In the context of global energy transition and the pursuit of carbon neutrality goals, the integration of renewable energy sources into power grids has accelerated significantly. The battery energy storage system stands as a critical technological enabler, enhancing grid stability, facilitating renewable energy penetration, and providing ancillary services. As a fundamental integration unit within such systems, the energy storage battery pack is a complex multi-physical field system involving mechanical, electrical, and thermal interactions. Its structural integrity directly impacts the overall safety, reliability, and lifespan of the entire battery energy storage system. Unlike automotive applications where dynamic loads are frequent, stationary energy storage packs primarily face mechanical challenges during transportation—from manufacturing facilities to installation sites. Therefore, a comprehensive assessment of the structural response under transportation-induced vibrations and shocks is paramount for ensuring the operational reliability of the battery energy storage system.
This study focuses on the structural safety design of a specific energy storage battery pack intended for large-scale stationary applications. We employ finite element analysis (FEA) as the primary investigative tool, utilizing Altair’s OptiStruct solver for its robust capabilities in linear and nonlinear structural analysis. The core objective is to develop a refined simulation model that accurately captures the pack’s dynamic characteristics and stress distributions under prescribed transportation极限 conditions. Subsequently, we validate the model through experimental modal analysis, identify critical structural weaknesses, and propose and validate design optimizations. The overarching goal is to ensure that the pack’s fundamental frequency is sufficiently high to avoid resonance with typical vehicle excitations and that its strength meets safety factors under anticipated inertial loads, thereby contributing to the robustness of the deployed battery energy storage system.
The battery pack under investigation comprises several key sub-assemblies. Its primary structure includes a lower carrier (or base tray), numerous lithium-ion battery cells, thermally conductive interstitial plates between cells, aluminum busbars (Al-Bus) for electrical interconnection, end covers, end plates, and a supportive transport frame. The cells are arranged into two modules, each containing 11 cells in series. The cells are bonded to the interstitial plates and the lower carrier using adhesive. The end plates are welded and bolted to the lower carrier, providing lateral constraint. The entire assembly is secured to a transport frame via bolts. A detailed bill of materials and component functions is summarized in Table 1.
| Component | Primary Material | Key Function |
|---|---|---|
| Lower Carrier / Base Tray | Steel (DC01) | Primary structural base, supports all internal components, interfaces with transport frame. |
| Battery Cell (Cathode: LiNiMnCoO2) | Homogenized Electrode-Separator-Electrolyte Composite | Energy storage element. Modeled as a homogeneous continuum for structural analysis. |
| Interstitial Plate | Aluminum (Al3003) | Provides thermal conduction pathway and maintains cell spacing. |
| Aluminum Busbar (Al-Bus) | Aluminum (Al1060) | Electrical series/parallel connection between cell terminals. |
| End Plate & Assembly | Aluminum Die Casting (A380) | Provides lateral stiffness and constraint to the cell stack, connects to lower carrier. |
| End Cover | Steel | Protects busbar connections and provides a mounting surface. |
| Transport Frame | Steel | Facilitates handling and shipping, provides mounting points to the transport vehicle. |
| Adhesive Layers | Structural Epoxy | Bonds cells to interstitial plates and lower carrier, transferring shear loads. |
| Spot Welds & Seam Welds | — | Joins sheet metal components (e.g., end plate to lower carrier). |
| Bolts | Steel | Provides clamping force for end cover and frame attachment. |
Developing a high-fidelity finite element model is the cornerstone of this analysis. We constructed a detailed 3D model where components are discretized using predominantly hexahedral and shell elements, balancing accuracy and computational cost. The total model consists of approximately 2.3 million elements. The material properties assigned are critical for predictive accuracy. The battery cells pose a particular challenge as they are heterogeneous assemblies. For structural dynamics, we treat them as a linear, isotropic, homogeneous solid with equivalent properties derived from micromechanics and literature. The equivalent Young’s modulus ($E_{eq}$), density ($\rho_{eq}$), and Poisson’s ratio ($\nu_{eq}$) are calculated based on the mass, volume, and constituent properties of the jellyroll (electrodes, separator). A widely used formula for the equivalent modulus of a wound cell under compression is given by:
$$ E_{eq} = \frac{E_{active} \cdot V_{active} + E_{sep} \cdot V_{sep} + E_{case} \cdot V_{case}}{V_{total}} $$
where $E_{active}$, $E_{sep}$, $E_{case}$ are the moduli of the active material stack, separator, and casing, and $V$ denotes volume fractions. For our model, we adopted a simplified yet validated approach, using $E_{eq} = 75$ MPa, $\rho_{eq} = 2123$ kg/m³, and $\nu_{eq} = 0.01$. The properties of all structural materials are listed in Table 2.
| Component | Element Type | Material | Young’s Modulus, E (MPa) | Poisson’s Ratio, ν | Density, ρ (kg/m³) | Tensile Strength (MPa) |
|---|---|---|---|---|---|---|
| Battery Cell | Hexahedral | Homogenized | 75 | 0.01 | 2123 | N/A |
| Aluminum Busbar | Hexahedral | Al1060 | 70,000 | 0.33 | 2700 | 55 |
| Interstitial Plate | Shell | Al3003 | 69,000 | 0.33 | 2700 | 120 |
| Lower Carrier | Shell | Steel (DC01) | 210,000 | 0.30 | 7850 | 330 |
| End Plate Assembly | Hexahedral/Shell | Al Die Cast (A380) | 71,000 | 0.33 | 2700 | 330 |
| Transport Frame | Shell | Steel | 210,000 | 0.30 | 7850 | N/A |
| End Cover | Tetrahedral | Steel | 210,000 | 0.30 | 7850 | 320 |
The modeling of connections and contacts is equally vital. We represented spot welds and seam welds using solid hexahedral elements at the actual weld locations, connected to the parent shells via rigid body elements (RBE2) and interpolating elements (RBE3) to simulate force transfer. Adhesive bonds between cells and plates/carrier were modeled using thin solid layers with corresponding material stiffness, tied to the adjacent surfaces. Mechanical contacts were defined for assemblies that are bolted or in frictional interaction. For instance, the interface between the lower carrier and the transport frame was modeled with a frictional contact (coefficient µ=0.2). All bolted joints were simulated using pretension elements or tied contacts in the clamped regions. The governing equation for the linear static and dynamic analyses performed is the fundamental equation of motion:
$$ [M]\{\ddot{u}\} + [C]\{\dot{u}\} + [K]\{u\} = \{F(t)\} $$
where $[M]$, $[C]$, and $[K]$ are the global mass, damping, and stiffness matrices, $\{u\}$ is the displacement vector, and $\{F(t)\}$ is the force vector. For modal analysis, we neglect damping and external forces, solving the eigenvalue problem:
$$ ([K] – \omega_i^2 [M]) \{\phi_i\} = 0 $$
Here, $\omega_i$ is the i-th natural circular frequency (rad/s), related to frequency $f_i$ by $\omega_i = 2\pi f_i$, and $\{\phi_i\}$ is the corresponding mode shape vector. The boundary conditions applied replicate the pack’s mounting during road transport. The pack is constrained at eight locations: four on the end plates and four on the transport frame, with all six degrees of freedom (DOF) fixed (Figure 4 schematic). This represents a fully fixed connection to the vehicle bed.
Modal analysis serves a dual purpose: it reveals the inherent dynamic characteristics of the pack and provides a basis for model validation. We extracted the first 20 modes using the Lanczos algorithm in OptiStruct. The results indicated that the first few global modes are dominated by the vibration of the battery module mass relative to the carrier structure. The first three mode shapes and frequencies are detailed in Table 3. The first mode at 15.3 Hz involves lateral (X-direction) bending of the cell stack, the second at 20.5 Hz involves vertical (Z-direction) bending, and the third at 28.5 Hz is a torsional mode of the modules. These low frequencies are concerning as they fall within the typical excitation range of road vehicles (1-25 Hz), posing a resonance risk during transport, which could lead to accelerated fatigue damage or immediate overstress in the battery energy storage system components.
| Mode Number | Natural Frequency (Hz) | Primary Mode Shape Description | Dominant Component Motion |
|---|---|---|---|
| 1 | 15.3 | Lateral (X) bending of the battery modules | Cells and interstitial plates move laterally as a beam. |
| 2 | 20.5 | Vertical (Z) bending of the battery modules | Cells and interstitial plates move vertically as a beam. |
| 3 | 28.5 | Torsional motion of the battery modules about the longitudinal (Y) axis | One end of the module twists relative to the other. |
To validate the finite element model, we conducted experimental modal analysis via swept-sine vibration tests. The battery pack was mounted on a hydraulic shaker table replicating the transport boundary conditions. Three tri-axial accelerometers were attached to strategic points on one battery module. Logarithmic sine sweeps from 10 Hz to 200 Hz were executed in the lateral (X), longitudinal (Y), and vertical (Z) directions. The frequency response functions (FRFs) were analyzed to identify resonant peaks. The experimental natural frequencies correlated well with the FEA predictions, as shown in Table 4. The maximum deviation was less than 10%, which is within acceptable bounds for engineering purposes and confirms the fidelity of our mass distribution, stiffness representation, and boundary conditions. This validated model forms a reliable basis for subsequent strength assessment and optimization, crucial for the structural integrity of the battery energy storage system.
| Excitation Direction | Experimental Resonance 1 (Hz) | Correlated FEA Mode (Frequency, Hz) | Deviation (%) | Experimental Resonance 2 (Hz) | Correlated FEA Mode (Frequency, Hz) | Deviation (%) |
|---|---|---|---|---|---|---|
| Lateral (X) | 14.0 | Mode 1 (15.3) | +9.3% | 27.9 | Mode 3 (28.5) | +2.2% |
| Longitudinal (Y) | 16.7 | Mode 1/3 (15.3/28.5) | ~9% / ~6% | 28.7 | Mode 3 (28.5) | -0.7% |
| Vertical (Z) | 20.4 | Mode 2 (20.5) | +0.5% | 26.1 | Mode 3 (28.5) | +9.2% |
With a validated model, we proceeded to assess the structural strength under极限 transportation inertial loads. These loads simulate sudden maneuvers: sharp turns (lateral acceleration), emergency braking (longitudinal acceleration), and traversing severe bumps or potholes (vertical acceleration). Based on industry standards and empirical data from heavy-duty transport, the following极限 acceleration factors were applied as static body loads in the respective directions (Table 5). This quasi-static approach is conservative and standard for evaluating peak stresses in such non-impact scenarios.
| Load Case Description | Direction | Applied Acceleration (g = 9.81 m/s²) | Rationale |
|---|---|---|---|
| Sharp Turn (Cornering) | ±X (Lateral) | ±2.0 g | Simulates high centrifugal force during a tight turn. |
| Emergency Braking | +Y (Longitudinal, forward deceleration) | +2.0 g | Simulates maximum deceleration force. |
| Emergency Acceleration* | -Y (Longitudinal, rearward acceleration) | -2.0 g | Simulates sudden forward lurch (less critical). |
| Traversing a Severe Bump | +Z (Vertical Upward) | +3.0 g | Simulates the pack being thrown upward. |
| Dropping into a Pothole | -Z (Vertical Downward) | -3.0 g | Simulates maximum downward G-force upon impact. |
*Note: The -Y load case is often less severe due to the presence of restraints, but is analyzed for completeness.
The linear static stress analysis was performed for each load case. The governing equation simplifies to $[K]\{u\} = \{F\}$, where $\{F\}$ is the inertial load vector computed as $\{F\} = [M]\{a\}$, with $\{a\}$ being the applied acceleration vector. The most critical load cases were the lateral (±2g) and vertical downward (-3g) accelerations. The results revealed a significant structural vulnerability. High stress concentrations, far exceeding the material yield and tensile strength, were found in the connection regions between the end plates and the lower carrier, particularly at the rearmost weld points and sharp geometric corners. The von Mises stress in these localized areas reached up to 1594 MPa in the initial design, which is catastrophic. In contrast, other components like the aluminum busbars, interstitial plates, and end covers exhibited relatively low stress levels (mostly below 50 MPa). The failure mechanism is intuitive: under inertial loads, the massive battery modules (contributing ~85% of the total pack mass) tend to continue moving, exerting large bending moments on the end plates, which act as cantilever beams welded to the base. The weak link was the insufficient stiffness and strength of this connection detail. This identified a clear need for design optimization to ensure the survival of the battery energy storage system during harsh transport.
The optimization objectives were twofold: 1) Increase the first global natural frequency above 22 Hz to avoid resonance with dominant vehicle frequencies (considering a 10% separation margin from a typical 20 Hz vehicle frequency), and 2) Reduce the maximum von Mises stress in all components below the material tensile strength with an acceptable safety factor (target stress < yield strength). We employed a parametric and iterative simulation-driven design approach rather than a formal topology optimization due to the predefined packaging constraints. The key design modifications implemented were:
- End Plate Reinforcement: The original end plate had an incomplete flange and a pronounced “Z”-shape stiffener that created stress risers. We extended the top flanges to form a continuous box-section, increasing the bending moment of inertia. The “Z”-stiffener was replaced with localized, thicker reinforcement patches near the high-stress weld zones. The improvement in bending stiffness can be approximated by the area moment of inertia formula for a beam section: $I_{new} > I_{old}$. For a rectangular section, $I = \frac{b h^3}{12}$, so increasing the height (h) or adding material away from the neutral axis significantly boosts $I$.
- Rear Constraint Enhancement: The initial design lacked positive restraint at the rear (tail end) of the pack relative to the transport frame, allowing excessive relative displacement. We added a mechanical stop system at the rear of the transport frame. This included set-screws for lateral (X) and longitudinal (Y) constraint and an additional bolted connection between the lower carrier and the frame for vertical (Z) constraint. This effectively turns the rear support into a semi-rigid connection, reducing the cantilever length of the pack and shifting the load path.
The modified design was re-modeled and analyzed. The modal analysis results (Table 6) showed a remarkable improvement. The first natural frequency increased by over 50% to 23.4 Hz, successfully meeting the >22 Hz target. This significantly reduces the risk of resonance during transport, enhancing the durability of the battery energy storage system.
| Mode Number | Natural Frequency (Hz) | Frequency Increase vs. Initial | Primary Mode Shape Description |
|---|---|---|---|
| 1 | 23.4 | +52.9% | Lateral bending of modules (stiffer connection). |
| 2 | 38.6 | +88.3% | Vertical bending of modules. |
| 3 | 51.7 | +81.4% | Module torsion coupled with frame motion. |
The static strength analysis under the same极限 load cases demonstrated even more dramatic improvements. The maximum von Mises stress across the entire assembly plummeted from 1594 MPa to 321.2 MPa, which is below the tensile strength of the lower carrier steel (330 MPa). The critical stress now occurs in the lower carrier near a rear bolt hole under the -3g vertical load case, indicating a more desirable and distributed load path. The stress in the end plate assembly was reduced to 247.9 MPa under the emergency braking load, well within its capacity. A summary of the worst-case stresses for key components in the optimized design is presented in Table 7. The stress reduction factor ($R_f$) for the most critical location is calculated as:
$$ R_f = \frac{\sigma_{initial} – \sigma_{optimized}}{\sigma_{initial}} \times 100\% = \frac{1594 – 321.2}{1594} \times 100\% \approx 79.8\% $$
This denotes an 80% reduction in peak stress, a highly effective optimization outcome.
| Component | Critical Load Case | Max Von Mises Stress (MPa) | Material Tensile Strength (MPa) | Safety Margin (Strength/Stress) | Status |
|---|---|---|---|---|---|
| Lower Carrier | Vertical Downward (-3g) | 321.2 | 330 | ~1.03 | Acceptable (at limit) |
| End Plate Assembly | Emergency Braking (+2g Y) | 247.9 | 330 | ~1.33 | Safe |
| Aluminum Busbars | Sharp Turn (+2g X) | 18.5 | 55 | ~2.97 | Very Safe |
| Interstitial Plates | Vertical Downward (-3g) | 29.7 | 120 | ~4.04 | Very Safe |
| End Cover | Vertical Downward (-3g) | 165.0 | 320 | ~1.94 | Safe |
| Adhesive Bonds | All Cases | < 5.0 (Shear) | ~20 (Shear) | > 4.0 | Safe |
The process and findings presented herein offer a practical and validated framework for the structural design and validation of energy storage battery packs against transportation hazards. By employing a detailed finite element modeling strategy that carefully represents the mass of the cells and the intricacies of their mechanical connections, we can accurately predict both dynamic and static responses. The imperative to validate such models through physical testing, like sweep-frequency tests, cannot be overstated, as it closes the loop between simulation and reality. The identified failure mode—high stress at the interface between the massive battery modules and the structural frame—is a common challenge in the design of any battery energy storage system enclosure. The optimization solutions implemented, namely reinforcing the load-bearing end plates and adding strategic rear constraints, proved highly effective. These modifications not only resolved the strength issue but also concurrently raised the fundamental frequency, addressing the resonance concern—a dual benefit that is often sought in structural dynamics.
In conclusion, this work underscores the critical importance of rigorous mechanical analysis in the development phase of stationary battery energy storage system components. Transportation is a vulnerable phase in the lifecycle of these systems, and overlooking its mechanical demands can lead to field failures that compromise system availability and safety. The methodology outlined—from high-fidelity modeling and experimental correlation to targeted strength and frequency-based optimization—provides a robust blueprint for engineers. Future work could extend this approach to include random vibration fatigue analysis based on road spectral data, multi-axis simultaneous loading, and the exploration of lightweight materials and advanced joining techniques to further enhance the performance and cost-effectiveness of battery energy storage system enclosures. Ultimately, ensuring the structural resilience of battery packs directly contributes to the reliability, longevity, and economic viability of large-scale battery energy storage system installations, supporting a more sustainable and secure energy infrastructure.