FEATURE EXTRACTION AND SENSOR OPTIMIZATION FOR CONDITION MONITORING OF RECIRCULATION FANS AND FILTERS
Mike Gerdes, Dieter Scholz · 2009
Complex systems are commonly monitored by many sensors. When one of the sensors fails, the current state of the system can not be calculated or the information about the current state is not complete. For that reason sensor failures are one of the main error sources of a system. Thus sensors that deliver significant information about the system state need to be redundant. This paper shows how to calculate the significance of the information that a sensor gives about a system by using modern signal processing and artificial intelligence technologies. It also shows how significant features can be extracted, evaluated from a set of data samples, how difficult it is to find an optimal parameter and sensor set and that it is possible to reduce the size of needed data by 97%. The paper concludes analyzing the results of experiments that show that the methods can classify different errors with a 75% probability.