Study On Life Prediction Method Of External Spur-gear Pump Based On Adaptive Fuzzy Neural Network

The running state of gear pump is closely related to the operation of the whole system and even the whole mechanical equipment.In the traditional hydraulic system,the external gear pump is widely used because of its economical and durable characteristics.So it is very important to predict and evaluate the reliability life of external gear pump.The health and life information of the external gear pump can be characterized by the modulated vibration signal.The nonlinear characteristics of the vibration signal make it difficult to excavate the state of the external gear pump.Therefore,in order to excavate the state and life information contained in the vibration signal of the external gear pump more deeply,the life prediction and reliability assessment of the external gear pump can be realized accurately.In this paper,the wear and degradation principle of external gear pump is analyzed in depth,and based on the experimental data of accelerated degradation developed and designed in this paper,Based on the MEEMD,multi-scale reconstruction and de-noising of vibration data is carried out to realize the modulation and demodulation of vibration data.Compared with the EMD and the EEMD,the advantages of the MEEMD method are verified.Secondly,from the view of time domain,frequency domain and time frequency domain,several feature extraction methods based on vibration data are discussed.Based on the kernel principal component method,the multi-feature fusion of external meshing gear pump is realized,and the robustness of identifying network input space is enhanced.Secondly,the evaluation index of degradation of external meshing gear pump is established by smoothing the fused characteristic components.Then,the ANFIS is trained by using the evaluation index of gear pump degradation,and the residual life of the failure and non-failure samples is predicted by the obtained degradation assessment model of gear pump.The ANFIS method is applied to the life prediction of hydraulic external meshing gear pump for the first time and good prediction results are obtained.The feasibility of the method is verified.Finally,based on the inverse distance estimation method and Monte Carlo sample expansion method,the reliability evaluation of this type of external gear pump under the condition of stepping stress accelerated degradation test is realized in this paper.In this paper,the residual life prediction of single gear pump is studied deeply,and the total life assessment of specific gear pump is realized,which provides a new direction and idea for life prediction and evaluation.