IMPROVE provides an alarm management algorithm to prevent alarm floods and to enhance the handling of such a flood in case it still occurs
Central characteristics of our alarm management algorithm:
- Combining similarity measure learning, offline case-base construction, semi-supervised learning, online flood detection, and case-based-reasoning (CBR)
- Detecting alarm floods and attempting to identify them based on a knowledge base
- Suggesting solutions in case an alarm flood occurs (identi¬fication of the flood, repair instructions, etc.)
- Reducing complexity and hiding “follow-up alarms” from a previous (root) alarm
- Supporting the operator in finding the root cause for the alarm and to take the right action
For more information, please contact our coordinator Prof. Dr Oliver Niggemann.
You wish to receive information for special stakeholders? Please find the specific information linked below.
Marta Fullen (IOSB-INA) explains the IMPROVE result