A Method for Sensor Selection in Reconfigurable Process Monitoring

Abstract

Sensor selection is important for process monitoring. Each time a machining system is reconfigured, the corresponding monitoring system needs to be re-designed for effective detection of the faults of interest. To accomplish this, a sensor selection methodology is proposed in this paper. Fuzzy theory is utilized for the multi-criteria decision-making process. Evaluation criteria are selected based on a thorough study of fault characteristics. Under subjective criteria, the ratings of different sensors are evaluated using linguistic terms, which are converted to trapezoidal fuzzy numbers. Under objective criteria, the fuzzy performance of each sensor is obtained directly from its specifications. The suitable sensor or sensors can be selected by aggregating and ranking the fuzzy numbers. Sensor selection in the turning process is used to illustrate the proposed methodology.