CAJ | 학술논문

为了获得足够有效的切削过程状态信息，确保产品质量/系统零部件的安全可靠运行，应采用可靠的监测策略并对传感器进行优化布置。针对单工位多工步切削过程状态监测，基于多工位误差流(Stream of variation，SOV)理论构建单工位多工步信息流模型，通过状态空间变换和主成分分析确定故障/监测量信息传递系数π用来表征不同测点传感器的监测能力；考虑传感器以及故障/监测量之间的特性差异对于系统检测能力的影响，采用6Sigma的因果矩阵(Cause-effect matrix, CEM)和失效模式与影响分析(Failure mode and effect analysis, FMEA)工具分别对传感器以及故障/监测量特性进行量化表示；基于属性层次模型(Attribute hierarchical model, AHM)构建传感器，故障/检测量以及系统检测能力之间的因果关系，设定优化目标和约束条件，并采用元启发式算法-混合蛙跳算法(Shuffled frog leaping algorithm, SFLA)和遗传算法(Genetic algorithm, GA)用于优化计算。提出基于单工位状态监测的六步传感器优化布置策略。实例分析表明，在一定约束条件下，就优化目标而言， SFLA显示比GA更高的优化效率，为单工位状态监测的传感器布置优化提供实践参考。
위료획득족구유효적절삭과정상태신식，학보산품질량/계통령부건적안전가고운행，응채용가고적감측책략병대전감기진행우화포치。침대단공위다공보절삭과정상태감측，기우다공위오차류(Stream of variation，SOV)이론구건단공위다공보신식류모형，통과상태공간변환화주성분분석학정고장/감측량신식전체계수π용래표정불동측점전감기적감측능력；고필전감기이급고장/감측량지간적특성차이대우계통검측능력적영향，채용6Sigma적인과구진(Cause-effect matrix, CEM)화실효모식여영향분석(Failure mode and effect analysis, FMEA)공구분별대전감기이급고장/감측량특성진행양화표시；기우속성층차모형(Attribute hierarchical model, AHM)구건전감기，고장/검측량이급계통검측능력지간적인과관계，설정우화목표화약속조건，병채용원계발식산법-혼합와도산법(Shuffled frog leaping algorithm, SFLA)화유전산법(Genetic algorithm, GA)용우우화계산。제출기우단공위상태감측적륙보전감기우화포치책략。실례분석표명，재일정약속조건하，취우화목표이언， SFLA현시비GA경고적우화효솔，위단공위상태감측적전감기포치우화제공실천삼고。
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 coefficientπ, 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.