NDT in EV Safety - 2025 Trends

Table of Content

• The Role of NDT in EV Manufacturing
• AI-Powered Inspection and Data Analysis
• The Integration of IoT in NDT for EVs
• Key Takeaways
• References

Trending vehicles like the Tesla, have shifted on-road transportation for common people worldwide. This fascination and mass approval of Electric Vehicles (EVs) has transformed the automotive manufacturing process, especially with respect to the materials and components used in the industry. 

NDT technologies are vital in ensuring the integrity of critical components, such as battery packs, electrical wiring, and the overall structural framework of EVs. NDT allows manufacturers to detect potential faults and ensure the safety and performance of components.

The Role of NDT in EV Manufacturing

As EVs continue to replace traditional combustion engine vehicles, the need for effective Non-destructive Testing (NDT) techniques has become exponentially vital. In 2025, NDT is set to be pivotal in ensuring electric vehicle safety, especially in areas like battery performance, structural integrity, and electrical systems. The NDT Techniques used in EV Manufacturing include the following:

1. Ultrasonic Testing 

The battery packs in EVs are a major concern during their design, production, and operation. Despite being efficient, Lithium-ion batteries can pose significant risks if they are compromised. Ultrasonic Testing (UT) can be used by impinging the sound waves on them to detect internal flaws, such as delamination or voids. 

Lithium-ion battery cells require nanometre-level precision to avoid dendrite formation, electrode delamination, or electrolyte dry-out. UT’s role is critical in NDT EV safety to help ensure that the battery packs are safe and operate at their maximum efficiency. Within UT Technologies, the application of PAUT includes:

I. Frequencies & Probes

5–20 MHz linear arrays (e.g., Evident’s Olympus Omniscan MX3) map the anode and cathode of the EV batteries with axial resolution <0.1 mm. Dual-Element Probes isolate signals from copper-aluminium interfaces, reducing noise in tab welds.

Image Credit: Olympus

II. Beam Steering

• Sectorial scans detect delamination between graphite anode and copper foil, even with curved cell geometries (e.g., Tesla 4680 cells).
• Full Matrix Capture (FMC): Post-processing algorithms like the Total Focusing Method reconstruct 3D defects using synthetic aperture focusing.

In-Line UT Systems can be implemented in Gigafactories for inspection, in the following ways:

• Robotic PAUT: ABB’s IRB 6700 robots with water-jet-coupled UT heads can inspect 100% of cells at 120 ppm), achieving less than 0.5% false rejects.
• Laser Ultrasonics: Non-contact systems use pulsed lasers to generate broadband ultrasound up to 100 MHz, detecting micro-voids (voids less than 10 µm) in solid-state battery electrodes.

UT can also be used for structural inspections, like that of composite & adhesive bonding. EVs use multi-material designs with materials like CFRP, aluminium, and HSS to lower the weight, but adhesive bonds are prone to disbonds. Automotive NDT Trends in 2025 prioritise shear wave UT for these interfaces:

• Battery Enclosures:

1. Guided Wave UT: Low-frequency Lamb waves (50–500 kHz) inspect large-area CFRP enclosures (e.g., BMW iX) for impact damage or resin-rich zones.
2. EMAT (Electromagnetic Acoustic Transducer): Non-contact UT for coated surfaces, detecting cracks in aluminium battery trays under thermal cycling.

• Adhesive Joints:

1. Pulse-Echo TOFD: Time-of-flight diffraction (TOFD) can quantify disbond lengths in structural adhesives with ±0.1 mm accuracy.

• Air-coupled UT: 400 kHz transducers inspect adhesive thickness between 0.2 and 2 mm without couplant, critical for BMW/Mercedes BEV platforms.

UT can also be implemented to test high-speed components. Electric motor failures can stem from rotor lamination cracks or stator winding defects. Non-destructive Testing Technologies here include:

• Immersion Ultrasonic Testing (UT) for Rotors:

1. 10 MHz focused probes like the Sonatest MasterScan 680 can scan laminated steel stacks for interlayer cracks caused by electromagnetic stress.
2. Automated defect sizing: Software like Eddyfi Lyft uses TOFD to calculate crack depth-to-length ratios, predicting fatigue life.
3. Stator Hairpin Inspection: Angled beams at 45°–70° can detect incomplete copper welding in hairpin stators like in the Volkswagen MEB motors, ensuring less than 5% resistance variation.

Terahertz Time-Domain Spectroscopy (THz-TDS) can also be used for inspection, wherein 0.1–3 THz pulses penetrate ceramic-coated separators to detect electrolyte stratification with 30 µm resolution.

Digital Twins can also be used with UT for predictive NDT. The ANSYS Granta MI system links UT inspection data like porosity distribution to finite element models, simulating thermal runaway propagation in battery packs.

2. Thermographic Testing 

Thermographic Testing can detect electrical faults, overheating, and hotspots within complex electrical systems. This method uses infrared cameras to capture temperature variations, which are indicative of issues such as poor connections, insulation failures, or short circuits.

Thermography detects thermal anomalies via infrared (IR) radiation emitted by components, governed by Planck’s law. Important metrics for this method include the emissivity (ε) and thermal sensitivity. Two types of thermography exist, namely: 

• Active Thermography: Uses external heat sources to induce thermal gradients to reveal sub-surface defects.
• Passive Thermography: Monitors operational components under load, identifying hotspots without external stimulation.

For thermal management systems in batteries, Thermography can be used for:

• Lock-In Thermography (LIT): Modulated heat input (0.01–1 Hz) isolates delamination signals in multi-layer Li-ion cells. This can in turn, help detect pouch cell delamination. Infratec’s VarioCAM HDx with 1280 × 1024 IR resolution can detect 50 µm air gaps between anode-separator layers.

Image Credit: Blink Tech

• Pulsed Thermography: Short-duration flash heating for approximately 5 to 10ms using sources like Xenon arc lamps can help identify coolant channel blockages in BMW i4 battery packs.
• Differential Absolute Contrast (DAC): Algorithms can compare thermal profiles to CAD models, flagging Thermal Interface Material (TIM) voids that are greater than 3% area.

EV systems require upkeep and optimal performance by power electronics and inverters. Thermography can help inspect them in the following ways:

• IGBT/SiC MOSFET Modules:

1. Transient Thermal Analysis: Cameras of around 30 Hz frame rate are used to capture junction temperature spikes greater than 150°C during switching helping predict solder fatigue.
2. Phase-resolved passive Thermography: This correlates thermal cycles with Pulse Width Modulation (PWM) frequencies to detect cracked wire bonds.

• DC-Link Capacitors: IR cameras can identify equivalent series resistance (ESR) anomalies in capacitors.

High-voltage wiring & connectors can also be inspected using Thermography.

• Contact Resistance Hotspots:

1. MWIR Cameras: MWIR Cameras of around 3 to 5 µm wavelength can detect loose lugs in 400 V busbars like in the Porsche Taycan with 0.5°C accuracy under 100 A load.
2. AI-Driven Anomaly Detection: Siemens’ Simcenter Testlab uses convolutional neural networks (CNNs) to classify thermal patterns.

Generative Adversarial Networks (GANs) synthesise synthetic defect data like simulated short circuits to train AI models which reduces false positives.

Thermography can be used for Quantum IR Sensors as well:

• Quantum Well Infrared Photodetectors (QWIPs): These offer 2–3x higher sensitivity than microbolometers, critical for inspecting thin-film insulators in 800 V architectures.

Structural Integrity & Multi-Material Challenges can be mitigated using Thermography in the following ways:

• Composite Motor Housings: Step Heating Thermography detects fibre misalignment in carbon-fibre motor housings by analysing anisotropic heat diffusion.
• Adhesive Bond Inspection: Pulsed Phase Thermography (PPT) provides Fourier transforms of thermal decay curves to quantify disbond depth in structural adhesives (e.g., 3M Scotch-Weld) with ±0.2 mm precision.

 Industry 4.0 applications also include using Digital Twins to import IR data to simulate thermal runaway propagation, and optimising cooling plate designs for batteries. IoT-Enabled Predictive Maintenance helps use IR data from production-line cameras to predict inverter failures, reducing warranty costs by 15%.

NDT technologies like AI-driven thermography, quantum IR sensors, and hyperspectral fusion will dominate automotive NDT trends, ensuring structural integrity and compliance with ISO 21498. These address EV-specific risks like thermal runaway, contact degradation, and adhesive failure while aligning with Industry 4.0 frameworks for scalable, data-driven manufacturing.

3. X-ray and CT Scanning 

The structural integrity of electric vehicles is imperative, especially for critical parts like the battery housing and chassis. X-ray and Computed Tomography Scanning technologies provide information on cracks, corrosion, or improper welds. 

CT scanning produces 3D images of complex internal structures, allowing manufacturers to inspect even minuscule and inaccessible parts. Both X-ray and computed tomography (CT) internal structures use ionising radiation. Important factors in these methodologies include:

• Spatial Resolution: Microfocus X-rays which operate around 160–225 kV can achieve less than 50 µm resolution, which can help conduct structural inspections of thin welds and composites in EVs.
• Voxel Size: Industrial CT systems can resolve 1–5 µm voxels, which refer to a three-dimensional pixel, enabling the 3D reconstruction of prismatic cell electrode alignment.
• Contrast Sensitivity: Dual-energy CT (DECT) can differentiate between materials like Aluminium or CFRP by micro variations.

These techniques can be incorporated into battery pack and cell inspection:

• Cell-to-Pack (CTP) Welds:

1. Microfocus X-ray can inspect laser welds for porosity less than 100 µm voids.
2. AI Porosity Analysis classifies defects using ISO 6520-2 criteria, achieving fewer false alarms.

• Electrode Alignment:

1. In-Line CT: Nikon’s XT H 225 ST system can scan prismatic cells at 15 sec/unit which maps anode-cathode misalignment greater than 20 µm.
2. Laminography: Laminography, which is limited-angle CT can inspect pouch cell edges for delamination without full 360° rotation.

Image Credit: Nikon

Power Electronics and HV Components can also benefit from X-ray and CT scanning in ways that include:

• IGBT Solder Joints:

1. Digital Radiography (DR): Flat-panel detectors used in DR can image voids in solder layers.
2. Automated Defect Recognition (ADR): ADR software uses blob analysis to flag cracked joints in Silicon Carbide (SiC) modules.

• DC-DC Converters: Phase-contrast CT enhances crack visibility in aluminium casings using edge-enhancement algorithms.

In lightweight structural components, the uses of X-ray and CT scanning include:

• CFRP Chassis:

1. High-energy CT can penetrate 30 mm-thick CFRP to detect impact-induced fibre fractures with a resolution of 50 µm.
2. Porosity Quantification: Porosity/Inclusion modules can calculate void content of less than 1% tolerance in resin transfer moulded parts.

• Aluminium Die Castings: Real-time radiography (RTR) using 100 fps imaging monitor pore formation during the gigacasting of Tesla Model Y rear underbodies.

The Generative AI Defect Synthesis technique creates synthetic CT datasets to train deep learning models. Automated Metrology can also integrate CT data with CAD models, verifying battery enclosure tolerances. X-ray and CT technology can also overcome Multi-Material Challenges:

• Battery Housing Welds: Here, Neutron Radiography complements X-ray to inspect hydrogen-rich thermal interface materials (TIMs).
• Hybrid Material Joints: Multi-Sensor Fusion process combines CT to detect cracks and eddy current for conductivity to assess steel-CFRP rivets.

Digital Twins can be used to import CT-derived pore distributions to simulate the crash performance of battery trays. Techniques like Blockchain Traceability use CT data management to log inspection results in blockchain, complying with EU Battery Passport mandates.

3. Eddy Current Testing 

Electric motors are at the heart of EVs, and Eddy Current Testing (ECT) is a proven NDT method used to inspect its internal components comprising coils and windings. By detecting surface and near-surface flaws, ECT helps prevent motor failures that could lead to vehicle malfunctions or accidents.

Here, electromagnetic induction is used to detect surface/subsurface flaws in conductive components. Key parameters while assessing the test subject include the Frequency Range, Impedance Plane Analysis and Skin Depth represented by the symbol δ. Eddy Current Testing is a viable testing solution for Electric Motors, and its applications include:

• Hairpin Stator Inspection: Differential Probes with 1–10 kHz frequencies can detect incomplete laser welds in stators, ensuring a resistance deviation of less than 10%.

1. Array Probes: Evident’s Olympus Nortec 500S arrays scan 8 hairpins simultaneously at 5 mm/sec, mapping weld uniformity in Tesla Model 3 drive units.
2. Rotor Lamination Cracks: Absolute Probes can identify interlamination short circuits caused by electromagnetic stress in rotors.

• Battery components - Tab Welds & Foils:

1. High-Frequency ECT: These detect microcracks less than 20 µm in aluminium/copper tabs using phase-sensitive detection (PSD) to suppress lift-off noise.
2. Hybrid ECT/UT Probes: A combination of ECT and ultrasound to inspect ultrasonic-welded tabs which can spot voids less than 0.1 mm².

• Power Electronics

1. SiC MOSFET Solder Joints: Low-frequency ECT can identify Kirkendall voids by monitoring conductivity shifts (±0.5% MS/m).

• Portable & Robotic Systems

1. Handheld ECT Scanners: Scanners with Wi-Fi connectivity perform on-site stator inspections in motors, streaming data to PLM software.
2. Robotic Arm Integration: Robots equipped with Pulsed ECT technology can scan motor housings at around 15 m/min, achieving 100% coverage in 45 sec/unit.

• AI/ML-Driven Analytics

1. Real-Time Defect Classification: Products like NVIDIA’s Jetson AGX Orin process impedance signals using CNNs trained on flaw signatures like cracks or porosity, reducing false positives by 30%.
2. Digital Twin Correlation: Digital Twins import ECT data to simulate eddy current distributions in faulty windings, predicting insulation breakdown risks.

Image Credit: Developer.nvidia

AI-Powered Inspection and Data Analysis

AI and machine learning analyse vast amounts of data collected through various NDT methods to enhance the accuracy and speed of inspections. AI/ML transforms non-destructive testing technologies by automating defect recognition and enhancing NDT EV safety.

Convolutional neural networks (CNNs) analyse ultrasonic C-scans, thermograms, and CT reconstructions. Whereas recurrent neural networks (RNNs) correlate historical UT/ECT data with failure modes like rotor lamination cracks. 

GANs, on the other hand, synthesise realistic defects to train models on rare flaws, reducing false negatives in battery weld inspections.

AI-driven automotive NDT trends in 2025 cut inspection times by 50% while assuring structural integrity that aligns with ISO 20669 and EU Battery Passport mandates.

The Integration of IoT in NDT for EVs

The Internet of Things (IoT) is gradually but steadily being adopted into NDT Processes within the EV industry. IoT-enabled sensors can be embedded in EV components for continuous monitoring. 

These sensors can feed data into centralised systems for real-time monitoring and analysis which helps with predictive maintenance. The fusion of IoT with non-destructive testing technologies enables real-time structural health monitoring, critical for NDT EV safety and automotive testing in 2025. 1. The technologies comprise the following:

1. IoT-Enabled Sensor Networks

I. Embedded Piezoelectric Sensors:

• Battery Health Monitoring: Specialised sensors detect micro-deformations in Li-ion pouch cells, correlating strain with dendrite growth.
• Motor Condition Monitoring: Analog Devices like accelerometers of ±40 g range track rotor imbalance in axels, identifying bearing wear via FFT spectral shifts with harmonics greater than 2 kHz.

II. Wireless Data Transmission

LoRaWAN Protocols help transfer UT/ECT data from sensors to IoT solutions like the Siemens MindSphere, enabling a latency of less than 5 ms for cloud-based analytics.

2. Cloud-Based Analytics & Predictive Maintenance

I. AI-Driven Platforms:

• AWS IoT Greengrass: Processes thermographic data from IR cameras to predict TIM degradation in EV battery packs using LSTM networks.
• Digital Twins: Digital Twins can integrate IoT strain data with FEM models, simulating battery enclosure fatigue under 10G crash loads per ISO 20669.

II. Emerging NDT-IoT Synergies

• Laser Ultrasonics: This non-contact UT is combined with lasers to generate Lamb waves, mapping electrode delamination in cells.
• Terahertz Imaging: This is conducted using products like the TeraSense THz cameras to detect ceramic separator cracks in cells of products like CATL Qilin with 30 µm resolution and IoT-enabled edge processing for less than 1 sec/image.
• In-Line Robotic CT: Robotic systems perform 100% inspection of cylindrical components using their 6-axis automation and SIRT algorithms to reconstruct voxels.

These technologies evolve and fade away with time, but the resilience, curiosity, and immense hard work of the NDT Industry and the people associated with it have led to numerous innovations and a yearning for more. The goal is to keep learning, evolving, and competing to grow the industry as a whole so we can reap the benefits and build a world that’s safe, sustainable, and smart. 

Key Takeaways

• Ultrasonic testing and thermographic inspection ensure the safety and efficiency of EV batteries by detecting internal flaws and electrical malfunctions.
• X-ray and CT scanning is gaining popularity in EV chassis and battery housing inspection to ensure they meet safety standards.
• Eddy current testing is critical for inspecting the internal components of electric motors to prevent failures and ensure reliability.
• AI and IoT integration into NDT systems will enable faster, more accurate inspections, predictive maintenance, and real-time data analysis for improved EV safety.

References

Research Gate. (n.d.). Production of Tesla Model 3 is revealed Tesla model Y by mega-giga casting technology (Duckers, 2022). Retrieved from Research Gate

S&P Global. (n.d.). S&P Global. Retrieved from S&P Global