Pressure Sensing Recognition Based On FBG Flexible Sensor Array

In recent years,with the rise of optical fiber communication technology,various new optical fiber sensors have emerged in all walks of life.In this paper,a fiber Bragg grating(FBG)flexible sensing array is designed.Based on the mapping relationship between the wavelength drift of the reflection spectrum center and the change of the structural state characteristic parameters,a high-precision mathematical prediction model of the position static load is established,and the pressure sensing prediction algorithm of the FBG sensing array is studied.Firstly,the FBG flexible sensor model is established by COMSOL simulation software.When the friction coefficient of the contact interface between the packaging substrate and the fiber is set to 0.75,the error caused by the mutual slip between the two can be ignored.On this basis,the influence of packaging thickness and packaging position on FBG stress feedback is studied.The results show that the packaging substrate thickness is 4 mm,and the packaging position is half of the substrate thickness.At this time,the linearity of the positive load and the stress measured by the domain point probe is good.Based on the discussion of packaging thickness and packaging position,the packaging method of this flexible FBG sensor array is finally determined.Then the flexible FBG array is divided into 16 regions,four FBG layout modes are designed,and the flexible FBG sensing array model is established.Subsequently,the prediction models of random forests were established by combining the four FBG layout methods and the stress values of domain point probes.The stress measured by the domain point probe is taken as the input parameter,and the position coordinate X-Y and the static load F are taken as the output parameters.The prediction results show that compared with several different layout modes,the RMSEs of X coordinate,Y coordinate,static load F prediction point and real sample point of layout modes from 4-FBG to 12-FBG are decreased by 71.45%,77.81%and 97.52%,respectively.The determination coefficient R~2increased by 3%,2.8%and 14.6%,respectively.The maximum error of the distance between the predicted point of the comprehensive coordinate X-Y and the static load F and the real sample point decreases by 91.39%,and the average error decreases by 88.38%.With the increase of the number of FBG,the RMSE of coordinate X-Y and static load F load gradually decreases,the determination coefficient R~2is closer to 1,and the prediction accuracy is gradually improved.Finally,by comparing the prediction results of the four layout methods,the 12-FBG flexible sensor array is selected as the theoretical basis for subsequent experiments.Finally,the fabrication methods of flexible FBG sensing array are introduced,and the flexible FBG load identification system is built.The weight of 100 g to 1000g is applied to different regions of the flexible sensor array one by one according to a certain trajectory,and the MOI demodulation equipment is used to record the center wavelength change data of 5 s.The wavelength variation is used as the input of the random forest prediction model,and the coordinate position and static load F are used as the output to establish the prediction model.The prediction results of X-Y and F show that the RMSE of the three prediction parameters are less than 0.1,and the correlation coefficients are greater than 0.99.Combining the three prediction indexes,the distance between the actual point and the predicted point is calculated.The minimum distance error is 0.03491,the maximum is 0.2481,and the average error is0.1515.