FIELD: pipelines.
SUBSTANCE: invention relates to predicting the ovality of a steel pipe after the pipe expansion stage, which is part of the technological process of manufacturing a steel pipe, which includes: the U-forming stage, the O-forming stage, performed by molding a billet having a U-shaped section, and the pipe expansion stage. The initial data collection stage is carried out, in which the initial data is collected by performing a numerical calculation, in which the input data is a data set of operating conditions, which includes one or more operating parameters selected from the operating parameters of the U-forming stage, and one or more operating parameters selected from the operating parameters the O-forming stage. The output data is information about the ovality of the steel pipe after the pipe expansion stage. The numerical calculation is performed repeatedly with a change in the data set of operating conditions and the generation of multiple data pairs from the specified data set of operating conditions and the specified information about the ovality of the steel pipe after the pipe expansion stage. The model generation stage is carried out, at which an ovality prediction model is generated, for which the input data is the specified set of operating conditions, and the output data is the specified information about the ovality of the steel pipe after the pipe expansion stage. The generation of an ovality prediction model is performed using machine learning, using the specified set of training data pairs generated at the initial data collection stage.
EFFECT: accurate prediction of the ovality of the steel pipe after the pipe expansion stage and regulation of the ovality of the steel pipe is provided.
12 cl, 20 dwg, 1 tbl, 7 ex
