Predictive Maintenance
Monom, a company belonging to the Álava Group and specialized in IoT, aspired to create a platform capable of interpreting the data collected by its sensors. The objective was to analyze the behavior of industrial machinery, such as wind turbines and other rotating equipment, to implement predictive maintenance through anomaly detection and other advanced analysis functions.
At WhiteBox, we were responsible for developing the core of the platform's predictive algorithms. We adopted several approaches, starting with the use of native BigQuery models and moving towards the implementation of unsupervised models created by us, using open source tools such as scikit-learn. These models were able to automatically identify the different operating modes of each machine and detect abnormal operating conditions. They also offered a detailed explanation of the detected anomalies and the impact of each variable on them.
Thanks to the innovation in algorithms on the part of WhiteBox, Monom managed to transform its traditional methods of detecting anomalies. We went from a semi-manual system, dependent on human analysis, to a fully automated solution based on artificial intelligence. This improvement not only raised the quality of the services offered, but also optimized operating and maintenance costs.
RailVision AI
CANETIA
