Machine learning-based optimization workflow of the homogeneity of spunbond nonwovens with human validation

TL;DR

The work tackles optimizing the homogeneity of spunbond nonwovens by blending a computationally expensive microstructure simulator with data-driven surrogates trained on informed data and validated by human experts. It develops a five-stage workflow (parameter selection, knowledge-integrated data collection, model evaluation, training/testing, and human-in-the-loop optimization) to predict the CV-based homogeneity across multiple grid resolutions. Among several regression approaches, an artificial neural network provides the best accuracy and scalable surrogate performance, enabling near real-time exploration of the parameter space via a visualization tool, with offline human validation ensuring aesthetics. The approach reduces computational costs, preserves physical validity with expert knowledge, and demonstrates potential for real-time optimization in manufacturing settings where demand-driven production must be quickly adapted. The practical impact lies in faster, visually guided optimization of nonwoven quality with minimal trial-and-error and validated outputs.

Abstract

In the last ten years, the average annual growth rate of nonwoven production was 4%. In 2020 and 2021, nonwoven production has increased even further due to the huge demand for nonwoven products needed for protective clothing such as FFP2 masks to combat the COVID19 pandemic. Optimizing the production process is still a challenge due to its high nonlinearity. In this paper, we present a machine learning-based optimization workflow aimed at improving the homogeneity of spunbond nonwovens. The optimization workflow is based on a mathematical model that simulates the microstructures of nonwovens. Based on trainingy data coming from this simulator, different machine learning algorithms are trained in order to find a surrogate model for the time-consuming simulator. Human validation is employed to verify the outputs of machine learning algorithms by assessing the aesthetics of the nonwovens. We include scientific and expert knowledge into the training data to reduce the computational costs involved in the optimization process. We demonstrate the necessity and effectiveness of our workflow in optimizing the homogeneity of nonwovens.

Figures (10)

• Figure 1: Principle of a spunbond process and necessary simulation models (due to gramsch:2020).
• Figure 2: Information flow of the knowledge integration in the proposed machine learning method. Diagram adapted from von2019informed.
• Figure 3: Workflow of machine learning-based optimization of spunbond nonwovens. Different stages are colored corresponding to the section numbers in this chapter.
• Figure 4: Simulation construction based on $\sigma_1$ and $\sigma_2$ values for sample size 5 x 5cm: Orange area shows the sample extracted for analysis. Green and cyan filaments are simulated in order to compute the sample, brown filament is part of the nonwoven but not simulated.
• Figure 5: Histogram depicting the relative deviation in CV values due to discretization step size (blue) and statistical uncertainty (red).