The proliferation of electric vehicles (EVs) has fundamentally altered automotive architecture. To achieve a low center of gravity and maximize energy capacity, the high-voltage EV battery pack is almost universally mounted to the vehicle underbody. This integration is evolving further with trends like Cell-to-Body (CTB) and Cell-to-Chassis (CTC), deeply embedding the battery into the vehicle’s structural core. While beneficial for packaging and dynamics, this placement exposes the critical energy source to direct threats from road debris, curbs, and uneven surfaces. Bottom impact events, where obstacles strike the undercarriage, present a significant safety risk as mechanical damage can lead to internal short circuits, thermal runaway, and potential fire.
Statistical analysis of real-world accidents involving EVs highlights the prevalence of underbody collisions. Studies indicate a substantial portion of incidents involve upward impacts from obstacles like curbs or loose objects. Research has often focused on the cell-level safety or the EV battery pack enclosure in isolation. However, there is a growing need for a holistic, vehicle-level assessment methodology. Following such impacts, the precise extent of damage to the EV battery pack structure is difficult to quantify, often leading to costly full-pack replacements as a precaution. Therefore, developing accurate simulation tools to predict structural damage is of paramount practical and economic importance.
This article establishes a virtual test environment for a standardized vehicle bottom collision scenario. It introduces the Generalized Incremental Stress State Dependent Damage Model (GISSMO) to simulate and predict the failure of the EV battery pack enclosure structure. The simulation outcomes are then correlated with physical vehicle testing, exploring methodologies to enhance predictive accuracy and provide reliable guidance for the crash-safe design of EV battery pack systems.
Theoretical Foundation: The GISSMO Failure Model
During a real-world collision, metal components experience complex, multi-axial stress states far beyond simple uniaxial tension or compression. Traditional failure criteria often fall short in accurately predicting fracture under these combined loading conditions. The GISSMO failure model, an evolution of the Johnson-Cook fracture model, addresses this by incorporating a sophisticated damage accumulation and failure mechanism. It is particularly suited for simulating ductile metal failure under dynamic, non-linear loading paths typical of crash events.
The core of GISSMO is its dependence on the stress state, quantified by the stress triaxiality ($\eta$). This dimensionless parameter represents the ratio of hydrostatic stress to von Mises equivalent stress, defining the propensity for volume change versus shape distortion at a material point:
$$
\eta = \frac{\sigma_m}{\sigma_e} = \frac{\frac{1}{3}(\sigma_1 + \sigma_2 + \sigma_3)}{\sqrt{\frac{1}{2}[(\sigma_1-\sigma_2)^2 + (\sigma_2-\sigma_3)^2 + (\sigma_3-\sigma_1)^2]}}
$$
where $\sigma_m$ is the mean stress, $\sigma_e$ is the equivalent von Mises stress, and $\sigma_1, \sigma_2, \sigma_3$ are the principal stresses. The failure strain of a metal is highly sensitive to $\eta$; it is typically highest in pure shear ($\eta \approx 0$) and lower in tension ($\eta > 0$).
GISSMO models the progression to failure through a damage variable $D$, which accumulates non-linearly from 0 (virgin material) to 1 (complete failure). The incremental damage accumulation is governed by:
$$
\Delta D = \frac{n}{\varepsilon_f(\eta)} \left( \frac{D}{\varepsilon_f(\eta)} \right)^{1-\frac{1}{n}} \Delta \varepsilon_p
$$
where:
- $\Delta D$ is the increment of the damage variable.
- $n$ is a nonlinear damage accumulation exponent.
- $\Delta \varepsilon_p$ is the increment of equivalent plastic strain.
- $\varepsilon_f(\eta)$ is the failure strain as a function of stress triaxiality, determined from a suite of material tests (e.g., notched tension, shear, compression tests).
The function $\varepsilon_f(\eta)$ is a key material input. For the aluminum alloys (e.g., 6063-T6) commonly used in EV battery pack trays, this curve is derived from coupled experimental and numerical calibration. A representative curve is shown conceptually below, where failure strain decreases with increasing positive triaxiality.
| Stress State | Stress Triaxiality ($\eta$) | Relative Failure Strain $\varepsilon_f$ |
|---|---|---|
| Biaxial Compression | -0.6 to -0.3 | Very High |
| Pure Shear | ~0 | High |
| Uniaxial Tension | ~0.33 | Medium |
| Plane Strain Tension | ~0.58 | Low |
| Biaxial Tension | ~0.67 | Very Low |
In LS-DYNA, the GISSMO model is typically implemented via the *MAT_ADD_EROSION keyword in conjunction with a standard plastic material model like *MAT_024. During simulation, the damage $D$ is calculated for each integration point based on its instantaneous stress state and strain history. Once $D$ reaches 1, the corresponding finite element is deleted, allowing the model to realistically simulate crack initiation and propagation, as opposed to merely showing large, unrealistic plastic deformation. This capability significantly enhances the predictive accuracy for EV battery pack enclosure integrity assessment.
Vehicle-Level Bottom Impact Simulation Framework
Defining the Critical Collision Scenario
Based on accident statistics, curb strikes are identified as a predominant scenario for EV battery pack underbody damage. Standard road curbs have heights ranging from 100mm to 250mm. For this study, a simplified yet severe rectangular rigid obstacle is defined to represent a high curb or similar object. The geometry and positioning are designed to ensure direct contact with the EV battery pack.
A critical observation is that during forward motion, front subframes often act as a first line of defense. The most direct threat to the EV battery pack occurs during reverse motion, where the rear section of the pack is leading. Therefore, the established simulation scenario is a reverse motion impact.
| Parameter | Value | Description |
|---|---|---|
| Vehicle Mass | 1738 kg | Including all payload and fluids |
| Initial Velocity | 20 km/h (5.56 m/s) | Reverse direction |
| Obstacle Type | Rigid Rectangular Block | Fixed to ground |
| Obstacle Dimensions (LxWxH) | 500 mm x 120 mm x Hadj | Height is adjustable |
| Obstacle Overlap | 40 mm | Obstacle top surface exceeds pack bottom surface by 40mm |
| Simulation Time | 200 ms | Sufficient for full impact and rebound |
Finite Element Model Development
A full-vehicle finite element model of a compact electric vehicle is developed. The model, built in HyperMesh and solved with LS-DYNA, consists of approximately 2.8 million elements, predominantly shell elements with a characteristic size of 4-9mm, using the Belytschko-Lin-Tsay formulation for computational efficiency. Gravity is applied, and the vehicle is given an initial velocity towards the fixed obstacle.
The EV battery pack model is highly detailed. The enclosure consists of an aluminum lower tray (the focus of the GISSMO failure analysis) and a composite upper cover. The internal modules, containing the individual cells, are modeled using a homogenized crushable foam material (*MAT_LOW_DENSITY_FOAM) to represent their bulk mechanical response under compression. The connections between the EV battery pack and the vehicle body via mounting brackets are modeled using rigid body elements (RBE2) to simulate the bolted joints accurately.
| Component | Material | LS-DYNA Material Model | Key Properties / Notes |
|---|---|---|---|
| Upper Cover | SCR2252 Composite | *MAT_024 (Piecewise Linear Plasticity) | Defined with failure strain; GISSMO not typically applied to composites here. |
| Cell Modules | Homogenized Cell Stack | *MAT_063 (Crushable Foam) | Stress-Strain curve defines compressive behavior. Represents bulk energy absorption. |
| Module Housing | Aluminum 6061-T6 | *MAT_024 + *MAT_ADD_EROSION | Elasto-plastic properties. GISSMO failure parameters for Aluminum 6061 are defined here. |
| Cross Members / Frame | Aluminum 6063-T6 | *MAT_024 + *MAT_ADD_EROSION | Primary structural members of the tray. GISSMO failure parameters for Aluminum 6063 are critical. |
| Lower Tray / Enclosure | Aluminum 6063-T6 | *MAT_024 + *MAT_ADD_EROSION | Main component of interest. GISSMO model predicts tearing and fracture in this panel. |
Simulation Results and Analysis with GISSMO
The energy history of the simulation is monitored to ensure validity. The total energy remains constant, with kinetic energy converting into internal energy (plastic deformation). The hourglass energy is controlled to be less than 5% of the total energy, confirming that the simulation is not corrupted by numerical artifacts.
| Time (ms) | Kinetic Energy (kJ) | Internal Energy (kJ) | Hourglass Energy (kJ) | Total Energy (kJ) |
|---|---|---|---|---|
| 0 | 26.8 | 0.0 | 0.0 | 26.8 |
| 10 | 21.4 | 5.1 | 0.2 | 26.7 |
| 20 | 14.7 | 11.6 | 0.4 | 26.7 |
| 30 | 10.5 | 15.8 | 0.4 | 26.7 |
| 200 | 5.2 | 21.3 | 0.4 | 26.9 |
The deformation sequence clearly illustrates the advantage of the GISSMO model:
- t=0-10ms: The rigid obstacle engages the rear corner of the EV battery pack lower tray. High local bending stresses develop.
- t=10-20ms: Severe intrusion occurs. The lower tray deforms plastically. Without failure, the mesh would distort excessively. With GISSMO active, elements reaching the critical damage threshold (D=1) are deleted, initiating a realistic tear in the aluminum sheet.
- t=20-30ms: Intrusion peaks and rebound begins. The tear propagates, and the deletion of failed elements prevents unrealistic “stretching” of the material, revealing the final ruptured state of the enclosure.
The analysis of structural forces shows high stress concentrations not only at the direct impact zone but also at the forward mounting points of the EV battery pack. This indicates a global transfer of load through the pack structure to the vehicle body, highlighting the importance of robust mounting system design. The GISSMO-based simulation successfully predicts the specific location and mode of failure—a localized puncture and tearing of the tray—which is a critical output for design improvement.
Physical Validation: Full-Vehicle Bottom Impact Test
To validate the simulation methodology, a physical test was conducted according to the defined scenario. A vehicle, ballasted to the target mass, was reversed at 20 km/h into a rigid obstacle with identical geometry and 40mm of overhang relative to the EV battery pack underside.
| Aspect | Test Configuration / Observation |
|---|---|
| Test Article | Production Electric Vehicle |
| Impact Speed | 20.1 km/h (measured) |
| Obstacle | 500mm x 120mm x [Height] Steel Block, anchored to ground |
| Primary Damage | Severe local crushing and puncture of the EV battery pack rear lower tray. |
| Secondary Effects | Visible deformation of internal module structures; cooling lines compromised. |
| Post-Test Assessment | EV battery pack integrity breached. Pack deemed unsafe for operation without major repair/replacement. |
The physical damage pattern showed remarkable congruence with the simulation prediction. The location of the primary puncture, the direction of tear propagation, and the overall deformation shape of the EV battery pack enclosure matched the GISSMO simulation output closely. This correlation confirms that the model can accurately predict both the occurrence and the morphology of structural failure.
A quantitative comparison was made using acceleration data. An accelerometer was placed on the EV battery pack tray near the impact zone. The simulation accurately captured the timing and magnitude of the pulse.
| Metric | Physical Test | Simulation (GISSMO) | Deviation |
|---|---|---|---|
| Peak Acceleration (g) | 12.84 | 12.27 | -4.4% |
| Time of Peak (ms) | 23.5 | 22.8 | -3.0% |
| Pulse Duration (ms) | ~35 | ~33 | -5.7% |
The minor deviations are within acceptable limits for crash simulation, attributable to factors like slight variations in material properties, modeling simplifications of joints, and test measurement uncertainty. The close match in both qualitative damage and quantitative acceleration response validates the integrated approach of using a detailed vehicle model with the GISSMO failure criterion for EV battery pack bottom impact assessment.
Discussion: Implications for EV Battery Pack Design and Safety
The demonstrated methodology provides a powerful tool for the proactive design of crash-resistant EV battery pack systems. The ability to predict not just deformation but actual fracture allows engineers to identify weak points in the enclosure design—such as sharp geometric transitions, areas with unfavorable stress triaxiality, or insufficient material gauge—before physical prototyping.
Key design strategies informed by such analysis include:
- Geometric Optimization: Shaping the underbody of the EV battery pack to deflect obstacles or create controlled crush zones that absorb energy away from critical areas.
- Material Selection and Grade: Using the GISSMO $\varepsilon_f(\eta)$ curves to select aluminum alloys or other materials with superior ductility under the specific stress states (e.g., shear-dominated) expected in bottom impact.
- Strategic Reinforcement: Adding localized reinforcement (e.g., extrusions, additional sheet metal layers) in high-risk areas predicted by the simulation, such as near mounting points that experience high load transfer.
- Integrated Protection Systems: Designing dedicated underbody protection structures (skid plates, ramps) that are the first point of contact, designed to deform and absorb energy, thereby reducing the load transferred to the primary EV battery pack enclosure.
Furthermore, this vehicle-level approach is crucial. It captures system-level interactions, such as how the vehicle’s suspension compliance, body stiffness, and EV battery pack mounting strategy influence the final load path and deformation mode of the pack itself—an insight impossible to gain from testing the pack in isolation.
Conclusion
This work successfully establishes a robust simulation framework for assessing the structural integrity of EV battery pack systems during bottom impact events. By integrating the advanced GISSMO failure model into a full-vehicle finite element analysis, the method moves beyond simple deformation prediction to accurately forecast the onset and progression of ductile fracture in the battery enclosure. The strong correlation between the simulation results and physical vehicle testing validates the model’s predictive capability.
The implications for the automotive industry are significant. This approach enables a more nuanced safety development process for EV battery packs, allowing for targeted, weight-efficient design improvements that enhance crashworthiness without resorting to excessive over-engineering. It also provides a scientific basis for post-collision damage assessment, potentially informing repair-or-replace decisions. As EV architectures continue to evolve towards greater integration, such high-fidelity, system-aware simulation tools will become indispensable for ensuring the safety and reliability of the vehicle’s most critical—and vulnerable—component: the high-voltage EV battery pack.