在分析航空发动机早期故障特点和数据融合的基础上,将证据理论应用于发动机故障征兆级别的数据融合中,提出了构造基本信任指派函数的方法,从而得到了更准确的早期故障征兆.在此基础上,应用模糊聚类分析方法,依据最大隶属度原则得到发动机早期故障源.该方法考虑了发动机早期故障征兆不明确、不精确、不肯定等特点,应用模糊测度来表示征兆,更符合发动机早期故障的实际情况. Based on the analysis of the characteristics and data fusion of early stage aero-engine faults, the evidence theory is applied to the data fusion of engine fault symptom level, and a method of constructing the basic trust assignment function is proposed to get a more accurate early fault symptom. Based on the fuzzy clustering analysis method, according to the principle of maximum membership, the early fault source of the engine is obtained.The method takes into account the characteristics of the engine early fault symptom is not clear, not accurate, not sure, the application of fuzzy measures to indicate signs, more in line with the engine The actual situation of the early failure.