Universal Time (ut) is a non-uniform time-scale based on the rotation of the Earth around its axis. Another time-scale based on the duration of the SI second that was originally defined by the revolution of the Earth in its orbit around the Sun, is called Terrestrial Time (tt). As tt is a uniform time designed to provide continuity with the former Ephemeris Time (et), the difference between tt and ut is known as delta-T (ΔT). The time series of ΔT values shows the variations in Earth rotation. There are several mathematical models to represent ΔT used in astronomical software and calculations.
This paper discusses a new model for values of ΔT from 1620 to 2010 that uses artificial neural networks. The maximum absolute error in this model is about 0.5 SI second. The accuracy of the new model is compared with that of four prior models.