Research On Dual Wavelength Coaxial Optical Fiber Sensor Used For Detecting Steel Ball Surface Defects

Rolling bearing is widely used as important basic components in mechanical equipment.As the rolling body of ball bearing,the quality of steel ball has an important effect on the motion performance,precision and life span of bearing.According to statistics,the failure rate caused by surface defects of steel ball is over 58.8% among the factors that cause bearing failure.Therefore,the surface quality of steel ball is the key factor of bearing comprehensive performance,and strict control of steel ball surface defect detection process is the key to improve the surface quality of finished steel ball.At present,the domestic steel ball surface defect detection equipment,in the measurement accuracy,detection efficiency,accuracy and so on has the big disparity compared with the foreign equipment.In this paper,through the in-depth study of the optical principle on surface defect detection of steel ball,a new optical fiber sensor is designed to detect the surface defect of steel ball.According to the types of steel ball surface defects and the evaluation standard of steel ball surface defects in mechanical industry,the defects are analyzed parameterized,and the displacements and roughness parameters are determined to evaluate the defects.The principle and feasibility of optical method for detecting defects characteristic parameters are analyzed.Combined with the type,application and detection principle of optical fiber sensor,the model of reflective intensity modulated fiber optic sensor is determined.On the basis of detection need,a scheme for measuring the surface displacement and roughness characteristic parameters of steel ball by dual-wavelength coaxial optical fiber sensor is proposed and formulated.This paper analyzes the distribution model of optical fiber output intensity,the effective receiving area of optical fiber,the structure of sensor probe,and so on,which influence the modulation characteristics of light intensity.The optical fiber intensity is selected as the quasi Gauss distribution model,and the optical fiber sensor probe is a double loop coaxial structure.Through the analysis of spherical measurement of optical fiber sensor,the measurement model of displacement and roughness and the expression of intensity modulation function of optical fiber sensor are established,which lays a theoretical foundation for the establishment of model and simulation analysis of optical fiber sensor.Based on the light intensity modulation function,the mathematical model of the sensor is established by using Matlab software.The factors such as the core diameter of transmitting fiber,the core diameter of the receiving fiber,the numerical aperture and the fiber spacing are simulated and analyzed.It is confirmed that the same size parameters are used for transmitting and receiving optical fibers.The core / cladding diameter is 105/125 ?m,the numerical aperture is 0.11.Every circle fiber on the end surface of the sensor is arranged closely,and the initial installation distance is 0.8mm.The influence of wavelength and roughness parameters on the characteristics of light intensity modulation system is analyzed,then two wavelength light sources,1310 nm and 1550 nm,are selected.A optical fiber sensor is made.The sensor calibration test platform and the steel ball detection platform were set up to test the dual wavelength coaxial optical fiber sensor.Based on the experimental analysis of displacement and roughness calibration of optical fiber sensor,the results show that the maximum error is 1.67% and 2.32%,respectively,which mean the test results are very good.In the testing experiment of steel ball,the displacement and roughness output curves of different types of defective ball are different,according to the difference of the curve,the surface condition of steel ball can be judged.The neural network is used to process the detection data.The result shows that the recognition rate of defect ball is 88.5%,which can meet the requirements of automatic detection data processing and defect ball proportion statistics.It lays a foundation for the development of steel ball surface defects detector.