Control Strategy Research On Multi-exciter Vibration Test System

Vibration test is intended to examine the anti-vibration capability of products and their performance in vibration environment so that design faults of products can be found out in vibration test accompanying with improving design and quality of these products. So far, Single-exciter vibration test has been widely adopted in the vibration testing, however, its shortages also attract people's attention since the structure of product is becoming more and more complicated. Multi-exciter vibration test actually can simulate the practical environment better than single-exciter vibration test, the latter bears some disadvantages, such as pushing force insufficiency, non-uniform vibration stress distribution, under-test and over-test and so on, while the former can overcome those drawbacks. The deep investigation to the control strategy of multi-exciter vibration test is conducted, and research findings are also applied into the design of the control system, which provides theoretical support to the development of multi-exciter vibration test control system.Identification of multi-exciter vibration test control system. To identify system and understand system characteristics should be finished before achieving effective system control. Control system for multi-exciter vibration test is complicated in algorithm, it is confirmed that frequency response function model is a suitable way in indentifying control system of multi-exciter vibration test based on the investigation to system identification model and consideration to the system characteristics. Partial coherence residual method is proposed to conduct frequency response estimation, meanwhile, calculation procedure of frequency response matrix estimation for the control system is also summarized and applied into system control. Compared with traditional way, frequency response function estimated by making use of the Partial coherence residual method is achieved through calculating inverse matrix to the upper triangular matrices, which leads to less calculation works and therefore is suitable to multi-exciter vibration test where harshly requires real-time capability, furthermore, more coherent information between input and output of the control system is provided by the method mentioned above.Optimal control strategies of multi-exciter sine vibration test. An optimal control strategy is proposed so that the drawbacks in traditional multi-exciter sine control algorithm adopting constant iterative step length and no updating for frequency response function can be avoided. The optimal control strategy whose iterative step length and frequency response function can update in real time according to system characteristic is able to speed up convolution speed of the control system on the basis of guaranteeing stability of the system, the way to estimate the amplitude and phrase of the multi-exciter sine signal is investigated in this thesis, the least square method is adopted to estimate the amplitude and phrase of the multi-exciter sine signal.Control strategies of multi-exciter random vibration test. The way to generate the multi-exciter random drive signal is studied, the true random signal which satisfies the requirement of the system is generated by using time domain randomization technology. Setting conditions to the reference spectrum matrix of the multi-exciter random vibration test is explored, and whether the reference spectrum setting is reasonable or not can be judged by analyzing positive definiteness of the spectrum matrix. Control algorithm of the multi-exciter random vibration test is investigated, a comprehensive equalization control algorithm is proposed based on the analysis to the traditional control algorithm, which contributes to overcome the shortages existed in traditional control strategy and improve the performance of control system. Non-Gaussian random vibration test has become the hot field in vibration test since it can simulate realistic environment better, Concurrent control strategy research for non-Gaussian random vibration test is proposed, which achieves independent control to the kurtosis control and power spectrum control respectively. Generation algorithm of non-Gaussian random signal is investigated, and no-Gaussian random signal generation algorithm based on the Poisson process is given in paper.Control system building and its experiment for multi-exciter vibration test. In conclusion, the whole system design is finished in this study, and hardware and software platform of the control system are built on the basis of the critical technology of the control system in multi-exciter random vibration test. The experiments concerned are conducted. The experiments including multi-exciter sine vibration test and multi-exciter random vibration test as well as non-Gaussian vibration test are carried out on the mentioned platforms. The comparison between optimal control strategy and traditional control algorithm in the same experimental conditions is made in multi-exciter sine vibration test, the former has better control performance and bears less fluctuation than latter in singular points nearby. self-closed loop control for multi-exciter random vibration test indicates that dynamic range for control system reaches 90dB, and control response power spectrum in multi-exciter random vibration test is under control within±1.5dB. Self-closed loop control for non-Gaussian random vibration test indicates that dynamic range for control system reaches 90dB, and all the points contained in the range of frequency for control response power spectrum in non-Gaussian random vibration test is also under control within±1.5dB. Non-Gaussian random kurtosis value is fluctuating nearby target, the amplitudes of non-Gaussian random signal collected from the shaker changes dramatically and bears distinct non-Gaussian characteristics.