Andon systems are one of the key components for manufacturers to conduct root cause analyses. They are used by the most advanced manufacturers in the world to identify issues on the production floor and stop the production line to get to the root cause of the issues thereby preventing them from reoccurring. Because quality and efficiency is key to successful manufacturing, andon systems have been used in manufacturing to alert quality issues since the mid 20th century, and are largely integrated into manufacturing processes.
While traditional andon systems have been useful, their reliance on manual intervention makes them prone to error. Using computer vision to activate andon systems is increasingly recognized as the superior approach for generating maximum returns within manufacturing.
What is an andon system
Andon (English: ‘Sign’ or ‘Signal’): A visual aid that highlights where action is required. In manufacturing issues occur quite often during the production process and require immediate action from multiple stakeholders on the shop floor. As a result companies often use a visual aid to point out the problem in the process that needs an action. This visual aid is called an Andon, and refers to the illuminated signal notifying shop floor personnel of quality control issues. Originating from the word for a paper lantern, Andon is a term that refers to an illuminated signal notifying others of a problem within the quality-control or production streams.
How andon systems are currently used for identifying production problems
While Toyota is known for making the Andon practice famous and ubiquitous within auto manufacturing, different versions of the same approach are used throughout the world. A commonly used method is having workers on the shop floor pull a cord or push a button when they spot a potential defect that can impact product quality and a warning light turns on at the impacted station alerting the line supervisor about the problem. In some cases the production line automatically stops and in other cases a line supervisor decides whether to stop the production line or not depending on the issue detected.
Inefficiencies with current andon solutions
While andons are crucial in manufacturing so that plants don’t produce goods with quality issues, the current system in place has some challenges. Modern day andons rely on human intervention as well as large scale infrastructure adjustments to make it work. For example, a part could move from Station A to Station B and wait at Station B for over 30 minutes until it is worked on. Unless someone is timing Station B, no one would know that there was significant downtime on Station B unless they happen to walk by it a few times and notice that the part is still sitting there or notice that they didn’t meet their production target for the shift.
Using computer vision to improve andon processes
In order to use andon processes while preventing bottlenecks and maintaining ideal efficiency, manufacturing companies are increasingly leveraging AI technology. At i-5O the concept of andon systems is central to our computer vision system as it is how our customers generate the maximum return on investment. When our system detects a potential problem on an assembly line, it automatically notifies the line management/supervisors/workers in real-time via SMS, email, warning light and/or notification on a local monitor. This allows our clients to take action immediately when a production problem occurs to minimize any downtime and maximize their revenue.
i-5O’s andon system: Fully automated and autonomous
Since i-5O’s vision system is non-invasive, it requires no production downtime to install since the cameras can easily be installed during non production hours. Furthermore, the system itself operates on set rules thus leaving no room for human interpretation i.e. the system is completely automated and autonomous. For example, a part waiting in the production line for 5 minutes without being worked on results in an alert sent to the production supervisor so they can immediately get someone to work on it and avoid revenue loss or overtime. Lastly, the system is always running so it can catch problems at anytime during production without human intervention. Production supervisors can be far away from the production line and still get notified of the issue to take action immediately.