产生和保持一个稳定、准确、可靠的时间尺度是所有时间实验室追求的目标.传统的ALGOS算法主要考虑时间尺度的长期稳定度,而地方原子时尺度需兼顾长、短期的稳定度.通过对原子钟噪声模型的分析研究,在保证地方原子时尺度长期稳定度不降低的条件下,提出适合中国科学院国家授时中心(National Time Service Center,NTSC)守时钟类型单一、钟性能相近的时间实验室计算地方时间尺度TA(NTSC)的一套完整算法.应用NTSC2008年全年所有参加国际原子时(International Atomic Time,TAI)计算的钟的数据进行新算法的验证计算,得到的TA(NTSC)的短期稳定度指标与长期稳定度指标均有提高.研究结果适用于与NTSC守时系统结构相似的时间实验室的原子时尺度计算.
It is always an endless target for all time and frequency laboratories to develop, own and keep a stable, accurate and reliable time scale. As a comparatively mature algorithm, ALGOS, which has been concerned about the long-term stability of the time scale, is widely used by the majority of time laboratories. For ALGOS, the weights are assumed on the basis of the frequencies of 12 months and the present month interval is included in the computation. This procedure uses clock measurements covering 12 months, so annual frequency variations and long-term drifts can lead to de-weight. This helps to decrease the seasonal variation of the time scale and improve its long-term stability. However, the local atomic time scale is primarily concerned with long-term stability not more than 60 days. So when the local time scale is computed with ALGOS in time laboratories, it is necessary to modify ALGOS correspondingly according to the performances of contributing clocks, the requirement of stability for local time scale and so on. There are 22 high performance atomic clocks at National Time Service Center, Chinese Academy of Sciences (NTSC). They include 18 cesium standards and 4 hydrogen masers. Because hydrogen masers behave poor, we only regard an ensemble of 18 cesium clocks in our improved algorithm. The performances of these clocks are very similar, and the number is less than 20. By analyzing and studying the noise models of atomic clocks, this paper presents a complete improved algorithm of TA(NTSC). This improved TA(NTSC) algorithm includes three aspects: the selection of the maximum weight, the selection of clocks taking part in TA(NTSC) computation and the estimation of the weights of contributing clocks. We validate the new algorithm with the annually atomic clock comparative data of NTSC taking part in TAI computation in 2008. The results show that the long-term and short-term stabilities of TA(NTSC) are all improved. This conclusion is based on the clock measurement data at NTSC.