The plant productivity data stream talks, but we humans don't listen. In the future, artificial intelligence will be able to identify the signal in noise, i.e. report the deviation from the normal state, without a manufacturing engineer having explicitly stated what “normal” is beforehand. The algorithm makes suggestions to humans as to where to start optimization.
The OEE development path
OEE has developed in companies. It started with a checklist written by hand. The first stage of development was Using Microsoft Excel. If you set up an OEE management system step by step, the employees have certainly understood it and can argue with the key figure. But then the next steps must also follow. Sending an Excel list once a week to all those responsible for operations is not 2018. There is much more to it.
Anyone using oee.ai has taken the further development steps: The data is available on-line and is displayed on Andons above the system and in offices. In addition to availability and quality loss, the causes of performance losses are also recorded. The background detection is digital and precise; graphics show the background distribution and curves. If the system does not perform as expected, a responsible person is immediately informed by email, SMS or push notification.
If targeted improvements are now being worked on using the findings from the data, we have plant productivity management using methods that comply with the current state of knowledge and technology.
The OEE data stream is talking
The OEE data stream, i.e. the actual system speed, the default speed and the causes of the discrepancy between the two data, contains a lot of information that has not yet been used specifically for process improvements.
Which production manager can answer the following questions:
- Does the morning shift systematically get ready faster than the late shift?
- Is it better to drive the system faster with more short shutdowns rather than slower and at the same time more stable?
- Was the OEE good for the ABC contract or is it usually manufactured at a lower cost?
- What is the good time from the end of the renovation to reaching the production ridge line?
- Which products cause a particularly large number of short shutdowns?
And these are just 5 of the many questions that no one in a manufacturing company can answer today.
Artificial intelligence in the analytics cockpit
In future, an analysis cockpit will actively alert the manufacturing engineer to anomalies in the data stream. Artificial intelligence algorithms search the data stream for patterns and deviations from them.
The data for the evaluations will continue to be collected by the proven minimally invasive sensor and tablets, so that no intervention in the system is necessary for this new level of OEE management. If the data is available in a networked controller, it can of course also be fed directly to oee.ai.
The anonymized data is used across companies. For example, the algorithms can be trained together for all operators of filling plants. Because we can already see that today: Whether mineral water, coffee, milk or spread is bottled makes no difference to us in the data patterns. OEE analysis of machine tools is the next frontier to be overcome.
If you are interested in staying up to date with the latest developments in the use of artificial intelligence in OEE management, you can contact us at info@oee.ai or sign up for our newsletter below. Please feel free to contact us.