Neural network model for wing strain-load relationship based on fusion of real and virtual data

When establishing a strain-load relationship model for aircraft structures, ground calibration tests can obtain high-fidelity data but are trapped with limited test ranges, while finite element simulations are not limited by test ranges but the data fidelity is low. This leads to difficulties in achieving win-win situation of accuracy and applicability based solely on either ground calibration test data or finite element simulation data. To address the above issue, two multilevel neural network models fusing real and virtual data are put forward, a mapping-based model and a compensationbased model. A method for measuring the model’s cognitive degree based on the variance of base learners is estab⁃ lished and embedded into the compensation-based model. A neural network model with high accuracy, wide applica⁃ bility, and the capability to forewarn unreliable prediction results is then developed. This developed model is validated using a scaled-down wing. Compared with complete reliance on real data from ground calibration tests, the load mod⁃ els based on fusion of multi-source data demonstrate superior capabilities, and the compensation-based model is bet⁃ ter than the mapping-based one. Moreover, the compensation-based model can effectively identify the data samples with poor cognitive degree of the load model and thereby provide warnings for unreliable prediction results.