University of Glasgow Edge AI Digital Twins Aim to Cut Operating Costs in Smart Buildings
University of Glasgow Edge AI Digital Twins Aim to Cut Operating Costs in Smart Buildings
Synopsis
- Edge AI-driven digital twins are being tested to reduce hidden energy waste in commercial buildings
- A University of Glasgow prototype shows significant reductions in standby power consumption
- The approach blends embedded AI, IoT connectivity and facilities management systems
Estimated reading time: 3 mins read
Edge AI-powered digital twins are emerging as a practical tool for tackling one of the most persistent and underestimated sources of energy waste in commercial buildings: so-called “phantom load.” This idle electricity consumption, generated by devices left in standby mode, quietly inflates operational expenditure and undermines sustainability targets. New research from the University of Glasgow suggests that shifting intelligence closer to the edge could make these losses both visible and manageable.
According to reporting by eeNews Europe, researchers at the university’s James Watt School of Engineering have designed and tested a prototype system that combines smart plugs, environmental sensors and local AI decision-making. Instead of relying on centralised cloud analytics or blunt timer-based shutdowns, the system uses edge-hosted digital twins to model individual assets and manage their power states with greater nuance. Early trials indicate that this architecture could deliver meaningful OpEx savings for facilities teams while avoiding the user frustration that often derails automation projects.
The work highlights how edge AI-powered digital twins could shape the next generation of building automation systems. Choices around IoT connectivity, embedded compute and data orchestration become tightly linked to how energy dashboards feed into corporate sustainability metrics and compliance reporting. It also points to a growing convergence between embedded AI, power electronics and facilities IT, an intersection that is increasingly relevant for engineers involved in future design-in decisions.
In office buildings and campus environments, phantom load from devices left in standby can account for close to a third of total electricity use. Despite its scale, this consumption is frequently treated as background noise rather than a controllable cost. The Glasgow team’s approach challenges that assumption by instrumenting power use at the plug level. Smart meters and environmental sensors are connected via LoRaWAN to an on-site edge server, where a digital twin of each monitored device is maintained.
At the heart of the system is a local AI layer that replaces simple time-based rules with a fuzzy-logic decision framework. Instead of cutting power after a fixed period of inactivity, the software evaluates multiple signals, including a user habit score, a device activity score and an overall confidence score. Based on these inputs, the system can keep a device powered, delay a shutdown decision, switch it off entirely, or prompt the user to confirm that a background task is still running. The objective is to curb unnecessary energy use without disrupting legitimate workflows, reducing the likelihood that users override or disable the automation.
Dr Ahmad Taha, Lecturer for Autonomous Systems & Connectivity at the James Watt School of Engineering and lead investigator on the project, framed the work in broader terms. He noted that small, collective actions on climate challenges can have outsized effects, and that phantom power consumption is a clear candidate for this kind of intervention. By embedding intelligence at the edge, the system aims to translate that principle into day-to-day operational practice.
To validate the concept, the researchers deployed the edge AI-powered digital twin system in a university laboratory, equipping workstations with smart plugs and LoRaWAN-connected sensors. In this controlled environment, weekly power consumption per monitored workstation fell by around 40 percent. More strikingly, phantom loads themselves were reduced by up to 82 percent. When these figures are extrapolated to a larger deployment of 500 devices and aligned with current UK electricity price caps, the model points to potential annual savings in excess of £9,000.
From a technical perspective, the implementation uses a containerised software stack that will be familiar to many embedded and IoT developers. Docker hosts core services, including an MQTT broker for messaging, Node-RED for data handling and orchestration, and InfluxDB for time-series data storage. On top of this, a forecasting module based on a Long Short-Term Memory neural network trains on short histories of consumption data to predict the following day’s demand profile. This gives facilities teams a forward-looking view of expected peaks, enabling proactive management rather than retrospective analysis.
User acceptance is recognised as a critical constraint in any automated power management scheme. To address this, the architecture incorporates an anti-oscillation filter designed to prevent rapid on-off cycling that could irritate occupants or place unnecessary stress on hardware. Beyond immediate energy savings, the researchers also point to a secondary benefit: extending asset lifetimes. By reducing overall electricity use, organisations may be able to delay the replacement of older equipment with newer, more power-efficient models, easing capital expenditure pressures in a challenging economic climate.
As outlined in the eeNews Europe report, the Glasgow prototype illustrates how edge AI-powered digital twins could move from academic experimentation to practical deployment in smart buildings. By combining granular sensing, local intelligence and digital representations of physical assets, the approach offers a pathway to measurable OpEx reductions without sacrificing usability. For facilities managers, embedded engineers and system designers alike, it signals a shift toward more adaptive, human-aware automation at the edge.
Source Here – Have a Story? Address it to the Editor and submit it here
About The University of Glasgow
The University of Glasgow, founded in 1451, is one of the world’s oldest and most respected research universities, consistently ranked among the top institutions globally for engineering, science, and technology innovation. Its James Watt School of Engineering is a major hub for research in autonomous systems, connectivity, embedded AI, and sustainable infrastructure, with strong links to industry and government bodies across the UK and Europe.
The school is named after James Watt, whose work on steam power helped shape the Industrial Revolution, reflecting a long-standing focus on practical, impact-driven engineering. In recent years, the University of Glasgow has expanded its research into edge AI, IoT systems, digital twins, and energy-efficient technologies, targeting real-world challenges such as climate change, smart buildings, and operational efficiency. Through applied prototypes, field trials, and interdisciplinary collaboration, the university plays a significant role in translating advanced engineering research into deployable systems that influence future industrial and urban technology design.
Featured image Source: Cintoo
Disclaimer
The information provided in this article is for general informational purposes only and from publicly available sources. While we strive for accuracy, we do not make any representations or warranties, express or implied, regarding the completeness, reliability, or validity of the content. This article does not make any direct claims about specific companies, individuals, or organizations. Any references to reports or external sources are for context and do not imply endorsement or verification of any specific allegations. Readers are encouraged to conduct their own research and seek professional advice before making business decisions. We disclaim any liability for any losses or damages incurred as a result of reliance on the information provided.