Research on Fault Diagnosis Technology of Thermal Test Measurement and Control System Based on Data Driven

Abstract

Temperature data, as one of the most important parameters in spacecraft vacuum thermal test, is also an important characteristic parameter for fault detection of spacecraft thermal test system. From the perspective of numerical heat transfer, the physical and mathematical models of thermal test specimens are established by using finite difference method. Based on the data driving of specimen measurement temperature and heating power, the mathematical model coefficients are fitted by the combination of least squares and sliding window filtering, and then realize the on-line prediction of temperature data and fault detection of thermal test measurement and control system. Based on the thermal test data of the solar wing of a certain type of spacecraft, the above-mentioned method was verified by physical experiments. The experimental result shows that the method can accurately and effectively realize the temperature prediction not less than 2 periods, and the prediction accuracy is less than ±2°C. It can provide an effective fault detection means for the thermal test measurement and control system.