OPTIMUM - INJECTION PROCESS OPTIMIZATION OF MOULDING MACHINES
I4MS, ICT Innovation for Manufacturing SMEs
Date: September 2019
OPTIMUM aims to cloudify data-driven and visual analytics tools for sustainable process optimization in the automotive plastic parts production, exploiting the HPC potential by applying model-based predictive functions (available at Pragma platform) to the injection process. Experiment-specific machine learning models, cognitive models and optimizations techniques will be coupled with advanced visual analytics leading to cost and energy savings and improved customer-focused flexibility. Thanks to that, new knowledge will be generated to feed product differentiation and rise of software-enabled offerings for smart factories.
The OPTIMUM experiment proposes to perform a process optimization on the most critical “BSI-Housing and Top” production line of GLN Plast Plant in Portugal for the production of Automotive Plastic parts and associated injection processes. The manufacturing processes which build up the experimental setting are based on the production flow of the line that start from the dried raw material (granulated plastic). After that, the dried raw material follows the machine hopper and then enters into the screw of the injection machine starting the execution of an injection moulding process through the use of the injection machine and the mould. Indeed, the plastic material is melted in the injection machine and then injected into the mould where it cools and solidifies into the final part.
In OPTIMUM, the manufacturing process will be modelled and through specific evaluation processes (using specific KPIs) and machine learning algorithms and cognitive models, it will be possible to evaluate the behaviour of the key production parameters (identified in the following sections) in relation to the internal and external production variables. To this end, the Pragma platform will be utilized, which is equipped with trend analysis and anomalies detection modules, which will be used to detect and process malfunction diagnosis and instability together with machine learning techniques that will analyse the data stemming from the various operation of the plant to provide support for the decision of the production line manager through real-time data monitoring and data intensive visual analytics. Pragma platform is a powerful white label IoT platform, which is protocol and device agnostic, with a dynamic API for the interconnection with 3rd party systems. It supports multi-level user authentication, dynamic dashboards and widgets, reports, rules, etc, and above all embedded artificial intelligence algorithms, such as prediction, anomalies and fault detection, abnormal behavioural detection, recommendation engine, etc. for real-time data analysis.
The project CloudiFacturing receives funding from the European Union’s Horizon2020 research and innovation programme (Grant No. 768892).