航空发动机涡轮榫接结构虚拟试验技术

To solve the problems of high cost，long cycle and difficulty in monitoring in real time the test status of the fatigue test of aero-engine turbine joint structure，the virtual fatigue test technology was studied. The fatigue test of turbine joint structure was carried out to obtain load-displacement data，which were used to construct the NARX (nonlinear auto regressive model with exogenous inputs) neural network to carry out preliminary displacement prediction. On that basis，Kalman filtering was used to correct the predicted state with the measured data， and real-time prediction and updating of virtual fatigue test displacement were realized with the prediction error less than 5%. Finally，based on 3D MAX and Unity 3D，a high-fidelity digital model and virtual environment of the turbine joint structure were constructed to realize the visual display and data visualization of virtual fatigue test process of turbine joint structure.