The idea for a low-cost UPS monitoring system at IIIT Hyderabad began not in a laboratory or funding proposal, but with a practical frustration. Prakash Nayak, a campus IT staffer, was dealing with repeated equipment failures during power outages, with no reliable way to determine how UPS systems were behaving when the power went out.

Servers were restarting and outages were occurring, yet despite widespread UPS installations, there was no clear visibility into their performance during disruptions. This operational gap became the foundation of a research project that has now resulted in Rs 2,000 IoT-based device capable of tracking UPS behaviour during outages with near-second precision.

The work is documented in a paper titled Low-cost IoT-based Downtime Detection for UPS and Behaviour Analysis by Sannidhya Gupta, Prakash Nayak, and Prof Sachin Chaudhari. The paper received the Best Paper award at the 18th International Conference on Communication System and Networks (COMSNETS-2026) Workshop on AI of Things, held recently in Bengaluru.

Frequent power outages in developing regions lead to equipment damage, operational downtime, and data loss, notes Sannidhya Gupta. While UPS systems are designed to mitigate these risks, affordable monitoring solutions remain limited. Commercial UPS monitoring tools, typically SNMP cards, were available but impractical. Costing over Rs 20,000 per unit, they were expensive, manufacturer-specific, dependent on network infrastructure, and often failed to record data when the UPS itself lost power.

In response, the research team designed a non-intrusive, battery-backed current-monitoring device. Rather than interfacing with UPS internals, it clamps onto input and output lines to observe current flow before, during, and after outages. The system continues recording even when both mains power and internet connectivity are unavailable.

The device was deployed across four UPS installations on campus, including one suspected to be faulty. Over a month, it collected approximately 3.7 million data points and automatically detected 61 outage events. The data confirmed that one UPS failed to exhibit proper charging behaviour after outages, despite briefly supporting loads.

Backend analytics automatically classify events into phases—normal operation, outage, stabilisation, and battery charging—using thresholds based on locally estimated baselines. This enables automatic adaptation across installations. The system recorded no missed outages or false alarms, with timing errors typically within three seconds.

A web-based dashboard now provides IT staff with real-time visibility and historical analysis of UPS performance. Built using off-the-shelf components, the device costs about Rs 2,000, nearly one-tenth the cost of commercial alternatives. Its affordability, power independence, and portability make it suitable for cost-constrained environments.

Prof Sachin Chaudhari emphasises that this project stands out for its origin in real operational needs. Prakash Nayak is both a co-author of the paper and a co-inventor on the recently filed patent, highlighting the value of involving campus operations teams as co-creators of research problems. The project demonstrates how research grounded in real-world challenges can deliver practical and impactful outcomes.