Early warnings and predictive maintenance on critical assets

Challenge

Equipment can break or reduce in efficiency over time. Machine learning algorithms can learn how equipment is supposed to operate under various modes and seasons. It can then warn maintenance engineers when an asset or process is not behaving as expected. However, this requires data science competency and often rather long and complex implementation projects.

Solution

With Intelecy, the maintenance engineers build and deploy these predictive models themselves in a few simple clicks, making it possible to monitor hundreds of assets without any involvement from specialists, such as data scientists. The results are fewer failures, outages, and ad-hoc maintenance jobs. A “simple” example can be that the temperature in circuit breakers or transformers is often monitored with static alarm limits. A static alarm limit doesn’t consider ambient temperature and the load.

Intelecy models will have dynamic alarm thresholds based on multiple variables, so 60 degrees in the summer may be ok, but not in the winter. This limits false alarms and ensures operators can take actions based actual events.

Result

• Reduced emissions
• Early warnings provided
• Enhanced process efficiency