Research Experience

Welcome to my research portfolio, showcasing my diverse projects. These applied research initiatives primarily focus on design, experimentation and analysis, bridging theory and real-world applications. Committed to excellence, I aim to make a meaningful impact in various fields.

Automatic Nano-porosities Detection in TEM Images Using Computer Vision Techniques

I'm involved in a project that focuses on automating the identification and pixel-wise segmentation of nano-porosities in Transmission Electron Microscopy (TEM) images, specifically for analysing cladding materials used in nuclear reactors. It addresses the critical challenge of accurately characterizing these nanoscale features to understand material integrity and performance. The pipeline leverages from classical ML models, deep learning (CNN) models such as Unet to Vision Transformer (ViT), for high-accuracy feature detection and its correlations.

Friction Stir Welding of Aluminium Alloys and Advanced Steel for Lightweight, Sustainable Automotive and Aerospace Applications

This research focuses on friction stir welding (FSW) of dissimilar materials with different properties, such as aluminium and Advanced steel. FSW is a well-known advanced solid state welding technique suitable to join aluminium alloys and now being explored for other materials such as advanced materials, composites and polymers, etc. The FSW is widely used in automotive and aerospace applications (click here to know more). The current research explores the potential of FSW in joining dissimilar alloys and its feasibility by analysing the mechanical and microstructural properties and their correlations.

Analysis and Prediction of Properties of Friction Stir Welded Dissimilar Materials Using Machine Learning Algorithms

This research complements the experimental work on FSW by leveraging machine learning (ML) to model and predict material properties. The primary objective was to predict the strength of welded joints based on key process parameters. I conducted a thorough Exploratory Data Analysis (EDA) using Python libraries to identify relationships within the dataset. Subsequently, I developed and evaluated ML models, including regression, decision tree, random forest, and support vector machine. This predictive framework, validated using the performance metrics, supports process optimization and provides deeper insights into the complex process-structure-property relationships inherent in FSW.

Page updated

Google Sites

Report abuse