Showing Off: Visual Insights
A demo designed to support end to end cognitive manufacturing
(demo at bottom)
It’s my pleasure to introduce you to our demo of the Visual Insights Solution built collaboratively by Watson IoT and the team at IBM’s Dallas Global Solution Center (more on the GSC at the end of post). Before we unveil it, let me tell you a little about it.
Many manufacturers have taken to instrumenting their plants with IoT. With sensors everywhere, they want to find the fast path to KPI improvements and delivery of a more productive plant floor. When IBM’s Institute of Business Value (IBV) asked electronics manufacturers which performance indicators were most important for their plants, COOs and manufacturing leaders indicated that throughput, uptime and defect reduction, in nearly equal measures (see figure 1).
That’s why this Visual Insights demo showcases how the layered value of IoT, Artificial Intelligence (AI) and S/4 HANA meet those specific needs. It’s about helping industrial companies reduce cost and downtime, improve quality and streamline operations by analyzing machines, predicting outages, handling equipment repairs and automate equipment maintenance.
By combining a very broad range of IBM offerings including AI, and IoT with Visual Insights (https://www.ibm.com/us-en/marketplace/visual-inspection-for-quality) and Acoustic Insights, we go beyond basic Predictive Maintenance when a piece may be defective, or a part needs replacing. We develop deeper understanding of non-obvious or interconnected issues. In addition, we can use Watson Equipment Advisor, Watson Personal Assistant, and Business Process Analytics to enable better interactions and more consistent outcomes. We can create a holistic view of a manufacturing process (see figure 2).
Resolve quality problems faster and more accurately. The as-is process in manufacturing is heavily reliant on manual inspection for detecting defects in every component of a product. This solution allows you to resolve quality problems faster and more accurately by using deep learning for image analysis that can be trained centrally and execute on low cost hardware at the inspection station. Automating this requires a Visual Inspection that can filter good parts with high confidence and speed, learn effortlessly, and scale across hundreds of defect types. It performs a rich defect analysis and image library to provide significantly more consistent and accurate inspection compared to manual methods via cognitive classification of defects, including confidence thresholds and is applicable across multiple industries impacted by manufacturing quality.
Solution flow. This Visual Insights demo is powered by IBM Watson and fully integrated with S/4 HANA. The Visual Insights system monitors production without human intervention, improving manufacturing yield. Exception incidences are put aside for manual inspection. The inspection supervisor is notified to verify defects and pays close attention to those with low confidence levels. Input is sent to S/4 HANA in real-time for further analytics. This process surfaces dark data and allows you to work with new patterns of data and automates the equipment maintenance and quality management processes, delivering results instantly. Data is analyzed using predictive analytics, in which future trends are predicted for relevant visual inspection KPIs. Critical events are handled automatically by triggering a Quality Management (QM) notification in the backend S/4 HANA system, which initiates actions such as an equipment maintenance order, schedule a recall campaign, or raise a quality claim to a supplier. After completion of the job and the elimination of the defect cause, the QM record is closed (see figure 3).
The Demo
Connect with the author, Charla Stracener . For more information about available demos and working with the Dallas GSC, please contact Charla.
To get the Cognitive Manufacturing Study, please click here.
For more information about IBM Visual Insights, please click here.
For more information about IBM Watson IoT, please click here.
For more intelligence on IIOT Security, please see the new IBV Expert Insight: The Internet of Threats.
