Non-destructive measurement of surface roughness of 3D printed sample

• Main Author: Ong, Jun Hao
• Other Authors: Fan Zheng, David
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Mechanical engineering
• Online Access: https://hdl.handle.net/10356/157918
• Institution: Nanyang Technological University

Additive manufacturing has since gained popularity in the manufacturing industry. However, achieving high surface finishing while producing complex designs, such as those with internal features and multiple curvatures is a challenge that manufacturers might face these days. A well-established method of surface finishing known as abrasive flow machining (AFM) can be used to tackle this issue. Although AFM is an advanced technology, it is still not able to track the surface roughness of the object during the polishing process. Operators have to resort to multiple sessions of polishing which is tedious and time-consuming. Therefore, this report serves to present an online monitoring method for the internal surface roughness of metallic components through the use of a frequency-based ultrasonic testing method. The method used consists of using an auto-alignment rig with a delay line transducer attached. Along with the help of MATLAB software, the monitoring process will be automated throughout the process of surface finishing using the AFM process. The reconstructed roughness obtained from the tests will then be validated using measurements taken from a stylus profilometer. Although the results obtained were inconsistent and inaccurate due to overestimations, the proposed method was able to capture the rate at which surface roughness decreases during the polishing process. With further calibrations and adjustments made, it is possible to improve the results. Lastly, to simulate real-world scenarios, offline ultrasonic testing was done on a rounded channel sample with a curved target surface and an aluminium plate with a flat target surface for comparison. Results have shown that a transducer at a low frequency is more suitable for roughness measurements of curved target surfaces and due to the effects of curvatures, results are more susceptible to overestimations if not well-compensated.