The Spectrum Analysis And Research Of Random Nuclear Signals In Nuclear Information Systems

The dependency relation of fission neutron (²⁵²Cf)-fission chain (²³⁵U system) is the foundation for reactivity measurement of nuclear information systems in frequency domain and time domain. By frequency domain high speed measurement system whose capability up to 1 GHz sampling rates and up to three input data channels which can detect random nuclear signals simultaneously, blocks of random nuclear signals in the system could be acquired, on this time-frequency analysis would be performed to get reactivity characteristics of fission nuclear information system, such as fission mass, configuration of uranium systems and so on, thus identifying whether the reactor is for military or civil.This paper focuses on the constructed nuclear signal high speed data acquisition system, on the basis of design principle of nuclear signal high speed measurement system, stands on the research work that includes spectrum analysis of random nuclear signals in high speed measurement system and spectrum analysis of random nuclear signals based on wavelet transform.The paper starts from power spectrum density, with the aid of Matlab, taking full use of block-average method, focusing on blocks acquired from HSDA, doing auto-correlation, cross-correlation, auto power spectrum density and cross power spectrum density analysis on random nuclear signals in three detection channels of system respectively, this is implementing correlation calculations first to take the total correlation function, and then FFT on total correlation would be done for total power spectrum density, finally the power spectrum density function we want would be got by average. Therefore, by observation and analysis on local characteristics of obtained spectrograms of nuclear information system, the prompt neutron decay constant and the count rate obtained from correcting the dead time would definitely be hold, which lay a solid foundation for the discussion of physical meaning of random signals in nuclear information system.In order to validate the performance of constructed nuclear signal high speed data acquisition system, data from Multi-channel Analyzer would be introduced, and random nuclear signal spectrum analysis based on wavelet transform should be developed as well. Focusing on the characteristics of random nuclear signals, from wavelet transform, a variety of de-noise methods have been carried out on these obtained random nuclear signals. By comparison analysis and demonstration, SURE method is regarded as the best one. Based on this, Discrete Wavelet Transform and Static Wavelet Transform are respectively used for the analysis on the noise that has been denoised. Absolutely experimental results present the Gaussian noise is the main noise of original signals, and what is more, results from Static Wavelet Transform are much better. The principle and approach that the 4th level coefficient of original signal is reconstructed by wavelet packet should be explored to get the characteristics frequency of random nuclear signal. Then, by taking advantage of least square fitting on frequency curves of denoised signals, the prompt neutron decay constant can be obtained for reactivity characteristics parameters of system normally.Finally, the comparison should be done between correlation curves and power spectral density curves based on Matlab and those of constructed nuclear signal high speed measurement system for some results of physical meaning of nuclear information system, all of them create conditions for further research work.