This task is to use sensor measurements and rule-based alarms to predict failures in the electrical grid to alarm the operators of a real-time operation centre. A task is to analyze the available data to identify how to tackle the problem. Could the rule-based alarms be used as symptoms that indicate problems ahead of time? Is it necessary to find new patterns in the sensor measurements instead of using the rule-based alarms?
The research is done in collaboration with TrønderEnergi. It is possible to get a working space at their offices at Lerkendal, sit together with the data science team and work closely with the domain experts.