提出了以产品结构树为诊断知识的组织框架和利用FMEA(Failure Mode and E ffect Analysis)信息建立父节点与子节点间的诊断规则的故障诊断方法。研究了基于产品结构树模型的框架知识组织方法,并探讨了由FMEA信息建立诊断推理规则及推理...提出了以产品结构树为诊断知识的组织框架和利用FMEA(Failure Mode and E ffect Analysis)信息建立父节点与子节点间的诊断规则的故障诊断方法。研究了基于产品结构树模型的框架知识组织方法,并探讨了由FMEA信息建立诊断推理规则及推理策略,最后以某列车自动门系统为例进行诊断分析。展开更多
研究由产品设计知识生成诊断知识是非常有意义的,它能很好地缓解诊断知识匮乏带来的窘境。在此思想的指导下,分析了基于“结构FMEA(Failure Mode and EffectAnalysis)”的诊断模型构建的可行性,并以发动机润滑系统为对象研究从“结构FME...研究由产品设计知识生成诊断知识是非常有意义的,它能很好地缓解诊断知识匮乏带来的窘境。在此思想的指导下,分析了基于“结构FMEA(Failure Mode and EffectAnalysis)”的诊断模型构建的可行性,并以发动机润滑系统为对象研究从“结构FMEA”自动生成故障树的方法,从而最终实现设计知识到诊断知识的转化。与手工建树相比,这种方法可以高效、准确地得到具有较完备诊断知识的故障树。展开更多
针对运动想象脑电信号策略缺乏多样性的问题,提出一种在肢体运动的同时大脑想象运动过程与完全想象运动过程相结合的运动想象脑电信号策略。结合小波变换(wavelet transform,WT)与奇异值分解(singular value decomposition,SVD)提取运...针对运动想象脑电信号策略缺乏多样性的问题,提出一种在肢体运动的同时大脑想象运动过程与完全想象运动过程相结合的运动想象脑电信号策略。结合小波变换(wavelet transform,WT)与奇异值分解(singular value decomposition,SVD)提取运动想象脑电信号特征,用支持向量机(support vector machine,SVM)与K最近邻(K-nearest neighbor,KNN)分类器对不同运动想象策略进行分类。针对脑电信号采集过程中常出现信号失真的情况,提出了自回归(auto-regression,AR)模型结合奇异值分解的规避方法,结果表明此方法能够有效排除信号采集过程中出现的失真情况。通过比较方法WT-SVD-SVM与WT-SVD-KNN的特征提取和分类算法,结果表明,WTSVD-SVM方法在单一策略和两者组合策略中最低分类精度达到90.00%,并且该方法在想象箭头向上、箭头向右以及左拳右摆策略下精度能够达到91.11%。展开更多
The knitting industry CAD/CAM system's development needs a solution of computer representation of the designed objects.One of them is a transformation of knitting rows and stitches form knitting pattern and its au...The knitting industry CAD/CAM system's development needs a solution of computer representation of the designed objects.One of them is a transformation of knitting rows and stitches form knitting pattern and its automatic analysis is to control data from the knitting design data.A method of the pattern CAD system including the analyzing and the compiling of the pattern data was developed in this paper.By rasterizing the pattern map,the drafted pattern could be compiled into knitting row data and stitches.The storage format of knitting rows data and stitches data could be defined by analyzing the motion of computerized fiat knitting machine.By reassembling the control data,it could be analyzed into the execute data of the controller in computerized flat knitting machine.This method mainly focused on the weak areas of multi-needle-bed technology and fully-fashioned knitting technology in inland and could be applied to most of control systems in the master-slave mode.The knitting row data and pattern row data in a " *.ptn" format file were integrated.Giving a corresponding analytic calculation process of control data and realizing independent development of the upper machine and the lower machine,the final knitting control data not only have low computation and transmission cost but also have fast transfer speeds.展开更多
文摘提出了以产品结构树为诊断知识的组织框架和利用FMEA(Failure Mode and E ffect Analysis)信息建立父节点与子节点间的诊断规则的故障诊断方法。研究了基于产品结构树模型的框架知识组织方法,并探讨了由FMEA信息建立诊断推理规则及推理策略,最后以某列车自动门系统为例进行诊断分析。
文摘研究由产品设计知识生成诊断知识是非常有意义的,它能很好地缓解诊断知识匮乏带来的窘境。在此思想的指导下,分析了基于“结构FMEA(Failure Mode and EffectAnalysis)”的诊断模型构建的可行性,并以发动机润滑系统为对象研究从“结构FMEA”自动生成故障树的方法,从而最终实现设计知识到诊断知识的转化。与手工建树相比,这种方法可以高效、准确地得到具有较完备诊断知识的故障树。
基金National Natureal Science Foundation of China(No.51305076)
文摘The knitting industry CAD/CAM system's development needs a solution of computer representation of the designed objects.One of them is a transformation of knitting rows and stitches form knitting pattern and its automatic analysis is to control data from the knitting design data.A method of the pattern CAD system including the analyzing and the compiling of the pattern data was developed in this paper.By rasterizing the pattern map,the drafted pattern could be compiled into knitting row data and stitches.The storage format of knitting rows data and stitches data could be defined by analyzing the motion of computerized fiat knitting machine.By reassembling the control data,it could be analyzed into the execute data of the controller in computerized flat knitting machine.This method mainly focused on the weak areas of multi-needle-bed technology and fully-fashioned knitting technology in inland and could be applied to most of control systems in the master-slave mode.The knitting row data and pattern row data in a " *.ptn" format file were integrated.Giving a corresponding analytic calculation process of control data and realizing independent development of the upper machine and the lower machine,the final knitting control data not only have low computation and transmission cost but also have fast transfer speeds.
