Scribd
Upload
1 views3 pages

Project Synopsis

This project focuses on integrating computer vision and machine learning algorithms with 3D printing to enhance print quality and detect errors in Fused Deposition Modeling (FDM). The research aims to develop predictive models by analyzing various 3D printing data and comparing different machine learning techniques. Expected outcomes include a fault detection system and a publication detailing the findings of the study.

Uploaded by

Shikhar N
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
1 views3 pages

Project Synopsis

This project focuses on integrating computer vision and machine learning algorithms with 3D printing to enhance print quality and detect errors in Fused Deposition Modeling (FDM). The research aims to develop predictive models by analyzing various 3D printing data and comparing different machine learning techniques. Expected outcomes include a fault detection system and a publication detailing the findings of the study.

Uploaded by

Shikhar N
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

MECHANICAL ENGINEERING PROJECT 1: - MEH701

Title: Integration of computer vision and machine learning algorithms in 3D printing for improved
print quality and error detection.

Project Supervisor: Dr. Ankit Sahai


Submitted by: Saloni Upadhyay
B. Tech (Mechanical with C.S. Specialization) – 4th year.
Dayal Bagh Educational Institute

Abstract
Additive manufacturing (AM) offers significant advantages over traditional manufacturing
methods, including material efficiency and design flexibility, yet it faces challenges in ensuring
consistent part quality. Starting a new 3D print from scratch when printing errors are detected
in comparison to the intended design wastes material, depletes machine and human resources,
and consumes more energy. Its inconsistency and intrinsic defects (surface roughness and
geometric inaccuracies) also hinder its application in several areas including the medical and
aerospace industries. This report focuses on developing a prediction model for defect detection
of components produced through Fused Deposition Modeling (FDM), a popular AM technique.
Utilizing machine learning (ML), deep learning (DL) algorithms, and Transfer learning(TL), the
study aims to analyze different forms of 3D Printing data (like data from environmental sensors
and image datasets of defective and non-defective printed parts) and develop predictive
models for fault detection in FDM printed parts. By comparing various ML, DL, and TL models,
the study aims to demonstrate the superior performance of these approaches in predicting
defects and ensuring high-quality outputs. The integration of these advanced algorithms with
real-time monitoring systems will also enhance the accuracy of quality predictions, enabling
dynamic adjustments to the printing process.

Objective
The objective of this research is to develop a predictive model for detecting defects in parts
produced using Fused Deposition Modeling (FDM), a prominent additive manufacturing (AM)
technique. The study aims to:
1. Analyze multiple forms of 3D printing data, including environmental sensor readings
and image datasets of defective and non-defective printed parts, to identify key
indicators of printing errors.
2. Utilize machine learning (ML), deep learning (DL), and transfer learning (TL) algorithms
to create robust models capable of predicting defects such as surface roughness and
geometric inaccuracies in FDM printed parts.
3. Compare the performance of various ML, DL, and TL models in detecting defects and
ensuring the production of high-quality outputs.

Methodology (with Expected timeline)


Selecting appropriate dataset from publicly
available datasets, and also specify the
exact problem statement (which defects to
be predicted). [3/09/24 - 8/09/24]

Data Extraction, Cleaning and Analysis


[9/09/24 - 22/09/24]

Selection of different ML, DL, and TL


Techniques to be applied, according to the
selected dataset. [23/09/24 - 30/09/24]

Implementation of different algorithms


using Python. [01/10/24 - 31/10/24]

Calculating performance metrices and


comparison of different algorithms.
[01/11/24 - 15/15/24]

Result Analysis, Conclusion and


Documentation [16/11/24 - 16/12/24]
Expected Project Outcome
- A Predictive modeling system for fault detection in AM-manufactured parts.
- A publication in a journal/conference paper detailing the comparison and results of the
different algorithms used.

You might also like