Melvin Mathew Avatar

Melvin Mathew

Research Scientist

Proactive Applied Research and Development Ph. D Mechanical Engineer, and avid learner, with 5 years of mechanical or material testing, finite element modeling (FEM) and 3D Manufacturing experience. Aiming to leverage a proven knowledge of experimental design, testing, failure analysis, image...
See More


Mechanical and Material NDE Testing Engineer / Doctoral Candidate  
Drexel University, March 2016 to September 2020, Philadelphia, Pennsylvania United States
(Applied Mechanics)
• Designed and conducted mechanical (>1500) tests (Design of Experiments) for capturing damage progression, failure modes, fatigue, and stress behaviors for a variety of materials and structural components following ASTM standards using universal testing machines. • Created and streamlined an optimization strategy to reduce errors (by 30%) using computer vision-based damage detection in comparison to industry standards. • Developed a finite element model (FEM) updating strategy from experimental sensors data while performing monotonic, fatigue or vibration testing using acoustic emission and photogrammetric (RGB, IR) techniques. • Analyzed failure mechanisms (FEMA) in systems, complex materials and structures including optimized 3D printed plastic / resin / metal prints.

Product Development Engineer  
Boehringer Laboratories, August 2014 to January 2016, Phoenixville, Pennsylvania United States
(Surgical Tools)
• Conceived and developed a mechanical laparoscopic surgical device, reducing the risk of uterine cancer proliferation, through product ideation, and design. • Prototyped & Piloted in-field testing with SME and surgeons to improve usability and effectiveness of the prototypes and streamline product training • Developed business development insights through sales tracking systems enabling ground level transparency while identifying key sectors for expansion.

Research Assistant  
Intelligent Maintenance Lab, Drexel University, June 2013 to July 2014, Philadelphia, Pennsylvania United States
• Conducted research focused on providing intelligent aircraft system health knowledge based on active condition and usage monitoring of structures via multisensory analysis to credibly improve reliability and preventive maintenance estimates.

Research Analyst  
Beroe Inc., June 2011 to July 2012, Chennai India
(Procurement Intelligence)
• Developed tailored benchmark procurement market intelligence reports for clients aimed at global cost saving opportunities (~$4.5 million) while cutting down internal costs through industry best practices and quantified negotiation opportunities.


Drexel Univ  
Doctor of Philosophy, Mechanical Engineering, Mar, 2016 to Sep, 2020

Drexel Univ  
Master Of Science, Mechanical Engineering, Sep, 2012 to Aug, 2014

National Institute of Technology-Calicut  
Bachelor of Engineering, Mechanical Engineering, May, 2007 to Apr, 2011


Multiscale Deformation Leveraging Multispectral Speckle Patterns  (16/289842)    
Inventors: Melvin Mathew, Andrew Ellenberg, Antonios Kontsos, Ivan Bartoli, Matthew McCarthy.  Issued March 01, 2019  in United States


A Bio-inspired Frequency-Based Approach for tailorable and Scalable Speckle Pattern Generation     
Published by (Experimental Mechanics)
Authors: M. Mathew, B. Wisner, S. Ridwan, M. McCarthy, I. Bartoli, A. Kontsos.  Published July 20, 2020

Background Digital Image Correlation (DIC) is a length scale independent surface pattern matching and tracking algorithm capable of providing full field deformation measurements. The confident registration of this pattern within the imaging system becomes key to the derived results. Practically, conventional speckling methods use non-reliable, non-repeatable patterning methodologies including spray paints and air brushing leading to increased systematic and random errors based on the user’s experience. Objective A methodology to develop a speckle pattern tailored to the imaging and experimental conditions of the given system is developed in this paper. Methods In this context, a novel bio-inspired speckle pattern development technique is introduced, leveraging spatial imaging parameters in addition to frequency characteristics of speckle patterns, enhancing the results obtained through DIC. This novel technique leverages gradient parameters in the frequency spectrum obtained from patterns fabricated using a bio-templating manufacturing technique. Results The analysis presented shows that optimized gradient features alongside tailored spatial characteristics reduce errors while increasing the usefulness of DIC results across the entire region of interest. With this new approach, gradient information is derived from the bio-templated pattern, extracted, optimized and then convolved with spatial properties of a numerically generated 2D point clouds which can then be transferred onto actual specimens. Numerical error analysis shows that the optimized patterns result in significant reduction in root mean square error compared to conventional speckling methods. Conclusions Physical experiments show the scalability of this optimized pattern to allow for varying working distances while consistently maintaining a lower error threshold compared to conventional speckling techniques.

Experimental Investigation of the Multiscale Mechanical Behavior of Knitted Textiles.     
Published by (Material Design and Processing Communications)
Authors: E. Tekerek, D. Liu, B. Wisner, M. Mathew, N. Castaneda, M. Oncul, A. Kontsos.  Published September 03, 2019

The nonlinear, anisotropic, and multiscale mechanical behavior of knitted textiles is investigated experimentally in this article. The approach is motivated by recent computational work by the authors that revealed, for the first time to their best knowledge, local‐global mechanical behavior effects related to the hierarchical, three‐dimensional structure of this type of materials. The investigation is carried out on single jersey knitted textile specimens. Mechanical testing consisting of tensile loading along the two principal directions was coupled with a noncontact, optical metrology method capable of providing deformation measurements. The effect of globally applied loading on yarn‐to‐yarn interactions was explored using measured data. The results validate the previously obtained computational findings that include the anisotropic behavior between course and wale directions, the pronounced out‐of‐plane motion observed when in‐plane loading is applied, as well as the characteristic nonlinear mechanical behavior of knitted textiles. These effects were linked to direct observations of the loop structure that demonstrated the coupling between local kinematics and kinetics with global mechanical behavior.

Multiscale Deformation Measurement using Multispectral Sensing     
Published by (Structural Control Health Monitoring)
Authors: M. Mathew, A. Ellenberg, S. Esola, M. McCarthy, I. Bartoli, A. Kontsos.  Published March 25, 2018

Remote measurements using images are particularly useful in structural health monitoring cases in which the installation of contact sensors is difficult. Some limitations, though, associated with photogrammetry‐type optical metrology involve the application of speckle patterns, which become even more important with variable working distance or when the required resolution and sensitivity are not a priori known which are both very practical common issues. In this context, multispectral sensing can circumvent some of the challenges of acquiring data at different working distances. The objective of this investigation is therefore to use multispectral imaging combined with the method of digital image correlation to demonstrate an approach for remote sensing related to deformation measurements at the structural level. To demonstrate this novel approach, two speckle patterns were designed for measurements at specified working distances; one was set to work in the visible and the second in the ultraviolet spectrum. The results show that the spectral specific reflectivity of the speckle patterns can be used for spatial overlay without affecting imaging in either spectral range considered, which allows the extraction of multiscale deformation information.