Sustainable Engineering Systems

Machine Learning & Embedded Hardware for Sustainability

This page showcases early-stage applied engineering and hardware-software co-design research aimed at optimizing sustainable energy generation, capture, and consumption.

1. Energy Management Using Machine Learning

• Affiliation: M. L. Khanna D.A.V. Public School

Engineered an integrated chain of technologies incorporating machine learning algorithms and embedded hardware to optimize real-time power consumption in local microgrids.

This system was specifically designed to address systemic electricity mismanagement and mitigate the 60% grid-level energy wastage reported globally by the International Energy Agency (IEA).

• Grid Balancing: Uses lightweight supervised learning algorithms to balance power supply and demand at local node levels.
• Anomalous Load Detection: Protects distributed electronics and prevents sudden distribution dropouts.

2. Open Gym Electricity Generator

• Affiliation: M. L. Khanna D.A.V. Public School

Formulated a zero-emission, dual-purpose sustainability project that captures and converts physical workout energy into regulated electrical current.

Designed for public parks and community centers, the hardware captures kinetic energy generated during workout sessions, regulates the electrical input using a basic neural network model, and stores it in high-capacity battery units to power public park lighting infrastructures at night.

• Kinetic Harvesting: Direct conversion of kinetic exercise energy into high-capacity batteries.
• Neural Network Modeling: Simple neural network models stabilize variable current inputs before storage.