real value is used as the output of the neural network,the error compensation models of the BP neural network and the RBF neural network are established respectively,and the validity and superiority of the two models are verified through experimental data.Comparing the error compensation results,it is concluded that the RBF neural network has higher error compensation accuracy,compared with the original error,its measurement accuracy is increased by 62.6%,92.4%,and 79.5% respectively,and meets the measurement accuracy requirements.In this paper,the dynamic measurement error compensation of the cylindrical roller diameter of the bearing is studied,and the high-speed and high-precision dynamic measurement of the cylindrical roller is realized,which provides theoretical and practical reference for the online dynamic measurement of the bearing cylindrical roller. |
PDF Full Text Request