| Servo actuator is a vital executive element in actuation control system,and its reliability directly affects the overall performance and safety of the system.Under the condition of long-term continuous working environment of high temperature and high pressure,the internal sealing performance of it plays a decisive role.Once the seal failure is triggered,it will affect the normal working efficiency,and may lead to unpredictable serious consequences.As one of the most commonly used seals in the sealing structure,the rubber sealing ring is particularly critical for its performance under complex working conditions.This paper takes the rubber sealing ring in a certain type of servo actuator as the research object to study the characteristics of its sealing performance and the application of fatigue life prediction methods.In order to study the sealing performance of rubber sealing rings in use,the finite element related theories and sealing judgment criteria of rubber sealing rings were analyzed.A two-dimensional rubber seal model was established using finite element analysis software ABAQUS,the variation and distribution of the stress on the rubber sealing ring under different working conditions are analyzed,as well as the influence of temperature on the sealing performance under the condition of mechanical stress,which has certain reference significance for the design and use of the rubber sealing ring.The analysis of fatigue life prediction draws on several fatigue life prediction methods that are widely used nowadays.According to the calculation results of finite element analysis,the dangerous points in the high-stress zone of the sealing ring are determined,and using the methods of fracture mechanics,the strain energy release rate of the rubber sealing ring is used as the damage parameter to establish the fatigue life prediction model,and the fatigue life of the rubber sealing ring under service conditions was approximately calculated.On the basis of traditional rubber fatigue life prediction methods,this paper applies the deep learning model to the field of fatigue life prediction.According to the relevant results of the finite element analysis,the prediction features are selected and constructed,and the fatigue life data set of the rubber seal ring is established after data preprocessing,and a recurrent neural network life prediction model was built and trained.The verification of the test set can prove that the model has the ability to predict fatigue life.In view of the possible deficiencies of the model,the long short-term memory network prediction model was designed and optimized.The parameters were adjusted through typical hyperparameter experiments.It was found that when the number of hidden layers is set to 2,the number of hidden layer units is set to 8,and the time step is set to 3,the dropout rate is set to 0.3 and the learning rate is set to 0.001,the model effect is the best,and the error is about 0.51%,which significantly improves the prediction accuracy of the prediction model.Finally,it is verified that the prediction model has good generalization ability in other data sets. |
