Study On Several Dynamic Signal Processing Techniques And Their Applications

Data acquisition, transformation and identification, are three main parts of the dynamic signalprocessing. This thesis is focused on several important issues in the digital singal processing and theirapplication in the operational modal analysis. The main contents are as follows:(1) Three methods for signal de-noising are studied: time domain filters, frequency domainfiltering, and time-frequency domain wavelet. The effect of the above methods iscompared with each other carefully.(2) Two methods for dynamic singal integration are studied: time domain integration andfrequency domain integration. The two methods are compared both in calculationprecision and speed by some simulation examples. The error sources of the two methodsare analyzed. The frequency domain method is successfully applied to the environmentevaluation of Nanjing metro line4.(3) The Zoom-FFT technique is introduced to the operational modal analysis. Theidentification precision is improved by the improvement of the frequency resolution as aresult of the select-band analysis. A two-story building model is used to invalidate thismethod.