By Yasser Khan, CEO, One Tech, Inc.
Partnering sensor output with existing data sources to improve overall equipment effectiveness (OEE) is being implemented with assets ranging from injection-molding machines to robotic welding arms. With the advent of artificial intelligence (AI)-based IoT technologies, this approach is increasingly important for delivering tangible business value to manufacturing environments by enabling full transparency of the overall performance and productivity of a manufacturing line.
One Tech's Yasser Khan
Perhaps lesser known is how AI-based IoT solutions can help streamline production-planning processes. Even before materials hit the line, data related to machine performance can be collected and used to analyze and evaluate equipment. Once lines are in production, AI-based edge solutions can identify and mitigate points of failure, poor performance and human error. We can then reduce unexpected downtime.
Consider, also, quality analysis courtesy of AI and IoT solutions. Can we accurately determine which lines are operating as expected and which are underperforming? Can we glean which products coming off the lines are of a lower quality? Can we figure out where quality issues are originating? And can we estimate how many more units can be added to a line (or if a line is already running at/above quality-assurance demands)?
Let’s look at a use case:
The automotive industry leverages robotic welders and robotic material-handlers. Robotic welding arms across an automotive assembly line were performing at a lower-than-expected OEE (~50%), caused by unexpected downtime and static maintenance schedules inducing unnecessary costs and avoidable service hours. There was a lack of visibility into assets showing signs of failure, along with the performance (primarily cycle times) of these assets. Productivity and uptime metrics were not within acceptable levels and the factory was leaving billable machine hours and production on the table.
By using AI at the edge:
- Maintenance and repair issues were discovered and mitigated before they became catastrophic. By sampling data points from assets at rates of once every 100 milliseconds, AI processes data, understands if there are anomalies or signs of failure and alerts stakeholders when detected.
- Predictions were utilized to continually optimize the maintenance schedule to avoid unnecessary machine servicing.
- OEE for the robotic welding arm assembly line improved to more than 80% due to:
- Less unexpected downtime
- Reduced maintenance visits
- Faster identification and remediation of critical machine issues leading to higher operational uptime
- Accurate and real-time status of machine health
Optimizing the manufacturing process is no longer just about downtime—it’s an end-to-end process that can be significantly improved by edge-AI IoT solutions. By planning all the way through to the finished project, manufacturers can improve performance and deliver better business outcomes.
