American University of Sharjah: Crack Detection Project Report

This report details a graduate project conducted at the American University of Sharjah, focusing on the detection of millimeter cracks in metallic objects. The project utilized near-field NDT/E technology and a polynomial classifier to identify and characterize small cracks. The report outlines the project's aims, objectives, and the student's role as a Graduate Teaching Assistant, including responsibilities in project design, simulation, and analysis. The methodology involved the use of waveguides for signal transmission and the development of mathematical models to correlate crack depth and width. Challenges included obtaining accurate sample data and mitigating noise in the system, which were addressed through digital signal processing and filtering techniques. The project successfully developed a system capable of accurately determining crack dimensions, demonstrating the effectiveness of the proposed techniques. The report includes details on team collaboration, technical problems faced, and the creative work strategies implemented. Overall, the project highlights the application of engineering skills in developing a precise crack detection system.

Competency Demonstration Report (CDR)
Career Episode 2

CE 2.1: Introduction to the Undertaken Project
Name of the Project : Detection of milli-meter cracks
Geographical Location : Sharjah, United Arab Emirates
Project Duration : August 2014 - June 2016
Organization : American University of Sharjah
Position in the Project : Graduate Teaching Assistant
CE 2.2: Background of the Project
CE 2.2.1: Features of the Undertaken Project
Whether harmonic or static loading, the metallic objects and structures generally bound
to ensure the amount of loading stressed over them. Though the harmonic loading break point are
comparatively less than the static loading. The structural integrity of metallic objects may
collapse within shorter time-period as compared to static loaded material. At certain point due to
metallic objects, the harmonic loading gets weakened which are generally known as the “stress
points”. Metallic defragmentation is the main cause of this stress point occurrence, while in the
manufacturing processes or in the presence of flaw materials in the process because of using the
manufacturing machinery in excessive terms. To form linear cracks these fatigues accumulates
which cannot be seen by naked eyes. The length, depth and parameter are micro-level sized and
most technology is not able to detect the linear cracks with proper resolution or accuracy.
Page 1 of 9

CE 2.2.2: Aims and Objectives
In this project, I have worked with the vital objective for evaluating and proposing the
sub millimeter system with the application of patter recognition technique. I have proposed for
the identification and characterization of small cracks on the metallic objects. For the completion
of the project while ensuring the application of proper standard for project, I have developed the
following objectives:
 To evaluate the various patter recognition technique that can be utilized in the project;
 To develop the appropriate theoretical concept for the pattern recognition techniques;
 To analyze the components based on the application of NDR technique;
 To simulate the developed system for evaluating the efficiency and accuracy;
CE 2.2.3: Characteristic of my work in the Undertaken Project
During the initial phase of the project, I have studied and observed that the formation of
fatigues or Sub-millimeter cracks weaken the metallic objects and life span of the elements. In
addition to that, the sub-millimeter cracks are responsible for wreaking the structural elements of
the components. In this project, I have further developed the probabilistic and mathematical
reasoning for examining the variation of the various polynomial class. Apart from that, I have
also developed the mobility features for providing the low cost system.
CE 2.2.4: Organizational Structure Associated with this Project
For the completion of this project, I was accompanied by three other students who have
worked alongside me for the completion of the crack detection system.
Page 2 of 9

CE 2.2.5: Responsibilities Undertaken by me
I had taken multidimensional responsibility in this project for ensuring the proper
implementation of the system. In addition to that, I have more responsibility of distributing the
work load among the tea members based on their engineering knowledge. Moreover, I had
developed the appropriate system within the proper application of theoretical concepts and
conducted the simulation study of the developed model.
Page 3 of 9

CE 2.3: Distinctive Engineering Activity
CE 2.3.1: Theory and Vital Components of the Project
The technology that I made effort to introduce in this project is the ‘near-field NDT/E’. I
have observed that this cutting technology has been the mostly researched and has its direct
implementation over the microwave technologies. This technology is still not old enough for
commercial testing. This approach is intended to determine the Sub-millimeter cracks physical
parameters that are located over the metallic objects that deal with the mechanical means. In this
project, I have used Polynomial Classifier to work in sub-dimension in the input matrix. I have
created more rooms for expansion of polynomial that accommodates the non-linear terms of
input and convert it to simpler term along with their respective expansion by spanning their
matrices. Eigen-vectors have been cross-multiplied in pseudo inverse the Eigen-vectors to get an
infinity number of iteration. The crack point that I have observed is very small and is negligible
margin of error.
CE 2.3.2: Application of Appropriate Engineering Skills and Knowledge
I have used the waveguide that allowed the distribution of the lobes that helps in the
transmission to distribute equally, along the facing point of the waveguide. This helped me to
allow the signal to hit the surface of the metal uniformly and directly with higher reactive
components.
Near Field < 2 D2
λ
Page 4 of 9

The transmitter and receiver is used by the rectangular waveguide mainly to obtain a DC
result along with the fluctuation while the micro-crack hits the signal. TE or TM propagation
mode could be operated by mode element that represents the propagation of a magnetic field or
any electrical field. I have observed that the nearest electrical field is used to maintain a
reflection coefficient of -1.
K band=18−26.5GHz
The K-band rectangular waveguide is used as a rectangular waveguide in the above
equation. The cut-off frequency is in which the dominant mode is allowed to wave to propagate
which is represented as TM10 :
f c=14.047 GHz
I have recorded the phenomena of overshooting that has occurred at the crack. This has
allowed me to keep the varying crack. Some mathematical relationship was developed to learn
about the system in every combination of depth and width cracks.
CE 2.3.3: Task Accomplishment and Delegation
I have evaluated and finalized the work design required for the development of the crack
detection and measurement system. I have utilized my knowledge and applied them efficiently
for the probabilistic reasoning, polynomial classifier, neural networking and various
mathematical models.
CE 2.3.4: Technical Problem and Difficulties Faced
Issue: I have categorized the sample size in two types based in the size of the cracks for
integrating the system learning capabilities for width and depth of the crack. I have therefore
Page 5 of 9

defined variant depths and constant width and another category for constant depths and variant
widths. I have ensured that each crack creates its individual overshoot when the waveguide form
orthogonal to the piece of metal. I have found significant issue when I was unable to obtain the
required value for the sample. I have examined the system and found that the cheap equipments
used in the project was including several noise within the crack.
Solution: For mitigating the issue, I have polished the digital signal processing and
implemented some basic but very effective filters and regression techniques. I have further
considered the overshoot for the maximum point an integrated the mathematical model
identifying the different cracks. Through this I was successful in obtaining the data from 100
cracks and develop the required matrix for using as input.
Figure: Analysis for the crack identification
CE 2.3.5: Creative Work Strategies
In order to develop creative work, I have created matrix in the training phase of the
project. I have used the first 15 columns for inputting the scan readings. I have further ensured
that each scan data is comprised with 3 different cracks and for each crack, 5 individual readings
have been considered. I have further utilized polymer vector classifier for evaluating the width
Page 6 of 9

of the crack. For examining the vector used in the system, I have used two different types of
input. For the first type, I have considered input vector obtained from the simulation phase. On
the other hand I have considered the second type for the computer generated vector data inputs.
After the analysis, I have also plotted the required error in corresponding to the y axis and x axis
and plotted the value in micrometers.
CE 2.3.6: Team Work Done
Three students were associated in this project along with me for the development of
detection of milli-meter cracks. Due to my past experience in various projects, I have developed
the appropriate plan for developing the test and conducting the analysis. Further I have utilized
communication matrix and allocated the tasks properly among the students for encouraging team
work within the project.
CE 2.4: Summary of Project
CE 2.4.1: Overall view of the Undertaken Project
Through this project I was able to develop the system and successfully conduct the
experiment. The proposed system was able to successfully and accurately determine the crack
while demonstrating the width and depth with great precision. With the application of various
reliable techniques, including the Polynomial Vector Classifier I was able to analyze the obtained
signal and data. Based on the use of training phase, I was able to train the system to accurately
identify and calculate the depth and width of the crack identified.
CE 2.4.2: My Personal Contribution to Project Activities
Page 7 of 9

I was involved in this project as the graduate teaching assistant for the design and
development of the detection of milli-meter cracks. With the application and proficient use of my
theoretical and practical knowledge I was able to successful conduct the experiment and carry
out the proper implementation. In addition to that, I have achieved success through the
application of the detailed technical understanding and managerial concepts.
Page 8 of 9