Iqra Jilani

Machine Learning, Artificial Intelligence, Renewable Energy, Photovoltaic Systems, Batteries (Lithium-ion and Lead-Acid), e-Cooking.

Iqra Jilani is a PhD student at Coventry University's "Center for Computational Science and Mathematical Modelling". There, she is working on "Predictive Maintenance and Smart Load Management for Optimization of Battery-aided Solar PV systems (application to e-cooking systems)". She started at the said center as a British Council Women in STEM Scholar for her MRes. Later, her MRes was upgraded to a PhD.  Iqra did her undergraduate studies in software engineering at UET Taxila, in Pakistan. After graduation, she worked in the corporate for 5 years. She represents a well-rounded profile with skills and knowledge from both industry as well as research. Iqra takes great interest in applied research in renewable energy

"Optimization 2D genetic algorithm (GA) via machine learning (ML) to solve the combinatorial optimization of load shedding scheduling problem in community-based solar PV systems"

Improving energy access in refugee camps can enable the provision of facilities like cooking, electrical lights, etc. Due to geopolitical reasons, standalone renewable energy management systems are more suitable in refugee camps, particularly Photovoltaic Systems (PV). These systems store the surplus energy (net energy remaining after satisfying the current electrical load in the given hour) in lead-acid batteries. Usually, batteries in such systems do not last as long as could be expected due to sub-optimal use. Due to demand and supply imbalance in such scenarios, load shedding often has to be implemented. We aim to optimize the load shedding schedule created by an Energy Management System (EMS) to
maximize battery lifetime, maximize Performance Ratio (PR), and increase user satisfaction. This presents the given problem as a "Combinatorial Optimization Problem (COP)".
Meta-Heuristic (MH)s, particularly Genetic Algorithm (GA), are usually used
to solve COPs. GA relies on its parameters, like the probabilities of mutation and crossover, to navigate the search space and find the optimal schedule. During the run of a GA, a lot of useful information is generated that can be utilized by machine learning (ML) algorithms to tune the parameters of the GA so that the objective function score, for a given instance of a load schedule, is maximized. In this project, we present our approach to utilizing machine learning to tune the parameters of the
GA solution to determine load shedding in an EMS.

"Predictive Maintenance and Smart Load Management of Solar PV-based Battery-aided e-cooking Systems".