Additive manufacturing, particularly fused deposition modeling, is transforming modern production by enabling rapid prototyping and complex part fabrication. However, its layer-by-layer process remains vulnerable to faults such as nozzle clogging, filament runout, and layer misalignment, which compromise print quality and reliability. Traditional inspection methods are costly, time-intensive, and often limited to post-process analysis, making them unsuitable for real-time intervention. In this current study, the authors developed a novel, low-cost, and portable faultdetection system that leverages multimodal sensor fusion and artificial intelligence for real-time monitoring in FDM-based 3D printing. The system integrates acoustic, vibration, and thermal sensing into a non-intrusive architecture, capturing complementary data streams that reflect both mechanical and process-related anomalies. Acoustic and thermal sensors operate in a fully contactless manner, while the vibration sensor requires minimal attachment such that it will not interfere with printer hardware, thereby preserving portability and ease of deployment. The multimodal signals are processed into spectrograms and time-frequency features, which are classified using convolutional neural networks for intelligent fault detection. The proposed system advances Industry 4.0 objectives by offering an affordable, scalable, and practical monitoring solution that improves faultdetection accuracy, reduces waste, and supports sustainable, adaptive manufacturing.
Additive manufacturing (AM), driven by advancements in digital design and automation, is increasingly revolutionizing the production landscape by enabling rapid prototyping, personalized fabrication, and material-efficient processes. Among the various AM modalities, fused deposition modeling (FDM) holds a prominent position, due to its cost-effectiveness, material flexibility, and widespread accessibility in both industrial and research contexts. Despite these advantages, FDM processes remain vulnerable to a spectrum of faults, ranging from nozzle clogging and filament runout to layer shifting and misalignment, which compromise geometrical fidelity, mechanical properties, and functional integrity of manufactured parts (Sampedro, Rachmawati, Kim & Lee, 2022;Deokar, Kumar & Singh, 2025).
- ———————————————— —————————————————————————————————–6 International Journal of Engineering Research and Innovation | v17, n2, Fall/Winter 2025 Artificial intelligence, particularly machine learning and deep learning algorithms, now plays a central role in sensor-based monitoring for AM. Convolutional neural networks (CNNs), SVMs, and deep adversarial learning models have been applied to spectrogram and timefrequency representations of multimodal sensor signals, yielding superior fault discrimination and process diagnostics. Kadam, Kumar, Bongale, Wazarkar, Kamat, and Patil (2021) achieved maximum accuracy in fault diagnosis by integrating SVM with pre-trained AlexNet models for layerwise defect detection, a strategy effective both in offline training and online implementation (Tan, Huang, Liu, Li & Wu, 2023). Despite promising results, several key limitations persist.
Many deployed systems remain non-portable, intrusive, or cost-prohibitive, and struggle to generalize across different printer architectures and materials. Moreover, few systems robustly support closed-loop control for real-time intervention and process optimization. Addressing these gaps is critical to advancing scalable, adaptive, and intelligent manufacturing (Behseresht, Love, Valdez Pastrana & Park, 2024). In this context, the authors of this current study introduce a novel, portable, and low-cost multimodal sensor fusion system for real-time fault detection in FDM processes. The system integrates acoustic, vibration, and thermal sensing in a combination of contactless and minimally intrusive configurations, enabling seamless deployment across heterogeneous printer environments without hardware modification. Unlike traditional monitoring systems that rely on expensive instrumentation or are restricted to specific platforms, this approach emphasizes accessibility, scalability, and adaptability. AI-driven classification of spectrogram and timefrequency features ensures robust anomaly detection and real-time process feedback, addressing common faults such as nozzle clogging, filament runout, and layer misalignment. By leveraging multimodal inputs, the system compensates for the limitations of individual sensors, achieving higher resilience to environmental noise and variability in operating conditions. Beyond technical improvements, the proposed framework directly supports Industry 4.0 objectives by advancing sustainable, automated, and data-driven manufacturing. It contributes to reducing material waste, lowering energy consumption, and improving production efficiency, thereby promoting wider adoption of additive manufacturing in industrial, research, and educational settings (Chen, Yao, Feng, Chew & Moon, 2023).
To address these challenges, the objective of this research study was to design and validate a portable, low-cost, and minimally intrusive fault-detection framework for FDM-based 3D printing using multimodal sensor fusion. Specifically, the goals were to: (1) develop an integrated sensing architecture to combine acoustic, vibration, and thermal data for comprehensive process monitoring;
(2) transform multimodal signals into time-frequency representations suitable for AI-based analysis; and, (3) implement and assess convolutional neural network models for real-time classification of common FDM faults. Through these goals, the authors sought to advance accessible, scalable, and intelligent quality-assurance solutions aligned with Industry 4.0 manufacturing environments.
Fused deposition modeling (FDM) presents a complex thermo-mechanical environment in which subtle variations in heat transfer, polymer flow, and toolpath execution influence part quality. In the Introduction section above, the authors outlines general challenges and common faults; the deeper technical mechanisms underlying these issues, however, warrant further examination. FDM stability depends strongly on the transient thermal field around the nozzle and build surface, the viscoelastic behavior of semimolten filament during deposition, and the synchronized operation of motion subsy
This content is AI-processed based on open access ArXiv data.