Session: Research Posters

Paper Number: 119823

119823 - Vibration Analysis for Fault Detection in Fused Filament Fabrication Printing 

This work examines the use of accelerometers to identify vibrational patterns that can effectively predict the state of a 3D printer, which could be helpful for predictive maintenance. The 3D printer had three 3-axis accelerometers positioned in different locations - near the nozzle, on the frame, and the print bed. These three accelerometers collected data via LabVIEW to observe the vibrational pattern as the nozzle transitioned from unclogged to wholly clogged. Prototypes using a simple rectangular shape and a more complex Octopus shape were fabricated and evaluated. As the nozzle condition changes, the printer exhibits a different vibrational pattern. After the collection of data, it underwent processing utilizing MATLAB software. To comprehensively analyze the collected data, techniques involving the Fast Fourier Transform (FFT) and machine learning models were employed, namely Spectrogram, Principal Component Analysis (PCA), and Support Vector Machine (SVM), which played integral roles in extracting meaningful insights and patterns from the dataset.
The findings from our research suggest that vibration analysis holds promise as a predictive tool for assessing the state of a 3D printer, specifically with regard to changes occurring in the nozzle. By monitoring and analyzing the vibrational patterns exhibited by the printer during operation, we have observed correlations between these patterns and the condition of the nozzle. These results indicate that alterations in the nozzle, such as clogs or blockages, can manifest as distinct vibrational signatures within the 3D printer. By detecting and interpreting these vibrations, it becomes possible to infer the state of the nozzle and identify potential issues before they significantly impact the printing process. However, the position of the accelerometers is crucial for vibration-based fault detection. Precisely, the sensor closest to the nozzle could predict the state of the 3D printer faster at 71% sensitive compared to accelerometers mounted on the frame and print bed. While the accelerometer mounted near the nozzle can quickly predict the condition of the nozzle, the accelerometer mounted on the bed can predict wobbling or wrapping during fabrication. Despite the suboptimal sensitivity of the accelerometer positioned on the frame, it still holds utility in verifying the presence of external factors contributing to the observed vibrations. While its positioning may limit its ability to capture subtle vibrations, it can effectively discern and differentiate between internal and external vibration sources. Therefore, the model presented in this study is appropriate for vibrational fault detection in 3D printers. Vibration analysis in 3D printing allows for developing real-time fault detection mechanisms, providing insight into the printer's health and operational integrity. Improving the reliability, efficiency, and quality of 3D printing processes is essential for any business that relies on this technology. One way to achieve this goal is using vibration analysis for predictive maintenance. It will reduce downtime and maintenance costs and improve product quality by detecting issues such as nozzle clogs or wobbling during fabrication. 

Presenting Author: Alexander Isiani Louisiana Tech University

Presenting Author Biography: I am a Ph.D. Student in Micro and Nanoscale Systems Engineering at the Louisiana Tech University, Ruston, Louisiana, USA. I am broadly interested in 3D printer-related research and real-time fault detection analysis. My research has focused on exploring the viability of vibration detection for identifying potential clogs in 3D printer nozzles. I am currently concentrating on detecting any banding and skewness that may occur during fabrication. My goal is to improve the reliability and quality of printed products. Feel free to reach out if you have similar interests and would like to collaborate on a project or have general questions about my work.

Authors:

Alexander Isiani Louisiana Tech University
Dr. Kelly Crittenden Louisiana Tech University
Dr. Leland Weiss Louisiana Tech University