Reducing the number of rejects in injection molding with the Meso Assistant

Greater quality and sustainability in injection molding production

In the plastics industry, injection molding is a widely used manufacturing process in which plastic granulate is liquefied and placed in a defined shape under high pressure. This process can be applied to produce plastic parts in complex shapes and in large quantities at low cost – as long as the reject rate remains low even at high cycle rates.

Fluctuations in material, tool wear and environmental factors such as temperature or humidity often result in typical process errors that lead to rejects. Not only does such waste increase material and energy consumption, it also results in additional work and production interruptions – both of which should be avoided for cost efficiency and sustainability reasons. Together with our partner SHS plus, we have developed the Meso Assistant, an assistance system for plastic injection molding that solves this problem.

Reduce production waste with AI

AI solutions such as the Meso Assistant can help to reduce waste in a variety of manufacturing processes. When it comes to plastic injection molding, there are two particular challenges:

• Product diversity is often extremely high.
• The machine fleets are often heterogeneous.

Specifically, this means: Various parts in widely differing shapes and with a broad range of granulates are produced on numerous machines over the course of a week or even a day. Yet there are only valid quality assessments for small samples of these manufactured parts. Moreover, these tests are often only carried out hours or days after production. This is too late, however, to correct the ongoing process should deviations occur. The Meso Assistant therefore works without historically validated quality data and predicts deviations before relevant rejects can occur.

Intelligent parameterization with real-time machine data

The Meso Assistant can be retrofitted for different tools on different machines. Once the injection molding machine is set up for a mold, the normal state of the machine is recorded and continuous anomaly detection is performed using real-time data. The parameterization with this data makes it possible to cater for the heterogeneity of the machinery and the variety of parts to be produced.

When an anomaly in the process is reliably detected, it is indicated to the machine operators. The Meso Assistant also provides possible sources of the anomaly, which simplifies and speeds up the search for the cause and rectification. This makes it possible to adjust production and material composition at an early stage, which reduces faulty results and therefore rejects. For many manufacturing companies, this is not only interesting for cost reasons, but also in terms of sustainability and ESG obligations.