Analysis Of The Noise Type Of Atomic Clock And The Degree Of Freedom Of Frequency Stability

Atomic clock serves as the basis for atomic time keeping. Hydrogen atomic clock and cesium clock are commonly used for generating and keeping standard time. The clock model which is able to precisely describe the characteristics of atomic clock must be established at first in order to obtain even time scale. It is the top priority to analyze the stability and noise type of atomic clock in a bid to master the characteristics of atomic clock and establish clock model. The corresponding confidence interval must be given in the process of estimating frequency stability variance. The key to estimate frequency stability variance is to determine its degree freedom. Therefore, the analysis of noise type and the degree of freedom for the stability of atomic clock is of great importance in time keeping. This thesis makes an initial analysis and exploration upon the noise type of atomic clock and the degree of freedom for frequency stability estimation.The means to describe frequency stability is changing and developing. The frequency stability may be characterized either from time domain or frequency domain. This thesis studies means to describe signal nature of atomic clock and to characterize the stability of atomic clock from the aforesaid two aspects. First, this thesis introduces the signal and model of clock, studies the various variances and usage scope of time domain frequency stability on the basis of atomic clock frequency stability, and makes an initial analysis upon frequency stability and the transformation between domains.Partial Least Squares is employed to undertake linear fit upon short-term, mid-term, and long-term time domain data in analyzing the signal model of atomic clock and the experimental data of time domain. The main noise types of two cesium atomic clocks and two hydrogen atomic clocks (one is US-made hydrogen clock, the other is China-made one) are differentiated in accordance with the linear fit rate. This thesis studies the computational methods for confidence interval of variance estimate based on time-domain frequency stability, undertakes further exploration upon the formula for equivalent degree of freedom for variances, analyzes the expression for the value of degree of freedom to determine the precise value of degree of freedom or estimated value as well as the equivalent degree of freedom under various circumstances.