Operational Optimization of the Lithium-ion Batteries of TerraSAR-X/TanDEM-X

Abstract

After a decade of the TanDEM-X mission operations, the degradation of the battery capacity due to ageing defined a new challenge that needed to be addressed by the mission operations team. A novel machine-learning strategy has been gradually deployed in the Mission Planning System in order to optimize the battery utilization. The objective is twofold: a) to protect the operational state of the battery, while b) maximizing the executed SAR acquisitions under the new planning restrictions. The resulting optimal battery utilization thresholds have been communicated to the customers in a user-friendly way in order to assist their future planning, minimizing the ratio of not-executed requests due to the new energy and power constraints in the joint TerraSAR-X/TanDEM-X Mission Planning System. In this paper, we describe step by step the quantitative approach to model the satellites’ batteries in comparison to the previously used physical models; we outline the process of the new machine learning model in the Mission Planning System; we present the operational results of the model in comparison to the telemetry and we discuss the evolution of the machine learning model towards higher accuracy telemetry estimations.