Predictive MPPT AI algorithm for fast response during tumbling conditions

Efficient power harvesting for small satellites is critical for mission success, especially during the initial detumbling phase and in case of attitude control anomalies. In these scenarios, the illumination conditions of the solar panels can vary at high frequency, and conventional Maximum Power Point Tracking (MPPT) techniques may struggle to reach a point of good power conversion under fast transients. If detumbling issues arise and cannot be promptly resolved, the batteries may discharge before ground intervention is possible, leading to the loss of the spacecraft and an early end of the mission.

In this work, we introduce a novel MPPT approach based on onboard artificial intelligence. The method adapts continuously to the spacecraft’s electrical and environmental conditions with real-time telemetry from the photovoltaic modules as input. By predicting short-term trends in illumination and electrical behavior, the controller is capable of anticipating cycling changes and adjusting the duty cycle proactively, rather than only reacting as happens with a standard Perturb and Observe method.

Since the controller adapts autonomously to onboard telemetry rather than relying on mission-specific datasets, it can be integrated into different spacecraft and photovoltaic configurations without extensive tuning or characterization before launch. The approach reduces the amount of power that would otherwise be lost during tumbling, deployment, or attitude disturbances. By maintaining a more favorable operating point under non-stationary conditions, the system helps preserve power availability during critical mission phases.

Early simulation results support the feasibility of the approach and demonstrate potential advantages in maintaining a fast reaching of the point of maximum power in cycling conditions of illumination, particularly during tumbling and post-deployment phases. The simulation results show the potential of AI-assisted power management for future technological development, suggesting that intelligent control can be integrated into power electronics without compromising safety or resource constraints.