Development Of Deformation Measurement System Based On Fiber Bragg Grating

With the development of the aerospace industry,drones are widely used in military reconnaissance,disaster relief,and agricultural production.The high-precision measurement in real time deformation is essential for guaranteeing the airfoil long-baseline antenna performance.The thesis researches on a deformation measurement system based on fiber Bragg grating.Wavelength demodulation of the fiber grating sensing system is achieved by an optical tunable fiber Fabry-Perot(FFP)filter.Because of the fuzzy systems can approximate the arbitrary nonlinear function,two measurement strain modification methods with fuzzy network are proposed.The proposed method is wrapped in the self-developed demonstration software,and the platform can display the real-time measurement results which is accurate proved by the experiment.It provides a guarantee for the practical application of the deformation measurement system in the structural deformation measurement.Compared the several common FBG demodulation methods,this paper proposes the FBG demodulation method with a tunable FFP filter.Determining the specifications of the demodulation system according to the requirements of actual measurement.To accomplish the demodulation of the FBG sensing system,the device selection,design of the optical system and basic circuits in the demodulation system are completed.According to the analysis of the strain measurement error and the universal approximating characteristics of fuzzy logic systems,a type-1 self-structuring fuzzy network(T1-SSFN)algorithm is proposed to update the measurement strain error through combining the inverse finite element theory and fuzzy network.T1-SSFN adopts the T-S fuzzy rules and completes the autonomous adjustment of the membership functions distribution according to the training data.In addition,the criterion of adding and deleting inference rules autonomously is adopted to the network,and ensure that a fuzzy network with simple structure and high approximation accuracy can be obtained from training phase.In the network training phase,the T1-SSFN is on the basis of training error minimization.So that it is possible to lead over-fitting phenomenon in the network building phase to make the system with a terrible approximation performance in application phase.Therefore,the self-structuring fuzzy network with SVR(SSFN-SVR)is proposed in the paper.Firstly,the construction of network adopts the proposed self-structuring algorithm in previous.According to the e-SVR theory,optimizing the parameters of fuzzy rules iteratively,and it avoids the influence of rule parameters that are not independent of each other in the optimization process.For the same nonlinear function approximation problem,simulation results of the two algorithms are compared.In addition,the measurement strain modification method is applied to the experiment on frame structure.The results illustrates the measurement strain modification method can improve the accuracy of deformation reconstruction effectively.For the own-developed framework measurement experiment system,an experiment demonstration software platform is completed in final.The proposed measurement strain modification method and the Deformation reconstruction algorithm based on IFEM theory are applied to the platform.Then the reconstruction deformation displacements of the measured structure are displayed on the demonstration interface.Comparing the reconstruction deformation displacements with the measured displacements of NDI optical measuring instrument in the process of the deformation measurement experiment.And the experiment results verify the measurement strain modification method by fuzzy network can improve deformation measurement accuracy.