In this episode of Visions: A Machine Vision and Automation Solutions Podcast, host Jim Tatum explores GM's BVSeal — an AI-driven, CO₂-based gas imaging system that detects and pinpoints battery pack leaks in real time on the production line. Learn how collaborative robots, edge computing, and specialized cameras combine to reduce scrap, speed decision-making, and improve manufacturing quality.

Visions: A Machine Vision and Automation Solutions Podcast, is the podcast for engineers, designers, integrators, and end users who want to keep an informed eye on the imaging and machine vision industry. Every Tuesday we will explore the latest in imaging trends, developments and solutions. Here you will find interesting, useful insights and observations from expert interviews, solo episodes, even the occasional panel discussion, all of which aim to expand your knowledge on imaging and machine vision.

Related: Embedded AI and Smart Cameras: The Next Decade of Machine Vision

Transcript

Well, hello and welcome to Visions, A Machine Vision and Automation Solutions Podcast. I'm your host, Jim Tatum, senior editor of Vision Systems design and visions is an endeavor business media production from your friends at Vision Systems Design. Here you'll find the latest on everything from end user machine vision solutions to trends, developments, and perspectives on all things machine vision and imaging. Whether you've been working in the industry for a while or you're just starting to take a closer look at it, this podcast is designed to grow your knowledge and bring greater focus to your understanding of the imaging and machine vision industry. And now on to our show. 

Well, hi everyone, and welcome to visions. I'm Jim Tatum, and today we're going to look at some very interesting work being done in the realm of electric vehicles. To be a little more specific, we're going to get into an important and somewhat expensive component of electric vehicles. That is the batteries. You know, when we talk about electric vehicle batteries, most people focus on range, charging, speed or performance. But behind the scenes, manufacturers are working diligently with something much more fundamental, keeping battery packs sealed and leak free. At General Motors, that challenge led to the development of an in-house automated inspection system called B.V. seal, a technology that can identify and pinpoint battery pack leaks in real time, right on the production line. To understand why that's important, let's start with the battery itself. Reducing weight is one of the keys to improving EV range and efficiency. To make battery trays and packs lighter, GM uses advanced high strength steels, new aluminum alloys and resistance spot welding during assembly. But there's a tradeoff. The rapid heating and cooling involved in resistance. Spot welding can sometimes create microscopic cracks that become leak paths. Traditionally, those leaks were found only at the end of the manufacturing process. The problem was an end of line test could tell engineers that a leak existed, but it couldn't say exactly where it was or how serious it might be. By the time the defect was discovered, the battery tray or pack was already fully assembled, often leaving scrapping the entire item as the only option left. According to Blair Carlson, Future Factory Program lead and senior technical fellow at GM Research and Development, the project began with a simple manufacturing question that is could cracks and resistance spot welds on high strength steel create leak paths? The engineering team explored several possible solutions. They looked at soap bubble dunk tanks, for example. This is one of the oldest and most tried and true leak detection methods out there. It's used often in automotive manufacturing. They looked at robotic helium sniffers, which is a much more sensitive and quite a bit more expensive methodology. And they looked at various acoustic based leak detection methods. All of these methods are effective detection methods, but none of them could reliably identify the precise location of a leak. In fact, even advanced CT scans of leaking welds weren't providing definitive answers. The breakthrough came through when the team shifted its thinking. Instead of trying to image the crack itself, they focused instead on imaging the escaping gas. The technical discovery was that weld cracks and defects are three dimensional and not practical to image in a high-volume production environment. So the breakthrough came from combining specialized equipment that could image trace gas with an algorithm capable of distinguishing gas plumes from the surrounding atmosphere. That insight led GM to a carbon dioxide imaging approach. The company selected CO2 because it's relatively inexpensive, widely available, and capable of revealing even very small leaks. While helium is commonly used in leak detection, Carlson said its higher cost and inability to provide precise leak point identification made CO2 the better choice for BV Seal. So how does the system actually work? Well, a battery pack arrives at an inspection station on the line via an autonomous mobile robot. On either side of that station are Fanuc CR collaborative robots, each equipped with a specialized gas imaging camera. The system also includes an internally developed gas excitation unit and Advantech GPU based edge computer and AI software developed by GM. One of the advantages of the robot-based approach is flexibility. Because collaborative robots have built in safety systems, they can operate without the fixed fencing required by traditional industrial robots. That makes it easier to move large battery packs into and out of the inspection station. As the cameras observe the pack, a sensor activates and highlights any escaping trace gas. The gas imaging cameras capture the resulting plume, and those images are sent to the edge computer, where an AI model analyzes the flow pattern within seconds. The software tracks the gas movement back to its source and pinpoints the exact location of the leak. At that point, a human operator can examine the defect and determine whether it can be repaired or whether the battery pack should be scrapped. Notably, the system doesn't require specialized production lighting. It operates using the standard overhead lighting already present in a factory. While the inspection itself happens at the edge for real time decision making, the data is also stored in the cloud, enabling long term analytics, reporting and process improvement efforts. So did it work? Well, according to Carlson, the answer is yes. By detecting defects before battery packs are fully completed, B.V. seal is reducing rebuilds, tear downs, and scrap rates that translate directly into savings on labor materials, production time and overall manufacturing costs. Perhaps more importantly, GM designed B.V. seal is more than a single purpose solution. The company sees the technology as a platform that could be applied anywhere. Seal integrity matters. Future applications could easily include engine assembly, powertrain, manufacturing, and other production environments where even small leaks can create major quality issues. So what began as a question about microscopic weld cracks has evolved into an AI powered inspection platform that helps manufacturers identify defects in real time, make faster decisions, and reduce waste before products ever leave the line. Pretty ingenious, we'd say. And that's yet another glimpse into how machine vision, AI, and robotics are shaping quality inspection on the factory floor.

Well, that's a wrap for this episode of visions, produced by Endeavor Business Media, a division of endeavor B2B. Thanks very much for tuning in. If you enjoyed today's show, be sure to subscribe to the podcast and share this episode with a colleague who would find it helpful. Until our next episode, you can find us at vision dash systems dot com or on LinkedIn, Facebook or X for more insights, updates, and breaking news to keep you in the know. Thanks for tuning in. Until next time, stay focused on your visions.