ADAPTIVE CONTROL OF ADDITIVE PRODUCTION PROCESSES IN REAL TIME USING MACHINE LEARNING Russian patent published in 2020 - IPC G06N20/00 B29C64/393 B33Y50/02 

FIELD: control systems.

SUBSTANCE: invention relates to a method and a system for controlling application process for making random-shaped articles after designing or a process of connecting after designing. Method comprises providing an input design geometry for an object, a set of training data, sensors which output real-time data for object properties when the object is physically made, and providing a processor programmed to predict an optimum set of process control parameters for initiating an application process for making articles of an arbitrary shape or a connection process, which is obtained using machine learning algorithm, removing noise from data on object properties, issued by sensors, before their inclusion in algorithm of machine learning, real-time classification of detected object defects and provision of commands for application process implementation for production of free-form products after design or connection process after designing to make object.

EFFECT: technical result consists in control of product manufacture.

19 cl, 25 dwg