New Wireless Charging Safety Breakthrough Detects Metal Debris with Precision
In the rapidly evolving landscape of electric vehicle (EV) technology, one persistent challenge has remained at the forefront of engineering innovation: ensuring safe and efficient wireless charging. While the convenience of charging without plugging in is a major selling point for consumers, the technology comes with a hidden risk—metal foreign objects inadvertently left on or near the charging surface can overheat due to induced eddy currents, posing safety hazards and reducing system efficiency. Now, a team of researchers from State Grid Tonghua Power Supply Company and North China Electric Power University has introduced a novel detection method that could significantly enhance the safety and reliability of EV wireless charging systems.
The breakthrough, detailed in a recent study published in Electrical Measurement & Instrumentation, presents a cost-effective, non-invasive solution leveraging a planar coil array and electromagnetic tomography (EMT) to detect, locate, and visualize metal debris with high accuracy. Led by Sun Dong, Gao Zichen, Han Xiaojuan, and Zhang Wenbiao, the research team has developed a system that not only identifies the presence of metallic contaminants such as coins, screws, and aluminum cans but also reconstructs their shape, position, and quantity—offering a comprehensive diagnostic capability previously unavailable in mainstream wireless charging setups.
As EV adoption accelerates globally, automakers and infrastructure providers are increasingly investing in wireless charging solutions to improve user experience and streamline urban mobility. Standards such as SAE J2954 and China’s GB/T38775.3-2020 have set rigorous safety benchmarks, including mandatory testing for metal foreign object detection (FOD). However, existing FOD methods have significant limitations. Some rely on monitoring changes in system impedance, which can be influenced by vehicle load variations and fail to pinpoint object location. Others employ additional sensor coils or magnetic field sensors, which increase system complexity and cost. Vision-based systems struggle in poor weather, while infrared thermal imaging requires waiting for objects to heat up—delaying detection and increasing risk.
The new approach addresses these shortcomings by integrating a 4×4 planar coil array directly above the transmitter coil of a magnetic resonant wireless charging system. Each coil in the array serves dual purposes: as an exciter and a detector. By sequentially energizing each coil and measuring mutual inductance with the others, the system collects a rich dataset reflecting distortions in the electromagnetic field caused by conductive materials. These measurements are then processed using electromagnetic tomography algorithms to generate a two-dimensional conductivity map of the charging zone.
Unlike conventional detection systems that merely flag the presence of metal, this method reconstructs an image of the disturbance, effectively “seeing” the foreign object. The core innovation lies in the application of EMT—a technique historically used in industrial process monitoring and biomedical imaging—to the domain of EV charging. By treating the charging surface as a tomographic imaging field, the researchers are able to visualize anomalies with spatial resolution, enabling precise localization and shape estimation.
The team utilized the Landweber iterative algorithm for image reconstruction, a method known for balancing computational efficiency with reconstruction accuracy. This algorithm progressively refines the conductivity distribution image by minimizing the difference between measured and predicted mutual inductance values. To enhance detection clarity, post-processing techniques such as Otsu thresholding were applied to segment the reconstructed images, isolating metal objects from background noise. From these segmented regions, the system calculates the centroid of each detected object, providing precise coordinates for its location.
To validate the system, the researchers conducted both finite element simulations and physical experiments. In simulation, various metallic objects—including a single coin, two coins, a screw, and a key—were virtually placed above the coil array. The reconstructed images showed clear representations of each object, with evaluation metrics such as Intersection over Union (IoU) and Location Error (LE) used to quantify accuracy. The average IoU across test cases was 0.56, indicating strong overlap between the reconstructed and actual object areas, while the average LE was 2.52 mm, demonstrating sub-centimeter precision in positioning.
In real-world experiments, the team constructed a physical prototype using copper enameled wire coils with an outer diameter of 24 mm, inner diameter of 20 mm, and 500 turns per coil. These were mounted on a custom plastic substrate to form the 4×4 array. Using a precision LCR meter, mutual inductance data was collected at a frequency of 100 kHz with a 1V sinusoidal excitation signal. Test objects included a nickel-plated steel coin, a zinc-alloy screw, crumpled copper foil, and a crushed aluminum can—all common items that could realistically be left on a charging pad.
The experimental results were compelling. The system successfully detected all tested objects, with an average IoU of 0.61 and an average location error of 2.33 mm. Circular objects like coins produced the most accurate reconstructions, with minimal distortion and high IoU values. Linear or irregularly shaped items such as screws and crumpled foil showed broader reconstructed profiles—particularly in width—due to the physical limitations of the coil spacing and field sensitivity. Nevertheless, their centroids were accurately determined, ensuring reliable positioning.
One notable finding was that the system could distinguish between single and multiple objects. In tests with two coins placed side by side, the reconstruction clearly showed two distinct high-conductivity regions, allowing the system to count and locate both items. This capability is crucial for real-world deployment, where multiple metallic objects may be present simultaneously.
The practical implications of this technology are significant. By enabling real-time, high-resolution imaging of the charging surface, the system allows for immediate intervention when foreign objects are detected. This could trigger an automatic shutdown of the charging process, alert the vehicle owner via a mobile app, or even activate a robotic cleaning mechanism in automated parking systems. Moreover, because the coil array is passive and requires only low-power electronics for data acquisition, it can be integrated into existing wireless charging pads with minimal modification and cost.
From a safety standpoint, the ability to detect objects before they begin to overheat is a major advantage over thermal-based methods. Infrared systems, while effective, require time for metal to heat up under electromagnetic exposure—during which energy is wasted and risk accumulates. In contrast, the EMT-based method detects the object at the moment it enters the field, enabling proactive rather than reactive safety measures.
The research also highlights the scalability of the approach. While the current prototype uses a 4×4 array, the methodology can be adapted to larger or denser coil configurations to cover bigger charging zones or achieve higher resolution. The use of a plastic substrate makes the array lightweight and compatible with outdoor installation, while its non-contact nature ensures durability and resistance to environmental wear.
Another advantage is the system’s robustness in adverse conditions. Unlike camera-based systems that can be blinded by rain, snow, or dirt, the electromagnetic sensing method operates independently of visual clarity. It is also unaffected by ambient temperature fluctuations, making it suitable for deployment in diverse climates—from freezing winters to humid summers.
The team emphasizes that their solution strikes an optimal balance between performance, cost, and complexity. Magnetic sensor arrays, while capable of high-resolution detection, require dozens of expensive sensors and complex calibration. The planar coil array, in contrast, uses off-the-shelf components and a single LCR measurement device, reducing both hardware and maintenance costs. The signal processing is handled by standard computing hardware, making the system accessible for mass production.
Integration with vehicle-to-infrastructure (V2I) communication protocols is another potential avenue. The detection data could be transmitted to the vehicle’s onboard system, enabling coordinated actions such as adjusting alignment, pausing charging, or notifying the driver. In fleet operations or autonomous valet parking scenarios, this information could be used to optimize parking precision and prevent charging interruptions.
The study also opens new possibilities for predictive maintenance. By continuously monitoring the charging surface, the system could log foreign object incidents, helping operators identify recurring contamination sources—such as loose change from drivers or debris from nearby construction. This data could inform design improvements, such as adding protective covers or installing debris traps around charging stations.
From a regulatory perspective, the method aligns well with international safety standards. The ability to detect small metallic objects like screws and coins meets the requirements set by SAE and GB/T standards, which specify testing with precisely these types of items. The quantitative metrics—IoU and LE—provide objective benchmarks for certification, moving beyond binary pass/fail tests to a more nuanced assessment of detection capability.
While the current implementation focuses on stationary wireless charging, the underlying principles could be extended to dynamic charging systems, where vehicles are charged while in motion. In such applications, rapid and accurate detection of road debris would be even more critical, given the higher speeds and reduced opportunity for manual inspection.
The researchers acknowledge that the system has limitations. The spatial resolution is constrained by the coil size and spacing, making it difficult to perfectly reconstruct very thin or elongated objects. Additionally, the current setup assumes a flat, uniform surface; uneven terrain or metallic structures beneath the charging pad could introduce interference. Future work will focus on improving resolution through advanced algorithms, optimizing coil geometry, and testing in more complex environments.
Nonetheless, the demonstrated performance marks a significant step forward in wireless charging safety. The ability to not only detect but also visualize and quantify foreign metal objects transforms FOD from a binary alert system into a smart diagnostic tool. This level of insight empowers both users and operators to make informed decisions, enhancing safety, efficiency, and trust in wireless charging technology.
As automakers like BMW, Hyundai, and Tesla move toward commercializing wireless charging for consumer vehicles, solutions like this will become increasingly vital. The technology not only protects vehicles and users but also safeguards the reputation of wireless charging as a reliable and safe alternative to plug-in methods. With public charging infrastructure expanding worldwide, ensuring the integrity of every charging session is paramount.
The work also reflects a broader trend in engineering: the cross-pollination of techniques from different domains. By applying industrial tomography methods to automotive systems, the researchers have demonstrated how innovation often arises at the intersection of disciplines. This interdisciplinary mindset—combining power electronics, sensor design, signal processing, and image analysis—is likely to drive future breakthroughs in smart mobility.
In conclusion, the planar coil array-based EMT system represents a promising advancement in EV wireless charging safety. It offers a practical, accurate, and scalable solution to a critical problem, combining scientific rigor with real-world applicability. As the automotive industry continues its transition to electrification and automation, technologies like this will play a foundational role in building a safer, more efficient, and more user-friendly transportation ecosystem.
Sun Dong, Gao Zichen, Han Xiaojuan, Zhang Wenbiao, North China Electric Power University, State Grid Tonghua Power Supply Company, Electrical Measurement & Instrumentation, DOI: 10.19753/j.issn1001-1390.2024.10.025