A factory which produces over 100 million liters of paint and speciality coatings
The time-based maintenance schedule was becoming too costly. Machines require frequent checks in order to prevent them breaking down, however this frequently resulted in checks for machines that were in a very good condition. At the same time, lowering maintenance costs by reducing the frequency of the maintenance schedule was likely to increase breakdowns.
All the production settings for each machine were recorded and stored for 6 months. This data, combined with the maintenance and breakdown reports, was used to train an AI model that predicts which machine needs maintenance. This type of model is capable of handling large volumes of multivariate, time-series data.
- The model was able to detect which machines needed maintenance with 89% accuracy
- This allowed for a much less frequent maintenance schedule for each machine that had not been selected
- Lowered maintenance costs by 7% while reducing downtime