NIT Rourkela Develops AI System For Autonomous Solar Cleaning
The system employs Federated Learning to enable intelligent monitoring while preserving data privacy by sharing encrypted model updates rather than raw operational data, thereby reducing cybersecurity risk and bandwidth usage. It integrates edge computing, predictive maintenance and autonomous fault detection to optimise operations and to identify panels that require selective cleaning. The technology aims to address efficiency losses caused by dust, bird droppings and industrial pollutants, which researchers say can reduce panel output by up to 40 per cent in arid regions.
Existing cleaning regimes are commonly labour intensive, water intensive and scheduled irrespective of panel condition, whereas the new system detects dirt accumulation and orchestrates selective cleaning only where necessary, which is expected to conserve resources. The approach has been validated through simulations at Technology Readiness Level (TRL) three and the team plans to develop a hardware prototype with Internet of Things (IoT) sensors for pilot deployments with government agencies and industry partners.
Researchers estimate that, once commercialised, the system could offer advanced AI driven capabilities at nearly 10 per cent of the cost of comparable solutions, which would support India’s National Solar Mission and net zero aims by enabling smarter maintenance of solar infrastructure. The technology is intended for deployment across utility scale parks, floating and rooftop photovoltaic installations, industrial facilities, smart city infrastructure and remote off grid renewable energy systems.