The common approaches for maintenance are scheduled maintenance where the parts are replaced at fixed intervals, or corrective maintenance where the parts are replaced after being broken. An alternative is condition-based maintenance where the condition of the system is monitored through sensor measurements, and by using advanced mathematical methods the current state of the system as well as it remaining useful life (RUL) can be determined. By continuously monitoring the condition of the system and its RUL, it enables better planning of the maintenance, reduces the required amount of spare parts, increases the systems availability, and allows for large cost savings. This project aims at establishing Teknova AS as a leading research institute within this field.
- Contribute to more effective industrial processes through optimizing reliability, maintainability and safety.
- Develop competency on intelligent predictive maintenance systems that exploit all events, sensor measurements and other knowledge.
- Form a leading research group in Agder to respond to industrial R&D need on condition based maintenance
- Establish a common networking arena for exchange of knowledge and experience
- Recruit high quality research personnel to Agder
- Develop portfolio of collaborative projects.
The project takes a modular hybrid approach to the field of condition based maintenance and focuses on the following selected topics.
WP 1 Condition Monitoring
- Online, contact free condition monitoring system
- Based on optics, electromagnetics, acoustics
- Combination of different techniques
Laser scanning - Computer vision - Triangulation - 3D Cameras - Structured light - Temperature measurements - Spectroscopy - Eddy current sensing - Acoustic emissions - Vibration -
WP 3 Fault Detection & Diagnosis
- Inferential sensing to complement the physical measurements
- Advanced and reliable state estimation
- Fault management
- Kalman filter - Modeling - Root cause analysis - Multivatiate statistics - Fault tree analysis - Data fusion - Feature extraction - Feature classification -
WP 3 Intelligent Decision Support
- Active, cooperative decision support systems
- Help to identify and solve problems
- Learn and understand from experience
- User-friendly control displays with minimum of false alarms
- Expert systems - Fuzzy logic - Machine learning - Data mining - Data visualization - Pattern recognition - Case-based reasoning - Genetic algorithms -