针对传统失效模式与效应分析（FMEA）方法的不足，构建了结合区间二元语义和ELECTRE方法的改进FMEA模型，以解决不确定语言环境下失效模式的风险排序问题。研究表明：该模型可以有效地将模糊语言表达转化为二元语义量化数据进行计算；通过对专家成员和风险因素设置权重有效减少了极端偏见等对结果的影响；同时通过多准则决策ELECTRE方法对数据项间的越级关系分析，增强了模型的灵活性和适用性，提高了排序结果的准确性。研究结果表明，E-Learning系统质量失效模式风险指数最高的是效率性，企业在产品设计和质量改进时，应对与效率性相关的内容重点关注和优先考虑。

Aiming at the shortcomings of traditional failure mode and effect analysis (FMEA) methods, an improved FMEA model combined with interval 2-tuple linguistic and ELECTRE method is constructed, to solve the risk ranking of failure modes under uncertain linguistic environments. Results show that, the model can effectively transform the fuzzy language expression into quantitative data by interval 2-tuple linguistic; reduce the impact of extreme bias on the results by setting the weights of expert members and risk factors effectively; and enhances the flexibility and applicability of the model and improves the accuracy of sorting results by using the ELECTRE method of multi-criteria decision-making to analyze the non-outranked relationship of data items. In the case of E-Learning system, the highest risk index of quality failure mode is efficiency, so enterprises should focus on and give priority to efficiency-related content in product design and quality improvement.

F27;C931

国家自然科学基金项目“复杂不确定环境下基于FMEA的产品质量与可靠性管理方法研究”，（71671125）；同济大学中央高校基本科研业务专项基金“数字语言学习系统精准学习服务策略研究”（22120170446）。