Deodatta Dixit Email & Phone Number

Aschaffenburg, BY, DE

Aachen, North Rhine-Westphalia, Germany

• Application of deep learning in the automated cutting of reinforcing textiles for parameter prediction and image classification.
• Creation of datasets: Created image and parameter datasets by cutting different textiles with various cutting parameters with an automated CNC cutter.
• Image Classification Pipeline: Implemented a Convolutional Neural Network with Transfer Learning (pre-trained Efficientnet B0 network) with a classification accuracy of 92% to classify the images of textiles according.
• Siamese Neural Network (SNN) pipeline: Applied the deep learning algorithm (SNN) to investigate its performance on a small image dataset. Made the architecture using ResNet 50 pre-trained network (Transfer Learning).
• Optimized parameter prediction: Employed ensemble machine learning algorithms like Random Forest and Gradient Boosting for predicting optimized parameters used for the cutting process of textiles.

Aachen, North Rhine-Westphalia, Germany

• Bilinear Fatigue degradation model for life estimation of FRP
• Modelled the degradation law to calculate the accumulated damage in unidirectional composite material.
• Performed lifetime prediction analysis of the composite material.
• Developed a User-defined Material Model (UMAT) subroutine in FORTRAN.
• The results corroborated the fact that non-linear damage behavior of the composite material can be explained through the linear damage law for low strain rates.

M.Sc, Mechanical Engineering

Bachelor Of Engineering - Be, Mechanical Engineering