Abstract: Aiming at problems that there were several types of uncertainty namely fuzziness,randomness,discrepancy in the course of product's Failure Mode and Effects Analysis (FMEA),and that there were coupling propagation relations between failure modes,a novel integrated FMEA method based on cloud model,generalized evidence theory and interpretive structure modeling was proposed.Cloud model was adopted to model experts'uncertain evaluation information on risk factors,and then relevant cloud operators were constructed to compute risk priority number.Generalized evidence theory was introduced to fuse all experts'evaluation opinions,so that the whole team comprised of experts could reach collaborative decision result of risk priority number.The correlative coupling relation between failure modes was analyzed through interpretive structure modeling,and coupling propagation index of failures modes which was built on the basis of correlative coupling relation was constructed.Coupling propagation index was utilized to amend collaborative decision result of risk priority number,so that perfect comprehensive risk level ordering of failures modes was resorted.A case study of large scale numerical control machine was offered to illustrate the practicability and validation of the proposed method.

Key words: failure mode and effects analysis, cloud model, generalized evidence theory, coupling propagation, interpretive structure model

摘要： 针对产品失效模式与影响分析(FMEA)的风险评估过程中,专家给出的信息具有模糊性、随机性、不一致性等多重不确定性问题,同时考虑到失效模式之间的耦合传播效应,提出基于云模型、广义证据理论和解析结构模型相集成的FMEA风险评估新方法。首先,利用云模型对每个专家关于风险因子的不确定性评价信息建模,通过构建相关的云算子计算风险优先数;其次,引入广义证据理论融合所有专家的评价意见得到团队协同性风险评价数值;最后,通过解析结构模型分析失效模式间的关联耦合关系,并通过定义耦合传播指数对团队协同性决策结果进行修正,进而获得失效模式风险程度全面完善的排序。以大型数控机床为例,验证了该方法的可行性与有效性。

关键词: 失效模式与影响分析, 云模型, 广义证据理论, 耦合传播, 解析结构模型