Manufacturing quality control is essential for products that are defect free, safe, and meet customer expectations. As products become more complex, connected, and personalized, manual inspection processes are becoming time-consuming and error prone. The downstream impacts of quality issues and defects are many and include: waste, scrap, failure analysis, and rework – all which negatively impact factory productivity and profitability.
In addition, organizations face additional costs for repairs, warranty claims, returns as well as customer complaints and possible damage to reputation. According to the American Society of Quality, the cost of quality in manual product line inspections is 15% to 20% of annual sales , or 10 to 15% of operations cost and up to $90 billion in savings if the top 100 manufacturers cut scrap and rework by 50%.
Reduce defects with automated manufacturing quality analysis
The future of manufacturing is connected, data-driven, autonomous, and secure. Thanks to the IoT, cameras, sensor data and analytics, plant operational technologies are converging with IT. Digitized factory operations and AI-enabled processes provide greater predictability and agility to respond when things go wrong. Artificial intelligence (AI)-based video analytics used in manufacturing harnesses the power of AI, computer vision technologies and edge computing to help improve product quality. An AI-enabled, automated manufacturing inspection process may:
- Improve quality assurance with AI-based inspection processes that are fast and precise
- Help reduce scrap, waste, and re-work by catching defects sooner (in near real-time)
- Result in better energy efficiency due to fewer production re-runs
- Improve overall equipment effectiveness (OEE), factory uptime and worker productivity
HPE integrates AI-based video analytics in its European manufacturing facility
Relimetrics, an HPE OEM and NVIDIA Metropolis partner, enables quality automation and smart manufacturing. HPE’s Data, Analytics and IoT practice partnered with Relimetrics to capture and analyze data on the shop floor assembly line in real time to automate the inspection process. With over 1,000 possible configurations to validate for each server coming down the line, manual inspection processes could no longer keep pace. The solution automates HPE’s production line’s quality audits with video analytics and computer vision at the edge. According to Carolyn Cairns, HPE Global Industry Marketing Lead for Manufacturing, “Real-time insights needed for smarter factory operations require bringing compute to the edge closest to where data is generated. Our edge-to-cloud architecture saves time and prevents latency problems.” Results of implementing the AI-based Relimetrics solution at the HPE manufacturing facility include:
- Reduced out of box quality issues by 25%
- Improved inspection speed by 96 seconds per server
- Edge computing solution processes images 10X faster than in the public cloud
- Edge to cloud solution decreased training time of deep learning models from three weeks to two
HPE and NVIDIA enhance modern manufacturing processes
HPE and NVIDIA® are trusted partners for integrating AI into manufacturing processes. HPE and NVIDIA solutions enhance quality control (QC) using the power of AI, computer vision technologies and edge computing. Scott Wieder, Senior Alliance Marketing Manager for HPE states, “By extending real time insights from edge to cloud, manufacturers are transforming quality assurance by fully digitizing their quality audit cycles. The full-stack video and data analytics software applies video analytics and object detection technology to inspect the configurations and properties of product components, improving overall inspection accuracy.”
Getting started with AI video analytics
HPE – NVIDIA solutions deliver powerful AI-platform based video analytics applications for smart factories. The solution offers a scalable AI platform built on HPE systems that are NVIDIA-Certified, enabling GPU-accelerated applications. The platform combines compute, storage, interconnects, software, and services for an end-to-end solution. The AI platform features turnkey enterprise-grade NVIDIA Metropolis application frameworks and toolkits including a software stack which offers pretrained models, optimization tools, deployment SDKs, and CUDA-X libraries. The Metropolis extensive developer ecosystem lets developers simplify the development and scaling of AI-enabled video analytics applications.
Manufacturers are using the power of AI and video analytics to enable better quality control and traceability of quality issues, bringing them one step closer to achieving zero defects and reducing the downstream impacts of poor product quality. Together, HPE, NVIDIA and Relimetrics enable customers to adapt to high production variability and velocity with faster, more accurate, automated inspection processes.
Call to Action
Read the case study: Making zero defects a reality