Thesis Format

Integrated Article

Improving Fused Filament Fabrication Additive Manufacturing through Computer Vision Analysis and Fabrication Optimization

Aliaksei Petsiuk, Western UniversityFollow

Degree

Doctor of Philosophy

Program

Electrical and Computer Engineering

Supervisor

Pearce, Joshua M.

Abstract

Additive manufacturing (AM), also known as 3-D printing, is one of the fundamental elements of Industry 4.0. According to ASTM standards, AM can be classified by production principles, types of raw materials, energy sources, and fabrication volumes. Fused filament fabrication (FFF) is one of the most accessible technologies that offers independent manufacturers great opportunities due to its simplicity, scalability, and low cost.

Modern 3-D printing is moving from single-material prototyping to complex multi-material product creation. It is firmly established in a wide range of applications, significantly expanding manufacturing horizons, providing innovative design capabilities, and improving product quality through the optimal combination of properties often impossible to achieve with traditional methods.

Despite the great potential and current exponential growth in production, AM, however, faces challenges that affect its adoption, efficiency, and product quality. An analysis of user databases shows the average failure rate is about 20 percent. The likelihood of manufacturing defects grows with the size of the object and the time required to print it, which can lead to increased material waste resulting from even a small failure rate. The ability to automatically detect anomalies in AM will greatly help reduce the wastage of material and time spent reproducing failed prints.

To strengthen the capabilities of AM technology, it is necessary to optimize the process of preparing a part for 3-D printing (slicing) and provide analysis systems that can detect and minimize the impact of emerging defects. The increasing complexity of geometric shapes and the number of materials used require optimization of fabrication processes and layer-by-layer monitoring of production processes for timely response.

This work presents several conceptually new approaches to FFF AM 3-D printer work volume monitoring and anomaly detection and localization based on monocular computer vision, machine learning, and synthetic data, as well as to increasing efficiency and reducing production waste in multi-color fabrication. Thus, a system for automatically creating a labeled G-code-based synthetic 3-D printing dataset was developed, providing layer-by-layer semantic segmentation of a printing part and its structural elements during the manufacturing process. A method has been developed for the procedural simulation of ideal fabrication by generating layer-wise photorealistic images of the manufactured part for further use as references for visual analysis at each manufacturing stage. To monitor the height, external contour, and internal structure of the manufactured object, a multi-stage approach based on computer vision has been developed, which allows analyzing images of each printed layer for compliance with the source 3-D model. A new fabrication method has been developed for multi-color printing to reduce energy and material costs for single-nozzle systems. The presented developments formed the basis for the concept of multifaceted visual analysis of 3-D printing processes. This will help improve FFF AM technology and reduce the amount of time, materials, and energy required to fabricate physical objects.

Summary for Lay Audience

Additive manufacturing (AM), or 3-D printing, is the cornerstone of the next-generation industry. Modern 3-D printing is moving beyond single-material prototyping to the creation of complex, multi-material products, expanding manufacturing capabilities across multiple fields, unlocking innovative design potential, and enhancing product quality through unique combinations of properties and geometries unachievable by traditional methods.

Exponential production growth, however, demonstrates a significant increase in plastic waste, which is also exacerbated by a high failure rate averaging 20%. This highlights the need for automatic failure detection systems in AM processes to reduce material, time, and energy losses.

This work demonstrates approaches to detecting and localizing 3-D printing defects based on computer vision, machine learning, and synthetic data, as well as to increasing efficiency and reducing production waste in multi-color fabrication. The presented methods will help improve additive manufacturing technology and reduce the amount of time, materials, and energy required to fabricate three-dimensional objects.

Recommended Citation

Petsiuk, Aliaksei, "Improving Fused Filament Fabrication Additive Manufacturing through Computer Vision Analysis and Fabrication Optimization" (2024). Electronic Thesis and Dissertation Repository. 10147.
https://ir.lib.uwo.ca/etd/10147