Analysis Of Non-linear Vibration Of Concrete Beams By Using Blind Source Separation

In recent years, people have done a great deal of research on structural damage detection based on linear vibration, and achieved fruitful results. However, the deviations of material properties or geometric, crack generating and developing before and after the structural damages would affect the changes of the modal characteristics. Therefore, the actual dynamic characteristics usually show some degree of nonlinearity. If structures before and after damages both are assumed to be linear system to describe its vibration characteristics, some important damage information can be lost, and may also affect structural damage identification.Modern engineering structures numerously adopt steel reinforced concrete structure. Several researchers have taken some researches on the damage identification based on the nonlinear vibration characteristics of concrete structures. The most common method is to cause man-made damage to the structure. And then the free vibration signals of the structure under different damage degree can be recorded, which are processed by time-frequency analysis to get the curve of frequency and amplitude. Existing reinforced concrete structural nonlinear vibration analysis mainly has the following two problems: firstly, the trandional method by which only a single channel used is greatly influenced by the noise. And it is difficult to solve; Secondly, the signal is firstly band-pass filtering or smoothing window processing. It may reduce the data amount of information, and some local vibration characteristics of the signal can be lost. This paper tries to introduce blind source separation technology into the nonlinear vibration analysis of reinforced concrete structural to solve the above problems.Blind source separation refers to the estimate of each component of the process of source signals only by source signals or some prior knowledge. Blind source separation technique is mainly used to solve mixed-signal from the detection to recover or estimate the source signals. Nowadaysit has been widely proposed in biomedical engineering, array signal processing, speech recognition, image processing, structure system identification, health monitoring and other fields. The application in structure engineering has just been beginning, it was mainly used for structural modal parameter identification. Blind source separation was used to analyze non-linear vibration of RC beams. The algorithm called second-order blind identification was applied to the structural vibration signals de-noised by wavelet soft-thresholding to acquire the first order vibration component, which was then processed by Hilbert transform to get the curve of frequency and amplitude. This paper tries to explore the feasibility of structural health monitoring by the use of nonlinear vibration characteristics of structural.The main work accomplished and conclusions of this paper are as follows:(1)Blind source separation technology is applied to the analysis of structure non-linear dynamic characteristics for the first time. The Simulation examples of dense frequency system and reinforced concrete simple beams were successfully identified by using second-order statistics identification algorithm by which can obtain each order vibration component. Simulation examples result shows that the feasibility of blind source separation algorithm, this method can be used for the analysis of the actual structure vibration signals. In addition, for densely frequency interference systems, blind source separation technology can obtain a good recognition result.(2) The algorithm called second-order blind identification was applied to the structural vibration signals de-noised by wavelet soft-thresholding to acquire the first order vibration component, which was then processed by Hilbert transform to get the curve of frequency and amplitude. Based on the verification by finite element simulation, the recorded vibration signals of a RC simple beam and a prestressed simple box-beam were analyzed by the above procedure. The results show that blind source separation is better than the short time Fourier transform regarding the ability of anti-noise and ease of implementation.Analysis found that:a)Blind source separation technology using multi-channel signals has strong anti-jamming capability.Combined with signals de-noised by wavelet soft- thresholding , it can better handle noise interference problems, and avoid the short-time Fourier transform principia window width choice.The robustness of his method is good.b)Damage identification by non-linear vibration sometimes is reliable than that by frequency change obtained by traditional modal analysis based on linear vibration.(3)The dynamic measurement signals of steel are analyzed, the results show that the nonlinear steel does not have obvious dynamic characteristics, nonlinear dynamic theory does not apply to the damage identification of steel.