---
title: "NIT Rourkela Patent Introduces Federated AI to Optimize Solar Panel Maintenance"
url: https://projectchintan.com/article/nit-rourkela-ai-solar-cleaning-federated-learning-patent-60c0z
publisher: Project Chintan
author: Project Chintan Newsroom
published: 2026-08-09T05:30:00.000Z
modified: 2026-08-09T08:00:49.210Z
language: en-IN
---

# NIT Rourkela Patent Introduces Federated AI to Optimize Solar Panel Maintenance

Researchers at NIT Rourkela have patented an autonomous system that utilizes federated learning to identify specific solar panels requiring maintenance. The technology aims to cut costs by 90% while significantly reducing water consumption at large-scale energy installations.

## Key takeaways

- NIT Rourkela researchers patented an AI system that identifies specific solar panels needing cleaning to prevent a 40% loss in energy generation.
- The technology uses Federated Learning to process data locally, enhancing cybersecurity and reducing internet bandwidth requirements.
- The system is expected to cost only 10% of current market solutions once it moves from simulation to large-scale deployment.

## Why It Matters

Accumulated dust and pollutants can slash solar energy yields by up to 40%, yet conventional cleaning schedules waste significant volumes of water and labor on clean panels. By applying Federated Learning (FL), the NIT Rourkela team offers a solution that identifies specific cleaning needs without compromising data security or wasting resources. This targeted approach is designed to function in dry, high-pollution regions where efficiency is often undermined by environmental debris.

## Key Facts

- The system was developed by Professor Bibhudatta Sahoo, Assistant Professor Arun Kumar, Dr. Lopamudra Hota, and Dr. Biraja Prasad Nayak.
- Researchers secured an Indian patent for the Federated Learning-based Autonomous System and Method for Monitoring and Cleaning Solar Plant.
- The technology integrates edge computing, AI, and predictive maintenance into a unified sandbox platform.
- Testing has reached Technology Readiness Level 3 (TRL-3), confirming the proof of concept via simulation.
- Projections suggest the system could operate at approximately 10% of the cost of current market alternatives.

## Background

Standard artificial intelligence models typically require raw data to be transmitted to a central server, raising concerns regarding cybersecurity and bandwidth. The NIT Rourkela architecture bypasses these issues by using encrypted updates rather than raw operational data. This privacy-preserving method allows the system to scale across sensitive sites, including defense installations and smart city infrastructure, while maintaining local data integrity.

## What Happens Next

Following the successful simulation phase, the research team is moving toward physical hardware development. The next stage involves integrating Internet of Things (IoT) sensors to collect real-time environmental and performance data. Once fully realized, the system is intended for diverse applications ranging from floating solar farms to remote off-grid renewable networks.

Source: The Better India

---
Canonical: https://projectchintan.com/article/nit-rourkela-ai-solar-cleaning-federated-learning-patent-60c0z
Reported from: The Better India