Realization Of Direct Digital Integration Method For Vibration Acceleration Signal Based On DSP

The mechanical vibration is the phenomenon which often occurs in the work of all kinds of machines. The vibration measurement plays important roles in the mechanical breakdown and condition monitor of large machine. Three parameters such as acceleration, velocity and displacement are usually used in the vibration measurement, but in fact the accelerometer is usually used to gain the signal in the vibration test, that is because that acceleration is more convenient and economical than velocity and displacement. Integral of acceleration is velocity and double integral of acceleration is displacement. Because integral operation is a gradually approximate process, the error occurs unavoidably. Especially after double integral of acceleration, the error becomes lager. In this paper elimination of the interference noise and the tendency item of the vibration acceleration signal is mainly researched so that the result of integral approaches to the real signal.This Thesis is centered on vibration acceleration signal processed, mainly concerns basic theories and its algorithms of direct digital integration about vibration acceleration signals and deeply expounds software system of taking the DSP and MATLAB as the central. Firstly the thesis briefly introduces DSP technology and its status quo and then outlines recent development of direct digital integration technology about the vibration acceleration signal at home and abroad. Finally, the thesis proposes research significance and its emphasis according to the reality and designs experimental scheme with the existing experimental conditions.Sensor is the most important tool obtaining signal in signal measurement as the acceleration signal measurement. The choice of sensor directly affects the efficiency and accuracy of measurement. So the sensor should be chosen carefully to achieve a better result of measurement. The thesis employs ICP (Integrated Circuits Piezoelectric) acceleration sensor to measure vibration acceleration signal by comparing different parameter of sensor. DSP processor selection is also very important because processor chip not only meets the speed and accuracy of processing requirements, but we also consider chip cost and the convenience and feasible of software development. The thesis selects TMS320C6000 series chip, which is not powerful for image processing, but also has a dedicated software development.Software design is the core of this thesis , the design software used CCS3.1 and MATLAB integration, between the two through CCS-Link interface data exchange. The executive software includes six modules: that is, to removing DC, wavelet de-noising, FFT filtering, oval of the trend, the number of integral and curve drawing. Wavelet de-noising is the most crucial part of the modules. The results of wavelet de-noising have direct influence the accuracy of integral. The thesis in combination with three common methods of de-noising: modulus maxima reconstruction filter, correlation filter in spatial domain, Wavelet threshold de-noising. Compared the three methods form principle and algorithm, and chose the more appropriate threshold de-noising method.Finally, the experimental were analysis by nonlinear method, nonlinear analysis using phase plane trajectory and Poincare section. Drawing the phase locus and Poincare section graph using the results of integral, and make use of relevant theory and criterion to carry out decision on the vibration state of vibration signal.