Theory And Applying Technique Study On Identification Method For Dynamic Loads In Time Domain

The back problems of dynamic loads in the vibration of engineering are studied in this dissertation of Doctor's Degree, which are the problems of dynamic loads identification.There are three terms that are excitation, the vibration system and response in a vibration problem. The excitation and the vibration system are the determinant factor in vibration. The response of the system to excitation is the purpose which vibration is studied. In most engineering, the response can be directly tested, but the excitation and the vibration system cannot. It is created the inverse problem of vibration or the back-problem of vibration. It is known as the â… ^th inverse problem of vibration that the vibration system is solved by the excitation and the respondence. It is also known as system identification or system reconstruction. Studying on the â… ^th inverse problem of vibration has been well-perfect and ripe, and it has been widely applied in practice now. The vibration problem that the excitation is solved in the respondence and system of vibration is known as the â…¡^th inverse problem. It is also known as dynamic load identification or load reconstruction. It is later that the â…¡^th inverse problem of vibration has been studied. And its theory is still not perfect and ripe. But it has been being studied and became the hot spot of studying.In the recently years, there are many scientists and engineers who are discussing the theory and technique of dynamic load identification with different way. Nowadays, the techniques of dynamic load identification are mainly classed into two classes: time domain identification method for dynamic load and frequency domain identification method. The frequency domain method is the means which dynamic loads spectra are identified by the response and the relationship between the matrix of frequency response function and response spectra in frequency domain, or modal force character in frequency domain is computed with modal coordinate transform. Because the input and output of math model of vibration system in frequency domain is linear and its inverse operation is easy, it was studied earlier, and a great many developments have occurred since it began to study. But the frequency response function matrix in frequency domain is ill-condition in the range of resonance, so there are some problems on numerical accuracy and stability. Its dynamic calibrating is simple, and its identifying accuracy is higher. But it demands that signal sample is certainly long, and it is only fit for steady or stationary random loads. The time domain method is the means that dynamic loads history is directly identified by response history from system dynamic equation in time domain. This method involve in a complicated back-analysis with higher difficulty that it related to the convolution relation between load and response. Because it is direct and is conveniently applied in engineering, it is widely accepted by engineering. Its advantage is not only that it can be fit for dynamic load identification of non- stationary and short sample random load, but also that it may be fit for real-time dynamic load identification. So this method has been being studied now. But current time domain method has some shortcoming, such as solving procedure being diverse and complicated, non-linear stronger, calculating amount being great, dependingon initial values, identifying accuracy low and its stability and robustness weak.The main work in this dissertation is studying the theory, means and applying techniques for dynamic load identification in time domain. And their correctness and reasonableness are verified by testing instance. Based on studying and analyzing the current theory of dynamic load identification, and ill-posedness in the identification process was fuller considered. It is the first to be posed that there are three aspect ill-posed problems that are in time history, physical model and space in dynamic load identification techniques. It is believed that the studying of dynamic load identification techniques begin with the three aspect ill-posed problems. And these three aspects being reasonably solved is the assurance of dynamic loads being correctly identified. This idea is originality. It is helpful to be clear the muddle about recognizing on dynamic loads identification techniques in this view. The idea has certainly directing significance for researching the techniques of dynamic loads identification. Although it is put forward in time domain, it can also be used in frequency domain. Under the idea directing, there are some innovations in this dissertation as follows:1) The ill-posed problems in dynamic loads identification were fuller analyzed. And The technical measures were put forward according to different ill-posed problems. It is the first to be posed that there are three aspect ill-posed problems that are in time history, physical model and space in dynamic load identification techniques. This idea is originality. It has directing significance for studying the techniques of dynamic loads identification.2) Ill-posed problem is not only occurred on the wording of mathematical & physical model, but also new ill-posed problem will be occurred during the process of dealing with numeric. It is a new point in this dissertation.3) A time domain identification method for dynamic loads was established, which base on the non-recursion-chain format. The depending on initial value and edge value is eliminated in this method. And it is not recursion-chain. So error delivered and accumulated can be avoided. Its numerical computation has a well-stable. It realized shifting back-analyzing into positive-analyzing. Its calculating efficiency is high and its calculating amount is little. And identifying accuracy is high and its stability and robustness is well. Furthermore it provides a theoretic base for real-time identifying dynamic loads.4) It was deeply studied that the identifying accuracy and stability is influenced by interval time in dynamic loads identification techniques. It is the first that the formula of interval time in dynamic loads identification techniques was deduced. And the correctness and reasonableness of the formula was verified by some instance. The formula is not only fit for time domain identification method but also fit for frequency domain identification method.5) It was pointed out that there is a great defect in the conception of modal-truncation in current dynamic loads identification techniques. It is the first to be posed that a conception of modality selection must be applied in dynamic loads identification techniques. The necessity and importance of the conception being applied in dynamic load identification was proved. And a new criterion for determining modality in dynamic load identificationand implementation method is proposed. A new formula for modality selection was deduced. Not only is the formula fit for time domain identification method, but also it is fit for frequency domain identification method. It can be also applied in positive problem analyzing.6) It is the first that a criterion for assigning testing points was clear presented. So that, it surmounts the shortcoming that testing points were assigned by one's experience before it. It gives a special guide to do it for practical engineering.7) Some new identification functions were submitted, and their conception, definition and relative formula of their sensitivity for system's dynamic character and sampling time were presented and analyzed. It concluded some importance result, which have importance mean to guide identifying dynamic loads.8) Studied some calculi methods for signal tested, a new way on calculus of signal tested was the first given, which use EFT technique and is modified by the character of signal tested. And a good effect is got. The difficulty, which one point only testing one among displacement, velocity and acceleration carry, was solved.9) The software for dynamic loads identification was worked out, and had been proved by many examples.