| As the core component of a time synchronisation system,the stability and accuracy of the output frequency of the frequency source directly determines the performance of the synchronisation system.In view of the development cost,this thesis selects an Oven Controlled Crystal Oscillator（OCXO）with high short-term stability as the reference clock of the synchronous system,but as OCXO is easily affected by factors such as ageing and ambient temperature,its output frequency drifts and cannot be directly applied in the field of timing,timekeeping and other high-precision time synchronisation.In addition,high-precision time interval measurement,as a prerequisite for improving the performance of synchronous systems,plays a vital role in improving the accuracy of synchronous systems.Therefore,in response to the above problems and needs,this thesis has researched and completed the specific design of a high-precision time interval measurement and time synchronization system,the main contents of its work and research results are as follows:1.Through an in-depth understanding and analysis of the time interval measurement principle,this thesis adopts a combined measurement method of FPGA pulse coarse counting and TDC-GPX2 fine measurement,and designs the serial transmission frame format of measurement data and measurement commands,and compensates for the transmission time delay of each channel measurement signal,so as to achieve a high-precision,multi-channel and large-range time interval measurement function.In the multi-channel measurement experiments,the system measured up to 70ps and was as accurate as the Stanford interval counter（SR620）in the USA in 10s tests,but slightly less stable and accurate than the SR620in nanosecond tests.2.When the system is in the taming mode,this thesis analyses the characteristics of GPS 1PPS and OCXO,and adopts a combined control algorithm of Savitzky-Golay filtering and PID,and designs an FPGA-based OCXO voltage control taming method to improve the synchronization performance of the system.After two hours of OCXO taming,the frequency accuracy of the OCXO improved from-5.81×10-8 to-3.25×10-10,and the average phase deviation between the local second pulse signal and the 1PPS signal was-0.2716ns and the standard deviation was 6.2276ns,achieved high precision synchronisation.3.In response to the problem of OCXO frequency drift after 1PPS signal failure,this thesis investigates the holding technique of the OCXO.By analysing the ageing characteristics,temperature characteristics and voltage control characteristics of the OCXO,a mathematical model and a BP neural network model of the OCXO are established.To address the problems of slow convergence and global optimal solution of the BP neural network model,this thesis designs an improved BP neural network model based on dynamic adjustment of the learning rate.Through test comparison and analysis,the improved BP neural network model is validated to be highly efficient and stronger prediction and retention capability.After the system was switched to the hold mode for one hour,the frequency accuracy of the OCXO improved from-5.81×10-8 to-1.003×10 9-,and achieved the synchronisation deviation between the local second pulse signal and the 1PPS signal better than 909ns.4.Based on the hardware structure scheme of FPGA+STM32,this thesis completes the specific design of the high-precision time interval measurement and time synchronization system,introduces the functions of each module in the system in detail,gives the relevant hardware and software design ideas and methods,and carries out simulation verification of the key modules in it.In the system performance test,it designed a homologation comparison test method and verified the system function indexes and met the design requirements of this thesis. |
PDF Full Text Request