Predictive Maintenance & Anomaly Detection
The optim*um service
- Data-based early forecast of necessary maintenance, particularly in manufacturing companies (predictive maintenance)
- A large, unannotated data volume serves as a model of the normal conditions so that it can classify significant deviations as an indication of a defect (anomaly detection)
- Capturing and assessing sensor data and machine parameters in large facilities and factories (e.g. vibration data, temperature values or operating voltages)
- Delivering alert systems and forecast models for assessing whether a facility or one of its components is in its normal condition, whether a defect is expected in the near future or whether a defect has already occurred.
What makes *um better?
Based on comprehensive experience in the industry sector, we can develop domain-specific solutions to the specification of our clients. We utilize state-of-the-art machine learning processes to detect and classify patterns in substantial sensor data and to develop forecast models and alert systems.
In cooperation with our internal data engineers, we can support our clients comprehensively in productive implementing such early warning systems, either as a remote service or as an on-premise solution. We advise our clients with a particular emphasis on data security. Following the analysis phase, we are also able to give clients recommendations for optimized sensor concepts. A specific recommendation could be, for example, installing more or fewer sensors in order to improve the database.