METHOD FOR SCIENTIFIC SELF-ADJUSTING MOULDING BASED ON REAL-TIME RECURRENT LEARNING

摘要

Technical description: method for scientific injection moulding, which combines processes, tools, materials and techniques for analysis with intelligent algorithms for eliminating unwanted variations in the injection process, the method being characterised in that it integrates: moulds containing instruments inside the physical geometry thereof; a network of sensors for capturing variables such as temperature and pressure, thermal images and images for dimensional analysis, which is characterised by deterministic redundant communication between sensors, actuation devices, intelligent process analysis and control devices and a user interface; analysis and intelligent control based on real-time recurrent learning, characterised by holistic self-adaptation of the process in real time, analysis and decision-making for the optimisation of injection quality by detecting critical areas and possible improvements in mould designs; electrical, hydraulic and pneumatic actuation devices; an interaction device with an operator with multiplatform connectivity to the moulding control and analysis device; and a device for automatically separating defective pieces, characterised in that it integrates with the intelligent control by means of the communication network, having at least one mechanism for controlling movement of at least two degrees of freedom.