Mechanical & Materials Engineering, Department of


First Advisor

Prahalada Rao

Date of this Version

Fall 10-2019


Published as:

[1] Montazeri, M., Yavari, R., Rao, P., and Boulware, P., 2018, "In-Process Monitoring of Material Cross-Contamination Defects in Laser Powder Bed Fusion," Journal of Manufacturing Science and Engineering, 140(11), pp. 111001-111001-111019. doi: 10.1115/1.4040543

[2] Montazeri, M., and Rao, P., 2018, "Sensor-Based Build Condition Monitoring in Laser Powder Bed Fusion Additive Manufacturing Process Using a Spectral Graph Theoretic Approach," Journal of Manufacturing Science and Engineering, 140(9), pp. 091002-091002-091016. doi: 10.1115/1.4040264

[3] Montazeri, M., Nassar, A. R., Stutzman, C. B., and Rao, P., 2019, "Heterogeneous sensor-based condition monitoring in directed energy deposition," Additive Manufacturing, 30, p. 100916. doi: 10.1016/j.addma.2019.100916

[4] Montazeri, M., Nassar, A. R., Dunbar, A. J., and Rao, P., 2019, "In-process monitoring of porosity in additive manufacturing using optical emission spectroscopy," IISE Transactions, pp. 1-16. doi: 10.1080/24725854.2019.1659525


A DISSERTATION Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Doctor of Philosophy, Major: Engineering (Materials Engineering), Under the Supervision of Prahalada Rao, Ph.D. Lincoln, Nebraska: October 2019

Copyright 2019 Mohammad Montazeri


The goal of this dissertation is to detect the incipient flaws in metal parts made using additive manufacturing processes (3D printing). The key idea is to embed sensors inside a 3D printing machine and conclude whether there are defects in the part as it is being built by analyzing the sensor data using artificial intelligence (machine learning). This is an important area of research, because, despite their revolutionary potential, additive manufacturing processes are yet to find wider acceptance in safety-critical industries, such as aerospace and biomedical, given their propensity to form defects. The presence of defects, such as porosity, can afflict as much as 20% of additive manufactured parts. This poor process consistency necessitates an approach wherein flaws are not only detected but also promptly corrected inside the machine. This dissertation takes the critical step in addressing the first of the above, i.e., detection of flaws using in-process sensor signatures.

Accordingly, the objective of this work is to develop and apply a new class of machine learning algorithms motivated from the domain of spectral graph theory to analyze the in-process sensor data, and subsequently, detect the formation of part defects. Defects in additive manufacturing originate due to four main reasons, namely, material, process parameters, part design, and machine kinematics. In this work, the efficacy of the graph theoretic approach is determined to detect defects that occur in all the above four contexts. As an example, in Chapter 4, flaws such as lack-of-fusion porosity due to poor choice of process parameters in additive manufacturing are identified with statistical accuracy exceeding 80%. As a comparison, the accuracy of existing conventional statistical methods is less than 65%.

Advisor: Prahalada Rao