Lithium Plating Detection: How Virtual Sensors Enable Smarter Fast Charging

Fast Charging Pushes Li-ion Batteries Toward a Hidden Risk

Fast charging has become a defining requirement for modern electric vehicles and battery-powered systems. Users expect shorter charging times, consistent performance, and reliable operation under all conditions. For simulation, battery control, and engineering teams, this shift raises a shared challenge: predicting and managing internal battery risk that cannot be directly measured.

However, increasing charging speed pushes Li-ion batteries closer to their physical limits. One of the most critical constraints that emerges under these conditions is lithium plating.

Lithium plating occurs when lithium starts to accumulate on the anode surface instead of being stored safely inside the battery. This can happen during fast charging, especially in cold temperatures, conditions that are common in real-world use.

It’s impact is significant:

• Capacity loss
• Accelerated aging
• Increased safety risk

What makes lithium plating especially challenging is that it remains invisible during operation. Standard battery management system signals, such as voltage, current, and temperature, may remain within normal ranges even as plating begins internally.

To address this hidden issue, better visibility into what is happening inside the battery is required. This is where physics-based models, such as GT-AutoLion, come in. These models provide insight into internal electrochemical behavior that cannot be measured directly.

Building on this, virtual sensors can convert these insights into real-time information. As shown in Figure 1, a virtual sensor based on electrochemical models can detect the risk of lithium plating as it occurs. This makes an invisible problem visible and, more importantly, controllable.

Figure 1: Virtual sensor estimating lithium plating risk in real time, making hidden battery behavior visible and easier to control

Why Traditional BMS Design Sacrifices Fast-Charging Performance

Because lithium plating cannot be measured directly, current battery systems rely on conservative safety margins.

While this approach ensures safe operation under worst-case conditions, it introduces important trade-offs:

• Charging power is restricted even when higher performance is possible
• Battery lifetime protection limits usable capability
• System-level overdesign increases cost and complexity

As a result, fast charging, durability, and efficiency become competing objectives. At the core of this limitation lies a fundamental issue: battery systems make decisions without direct visibility into internal electrochemical risk.

How Virtual Sensors Detect Lithium Plating in Real Time

To overcome this limitation, the internal state of the battery must be translated into a real-time, observable signal. This is enabled by virtual lithium plating sensors. These sensors estimate the anode potential, the key indicator of plating risk, by combining:

Physics-Based Modeling (GT-AutoLion)

• Captures lithium transport, diffusion, and reaction dynamics
• Simulates internal behavior of battery including Li plating potential

Machine Learning Assistant Tool

• Builds metamodels trained on GT-AutoLion simulation data
• Enables fast, real-time prediction of internal battery states

The result is a real-time capable virtual sensor that provides continuous insight into internal battery behavior, without additional hardware.

To make this predictive capability practical for control systems, the model output (anode potential) is converted into a Lithium Plating Risk Index (LPRI):

• LPRI = 0 → No risk (> 40 mV)
• LPRI = 1 → Low risk (20–40 mV)
• LPRI = 2 → Medium risk (0–20 mV)
• LPRI = 3 → High risk (≤ 0 mV)

As shown in Figure 2, the continuous electrochemical prediction is translated into discrete risk levels. This allows for simple interpretation and direct integration into battery control strategies.

Figure 2. Mapping of predicted anode potential to discrete LPRI levels for real-time risk assessment

Real-Time Fast Charging with LPRI Feedback

The real value comes when this predictive capability is integrated into charging control. Once lithium plating risk is available as a real-time signal, it becomes a control variable rather than a hidden failure mode. As shown in Figure 3, the LPRI signal can be integrated into a feedback control loop to adjust charging current dynamically. Charging current is then continuously adapted based on actual battery conditions:

• Under safe conditions, full fast charging power is applied
• As risk increases, current is smoothly reduced by using a control strategy (e.g., PID-based)
• Near critical conditions, charging is actively limited to stay within safe boundaries
• When conditions improve, higher charging power is automatically restored

Control strategies such as PID controllers ensure smooth transitions and stable operation. This approach allows the battery to operate close to its optimal limit, balancing performance, and safety in real time. Lithium plating is no longer detected after damage occurs but prevented before it happens.

Figure 3. PID-based charging control using LPRI to prevent lithium plating

Conclusion: Making Lithium Plating a Controllable, Not Hidden Variable

Fast charging is pushing Li-ion batteries to their operational limits, and traditional conservative approaches are no longer sufficient to balance performance and safety. By combining physics-based simulation using GT-AutoLion, a machine learning assistant tool, and real-time control, virtual sensors remove a critical blind spot in battery management. Lithium plating is no longer an invisible and uncertain phenomenon; it becomes a measurable and controllable variable.  This transformation enables a new generation of battery systems that are

• Faster, by maximizing charging performance
• Safer, by preventing degradation early
• More efficient, by avoiding unnecessary limitations

Fast charging is no longer just about speed; it is smart, condition-aware control, where performance and safety work together instead of competing.

To learn more about simulation-driven battery development, contact us and visit our Battery Simulation Solutions page to explore advanced battery modeling capabilities, read our previous blogs on “What is a Battery Management System” and “BMS Architecture” where we discuss the fundamentals of BMS functionality and system design in greater detail. Also, follow us on  LinkedIn to stay updated on the latest developments in simulation-driven engineering.

FAQ

What is lithium plating, and why is it a concern during fast charging?

Lithium plating occurs when lithium ions deposit as metallic lithium on the anode surface instead of intercalating into the anode material. This typically happens during aggressive fast charging, especially at low temperatures or high states of charge. Lithium plating can lead to permanent capacity loss, accelerated battery aging, reduced charging efficiency, and increased safety risks, making it one of the key challenges in fast-charging battery management.

How do virtual sensors detect lithium plating without additional hardware?

Virtual sensors estimate lithium plating risk using physics-based battery models and machine learning rather than physical sensors. In GT-AutoLion, electrochemical simulations predict internal battery states, while machine learning metamodels enable these predictions to run in real time. This allows battery management systems to monitor lithium plating risk continuously without adding extra sensors or increasing system cost.

How does real-time lithium plating detection improve fast-charging performance?

Real-time lithium plating detection allows the battery management system to adjust charging current dynamically based on the battery’s actual internal condition. Instead of relying on conservative safety margins, the system can maximize charging power when conditions are safe and reduce current only when lithium plating risk increases. This enables faster charging, improved battery life, enhanced safety, and more efficient battery operation.