Signal Preprocessing And Denoising Of Grid Structure Based On Wavelet Analysis

In the process of civil engineering construction,operation and monitoring,the pretreatment and denoising of relevant signals are the key technologies to obtain the corresponding information of civil engineering structures.In this paper,the signal preprocessing and denoising of spatial grid structures are studied on the basis of dynamic test signals.An experimental model of the full-length 14m-span spatial grid structure was made,and its modal mode was solved by using ANSYS.The layout of the acceleration sensor was set according to the modal mode,and the actual measured MAC value was used to evaluate the sensor layout,indicating the rationality of the sensor layout.The sensor can accurately measure the first four modes.The mother wavelet(Haar,Daubechies,SymN,CoifN and BiorN wavelet)was studied.Soft and hard thresholds;Threshold estimation(HeurSure,Rigrsure and Minimax threshold estimation);The influence of wavelet decomposition layers on signal noise reduction.According to the parameter study,the CoifN wavelet family has the most mother wavelets that meet the requirements,and the Sym8 wavelet is the most suitable for signal noise reduction in this experiment.The soft threshold function is more suitable for signal noise reduction in this experiment.Minimaxi and RigrSure threshold method have poor denoising effect,and HeurSure threshold method is suitable for signal denoising in this experiment.The noise reduction effect of two-layer decomposition is the worst,the number of other layers is better,and the optimal number of noise reduction layers is five-layer decomposition.Through the above research and analysis,the network structure experiment has collected the complete modal mode,and obtained the real structure response after denoising,which can meet the needs of the acceleration signal in the follow-up research.