Research On The Inductive Methods Of The Measured Data Of Random Vibration Environment

When equipment or products are using in practical engineering environment, the random vibration environment cannot be avoided. In order to improve their adaptability, the random vibration environment simulation test should be done before employing these products. The key to influence on the effectiveness of the test is to determinate the test level. According to the national standard, the measured data takes preference to determine the test level. So the induction of measured data has important significance.The methods of data induction in present national standards are not described in detail, and the application of these methods is limited to two methods of them. In this thesis, how to determine the test level by the measured data is studied well. This thesis is organized as follows:At first, since the most of the random vibrations in practical engineering environment can be considered as a signal which is stationary, ergodic and with the normal distribution, the distribution of the random signal and the method to characterize the features in frequency domain are studied. And the estimation of the distribution of power spectral density of the random vibration signals with normal distribution is obtained.Secondly, this thesis gives the derivation process of some kinds of inductive methods based on the mathematical statistics theory. The advantages and disadvantages of these methods are pointed out, and some improvements are proposed.Then, a multiple input multiple output(MIMO) simulation model of cantilever is established in Matlab, and its random vibration response data is obtained through simulation under the excitation of random signals which are simulated with the technology of generation of random signals. Comparing the measured spectra and the normalized spectra obtained from various data induction methods, two criteria are proposed according to the measured spectra. After this, the characters of these different methods are discussed.Finally, to design and build the experiment platform, then several groups of typical experimental data are obtained. The inductive methods discussed above are used here. And the conclusion can be made through comparing the results with the simulation.