Simulation App Speeds Up Electric Truck Battery Design: Multiphysics Modeling Optimizes Laser Welding

For electric trucks to truly replace diesel fleets on long-haul trunk routes, battery pack lifespan and robustness are unavoidable hurdles. Laser welding, as a key process for connecting cells to busbars, can extend battery life when done well—or directly plant the seeds of accelerated aging when done poorly. An EU industry-academia project called LaserBATMAN has turned this problem from "trial and error" into something "computable," using multiscale simulation and custom simulation apps.

LaserBATMAN: Full-Scale Modeling from Cell to Battery Pack

The LaserBATMAN project launched in May 2022 and concluded in May 2025 (Ref. 1), with participants including Aurobay, Resolvent, the European Commission, the Technical University of Denmark (DTU), Vinnova, Innovation Fund Denmark, and Volvo Group Trucks Operations. The project's goal was clear: optimize two core metrics of the battery pack—robustness and lifespan.

The company responsible for simulation app development was Resolvent, a Danish green energy engineering consultancy. According to Martin Refslund Nielsen, its former Chief Commercial Officer, for this kind of multiphysics, multiscale problem, "COMSOL Multiphysics is exactly the right tool."

The project's technical approach unfolded in three steps: first, build an electrochemical model of the cell, incorporating relevant chemical reaction kinetics equations; then use the model's output data to train a deep neural network (DNN), producing a reduced-order model for rapid insight into cell behavior; finally, build a custom simulation app based on the reduced-order model to optimize busbar tab configurations in the battery pack.

You might not have noticed, but the team's initial ambition wasn't particularly large—they only wanted to study how laser welding affects cell electrochemistry and degradation. But as Nielsen put it, "the goals kept expanding," and ultimately their modeling covered the full performance range from cell to battery pack, and they "turned it into a design tool."

Reduced-Order Model (ROM) is a key means of addressing the computational bottleneck in multiscale, multiphysics simulation. A complete 3D electrochemical-thermal-electrical coupled finite element model can take hours or even days for a single transient solve, which cannot support the repeated iterations needed during the design phase. The ROM approach approximates the high-dimensional dynamics of the original system with a small number of degrees of freedom: common methods include Proper Orthogonal Decomposition (POD), Krylov subspace methods, and the deep neural network surrogate model used in this article. DNN-based reduction is a data-driven route, meaning high-fidelity models are first sampled in bulk across the parameter space to generate input-output datasets, and then a network is trained to fit the mapping relationship. After training, a single inference typically takes milliseconds, making interactive simulation apps possible. The trade-off is that ROM accuracy is limited by the coverage of the training data, and errors grow significantly when extrapolating to operating conditions outside the training set—so sampling design and accuracy validation are steps that must be handled carefully in this type of method.

"A Tool Within a Tool": How Busbar Weld Design Is Implemented

The simulation app's functionality was designed around practical engineering needs. Users can input operating parameters such as battery pack geometry, number of cells, ambient temperature, and charge-discharge current profiles, or apply a simple charging current or custom cycle. The battery performance results corresponding to each set of inputs are presented intuitively.

What truly demonstrates engineering value is the Busbar Weld Design Tool button—which Nielsen calls "a tool within a tool." Clicking it opens a window that allows setting parameters such as weld seam shape, busbar thickness, and laser penetration depth.

"From a robust design perspective, this is very interesting," Nielsen said. Users can apply variability to input parameters and observe what happens if the weld doesn't penetrate deeply enough into the busbar on the production line. This kind of forward-looking analysis of manufacturing deviations is difficult to achieve systematically with traditional trial-and-error processes.

Nielsen further noted: "We did gain deep technical insight that allows us to set design dimensions in the optimal range. Deep technical insight is also key to optimizing product performance." One concrete example: if simulation results show that heat dissipation could be better, the team might reduce the design by two cells while maintaining equivalent performance, thereby easily quantifying the trade-off between heat dissipation and configuration.

Predicting Battery Pack Performance: From Temperature to SEI Layer Thickness

The Compute Battery button in the app interface allows viewing weld temperature, state of charge (SOC), and solid electrolyte interphase (SEI) layer thickness under a given state. These metrics help engineering teams determine whether there is still room for design optimization. For example, users can apply cooling airflow in the app to observe whether turbulence or stagnant air occurs.

SEI layer thickness is an important indicator of battery aging. If laser welding introduces excessive heat effects, it accelerates SEI layer growth and directly shortens battery life. Presenting this microscopic metric alongside pack-level temperature and SOC distributions in the same interface means engineers can assess the cascading effects of the welding process on battery life without switching back and forth between multiple standalone tools.

According to Nielsen, the simulation app also provides more consistent analysis and reporting, and its intuitiveness makes knowledge sharing easier. "We are in a better position to optimize performance and improve robustness," he said.

Solid Electrolyte Interphase (SEI) is a passivation film formed on the surface of the lithium-ion battery anode by electrolyte decomposition during the first charge-discharge cycle, composed mainly of lithium carbonate, lithium fluoride, and various organic lithium salts. This film is itself necessary: it prevents the electrolyte from continuously reacting with the anode, and is a prerequisite for the battery to cycle stably. But the SEI continues to thicken with cycling and high temperatures, consuming active lithium and increasing interfacial impedance, leading to capacity fade and rising internal resistance—one of the dominant mechanisms of lithium-ion battery aging. The effect of temperature on SEI growth rate approximately follows an Arrhenius relationship, with the growth rate potentially doubling for every ~10°C increase. This is precisely the coupling point between the laser welding process and battery life: excessive welding heat input causes local temperature rise, accelerating SEI growth near the weld seam and creating a lifespan weak point. Therefore, in battery pack design, SEI thickness is often used as a microscopic quantitative indicator for evaluating the merits of thermal management strategies and manufacturing processes.

Implications for the Electric Truck Industry

Long-haul trunk transport is the toughest scenario for electric trucks—high range requirements, frequent charge-discharge cycles, and extremely low tolerance for battery reliability issues. LaserBATMAN's approach offers a reusable engineering paradigm: connecting the physical process of a manufacturing technique (laser welding), the electrochemical reactions of the cell, and pack-level thermal management and electrical performance through multiscale modeling and reduced-order models, then packaging it into a design tool that frontline engineers can use directly.

The broader value of this approach lies in lowering the barrier to using multiphysics simulation. Traditionally, tools like COMSOL require professional simulation engineers to operate, but custom apps package complex models into buttons and input fields, allowing battery design engineers and even manufacturing process personnel to directly conduct parameter studies and robustness assessments.

From a broader perspective, if the battery lifespan problem for electric trucks is not solved, fleet operational economics cannot compete with diesel vehicles. The "simulation upfront, manufacturing optimization downstream" path represented by LaserBATMAN may be more fundamentally effective at advancing the long-haul adoption of electric trucks than simply stacking more battery capacity.

Multiphysics simulation refers to solving two or more interacting physical processes in a coupled manner within the same computational domain. A battery pack is a typical multiphysics object: electrochemical reactions generate current and heat, temperature in turn affects reaction rates and heat transfer characteristics, and volume changes caused by charge-discharge also introduce structural stress. Platforms like COMSOL Multiphysics use fully coupled or segregated solvers to handle such problems, allowing users to define custom partial differential equations. Their advantage is maintaining accuracy even with strong coupling between physical fields, at the cost of higher modeling complexity and computational expense, typically requiring simulation engineers with professional training to operate. This is precisely where the value of custom simulation apps lies: they package well-tuned multiphysics models into a restricted graphical interface, with the underlying equations and mesh locked by the developer, so users can only adjust preset parameter inputs—thereby reducing the usage barrier to a level acceptable to engineering personnel while maintaining model reliability.

One-line Summary

The EU's LaserBATMAN project uses multiscale simulation and custom apps to turn laser welding's impact on electric truck battery life into a computable design tool.

Paragraph Summary

The EU industry-academia project LaserBATMAN (2022-2025), jointly advanced by Resolvent, Volvo, DTU, and others, aims to optimize the robustness and lifespan of electric truck battery packs. The technical approach has three steps: first, build an electrochemical cell model; then train a reduced-order model using a deep neural network; finally, package it into a custom simulation app. The Busbar Weld Design Tool in the app can analyze parameters such as weld seam shape, busbar thickness, and laser penetration depth, supporting forward-looking robust design against manufacturing deviations; the Compute Battery function presents weld temperature, SOC, and SEI layer thickness on the same screen, connecting microscopic aging indicators with pack-level thermal-electrical performance. This method lowers the barrier to multiphysics simulation to a level usable by frontline engineers, quantifies the trade-off between heat dissipation and cell count, and provides a reusable engineering paradigm for the battery lifespan challenge of long-haul electric truck transport.