Khaberni - Batteries have become pivotal with the significant expansion in the use of electric vehicles, drones, and renewable energy systems.
However, traditional battery management systems remain limited in their functionality, merely offering indicators of remaining charge without addressing the crucial question: Will this energy suffice to complete the intended journey or task? A team of researchers from the University of California, Riverside, led by Professors Mihri and Cengiz Ozkan, developed an innovative solution to this challenge, represented in a new system called "State of Mission" (State of Mission – SOM).
Unlike current battery assessment systems, like State of Charge (SOC) or State of Health (SOH), SOM goes beyond mere numerical ratios. It is a "mission-aware" system that predicts whether a battery will accomplish its planned mission under realistic conditions, such as driving a car over rough terrain, flying a drone in harsh weather conditions, or powering a house on a cloudy day with low solar panel output.
Smart Integration
The primary innovation in this system lies in the intelligent integration of physics and artificial intelligence. Typically, engineers rely on physical models to interpret battery behavior, but these are complex and slow to compute and are not suitable for immediate applications. Conversely, machine learning-based systems can handle vast amounts of data quickly but sometimes lack causal understanding, which can lead to errors under changing conditions. The SOM system employs hybrid techniques like Neural ODEs and physics-informed neural networks (PINNs), blending the precision of scientific models with the flexibility of artificial intelligence.
To test the system's effectiveness, the team relied on extensive datasets from NASA and Oxford University, covering realistic charging and discharging cycles, temperature changes, current and voltage rates, and long-term degradation. The results were promising; prediction errors significantly decreased compared to traditional methods, giving the system a high capability to provide reliable battery performance estimates.
In practice, this means instead of just telling you that your car's battery is at 60%, the system can tell you whether this percentage is enough to get over mountains or through cold weather. It can prevent you from sending a drone on a hazardous flight or assist in managing power in electric grids, all based on accurate and realistic calculations.
Expected Spread
Although the system still requires advanced computing capabilities that might exceed current hardware capabilities, researchers expect it to become more widespread as computing hardware evolves. The team also plans to apply the system to different types of batteries like sodium and solid-state batteries, making it a scalable and flexible platform across various energy fields.
In summary, the SOM system not only offers a reading on battery condition but represents a qualitative leap in energy management, transforming data into actionable decisions and enhancing safety and efficiency in cars, aircraft, and homes. In an age where technology increasingly relies on renewable energy more than ever before, SOM seems exactly what we needed to use this energy in the smartest possible way.
