关键词: 航空发动机, 过程神经网络, 遗传算法, 图谱化预测, 健康状态

Abstract: To predict the health states of aero engines visually,quickly and systemically,a graphics prediction method based on Genetic Algorithm-Process Neural Network (GA-PNN) was presented.Aiming at the problems of Process Neural Network (PNN) model structure design and local minimum,the reconstructed phase space theory was studied to construct the training sample set,and genetic algorithm was discussed to optimize the structure parameters,the initial weights and the thresholds of PNN model.A prediction performance matrix of aero engine was presented by using GA-PNN algorithm.The prediction performance matrix was transformed into colorful prediction graphics for high-dimension,high coupling and non-linear features of the performance parameters.The health states prediction of aero engines was achieved from the implicit health information of the prediction performance matrix by colorful prediction graphics.Two cases study were described to demonstrate the effectiveness and feasibility of the proposed method.

Key words: aero engines, process neural network, genetic algorithms, graphics prediction, health states