Process control in a high-noise environment using a limited number of measurements
ABSTRACT In this paper, we develop a hybrid control algorithm that produces control values for processes where only a limited number of function evaluations are available for the control law generation. This situation arises, for example, in stencil printing processes in printed circuit board manufacturing, where the cost associated with multiple function evaluations is prohibitive: The proposed control algorithm is given by a modified version of a constrained conjugated-gradient method, transitioned into a windowed-smoothed block-form of the least-squares affine estimator.
- SourceAvailable from: gatech.edu[Show abstract] [Hide abstract]
ABSTRACT: This paper presents a neural network model for the stencil printing process (SPP) in surface-mount technology (SMT) manufacturing of printed circuit boards (PCBs). A practical model description that decomposes the overall steady-state process in independently modeled subspaces is provided. The neural network model can be updated in real-time procuring a method to control the process by dynamically searching the optimal set point of the control variables. The optimization is performed by minimizing the weighted mean squared error with respect to the desired solder brick height or volume; furthermore, in the case when multiple solutions exist, the set point that yields the lowest variance is used. The process simulator is mainly suitable for offline testing and debugging of more complex closed-loop control algorithms for the SPP optimization providing a common and realistic framework for algorithm performance evaluation. An important consideration in this paper is based on the fact that the estimation of the sampled moments of the probability distributions is made using a statistically significant number of data samples from each board, for each component type, for each printing direction, and for each pad orientation.IEEE Transactions on Electronics Packaging Manufacturing 02/2008; · 1.16 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: A pattern clustering algorithm is proposed in this paper as a statistical quality control technique for diagnosing the solder paste variability when a huge number of binary inspection outputs are involved. To accommodate this goal, a latent variable model is first introduced and incorporated into classical logistic regression model so that the interdependencies between measured physical characteristics and their relationship to the final solder defects can be explained. This probabilistic model also allows a maximum-likelihood principal component analysis (MLPCA) method to recognize the dimension of systematic causes contributing to solder paste variability. The correlated measurement variables are then projected onto the reduced latent space, followed by an appropriate clustering approach over the inspected solder pastes for variation interpretation and quality diagnosing. An application to a real stencil printing process demonstrates that this method facilitates in identifying the root causes of solder paste defects and thereby improving PCB assembly yield.IEEE Transactions on Electronics Packaging Manufacturing 11/2007; · 1.16 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The distillation of massive quantities of solder paste inspection data into relevant quality information allows rapid understanding of the low production yield in PCB assembly. The statistical diagnosis method proposed in this paper provides more meaningful insights into the defect mechanisms than traditional yield analysis methods, which can identify the assignable causes of defects and their effects on yield by integrating MLPCA and logistic regression model. This offers a systematic representation on the impacts of process condition changes to the variation of solder paste profile. The probabilistic latent variable model allows ML estimation to determine the latent space by iteratively maximizing the likelihood function. In contrast to standard PCA, this approach is also efficient for multivariate process analysis when some sample data are missing. The clustering algorithm over the projected regression coefficients onto the latent space is relatively easy to implement with affordable computational effort. Experimental study demonstrates that the statistical interpretation of solder defect distributions can be enhanced by intuitive pattern visualization for process fault identification and variation reduction.01/2009; , ISBN: 978-3-902613-53-0 · 1.16 Impact Factor
PROCESS CONTROL IN HIGH-NOISE ENVIRONMENTS
USING A LIMITED NUMBER OF MEASUREMENTS
The Academic Faculty
Leandro G. Barajas
In Partial Fulfillment
Of the Requirements for the Degree of
Doctor of Philosophy
School of Electrical and Computer Engineering
Georgia Institute of Technology
Copyright © 2003 by Leandro G. Barajas
S T I TUTE • OF • TECHNO LOGY •
S ERV I CE
Process Control in High-Noise Environments
Using a Limited Number of Measurements
Dr. Magnus Egerstedt, Advisor
Dr. David Taylor
Dr. George Vachtsevanos
Date Approved: 03/28/2003
This work is dedicated to my parents Edgar and Alba,
who gave me life and strength of character to become what I am…
… to my brothers Mauricio and Nicolas,
for always believing in and trusting me…
… to my Family,
who surrounded with love as I was growing up…
…and to my wife Elizabeth for a life of happiness.
I would like to express my deepest and sincerest gratitude towards my thesis advisors,
Dr. Magnus B. Egerstedt and Dr. Edward W. Kamen for instructing me in the paths of
knowledge and professional integrity, for their constant support and encouragement and
for giving me the highest standards to live up to as a researcher. Having a mentor of the
stature and trajectory of Professor Kamen provided me with the guidance, support, and
right mentality to be successful in my research career; somehow, he was always able to
foresee problem outcomes in ways I never even imagined; our different points of view
often originated very interesting and constructive practical and theoretical discussions.
With his enthusiasm, drive, and directed thinking, Professor Egerstedt provided
invaluable advice and contributions to this work. I am grateful to him for initiating me
into the field of hybrid systems, and for being able to focus my research in the overall
picture (rather than expending too much time in minute details as I used to). His
continual support, encouragement, and feedback always kept me going, especially during
the exhausting writing process of the papers generated by this work.
I am grateful to Professors David G. Taylor, George J. Vachtsevanos, Aaron D.
Lanterman, and Amy R. Ward for serving as committee members in the proposal
examination and the final dissertation defense, and for their extensive and constructive
comments and suggestions.
I would also like to express my gratitude to Dr. Allen R. Tannenbaum for introducing
me to the world of the computer vision and to the wonders of mathematical morphology.
In addition, for their dedication, help, and inspiration, I want to thank all my graduate
school professors but specially the following: Phillip E. Allen, John F. Dorsey, Paul E.
Hasler, Mark J.T. Smith, Erick I. Verriest, and Yorai Y. Wardi. Furthermore, I want to
express my gratitude to the Universidad Distrital Francisco José de Caldas Professors
Jaime Angulo, Edgar Betancourt, Rodrigo Herrera, and Jorge Pedraza that during my
undergraduate years encouraged and oriented me in the pursuit of a research career that
lead me to this point.
Special thanks to Alex Goldstein for his constant advisement and support during the
implementation and development of this research project; especially for always keeping
the laboratory facilities running smoothly and for managing the projects developed with
all different companies that were involved in this work.
My profound thanks and love to my family, for giving me their constant moral,
emotional, and economical support during these years. Finally, yet importantly, I thank
my wife, Elizabeth for showing me and giving me happiness in life.
This research was supported in part by Siemens Dematic EAS, Norcross, GA, USA,
and by The Julian T. Hightower Chair Professorship in Manufacturing Engineering.
I also give my sincere gratitude to the Center for Board Assembly Research (CBAR) at
the Manufacturing Research Center (MARC) for providing all facilities, equipment,
parts, and supplies and as well as human and technical resources for the successful
culmination of this research project.