In this section, a brief summary of some of the key research and implementation projects I have been involved in is given.

Post doctoral project and research work

• As part of a FUIP project, SPHEREAU (Solutions de Programmation Hiérarchisée pour l’Efficience des Réseaux d’EAU) which aims to improve efficiency of water distribution networks.
• The team in CRAN lab is involved in Leak detection. The other members are Vincent Laurain and Samir Aberkane (both Associate Profs).
• The main challenges associated with the problems: lack of labeled data and presence of faults on sensors.
• Key contributions include
  • Utilizing a kernel model trained using SVM-LS technique.
  • A novel approach to model the simultaneous leak and fault estimation problem using graphs.
  • An optimization approach that exploits the redundancies in the topological graph models to provide an estimate for the leaks and sensor faults.

PhD thesis work

• Title: State and Parameter Estimation and Identifiability of quasi-LPV models
• In this thesis, two relevant problems in the analysis of Linear Parameter Varying (LPV) models are considered:
  • Joint estimation of states and parameters
  • Identifiability of parameters.
• We deployed the Takagi-Sugeno polytopic approach to tackle the first problem. For the second problem we proposed an algorithm that exploits the ideas of parity space.
• The thesis was defended on June 28, 2018 and it can be accessed here.

Energy in Time

• Energy IN TIME was a Large-scale integrating project within the 7th Framework Programme FP7-NMP, Sub-programme EeB.NMP.2013-4: Integrated control systems and methodologies to monitor and improve building energy performance.
• The project involved working with 13 partners across Europe.
• The team at CRAN was responsible for work packages that involves, fault diagnosis, fault adaptive control and predictive maintenance.
• My contribution lay in the fault diagnosis and adaptive control modules where we proposed solutions and implemented them in MATLAB/Simulink using SIMBAD toolbox.

Non-Intrusive Load Monitoring (NILM)

• NILM refers to the ability to leverage the smart electrical power meter data to obtain insights about the consumption of individual (or a group of) electric appliances in a household or a building.
• This was a project I was involved in at the Innovation Labs, Tata Consultancy Services during the time when NILM was at its nascent stage.
• I developed a semi-supervised learning framework that was piloted in a few households in Netherlands.
• I worked as the responsible for algorithm development guiding another team member and worked with software development and field teams.
• The framework was later used for other pilot projects that involved different types of buildinds.

Fault diagnosis of Self-Powered Neutron Detectors (SPND)

• My Master's project at IIT Bombay. The project aims to detect faults in the SPNDs present inside neutron detectors. This project was part of a collaboration study with the Bhaba Atomic Research Center (BARC) in Mumbai.
• I developed model-based approaches to estimate the input of an SPND given its measurement.

Electromagnetic interference of a Digital phone daughter-board

• This was one of the several projects I participated in during my stay at a client location in Belleville, Canada as part of the team from Wipro Technologies.
• Problem: A daughter-board connected to a digital phone lead to a functional composite system but failed emission standards. This lead to several delayed shipments to key markets.
• I worked on this challenge and found legacy and new design issues that were leading to EM emission failures and proposed an easy-to-implement solution