Artificial intelligence in industrial welding produces near-real-time insights through virtually 100% sample sizes
In metal-to-metal welding applications, automating inspections to detect weld defects early and often can help avoid costly product recalls, excessive scrap, rework, and other expenses associated with poor quality. Companies have found it challenging to implement such automation.
IBM Smart Edge for Welding (SE4W) with AWS aggregates quality inspection capabilities based on artificial intelligence (AI), acoustic, and visual insights, into a novel solution that can tackle some of the most challenging pain points in any welding application. This provides customers with a virtual step change in quality performance. The solution facilitates inspection of virtually 100 percent of the weld sample during the welding process. Edge AI optimizes and automates welding inspections and gives weld operators near-real-time and actionable weld-quality insights for faster defect diagnosis and resolution. This reduces costs and improves efficiency.
Welding is the fusion of two compounds with heat. It’s a process that happens billions of times every day, and one which we all depend on. The chair you’re sitting in, for example, likely has dozens of welds. Your car has hundreds to thousands of welds. The electricity generated from hydroelectric dam travels hundreds of miles through transmission towers. These transmission towers were constructed with thousands of welds.
Unless something goes wrong, people rarely give any thought to welding. We tend to enjoy the many benefits of welding without being overtly aware of its presence in our world.
Manufacturers are the unsung heroes who ensure our safety. Work that requires close collaboration across different domains, including design, process engineering, technician support, quality control, and trusted networks of suppliers and equipment providers. The irony is that manufacturers tend to remain anonymous when they do their jobs well. However, when things go wrong as in cases of accidents, recalls, leaks, or even deaths, manufacturers are often the first to be scrutinized. On top of the reputational costs and risks of bad welds, the financial repercussions in the automotive industry alone can be up to $9.9 billion per year.
What are the specific challenges of welding inspection?
Take a moment to inspect the weld joint below. At first glance, can you determine whether this weld is good or bad?
Most likely you cannot. That’s all right, because almost nobody can tell from visual inspection. Just like an iceberg floating in the water, where only the clear white tip is visible and the danger lies invisible beneath the surface, many weld quality indicators are invisible to the human eye.
Figure 1 is a chart of the most common defects in arc welding. The color of the star next to each defect indicates level of visibility to subject matter experts.
Figure 1: Common arc-welding defects
Manufacturing processes use a combination of destructive and nondestructive quality-testing methods to determine whether a weld has a discontinuity or defect.
Destructive testing is costly and time consuming. It includes the mechanical disassembly of welds (such as by grinding) and chemical etching (such as by applying ethanol with citric acid) to measure fusion parameters. These methods are the most accurate form of quality evaluation. An advantage of this method is that it requires only a small number of samples, but once a defect is discovered, its remediation means discarding all welds that have taken place in the time between discovery and response.
Nondestructive testing is largely done by human visual inspection. It is occasionally augmented by ultrasound testing, which also relies on direct human involvement. Once a defect is discovered and remediated, each weld completed in the intervening time period must also be tested. These types of inspection are subjective, inconsistent, applicable to only a subset of defect types, expensive, and time-consuming.
The game changer
Equipment and sensor providers are trying to address this problem, and most manufacturers are attempting—with varying degrees of success, to bring advanced analytics and AI to the problem. Equipment providers focus on the data their components produce, while sensor providers focus on the information their sensors generate. Some of the challenges these approaches face are
- covering only a relatively small subset of failure modes;
- providing short-term accuracy while suffering from long-term model drift;
- not adapting to operational change;
- making use of only certain types of data;
- requiring large amounts of data.
Why is IBM Smart Edge for Welding with AWS better?
SE4W uses audio and visual capturing technology developed in collaboration with IBM Research. Using visual and audio recordings taken at the time of the weld, AI and machine learning models analyze the quality of the weld. If the quality doesn’t meet standards, alerts are sent and remediation action can be taken without any delay.
This solution substantially reduces the time between detection and remediation of defects and the number of defects on the manufacturing line. One of the results of this solution is overall cost reduction.
SE4W uses multimodality and IBM Research’s patented multimodal AI to provide accurate insights through a combination of the following features
- Visual Analytics
- IBM Maximo Visual Inspection (MVI) analyzes in-process welding videos in near-real-time with computer vision (with MVI Edge and MVI on AWS Cloud).
- Xiris Weld Cameras are purpose-built industrial optical cameras that provide unprecedented high resolution in-process videos of the weld pool, wire, workpiece, and more.
- Xiris Thermal Camera is a purpose-built industrial thermal camera that visualizes heating and cooling behavior of a weld as it is being produced.
- Acoustic Analytics
- IBM Acoustic Analytics is a proprietary, patented, purpose-built neural network that analyzes weld sounds.
- Xiris WeldMic is a purpose-built industrial microphone that listens to the arc sound in near-real-time—just like experienced weld technicians do.
- AWS Cloud and AWS for the Edge
- Industrial Edge Computing lets us seamlessly integrate into your manufacturing environment to generate near-real-time insights that are safe and secure, without any sensitive information ever leaving the plant.
- Cloud Computing, available as public, private or dedicated cloud deployment, facilitates scalability across production lines, plants, and even regions.
Seeing the defect is believing
The optical video clip below visualizes several components of a weld:
- size and shape of the weld pool and how it solidifies as it cools;
- behavior of the wire as it deposits filling material;
- spatter that is generated;
- turbulence in the shielding gas;
- holes forming from burns;
- annotations created by our AI models (in this case for burns) in near real time.
Figure 2: see Video of Common arc-welding defects
The infrared video clip below visualizes several additional components of a weld:
- thermal zones through color coding;
- uniformity of the trail;
- heat signatures and the size and purity of the weld pool;
- annotations created by our AI models (in this case for porosity) in near real time.
Figure 3: see Thermal video of welding
The image below is a translation of the welding sound into a sound wave and sound spectrum and identifies the following features:
- patterns of normal and abnormal behavior
- abnormalities classified by specific failure modes
Figure 4: Translation of welding sound into a sound wave and spectrum
By using a combination of optical, thermal, and acoustic insights during the weld inspection process; two key manufacturing personas can better determine whether a welding discontinuity may result in a defect that will cost time and money
- Weld technician—works on the shop floor and uses insights on weld performance in near-real-time to add, change, or optimize the process as needed. The solution can be integrated into any platform and device used on the shop floor, such as human-machine interface or mobile devices.
- Process engineer—wants to understand patterns and behavior across shifts, weeks, months, weld programs, and materials to improve the overall manufacturing process.
SE4W facilitates the following gains:
- Improved quality through inspection of 100% of welds.
- Reduction of time and optimization of setting up the weld program.
- Accelerated launch of new products or changes.
- Identification of trends as early warning signs of defects and other near-real-time insights.
- Reduction of time between identification and resolution of an issue.
- Cost reductions through reduced physical labor, human testing, material needed, and scrap material resulting from destructive testing, bad weld batches, and preventative remediation.
- Reduced warranty risk and recalls, because each weld is inspected and quality standards are consistently met.
A single factory has demonstrated potential savings of $18 million a year through these cost reduction benefits. Warranty costs and recalls, which cost the automotive industry alone an estimated $9.9 billion a year. These are due to bad welds and can be avoided or significantly reduced. Brand reputation is maintained through consistently high-quality and safe welds.
Smart Edge for Welding on AWS
IBM partnered with AWS to develop a solution to address the industry-wide manufacturing challenge of quickly identifying weld defects to facilitate fast remediation.
AWS has over 200 services that can be used to enhance, optimize, and further customize this solution. IBM’s AI models are trained in AWS and deployed to the edge for inferencing. All weld data is stored in the cloud in a low-cost storage environment for analysis and future model training. Amazon QuickSight, which powers data-driven organizations with unified business intelligence at hyperscale, can be used for process engineer dashboards and reporting. It facilitates an automated process of model deployment to edge endpoints.
The edge environment of this architecture runs on AWS IoT Greengrass, which lets businesses build intelligent Internet of Things devices faster. Data is ingested from shop floor sensors (such as cameras and microphones). It is preprocessed to remove excess noise from the audio data and blurred images from the video data. Then, model orchestration and inferencing is executed through a machine learning model using IBM Maximo Visual Inspection and IBM Acoustic Analytics; to identify the quality of the weld and determine if it meets the set standards. Postprocessing takes place from the moment of alert notification and reporting, and it includes transferring data to the cloud for further analysis, model training, compliance archiving, and other beneficial purposes.
Figure 5: SE4W—Reference Architecture
Figure 6: SE4W—Component Architecture with AWS Services
SE4W provides customers with an end-to-end, production-ready solution that generates bottom-line impact through the optimization of manufacturers’ welding processes. IBM offers the power of AI, from computer vision with IBM Maximo Visual Inspection to IBM Acoustic Analytics. The solution provides manufacturers with near-real-time weld defect insights for faster problem diagnosis and remediation through a weld-quality single pane of glass.
Welding technicians and process engineers can inspect up to virtually 100 percent of welds to determine the cause of welding defects in the earliest stages of the production process. This results in less repetitive defects and rework, along with reduced material waste and negative, high-risk exposure for a company’s brand.
Learn more about our IBM Smart Edge for Welding with AWS solution and other IBM Consulting services for AWS at the following locations.
The IBM and AWS partnership: https://www.ibm.com/aws
IBM Consulting services for AWS Cloud: https://www.ibm.com/consulting/aws
IBM on the AWS Partner Solutions Finder: https://partners.amazonaws.com/partners/001E000001IlLnmIAF/IBM
IBM Maximo Visual Inspection: https://www.ibm.com/products/maximo/visual-inspection
Special thanks to our contributors and collaborators, including Manoj Nair (IBM), Caio Padula (IBM), Wilson Xu (IBM), Ofir Shani (IBM), Nisha Sharma (IBM), Penny Chong (IBM), Tadanobu Inoue (IBM), and Ryan Keough (AWS).
IBM–AWS Partner Spotlight
IBM Consulting is an AWS Premier Tier Services Partner that helps customers who use AWS harness the power of innovation and drive their business transformation.
Tags: AWS Competency Partners, AWS MSP Partner Program, AWS Partner Guest Post, AWS Partner References, AWS Partner Solutions Architects (SA), AWS Partner Success Stories, AWS Premier Tier Services Partners, AWS Public Sector Partners, AWS Service Delivery Partners, AWS Solution Provider Partners, AWS Well-Architected Partners, Manufacturing, Internet of Things, Data Mesh, IBM, Managed Service Provider