New Wireless Charging Safety Breakthrough Detects Metal Debris with Precision
As electric vehicles (EVs) continue to gain traction across global markets, one of the most anticipated technologies—wireless charging—is inching closer to widespread adoption. While the convenience of simply parking over a charging pad appeals to consumers and automakers alike, a persistent safety concern has slowed its rollout: the risk posed by metal foreign objects inadvertently left on or near the charging surface. 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 study, titled Metal object detection for wireless charging of electric vehicles using planar coil array, presents a cost-effective, high-precision solution for identifying metallic debris such as coins, screws, keys, and aluminum cans that might enter the charging zone. These seemingly harmless items can become hazardous when exposed to the high-frequency electromagnetic fields generated during wireless power transfer. Under such conditions, eddy currents are induced within the metal, causing rapid heating that not only reduces charging efficiency but also poses fire and burn risks. In extreme cases, unattended metal objects can overheat to dangerous temperatures, threatening vehicle components and nearby individuals.
Current detection methods have limitations. Some rely on monitoring changes in system parameters like impedance or resonant frequency, but these are often influenced by the presence of the vehicle itself, making it difficult to distinguish between normal load variations and actual foreign objects. Other approaches use auxiliary detection coils, magnetoresistive sensors, infrared imaging, or even machine vision. While each offers certain advantages, they come with trade-offs—ranging from high cost and system complexity to environmental sensitivity and delayed response times. Infrared systems, for instance, require the metal to heat up before detection is possible, meaning the system may only react after a potential hazard has already developed.
The new method developed by Sun Dong, Gao Zichen, Han Xiaojuan, and Zhang Wenbiao takes a fundamentally different approach. Instead of relying on indirect indicators or thermal signatures, their system directly images the electrical conductivity distribution across the charging surface using a technique adapted from medical and industrial imaging: electromagnetic tomography (EMT). At the heart of the innovation is a 4×4 planar coil array positioned directly above the wireless charging transmitter. Each coil in the array serves dual roles—as both an excitation source and a sensing element—enabling the system to map the electromagnetic field distortions caused by conductive materials in real time.
The principle is elegant in its simplicity. When a metal object is present, it disturbs the magnetic field generated by the charging system. This disturbance alters the mutual inductance between the coils in the array. By systematically exciting each coil and measuring the response in all others, the system collects a comprehensive dataset of electromagnetic interactions. This data is then processed using an advanced image reconstruction algorithm known as the Landweber iteration method, which translates the raw inductance measurements into a two-dimensional conductivity map of the charging area.
What sets this method apart is its ability to not only detect the presence of metal but also to determine its exact location, size, and approximate shape. Unlike binary detection systems that merely signal “metal present” or “no metal,” this imaging-based approach provides actionable visual information. For example, the system can differentiate between a single coin and a cluster of metallic debris, or distinguish a small screw from a crumpled soda can. This level of detail is crucial for automated charging systems that may need to halt the charging process only when a genuine threat is confirmed, avoiding unnecessary interruptions.
To validate their approach, the research team conducted both finite element simulations and physical experiments. In simulation, they modeled various common metal objects—single and multiple coins, screws, and keys—placed at different positions over the coil array. The reconstructed images consistently revealed the presence and general shape of the objects, with quantitative evaluation using Intersection over Union (IoU) metrics showing an average score of 0.56. IoU is a widely accepted measure of object detection accuracy, comparing the overlap between the detected region and the true object area. A score above 0.5 is generally considered good, indicating substantial spatial agreement.
More importantly, the system demonstrated high positional accuracy. The average location error (LE), defined as the distance between the calculated center of the detected object and its actual position, was just 2.52 millimeters in simulations. This precision is more than sufficient for practical applications, where even a rough estimate within a few centimeters can trigger appropriate safety protocols.
Encouraged by the simulation results, the team moved to real-world testing. They constructed a physical prototype using copper enameled wire coils, each with an outer diameter of 24 mm, inner diameter of 20 mm, and 500 turns, mounted on a custom plastic substrate. Using a precision LCR meter, they applied a 100 kHz sinusoidal excitation signal and measured mutual inductance across all coil pairs. The experimental setup closely mirrored the simulation conditions, allowing for direct comparison.
In the lab, they tested a range of real-world metallic objects: a standard steel-core nickel-plated coin, a threaded alloy screw, a piece of crumpled copper foil, and a section of an aluminum beverage can. The results were striking. The reconstructed images clearly identified each object, with the coin—due to its simple, symmetrical shape—achieving the highest detection accuracy. The system successfully detected both a single coin placed at the center and two coins positioned off-center, demonstrating its ability to handle multiple objects.
Even complex, irregularly shaped items like crumpled copper foil and crushed aluminum cans were detected with confidence. While the reconstructed shapes were not perfect replicas—expected given the limited spatial resolution of a 4×4 coil array—the system reliably captured the general location and extent of the objects. The average IoU in experiments reached 0.61, slightly higher than in simulations, and the average location error dropped to 2.33 mm, indicating excellent repeatability and real-world performance.
One of the most compelling aspects of this technology is its scalability and adaptability. The 4×4 array used in the study represents a balance between resolution, cost, and complexity. However, the underlying methodology can be extended to larger arrays with more coils, potentially improving image fidelity for more intricate detection tasks. Conversely, for lower-cost applications, a smaller array could suffice for basic presence detection.
From a system integration standpoint, the solution is remarkably straightforward. The coil array can be embedded in the charging pad during manufacturing, adding minimal thickness or cost. The signal acquisition and processing can be handled by a microcontroller and a small embedded computer, making it feasible for integration into existing EV charging infrastructure. Because the system operates at the same frequency as the wireless charger—100 kHz in this case—it can share components such as oscillators and amplifiers, further reducing hardware overhead.
Safety is not the only benefit. By accurately identifying and localizing metal debris, the system enables smarter charging management. For instance, if a small coin is detected far from the primary coupling zone between the transmitter and receiver coils, the system might allow charging to proceed at a reduced power level rather than shutting down entirely. This flexibility enhances user experience while maintaining safety margins.
Moreover, the non-contact nature of the detection method makes it ideal for outdoor or harsh environments. Unlike camera-based systems that can be blinded by rain, snow, or dirt, or infrared sensors that struggle with ambient temperature fluctuations, the electromagnetic approach is largely immune to such conditions. It works equally well in bright sunlight, heavy rain, or freezing temperatures, ensuring consistent performance year-round.
The implications extend beyond consumer EVs. Commercial fleets, autonomous vehicles, and robotic systems that rely on automated wireless charging could benefit immensely from this level of situational awareness. In industrial settings, where metal shavings or tools might accidentally fall onto charging surfaces, the ability to detect and respond to foreign objects in real time could prevent costly equipment damage and downtime.
From a regulatory perspective, this technology aligns with emerging safety standards for wireless power transfer. Organizations such as SAE International and the International Electrotechnical Commission (IEC) have begun to define test procedures for foreign object detection (FOD) in EV charging systems. Standards like SAE J2954 require testing with various metal objects, including coins and screws—exactly the types of items evaluated in this study. By providing a method that not only detects but visualizes these objects, the planar coil array approach could set a new benchmark for compliance and safety certification.
The research also opens doors for future enhancements. One possibility is combining EMT with other sensing modalities—such as temperature monitoring or magnetic field mapping—to create a multi-layered safety system. Another avenue is the integration of machine learning to improve image reconstruction and object classification. With sufficient training data, a neural network could learn to recognize specific object types (e.g., a key versus a coin) and even estimate their mass or composition based on conductivity patterns.
Another promising direction is dynamic scanning. While the current system captures a snapshot of the charging surface, future versions could perform continuous monitoring during the entire charging cycle. This would allow the system to detect objects that are introduced after charging has started—such as a driver accidentally dropping a coin while exiting the vehicle—or to monitor for changes in object temperature over time.
The economic case for this technology is strong. While adding a coil array and processing unit increases the bill of materials, the cost is likely to be modest compared to the potential liabilities of undetected metal objects. Automakers and charging infrastructure providers face significant reputational and financial risks if a wireless charging incident leads to a fire or injury. A reliable, low-cost detection system could mitigate these risks while accelerating consumer adoption.
User trust remains a critical factor in the success of any new technology. Many drivers are still skeptical about the safety and efficiency of wireless charging. A transparent, reliable detection system that can visibly confirm the absence of hazards—perhaps through a dashboard alert or mobile app notification—could go a long way toward building confidence. Imagine a scenario where, upon parking, the vehicle displays a real-time image of the charging surface with a green checkmark indicating no metal debris. Such feedback would reassure users and reinforce the perception of safety.
The work also highlights the growing importance of interdisciplinary research in advancing EV technology. By borrowing concepts from medical imaging and adapting them to automotive applications, the team demonstrates how innovation often occurs at the intersection of fields. Electromagnetic tomography, originally developed for industrial process monitoring and biomedical diagnostics, has found a new and impactful application in sustainable transportation.
As wireless charging infrastructure expands—from private garages to public parking lots and even dynamic charging roads—safety systems like this will become increasingly essential. The transition to electric mobility is not just about replacing internal combustion engines with batteries; it requires rethinking every aspect of how vehicles interact with energy systems. This includes ensuring that the methods we use to deliver power are as intelligent and safe as the vehicles themselves.
In conclusion, the planar coil array-based metal detection method represents a significant step forward in making wireless EV charging both safer and more practical. By transforming a passive charging surface into an active sensing platform, the technology adds a crucial layer of intelligence to the charging process. It addresses a real-world safety concern with a solution that is elegant, effective, and scalable. As automakers and infrastructure providers look to deploy wireless charging at scale, innovations like this will play a vital role in ensuring that the technology delivers on its promise of convenience without compromising on safety.
The findings were published in the journal Electrical Measurement & Instrumentation by Sun Dong, Gao Zichen, Han Xiaojuan, and Zhang Wenbiao from State Grid Tonghua Power Supply Company and North China Electric Power University. DOI: 10.19753/j.issn1001-1390.2024.10.025