Virtual Validation of Lithium-Ion Battery Management Systems

Battery Management Systems (BMSs) play a critical role in ensuring lithium-ion battery packs operate safely, efficiently, and reliably across all operating conditions. As battery systems become more complex and application demands continue to increase, validating BMS algorithms has become significantly more challenging. Building on the previous blog’s overview of BMS architecture for lithium-ion batteries (LIBs), this blog explores how virtualization and predictive models can help engineers validate control strategies, optimize charging behavior, and reduce development effort. Ultimately, an optimized BMS ensures the battery pack operates within its safe operating window, as shown in Figure 1.

Figure 1: Operational window of a lithium-ion battery

One of the main challenges with designing a BMS is that the battery pack is a black box. The most critical indicators of the health of a battery pack cannot be directly measured. Instead, the BMS relies on typically three primary external inputs: voltage, current, and temperature. The BMS uses these inputs to calculate the State of Charge (SOC), State of Power (SOP), and State of Health (SOH). SOC and SOP are normalized values that are used to determine the remaining capacity and the power capability of the battery pack, respectively. While SOH is a normalized measure of its remaining useful life.

Figure 2: Problems (left column) and solutions (right column) for BMS design

This is where virtualization comes in. Figure 2 shows the challenges that a BMS designer has to face (left column) and how they can use tools within the GT-SUITE ecosystem to address those challenges (right column). Virtualization simplifies this process by allowing for the creation of offline digital shadows that can be used to predict battery behavior under all conditions. This can lead to a significant reduction in testing time and development costs. To illustrate how these capabilities translate into actional insights for BMS development, consider the following example of validating a charging strategy under extreme operating conditions.

Validating a CCCV Charging Strategy Under Extreme Operating Conditions

An optimal controller should be able to ensure the safety and optimal usage of the battery pack under all conditions. To showcase how a predictive plant model can be used to verify this, we have created an example of a battery pack being simulated through a constant-current constant-voltage charging protocol (CCCV) across different ambient temperatures and charge rates.

Figure 3: Constant-current charging strategy example

The 96s3p battery pack is made up of a scaled LG M50T cell model that has been calibrated from experimental data. The max voltage for this pack is set to be 400 V with a cutoff current of 20 A. The model is outputting the pack voltage, current, temperature, heat generation rate, SOC, and li-plating potential as shown below.

Figure 4: Monitor signals for relevant battery pack states

Here, the li-plating potential is a strong indicator of an aggressive charge current being applied. Li-plating plating occurs when the anode electrode becomes too lithiated or the diffusion of li-ions becomes faster than the rate of intercalation. Essentially, the li-ions that are not intercalated into the electrode become solid li-metal, leading to a drastic increase in internal resistance, decrease in capacity, and increasing the risk of safety critical events such as thermal runaway.

Figure 5: SEI and Li-plating formation

A strong indicator of the chances of li-plating is tracking the li-plating potential. For traditional LIB chemistries, when this state reaches or drops below a value of 0V, the equilibrium conditions are met for lithium to plate. This state cannot be measured in experiments but can be captured through a predictive physics-based model. We can use this to evaluate which charge current is best suited in different ambient conditions.

To evaluate the proper charging load applied during the CC phase of the protocol, we set up a Full Factorial DOE with the following bounds and levels.

Figure 6: Full factorial DOE setup for CCCV charging analysis

The battery pack was initialized with a SOC of 80%, and the initial temperature was the same as the ambient temperature of each case. Each design was simulated until the CV portion of the charging strategy was completed. The following plot shows the li-plating potential for the designs simulated at -10°C.

Figure 7: Li-plating potential output at -10°C

It can be observed that the chances of li-plating drastically increase when the applied charge current is 0.325C and 0.55C. This is most likely due to low conductivity and rate of diffusivity at this temperature, leading to higher internal resistance. This would suggest that if the battery reaches this temperature, the charge current should be limited to a value of 0.1C if using a CCCV strategy.

We can use the same output to evaluate the charging strategy at the reference temperature. The following plot shows the li-plating output at 25°C.

Figure 8: Li-plating potential output at 25°C

The above plot shows that the charging current can be pushed up to 0.55C. BMS engineers can use this information to further optimize their charging strategy and algorithms.

Aside from li-plating potential, physics-based models provide a deeper analysis of heat generation and temperature. The battery model used in this example combines kinetic heat, joule heating from electronic resistance, ionic resistance, and concentration overpotential to output an accurate representation of internal heat generation. This can be tied to high-fidelity thermal models to provide an accurate representation of the temperature of the battery pack. The following figure shows the li-plating potential output at 60°C.

Figure 9: Li-plating potential at 60°C

The figure shows low risk of li-plating for higher charge rates. However, if we evaluate the battery pack temperature across these different charge rates, we can see that for the cooling strategy selected, going over a 0.775C charge rate can increase the battery pack temperature by 4°C.

Figure 10: Battery pack temperature at 60°C charging

 Although this may seem negligible for the whole pack, it is a strong indicator of potential localized temperature differences that can lead to cell imbalances, non-homogenous aging, hot spots, and safety critical events. To mitigate this, the BMS designer may set 0.55C as the limiting charge current.

This example highlights how virtualization and predictive plant models can fundamentally change the way BMS development and validation are approached. By enabling engineers to evaluate control strategies across a wide range of operating conditions, virtualization provides deeper insight into battery behavior and system limitations.

Using physics-based models, critical internal states such as lithium plating potential and heat generation can be assessed with high fidelity, allowing for more informed decisions around charging limits, safety margins, and overall system performance. This not only improves the robustness of BMS algorithms but also reduces reliance on extensive physical testing campaigns.

Ultimately, integrating virtual tools within the BMS development workflow empowers teams to accelerate design iteration, reduce costs, and increase confidence in system performance across the full battery lifecycle. As battery technologies and application demands continue to evolve, virtualization will play an increasingly essential role in delivering safe, efficient, and optimized energy storage systems.

Advancing BMS Validation with Predictive Models

Virtualization and predictive plant models are transforming how Battery Management Systems are developed and validated. In this example, the models helped identify how charging limits must vary across operating temperatures to reduce the risk of lithium plating and thermal issues. By providing visibility into critical internal battery states that cannot be directly measured experimentally, physics-based models enable more informed charging and control strategy decisions. This reduces reliance on extensive physical testing while improving confidence in battery performance and safety. As battery systems continue to evolve, virtualization will play an increasingly important role in enabling faster, safer, and more optimized BMS development.

Continue Exploring BMS Development

To further explore Battery Management System development, 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. To learn more about simulation-driven battery development, contact us and visit our Battery Simulation Solutions page to explore advanced battery modeling capabilities. Also, follow us on  LinkedIn to stay updated on the latest developments in simulation-driven engineering.