Research On Damage Source Localization Of Three-dimensional Braided Composites Based On Carbon Nanowires

Because of the excellent specific strength and specific stifness,anti-fatigue performance and durability,composites have been widely used in aerospace,construction,automotive,high-speed rail and other fields.Efficient techniques of real-time non-destructive testing of composite structures damage and continuous on-line monitoring are become very important.In this thesis,carbon nano-sensing measurements and Singular Value Decomposition algorithm were applied to the three-dimensional braided composite specimen damage detection.Experiments of resin-based composite material damage existence and damage location determination were carried out.Three-dimensional braided composite material damage detection system based on carbon nanoyarn was set up.Firstly,the basic definition of composite material and its wide application in the aviation field were analyzed to introduce the importance of composite materials damage identification,then the status of domestic and foreign health monitoring of composites were summarized.Secondly,it described the sensing principle and preparation method of carbon nanoyarn,and the theoretical basis for damage detection of carbon nano-sensors for three-dimensional braided composites materials were given by analysing the sensing properties of carbon nanoyarn.Subsequently,the knitting method of carbon nanoyarn embedding in the three-dimensional braided composite specimens was studied.The specimens were produced by three-dimensional four-step six-direction method.For 3-D six-direction braided composites materials,based on the existing experiment observation and the movement traces of yarn carriers on the braiding machine bed,three unit-cell model of structure was established and the in-plane and spatial movement traces of yarns in the interior,surface and coener region were studied.The trend of carbon nanoyarn sensor inside is predicted.The data matrix collected by carbon nanoyarn sensor array was processed with two-dimensional wavelet threshold method.A novel dynamic threshold algorithm was proposed taking into account the discontinuity of the hard threshold function and theconstant bias between soft threshold and original image through modifying aparameter.This made the threshold dynamically changed with the decomposition level increasing and the novel threshold function had better noise decrease effect.This would lay a good foundation for the subsequent data processing.Considering the mass data processing problems in health monitoring system of composites,a quarter of Singular Value Decomposition algorithm was presented in order to simplify the data and ensure a good real-time system,at the same time the key features of the data matrix was extracted.By comparing the key features with healthy templates,the existence of injury was determined and the damage location was calculated.