Rheometer Quality Inspection system for smart factory based on deep learning

Abstract

Quality conformity inspection using rheometer testing is a critical process in the rubber manufacturing industry, particularly for rubber products intended for automotive applications. Traditional methods rely heavily on manual interpretation by skilled experts, which can introduce subjective variability, inefficiencies, and inconsistencies in defect identification. To overcome these limitations, this study proposes an automated inspection system based on deep learning, integrating Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNNs) to capture both temporal and spatial characteristics of rheometer-generated data. Furthermore, material composition information is embedded into both models as auxiliary inputs to provide product-specific contextual information, enabling the system to generalize across various rubber formulations. The proposed method was validated using a dataset of 30,000 samples collected from real industrial processes. Experimental results demonstrate a high level of classification performance, with the ensemble model achieving an average F1-score of 0.9940. These findings confirm that the system offers superior accuracy and robustness in quality assessment tasks, while significantly reducing reliance on manual labor. The proposed solution represents a scalable and reliable approach to smart factory implementation in the rubber manufacturing industry.