Digital Twin of Injection Molding

In 2018, around 360 million tons of plastic were produced worldwide, of which 61.8 million tons were manufactured in Europe. For Europe, this results in a positive trade balance of €9.4 billion in terms of imports and exports. In Germany alone, the plastics processing sector employs around 336,000 people in some 3,058 companies, which generate total annual sales of over €65.1 billion.

Just under a quarter of the tonnage produced annually is processed by injection molding, which is thus one of the most important fields of application. In addition to high quality and reliability, a significant unique selling point of injection molding machines manufactured in Germany is the high level of innovation with regard to integration and networking options. In 2018, for example, process innovations in the plastics processing industry alone resulted in a cost reduction of 2.3% compared to the previous year. The new networking options for injection molding machines, for example communication interfaces such as OPC-UA, allow machine and process variables to be recorded in high resolution. This data can be used for better quality monitoring, which can contribute to further cost reductions. Automated quality control and adjustment of process parameters can, for example, enable a higher proportion of recyclates (reused production waste) in production and thus further cost savings. For these reasons, injection molding has for decades sought to regulate the quality characteristics of manufactured components. However, control concepts currently implemented in commercially available injection molding machines merely regulate machine or process variables that correlate with the component properties.

True control of component properties requires both inline measurement of the corresponding quality variables and dynamic process models of the entire chain of effects, from the process parameters set on the machine to the resulting component quality. The dynamic models, i.e. the digital twin, can then be used for model-based closed-loop or open-loop control.

The goal of the DIM project is to generate competitive advantages for SMEs by enabling them to form Digital Twins of their production equipment and use them to optimize the production process. To achieve this goal, methods and algorithms for the data-driven modeling of the digital twin as well as methods and algorithms for the optimization of the production process based on the digital twin will be developed and made publicly available in a platform-independent Python library. On the other hand, beyond the mere provision of the tools, a transfer and preservation of the knowledge generated during the tool development and the developed technologies is aimed at. Through a demand-oriented knowledge and technology transfer, companies should be enabled to independently carry out the development of such systems in the future.