DIGITAL INVERSE ENGINEERING OF POLYMERS FOR 3D PRINTING: APPLICATION POSSIBILITIES AND CHALLENGES

The article considers a promising approach to the design of polymer materials for additive technologies using neural network engineering methods. The necessity of transition from traditional empirical selection to intelligent modeling based on the analysis of interrelations between properties and molecular structure of polymers is substantiated. Arguments are given in favor of using neural networks to predict the performance characteristics of materials and generate new formulations adapted to the specific technological parameters of 3D printing. The use of modern neural network architectures, such as transformers and NLP modules, makes it possible to automate the selection and generation of polymer compounds, which improves the quality and predictability of the final product. An important result is a significant reduction in the financial and time risks associated with experimental trials, which is especially important for small and medium-sized businesses.Special attention is paid to identifying potential limitations and risks of using neural network models, as well as strategies for minimizing them through multilevel filtering and data verification. The presented platform concept is considered as a key element of the digital infrastructure capable of stimulating innovative development and ensuring the economic sustainability of the industry through integration with industrial and research ecosystems.