| Popis: |
Carbon fiber reinforced composites applied to aerospace vehicles in a large amount are always provided with inhomogeneity and anisotropy. The existing theoretical calculation model, namely dynamic homogenization model, can only solve the problems of slow calculation and large memory consumption caused by inhomogeneity and anisotropy when the direction of electromagnetic wave electric field is perpendicular to carbon fiber. In order to meet the requirement of high precision extraction of electromagnetic parameters of aerospace composite materials, it is necessary to have a further study on the calculation model of electromagnetic properties of composite materials at arbitrary incident angles of electromagnetic waves. Our work mainly researches the high-precision calculation model of equivalent permittivity of anisotropic composites based on machine learning. Emphasis is laid on the study of the influence of multi-incident angle of electromagnetic wave on electromagnetic properties of anisotropic composites, the high-precision prediction model of electromagnetic parameters based on machine learning, model verification and optimization, etc.. First, the model training database is obtained through data processing based on the Helmholtz equation method; then, the machine learning high-precision calculation model is established by neutral network; finally, the remaining data is utilized into the machine learning model to verify the accuracy of the electromagnetic parameter model of the composite material. The calculation model of composite electromagnetic characteristics based on machine learning has the high-precision prediction ability of electromagnetic parameters of composite materials under arbitrary incident angle of electromagnetic wave. The extracted equivalent permittivity can be applied to electromagnetic properties of airframe structure in the aerospace vehicle modeling, which can be utilized to analyze shielding effectiveness and electromagnetic environment effects of aerospace vehicles. The purpose is to shorten the research and development cycle of aerospace vehicles, reduce the number of experiments and save the research and development funds. |
