Aero-generator life prediction is an important problem to solve in aviation industry. This paper proposes variable weight combination model, which is Based on MGM(1, n) model and LSSVM model, to predicts the life of aero-generator. By the use of special accelerating life test platform for a certain type of real aero-generator, test data of the input speed, fuel injection pressure, load, inlet oil temperature as well as other life characteristic parameters of generator are obtained. After an in-depth analysis of these parameters, the variable weight combination model is built, to solve the problem that the prediction accuracy of traditional single life prediction model is not satisfactory. Then use the combination model to predict the life parameters of aero-generator. Research results show the superiority of combination model that the prediction accuracy of the variable weight combination model is better than each single prediction model. The variable weight combination model may accurately achieve the purpose of aero-generator life prediction and has good practical engineering value.
Aero-generator Life Prediction MGM(1 n) LSSVM Variable Weight Combination Model
Jianguo Cui Jianqiang Shi Yuezhong Li Shiliang Dong Xiao Cui Tao Huang
School of Automation, Shenyang Aerospace University, Shenyang 110136, China School of Automation, Shenyang Aerospace University, Shenyang 110136, China; School of Northern Science and Technology, Shenyang Aerospace University, Shenyang 110136, China Shenyang Aircraft Design & Research Institute, Shenyang 110035, China School of Information Science and Engineering, Shenyang University of Technology, Shenyang 110870, C School of Management and Economics, Shenyang Aerospace University, Shenyang 110136, China