| 摘要 |
The method involves computing a rotating Gaussian bell curve as a model of a quasi-3D spray image, inputting fixed model parameters, inputting real physical parameters to a neural network trained with real input data that converts the additional parameters into model input parameters, feeding the parameters into the model, generating spray images, integrating spray image copies for the entire coating, outputting the thickness distribution. The method involves using a data processing system and computing a rotating Gaussian bell curve as a phenomenological model of a quasi-three-dimensional spray image, inputting specific fixed model parameters, inputting real physical parameters to a neural network previously trained using real input data that converts the additional parameters into model input parameters, feeding the input parameters into the model, generating spray images depending on spray device motion data, integrating the copies of the spray images for the entire coating and outputting the thickness distribution.
|
