Jaswant Jayacumaar

The University of Manchester

Graduate Teaching Assistant, Year 1 Physics Laboratory (PHYS10180)

Sep 2021 - Jul 2022

Vellore Institute of Technology

Teaching Assistant, Engineering Thermodynamics (MEE1003)

Sep 2019 - Nov 2019

Jaswant Jayacumaar, a former postgraduate researcher at the Jodrell Bank Centre for Astrophysics (Pulsars & Time Domain Astrophysics Research Group), obtained his Master of Science by Research (MScR) in Astronomy and Astrophysics from the University of Manchester, United Kingdom. His research focused on deriving coherent timing solutions and analyzing spin-down rate variations in regularly observed pulsars within the MeerTime Thousand Pulsar Array (TPA). Prior to this, he contributed to gravitational wave research at the Bose.X Center for Astrophysical Research in India. Jaswant holds a Bachelor’s degree in Mechanical Engineering from the Vellore Institute of Technology, India.

MScR Dissertation (Pulsar Timing)

Title: Timing Pulsars with the Thousand Pulsar Array on MeerKAT

Abstract: Pulsars are rotating neutron stars with rotational periods ranging from 1ms – 15s. This study employs the TEMPO2 package to establish coherent timing solutions for regularly observed pulsars within the MeerTime Thousand-Pulsar-Array. The central focus of this dissertation is to explore variations in the pulsars' spin-down rate by analyzing the obtained timing residuals. Employing two distinct methodologies — Bayesian and fitwaves analysis, the latter being a novel addition to the study — we delve into the variations in spin-down rate. All timing observations originate from the Lovell telescope at the Jodrell Bank observatory (JBO) and the MeerKAT interferometer. By merging the JBO and MeerKAT data, a third dataset is created, and its contribution is assessed in estimating spin-down rate variations through both methodologies. Comparative analysis aims to identify which dataset more effectively constrains spin-down rate variations. The results reveal consistent spin-down rate variations for the majority of pulsars between the two techniques. However, challenges arise in cases with limited observations or significant timing residual uncertainties along with irregularly sampled data, particularly impacting the fitwaves method. The creation of the third dataset has proven to be valuable. However, caution is advised in the utilization of data with high uncertainties, as it is preferable to exclude such data before conducting the analysis.

Keywords: Pulsar Timing, Bayesian Inference, FITWAVES, TEMPO2, MeerTime, Thousand-Pulsar-Array (TPA)

Gravitational Waves*

Title: Use of Deep Learning to Classify Lensed and Unlensed Gravitational Waves

Abstract: Gravitational waves (GWs) are ripples in the curvature of space-time, produced by energetic processes such as the collision of two compact objects orbiting each other and core-collapse supernovae. These GWs are lensed i.e., bent near massive astrophysical objects, similar to electromagnetic waves. Lensing of GWs produces multiple images at different times and it is hard to classify a lensed GW from an unlensed one. Hence, we have decided to use convolutional neural networks (CNN) to extract the complicated features of lensed GWs and use these features to classify lensed GWs from unlensed ones. Furthermore, these features are vital to study merger events as they possess physical significance.

Keywords: Gravitational Waves, Gravitational Lensing, Deep Learning, Convolutional Neural Networks

*This project is being conducted in collaboration with Bose.X Center for Astrophysical Research - A Division of Bose.X

Bachelor's Thesis (Renewable Energy)

Title: Parametric Study of Solar Pond and its Prediction of Performance Parameters Using Artificial Neural Networks

Abstract: There is a growing need to identify the adjustments required to enhance the functioning of a solar pond, especially when there are certain technical or geographical constraints. Some proven performance-enhancing techniques such as incorporating Low-Temperature Energy Storage (LTES) system in the Lower Convective Zone (LCZ) and by placing twisted tape inserts in the flow passage of the heat exchanger were incorporated. The readings were taken for both laminar & turbulent conditions. Developed an artificial neural network (ANN) model with a 5-15-3 neuron configuration to predict the performance parameters such as outlet water temperature, the efficiency of the solar pond, and the effectiveness of the heat exchanger. The data collected during experimentation is used as the training set, and random experimental data was used to validate the model. The geometry of a 5-pass heat exchanger and a twisted tape were developed using ANSYS Design Modeler 19.0 and computational fluid dynamics (CFD) analyses were carried to investigate the effect of twisted tape inserts in the heat exchanger. It is inferred from the experimentation that providing twisted tape inserts improves the thermal performance of the heat exchanger and LTES increases the period of operation by increasing the thermal energy storage capacity of the solar pond. By using the ANN model, the degree of accuracy for the prediction of outlet water temperature is 98.63%, 97.58% for the efficiency of the solar pond, and 97.05% for the effectiveness of the heat exchanger. The correlation coefficients (R²) are 0.9735, 0.9941, and 0.9341, respectively. The error values of the prediction model are well within the desired tolerance. Thus, ANN can be used to obtain the performance parameters for different conditions of SGSP. From the CFD results, it is conclusive that the twisted tape inserts have a significant effect in laminar flow than turbulent flow.

Keywords: Solar Pond, Twisted Tape Inserts, Nanoparticles, Artificial Neural Networks, CFD Simulation

Citation: Jayacumaar, J., Dominic, J., Karunamurthy, K., (2021) Performance Analysis of Solar Pond and its Prediction. Journal of Thermal Engineering (Manuscript ID: JTEN-2021-445, Minor revision)

Materials Science

Title: Development of Natural Fiber-Based Aluminum Composites for Electromagnetic Interference Shielding Applications

Abstract: Natural fiber mats, aluminum foil sheets, and epoxy resin have been compressed using a hydraulic compression molding press to fabricate hybrid composites. These hybrid composites consist of natural fiber-based shells and ultra-thin aluminum sheet core. Jute, Flax, and Kenaf have been the natural fibers used in fabricating these hybrid composites. Through natural fiber-based composites and aluminum sheets, the fabricated hybrid composites can show excellent mechanical properties and an exceptional electromagnetic interference (EMI) shielding action. The natural fiber mat (shell material) shields the aluminum sheet, which avoids corrosion and contact with the atmosphere. The exceptional EMI shielding performance and excellent mechanical properties allow the newly fabricated hybrid composites to be used in the EMI protection fields and electric devices. In this study, the EMI shielding effectiveness, the tensile property, and the flexural property of the hybrid composites have been examined.

Keywords: Aluminum, Electromagnetic Interference Shielding, Elevated Temperature, Hybrid Composites, Natural Fiber

Citation: Kshirsagar, A., Jayacumaar, J., Rohith, S. V., Subramanian, J. (2019) Development of Natural Fiber-Based Aluminum Composites for Electromagnetic Interference Shielding Applications. International Review of Mechanical Engineering, 13(6): 367-373 10.15866/ireme.v13i6.17465