At Kuteno 2023, the Plastics Center in Leipzig (KUZ) is presenting a groundbreaking AI-supported assistance system that revolutionizes the dosing of liquid paint in the injection molding process. This innovative technology not only promises significant time savings when coloring molded parts, but also enables seamless integration of coloring into the injection molding manufacturing process.
AI-based online prescription and dosing
The KUZ has developed an automated, AI-supported process that overshadows conventional methods of coloring in the injection molding process. By using inline color measurement and online dosing of liquid colors, the entire coloring of thermoplastics can be integrated directly into the ongoing injection molding process. This means that the formulation and dosage of the color takes place in real time on the machine. Such an innovation significantly reduces the time required and also enables continuous quality control.
Color measurement and adjustment in real-time process
As part of the “Online Color Mixing” research project, a test stand was developed that seamlessly links color measurement and dosing with the injection molding process. The user can specify the desired target color using a control computer. A color recipe for available color ranges is then calculated. It is sent as a control signal to the color dosing station. The molded parts produced are measured using a color sensor and compared with the target color. Through iterative adjustments, the shape color is automatically adjusted to the defined target color. Two different methods were explored. The first uses a physical color model, while the second is based on a neural network and relies on samples to calculate the correct proportions of the color channels for a defined target color.
Self-optimizing AI paint dispensing systems
The current research project “AI Color” builds on these findings and aims to develop a self-optimizing paint dosing system for liquid paints. This system will help users optimize the color recipe for a desired target color in just a few iteration steps. The goal is to minimize the perceived color difference ΔE with each iteration until it reaches an optimal level. This is important because a ΔE <1.0 is only visible to trained eyes, and a ΔE <0.5 is considered an almost imperceptible deviation. The use of a data-based model in the form of a neural network is particularly promising. It has proven to be a practical alternative to a physical color model. This model is constantly learning and expanding its database with each new cycle. This also makes the time-consuming creation of calibration series unnecessary and thus saves resources and working time.
Further focal points of the KUZ on the Kuteno
Apart from this groundbreaking AI technology, the KUZ is presenting its expert knowledge in the field of life cycle assessment at the trade fair in Rheda-Wiedenbrück. This analysis serves as the basis for calculating the ecological footprint and helps to make products and processes more sustainable. The KUZ shows how the targeted selection of materials has a significant impact on potential greenhouse gas emissions. It highlights which types of plastic can enable a significant reduction in emissions.
