In the pursuit of global carbon neutrality and the reduction of greenhouse gas emissions, the transition from traditional internal combustion engine vehicles to new energy vehicles has become a pivotal strategy. As the core component of electric vehicles, the performance and structural integrity of the power battery system directly determine the safety, reliability, and longevity of the entire vehicle. Among various environmental stresses, vibration, particularly random vibration encountered in harsh operating conditions such as mining and construction, poses a significant threat to the mechanical and electrical reliability of lithium ion battery packs. Failures induced by vibration can lead to issues like loose connections, electrolyte leakage, internal short circuits, and even thermal runaway. Therefore, a comprehensive vibration reliability study is essential during the design and validation phase of lithium ion battery systems for heavy-duty applications. This article presents a detailed investigation from a first-person research perspective, employing finite element simulation and experimental validation to evaluate the random vibration reliability of a lithium ion battery pack intended for heavy-duty mining vehicles. The focus is on modal analysis to avoid resonance and random vibration stress analysis to ensure structural integrity, with findings summarized through extensive tables and mathematical formulations.
The structural configuration of a typical lithium ion battery pack for heavy-duty vehicles is complex, comprising multiple subsystems that must withstand rigorous mechanical loads. The pack under study consists of a lower tray (or base), an upper cover, internal battery modules arranged in series and parallel configurations, mounting brackets, busbars, high-voltage connectors, a Manual Service Disconnect (MSD), and a Battery Management System (BMS) enclosure. The battery modules themselves contain numerous cylindrical or prismatic lithium ion cells, which are sensitive to mechanical deformation. The entire assembly is typically mounted onto the vehicle’s chassis frame, which transmits road-induced vibrations. The primary materials used include high-strength steel grades for structural parts like the tray and brackets, and aluminum for heat dissipation components. For simulation purposes, accurate material properties are crucial, as summarized in Table 1.
| Component | Material | Young’s Modulus (GPa) | Poisson’s Ratio | Density (kg/m³) | Yield Strength (MPa) | Tensile Strength (MPa) |
|---|---|---|---|---|---|---|
| Battery Tray & Cover | HC420LA Steel | 210 | 0.29 | 7800 | 420 | 470 |
| Mounting Brackets | HC420LA Steel | 210 | 0.29 | 7800 | 420 | 470 |
| Battery Module (simplified) | Equivalent Homogeneous Material | 1.25 | 0.32 | 1620 | N/A | N/A |
| Busbars | Copper C11000 | 110 | 0.34 | 8960 | 69 | 220 |
To assess the dynamic characteristics, a high-fidelity finite element model was developed. Using CAD software, the detailed geometry was imported into a pre-processing tool. The model was discretized using a combination of element types: Shell elements (SHELL181) for thin-walled structures like the tray and cover, solid elements (SOLID185) for the simplified battery module blocks, and rigid body elements (RBE2) and weld elements (CWELD, RBE3) to simulate bolted connections, spot welds, and seam welds, respectively. The lithium ion battery modules were modeled as equivalent solid blocks with homogenized material properties, matching the total mass and approximate stiffness. This simplification is valid for global structural analysis, though for localized cell stress, a more detailed model would be needed. Connections between modules and the tray, often made with structural adhesive, were simulated using bonded contact (CONTA174/TARGE170 with bonded behavior). Non-structural masses like the MSD and connectors were modeled using MASS21 elements. The final mesh comprised approximately 240,000 elements, ensuring a balance between computational accuracy and efficiency. The governing equation for the undamped free vibration, which forms the basis of modal analysis, is given by:
$$ [M]\{\ddot{u}\} + [K]\{u\} = \{0\} $$
where $[M]$ is the global mass matrix, $[K]$ is the global stiffness matrix, $\{u\}$ is the displacement vector, and $\{\ddot{u}\}$ is the acceleration vector. The solution to this eigenvalue problem yields the natural frequencies and mode shapes:
$$ ([K] – \omega_i^2 [M]) \{\phi_i\} = \{0\} $$
Here, $\omega_i$ is the i-th natural circular frequency (related to frequency $f_i$ by $\omega_i = 2\pi f_i$), and $\{\phi_i\}$ is the corresponding mode shape vector. The Block Lanczos method was employed in ANSYS to extract the first 10 modal frequencies. The results are critical for understanding whether the pack’s natural frequencies coincide with dominant excitation frequencies from the vehicle chassis, typically around 1-30 Hz for low-frequency road inputs. The extracted modal frequencies are listed in Table 2, and the first four mode shapes are described.
| Mode Number | Natural Frequency (Hz) | Primary Participating Component & Description |
|---|---|---|
| 1 | 33.85 | Upper Cover – Bending vibration along the longitudinal axis. |
| 2 | 37.90 | Battery Module Restraint Bars – Lateral rocking motion. |
| 3 | 39.10 | Battery Module Restraint Bars – Torsional motion. |
| 4 | 40.20 | Lower Tray Side Walls – Local panel bending. |
| 5 | 55.60 | Combined bending of tray and modules. |
| 6 | 62.70 | Mounting Bracket local vibration. |
| 7 | 70.80 | Upper Cover – Higher order bending. |
| 8 | 82.10 | Torsional mode of the entire pack. |
| 9 | 103.50 | Local vibration of internal busbar structures. |
| 10 | 145.70 | Complex combined mode involving tray stiffeners. |
The first modal frequency at 33.85 Hz is of paramount importance. Since the predominant external excitation frequency transmitted from the mining vehicle’s frame is estimated not to exceed 30 Hz under normal operating conditions, this indicates that the fundamental resonant frequency of the lithium ion battery pack is sufficiently separated from the primary excitation band. This frequency margin helps avoid low-order resonance, thereby enhancing the vibration reliability of the pack structure. However, modes 2 and 3 (around 38-39 Hz) involve the restraint bars that secure the lithium ion modules. Although their frequencies are above 30 Hz, prolonged exposure to vibrations near these frequencies could cause repetitive impact on the cells. A design improvement suggested is the application of high-strength structural adhesive between the bars and the modules to increase damping and stiffness, effectively raising these frequencies further and reducing relative motion.
Following modal analysis, random vibration analysis was performed to predict the stress response under real-world vibration spectra. Random vibration is characterized by its Power Spectral Density (PSD) function, $G_{xx}(f)$, which describes the distribution of mean square acceleration per unit frequency. The input PSD profiles for the three orthogonal axes (X-longitudinal, Y-lateral, Z-vertical) were defined according to the standard GB/T 31467.3-2015, which is aligned with typical automotive vibration test specifications. The PSD levels in $(m/s^2)^2/Hz$ are tabulated in Table 3 as piecewise linear segments.
| Axis | Frequency Range (Hz) | PSD Level $(m/s^2)^2/Hz$ | Slope (dB/octave) |
|---|---|---|---|
| Vertical (Z) | 5 – 10 | 0.05 to 0.06 | +3 dB/oct |
| 10 – 20 | 0.06 (constant) | 0 dB/oct | |
| 20 – 200 | 0.06 to 0.0008 | -12 dB/oct | |
| Overall Grms | Calculated: 1.22 $m/s^2$ | ||
| Lateral (Y) | 5 – 20 | 0.04 (constant) | 0 dB/oct |
| 20 – 200 | 0.04 to 0.0008 | -12 dB/oct | |
| Overall Grms | Calculated: 0.98 $m/s^2$ | ||
| Longitudinal (X) | 5 – 10 | 0.0125 to 0.03 | +12 dB/oct |
| 10 – 20 | 0.03 (constant) | 0 dB/oct | |
| 20 – 200 | 0.03 to 0.00025 | -12 dB/oct | |
| Overall Grms | Calculated: 0.71 $m/s^2$ | ||
The random vibration analysis is a spectral method based on the mode superposition approach. The process involves calculating the modal participation factors and then combining the responses from each mode within the frequency range of interest. The equation of motion for the system under base excitation with random acceleration is:
$$ [M]\{\ddot{u}\} + [C]\{\dot{u}\} + [K]\{u\} = -[M]\{R\} \ddot{u}_g(t) $$
where $[C]$ is the damping matrix (assumed as Rayleigh damping: $[C] = \alpha[M] + \beta[K]$), $\{R\}$ is the influence vector, and $\ddot{u}_g(t)$ is the base acceleration with a given PSD. The response PSD matrix $[S_{uu}(f)]$ of displacements can be derived from the input PSD $G_{aa}(f)$ using the transfer function $[H(f)]$:
$$ [S_{uu}(f)] = [H(f)]^* \{R\} G_{aa}(f) \{R\}^T [H(f)]^T $$
where $[H(f)] = (-\omega^2[M] + i\omega[C] + [K])^{-1}$, and $^*$ denotes complex conjugate. The stress PSD is then obtained from displacement PSD via the strain-displacement matrix $[B]$ and material elasticity matrix $[D]$: $[S_{\sigma\sigma}(f)] = [D][B] [S_{uu}(f)] [B]^T [D]^T$. The root mean square (RMS) stress $\sigma_{rms}$ is calculated by integrating the stress PSD over frequency:
$$ \sigma_{rms} = \sqrt{\int_{f_1}^{f_2} S_{\sigma\sigma}(f) \, df} $$
For reliability assessment, the peak stress (often taken as 3 times the RMS value for Gaussian processes) is compared to the material yield strength. The simulation results for von Mises stress in key components under the three independent directional inputs are consolidated in Table 4. The analysis considered stress distributions at critical locations such as the root of mounting brackets, corners of the lower tray, junctions of stiffeners, and areas near the battery module attachments.
| Key Structural Component | Maximum Stress – Z Direction (MPa) | Maximum Stress – Y Direction (MPa) | Maximum Stress – X Direction (MPa) | Material Yield Strength (MPa) | Safety Factor (Min based on Peak Stress ≈ 3×RMS) |
|---|---|---|---|---|---|
| Mounting Brackets (Root Section) | 302.9 | 109.2 | 8.5 | 420 | ~1.38 (Z-direction critical) |
| Lower Tray (Center & Corners) | 251.4 | 68.3 | 75.9 | 420 | ~1.67 |
| Longitudinal Stiffeners on Tray | 165.5 | 13.0 | 56.7 | 420 | ~2.54 |
| Upper Cover (Central Region) | 21.7 | 5.4 | 3.3 | 420 | >5.0 |
| Battery Module Attachment Points | 89.3 | 45.1 | 32.8 | N/A (Adhesive interface) | N/A |
The results clearly indicate that all structural components made of HC420LA steel exhibit maximum calculated stresses well below the material yield strength of 420 MPa, even when considering peak stress estimates. The mounting brackets experience the highest stress, primarily in the vertical (Z) direction, which is consistent with the dominant vertical road input. The safety factor, calculated as yield strength divided by the peak stress (approximated by $3 \times \sigma_{rms}$), remains above 1.3 for the most critical part, indicating a robust design. The stresses in the upper cover are negligible, confirming its non-critical structural role. Importantly, the simulated stresses at the interface between the simplified lithium ion battery module blocks and the tray are also low, suggesting that the adhesive bonding is sufficient to prevent debonding under the specified random vibration. However, it must be noted that this simulation does not capture potential fatigue damage over time. Fatigue life estimation requires an S-N curve for the material and a cycle counting method like Rainflow, applied to the stress time history derived from the PSD. A preliminary fatigue life estimate can be made using Miner’s rule:
$$ D = \sum_{i=1}^{k} \frac{n_i}{N_i} $$
where $D$ is the cumulative damage, $n_i$ is the number of cycles at a given stress range, and $N_i$ is the number of cycles to failure at that stress range from the S-N curve. For the stress levels observed, the predicted fatigue life for the steel components far exceeds the typical service life of a mining vehicle.
To validate the simulation results, a physical random vibration test was conducted on a prototype of the lithium ion battery pack. The pack was instrumented with accelerometers at strategic locations on the lower tray and mounting brackets to monitor the input and response vibrations. It was securely bolted to a high-performance electrodynamic shaker table, exactly replicating its in-vehicle mounting conditions. The test profile followed the same PSD specifications used in the simulation (Table 3), applied sequentially along the Z, Y, and X axes. Each axis test duration was 8 hours per the standard, making a total of 24 hours of continuous random vibration exposure. During and after the test, the pack was monitored for any electrical discontinuities (via continuous monitoring of insulation resistance and open-circuit voltage) and subjected to a thorough visual and dimensional inspection. The test outcomes are summarized in Table 5.
| Inspection Category | Pre-Test Condition | Post-Test Condition | Pass/Fail Criteria | Result |
|---|---|---|---|---|
| Structural Integrity | No cracks, permanent deformation. | No visible cracks, permanent deformation, or loose fasteners. | No permanent deformation or structural damage. | PASS |
| Electrical Function | Normal insulation resistance (>100 MΩ), stable voltage. | Insulation resistance maintained >100 MΩ; No voltage drop or interruption. | No loss of electrical continuity or isolation. | PASS |
| Mounting Points | Torque values as specified. | No loosening of mounting bolts; torque check within tolerance. | Bolts remain tight per specification. | PASS |
| Internal Components | Modules secured, connectors seated. | No displacement of modules; connectors intact. | No internal dislodgement or damage. | PASS |
| Sealing (if applicable) | IP67 rating verified. | No ingress of dust/water; seal integrity intact. | Maintains specified IP rating. | PASS |
The successful passage of the vibration test without any structural or electrical failures provides strong experimental validation for the finite element simulation results. It confirms that the lithium ion battery pack design possesses adequate strength and stiffness to withstand the random vibration environment of a heavy-duty mining vehicle. The correlation between test and simulation also builds confidence in using the finite element model for future design iterations and optimization studies.
In conclusion, this integrated simulation and experimental study demonstrates a robust methodology for assessing the random vibration reliability of lithium ion battery packs. The key findings are: Firstly, the fundamental natural frequency of the pack (33.85 Hz) is higher than the dominant vehicle frame excitation frequency (≤30 Hz), effectively avoiding detrimental low-frequency resonance. Secondly, the random vibration-induced stresses in all critical structural components, such as mounting brackets, lower tray, stiffeners, and cover, are significantly lower than the yield strength of the materials used, ensuring static strength reliability. Thirdly, the physical vibration test corroborated the simulation predictions, with no observable permanent deformation, cracking, or electrical malfunction. For future work, the model can be enhanced by incorporating more detailed representations of the individual lithium ion cells and their internal components to study cell-level stress and potential separator deformation. Furthermore, multi-axial random vibration analysis and combined thermal-mechanical vibration analysis would provide an even more comprehensive reliability assessment for these critical energy storage systems in demanding applications. The methodologies and results presented herein contribute to the foundational knowledge required for designing safer and more durable lithium ion battery systems for the heavy-duty electric vehicle industry.