Timing System Failure Prediction

The hydrogen maser and time comparison system (timing system for short) of the VLBI observatory provides a stable and reliable time and frequency reference for the observing station. The quality of the system signal will directly affect the quality of the observation data. Among them, the hydrogen maser is the key equipment of the timing system. Once it fails, its damage to data is fatal. Therefore, at the beginning of the design and construction of the VLBI station timing system, the state of the hydrogen maser and its surrounding parameters are monitored and recorded in real time. Once an abnormality is found, the technicians need to deal with it quickly. Chinese VLBI Network (CVN) has been established so far. There are currently five stations. The timing system of each station is equipped with two or more hydrogen masers. The system has accumulated a large amount of data since its long-term operation. Through the analysis of these data, we can study the relationship between state changes, performance changes, and environmental changes of the hydrogen maser, and then establish a failure prediction model to predict failures. This article describes the test process of creating a timing system data warehouse and using data mining technology to predict system failures, and we find the clustering model can be used to make good predictions in this project.