Выбранную модель рекомендуется к интеграции в существующую систему диспетчерского управления и сбора данных рассматриваемой производственной линии с целью контроля данных в режиме реального времени и выявления предаварийных и аварийных ситуаций. Это также может стать основой для разработки новых алгоритмов и режимов работы машины для сварки углепластиков.

The aim is to develop a data-driven diagnostic and predictive model of time series forecasting for obtained data of the carbon plastic elements welding machine. The following document is about the “Data analysis of the technological process of the production lines” with Machine Learning algorithms based on Neural Networks. This can be achieved by building a number of models and comparing their results with evaluation metrics in order to find the model with best performance. In this work we develop a number of Neural Network models in order to make a Time Series Forecast using the data received from the welding machine of carbon plastic elements. First, the collected data will be preprocessed and transformed to the form, which models require. Second, the structure of all models will be illustrated by their layers and their performance will be calculated with the defined metric. After building and analyzing the constructed models, the best one will be selected. The chosen model is recommended for integration into the existing supervisory control and data acquisition system of the considered production line in order to control the real-time data and detect pre-emergency and emergency situations. It also can become a foundation for the elaboration of the new algorithms and modes of the carbon plastic welding machine.