Aug 2026· International Journal of Precision Engineering and Manufacturing· 0 citations· 23 references
TL;DR
This study investigates reinforcement learning (RL) as a means of supporting process optimization in injection molding by constructing an environmental model that reflects the severe class imbalance and the requirement for continuous control.
Abstract
Injection molding is a widely used process in the plastics industry, yet its quality depends heavily on precise adjustment of interdependent process parameters, which has traditionally relied on trial-and-error by skilled engineers. Such reliance is inefficient, time-consuming, and difficult to scale, particularly as the number of controllable parameters increases. This study investigates reinforcement learning (RL) as a means of supporting process optimization in injection molding by constructing an environmental model that reflects the severe class imbalance and the requirement for continuous control. The environmental model for RL-based agents was built using the Injection Molding AI Dataset, where weight balancing was applied to classifiers and anomaly detection methods were separately employed to improve reliability under imbalanced conditions. On this basis, two RL-based agents, Deep Q-Network (DQN) and Multi-Agent Deep Deterministic Policy Gradient (MADDPG), were designed and evaluated for their ability to optimize key process parameters such as injection time, filling time, and screw positions. The comparative analysis showed that while DQN offered advantages in terms of lightweight architecture, stable training behavior, and ease of implementation in discretized settings, MADDPG achieved greater effectiveness in continuous action spaces with multiple interdependent variables. Importantly, the MADDPG-based agent consistently identified non-defective production conditions in a single step, within the simulated environment, suggesting promising industrial applicability in future injection molding systems.
Automated visual defect detection has become an important technology for improving quality control in modern industrial manufacturing. Traditional manual inspection is costly, inefficient, subjective, and difficult to sustain in high-speed production, while rule-based vision systems often fail under changing lighting conditions, surface variations, and new defect types. This paper reviews the development of AI-powered visual defect detection, focusing on machine learning, convolutional neural networks, and YOLO-based real-time detection methods. It explains how deep learning replaces hand-crafted feature design with data-driven representation learning, enabling more accurate recognition of scratches, cracks, pits, edge defects, and other surface anomalies. The paper further discusses a physics-assisted framework that combines heat conduction modelling for synthetic defect generation, anisotropic diffusion for structure-preserving image preprocessing, and gradient descent analysis for training stabilisation under imbalanced datasets. Although these methods improve detection accuracy, robustness, and real-time performance, challenges remain in defect data scarcity, environmental domain shifts, model interpretability, computational cost, and edge deployment. Future research should focus on lightweight models, open datasets, explainable AI, and physics-informed learning.
Xinhao Jiang· MATEC Web of Conferences· 0 citations
In the injection molding industry, the shift toward small-batch production has led to a greater variety of products and smaller batch sizes, necessitating frequent mold changes and efficient quality control, which still largely relies on human operators. This study proposes a comprehensive methodology for evaluating and comparing deep learning-based automatic optical inspection (AOI) strategies to detect complex surface defects in injection-molded parts. Three inspection setups were assessed: static frontal imaging, belt conveyor inspection, and robotic-assisted inspection. The findings reveal clear differences in defect detection capabilities among the methods, with the robotic-assisted approach demonstrating superior performance, achieving higher defect detection accuracy due to its flexibility in optimizing camera angles and positions. The proposed methodology serves as a workflow to systematically evaluate and optimize inspection setups across different parameters, enabling informed decisions about AOI systems design. This research contributes to narrowing the gap between the development of advanced detection algorithms and their industrial application, offering insights into the strategic implementation of AI technologies in quality control processes and enhancing the automatic detection of challenging defects.
Enrico Bovo, X. Wang, G. Lucchetta et al.· Scientific Reports· 0 citations
A data engine which gathers data and improves its performance while executing the task, and demonstrates the ability to learn and reduce the need for expensive verifications over time, while staying within the set error-rate.
Zebin Duan, Norbert Krüger, Juan Heredia et al.· arXiv.org· 0 citations
Additive Manufacturing (AM) plays a vital role in the ongoing industrial revolution. However, quality control remains crucial and challenging due to printing defects or potential cyber-physical intrusions. Image or video-based anomaly detection is a key effort towards addressing these challenges. Various approaches have been explored in this domain, including reconstruction-based, embedding-based, and flow-based methods. Though normalizing flow-based methods address some of the core challenges of unforeseen defects and generalization while maintaining detection performance, existing approaches struggle with tiny/stringing defects common in 3D printing. In a small-data setting, this poses a limitation in generalization. To address these limitations, we propose \textbf{GuidedFlow}, a novel attention-guided normalizing flow model for anomaly detection and localization. GuidedFlow employs a pre-trained ResNet model, fine-tuned on the domain dataset. An attention-guided spatial and temporal flow framework models the dynamics across multiple scales and frames. A Spatio-Temporal Attention Network (SAN) enables the flow model to prioritize relevant contextual cues from input frames. We evaluate GuidedFlow on our AM3D-AD dataset, consisting of benign and anomalous real 3D printed object images and videos. We also conduct a comparative study using the MVTec-AD industrial image anomaly detection dataset. Experimental results demonstrate that GuidedFlow outperforms most of the state-of-the-art models with enhanced detection accuracy and AUROC.
A transferred SISA (Sharded, Isolated, Sliced, and Aggregated) fault diagnosis framework is developed and applied to rolling bearing data, demonstrating a 84.32% decrease in retraining time compared to non-SISA full-retraining while restoring accuracy to the pre-poisoning SISA level.
Emily Yin, Jingyi Yan, Nanhong Liu et al.· 0 citations
Machinery fault detection (MFD) remains heavily reliant on supervised learning, which struggles with the scarcity of fault labels in real-world settings. While reinforcement learning (RL) offers a framework to model the sequential nature of degradation, current ``RL-based''MFD methods reduce the problem to a static contextual bandit (CB) formulation: by ignoring state transitions and discarding the temporal discount factor, they collapse to standard supervised classification. We propose an adversarial inverse reinforcement learning (AIRL) framework that treats MFD as an offline IRL problem. Unlike reconstruction-based approaches that rely on static error margins, or CBs that ignore dynamics, our method recovers an intrinsic"health"reward directly from observational state transitions, requiring neither manual reward engineering nor fault labels. On three run-to-failure benchmarks (HUMS2023, IMS, XJTU-SY), AIRL is the only method achieving non-saturated post-detection consistency across all datasets, while CB baselines fail to detect gradual degradation and reconstruction models collapse into always-anomalous states. Code and data: https://github.com/dhirajneupane/AIRL-MFD-DN.
Dhiraj Neupane, Mohamed Reda Bouadjenek, Richard Dazeley et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
