Structural Health Monitoring of Additively Manufactured Components

Authors

Alireza Modir, Florida International UniversityFollow

Document Type

Dissertation

Degree

Doctor of Philosophy (PhD)

Major/Program

Mechanical Engineering

First Advisor's Name

Ibrahim Nur Tansel

First Advisor's Committee Title

Committee chair

Second Advisor's Name

Yiding Cao

Second Advisor's Committee Title

committee member

Third Advisor's Name

Armin Mehrabi

Third Advisor's Committee Title

committee member

Fourth Advisor's Name

Pezhman Mardanpour

Fourth Advisor's Committee Title

committee member

Keywords

Deep Learning, CNN, Structural Health Monitoing, Additive Manufacturing, LSTM, SuRE Method

Date of Defense

3-6-2023

Abstract

Structural Health Monitoring (SHM) methods evaluate the condition of machine components by using permanently installed sensors to assure the safe operation of machines, reduce maintenance costs, and eliminate the need for periodic inspections. Additive manufacturing (AM) has become a new production tool for the creation of complex parts and the reduction of the number of assembled components. The customized design of the internal geometry of AM parts is another advantage of it over conventional methods. In the present study, Surface Response to Excitation (SuRE) method was applied to additively manufactured parts in order to detect, localize, and evaluate the severity of the structural defect or applied load on the structure. SuRE method excites the surface of the structure by mechanical waves using a piezoelectric element. One or multiple piezoelectric elements are used to monitor the dynamic response to excitation at desired locations. To quantify the changes in the frequency spectrum of the monitored signals, the sum of the squared differences (SSD) can be used.

In this dissertation, three deep learning models (1D-CNN, 2D-CNN, and LSTM) have been developed for the classification of the recorded data. While 1D-CNN and LSTM models work directly with time-domain data, 2D-CNN requires a signal processing step to convert the sensor data into images. Four major studies have been performed for condition monitoring of AM components. First, the wave propagation characteristics and fault localization in AM parts were examined with respect to infill type. In the second study, identical test specimens were fabricated with different print orientations, and three excitation signals were used to analyze the time domain response, compare the wave travel speed, and distinguish the data based on the print orientation and applied load. In the third study, the SSD index and the abovementioned DL algorithms are compared for estimating the damage length. In the fourth study, load sensing was investigated on conventionally and additively manufactured metallic bars using 2D-CNN and LSTM models. The results showed that the SSD metric successfully detects anomalies in the structure, while it is not as accurate as deep learning models in estimating the damage severity.

Identifier

FIDC011003

ORCID

https://orcid.org/0000-0002-2606-2424

Previously Published In

Modir, A., & Tansel, I. (2022). Structural Health Monitoring of Additively Manufactured Parts by Combining Infill Design, Multiple Pulse Width Excitation (MPWE), and Deep Learning. Journal of Vibration Engineering & Technologies, 10(8), 3227-3238.

Modir, A., & Tansel, I. (2022). Analysis of Force Sensing Accuracy by Using SHM Methods on Conventionally Manufactured and Additively Manufactured Small Polymer Parts. Polymers, 14(18), 3755.

Modir, A., Tansel, I., Perez, D., & Lamy, B. (2022, May). Loading detection in 3D printed polymer parts using the MWPE method and deep learning. In 2022 8th International Conference on Control, Decision and Information Technologies (CoDIT) (Vol. 1, pp. 561-564). IEEE.

Modir, A., & Tansel, I. (2021). Wave Propagation and Structural Health Monitoring Application on Parts Fabricated by Additive Manufacturing. Automation, 2(3), 173-186.

Tansel, I., & Modir, A. (2021). New Excitation (Multiple Width Pulse Excitation (MWPE)) Method for SHM Systems” Part 1: Visualization of Time-Frequency Domain Characteristics. STRUCTURAL HEALTH MONITORING 2021.

Modir, A., & Tansel, I. (2021). New Excitation (Multiple Width Pulse Excitation (MWPE)) Method for SHM Systems— Part 2: Classification of Time-Frequency Domain Characteristics with 2DSSD and CNN. STRUCTURAL HEALTH MONITORING 2021.

Modir, A., & Tansel, I. (2021). Implementation of the surface thickness on additively manufactured parts for estimation of the loading location. Smart Materials and Structures, 30(2), 025032.

Mohamed, A. F., Modir, A., Tansel, I. N., & Uragun, B. (2019, June). Detection of compressive forces applied to tubes and estimation of their locations with the surface response to excitation (SuRE) method. In 2019 9th International Conference on Recent Advances in Space Technologies (RAST) (pp. 83-88). IEEE.

Mohamed, A. F., Modir, A., Shah, K. Y., & Tansel, I. (2019). Control of the Building Parameters of Additively Manufactured Polymer Parts for More Effective Implementation of Structural Health Monitoring (SHM) Methods. Structural Health Monitoring 2019.

Recommended Citation

Modir, Alireza, "Structural Health Monitoring of Additively Manufactured Components" (2023). FIU Electronic Theses and Dissertations. 5320.
https://digitalcommons.fiu.edu/etd/5320