Abstract: The impact of optimal sensor placement on the access to status information of the cutting process, the product quality and the operation safety of mechanical parts in manufacturing systems is significant. Aiming at multi-step status monitoring in single station, an optimal sensor layout is proposed for troubleshooting. The stream of multi-step information model is proposed based on the stream of variation (SOV) theory. The information transfer coefficient , which characterizes detectability of sensors in different measuring points, is derived from the state space transform and main component analysis. Considering the influence of the characteristics of sensor and fault/object on detectability of the system, the six sigma tools, the C&E matrix (CEM) and the failure mode and effect analysis (FMEA), are employed to quantify the characteristics of sensor and fault/object, moreover, the causal relations between sensor, fault/object and detectability of system are developed based on the attribute hierarchical model (AHM).The optimization goals and constraints are determined. The shuffled frog leaping algorithm (SFLA) and genetic algorithm (GA) are used for calculation. Six steps of the sensor deployment are proposed. Case analysis shows that, under some constraints, the efficiency of SFLA is higher than that of GA for optimization goal, which provides a practical reference for the status monitoring in single station.

Key words: attribute hierarchical model, condition monitoring, sensor deployment, single station