CAJ | 학술논문

针对发动机性能监控过程中出现的单参数监控信息量不足和多参数容易矛盾的情况，提出一种发动机性能监控模糊指数融合方法。该方法基于滑动窗口参数均值和熵，采取有效的信息融合技术，建立表征发动机性能的模糊融合指数。利用神经网络方法，依据已确定的模糊融合规则推导出剩余决策规则。实例表明，模糊融合指数能很好地跟踪发动机性能缓慢衰退和突变两种情况。用正交基分解的方法对模糊指数进行重构预测，与线性拟合和二次拟合相结合的方法相比，其预测精度更高，能准确预测发动机的性能变化。
침대발동궤성능감공과정중출현적단삼수감공신식량불족화다삼수용역모순적정황，제출일충발동궤성능감공모호지수융합방법。해방법기우활동창구삼수균치화적，채취유효적신식융합기술，건립표정발동궤성능적모호융합지수。이용신경망락방법，의거이학정적모호융합규칙추도출잉여결책규칙。실례표명，모호융합지수능흔호지근종발동궤성능완만쇠퇴화돌변량충정황。용정교기분해적방법대모호지수진행중구예측，여선성의합화이차의합상결합적방법상비，기예측정도경고，능준학예측발동궤적성능변화。
Aiming at information shortage of single parameter and easy contradiction of multi-parameter in the process of engine performance monitor, a method of fuzzy index fusion in the engine performance moni-tor was presented. With effective information fusion technology, based on parameter mean and entropy of sliding window, performance fuzzy fusion index was established. Residual decision-making logic was de-duced by neural network according to pre-determined fuzzy fusion rule. Example indicated fuzzy index could track the engine performance slow deterioration and sudden changes. Fuzzy fusion index was recon-structed and predicted by adopting orthogonal basis decomposition method. Compared with combination of linear and quadratic prediction method, orthogonal basis decomposition method gets higher accuracy, and predicts performance change accurately.