Long-Life IoT Sensing and AI-Supported Analytics for Smart District Heating: Lessons from the LoLiPoP-IoT Project
TL;DR
This paper identifies the technical and organisational conditions required before AI-enabled district heating optimisation can be reliably implemented and proposes a layered architecture that connects energy-harvesting sensors, data acquisition, local preprocessing, AI-assisted interpretation, and human-centred decision support.
Abstract
Smart district heating systems require reliable data, scalable sensing infrastructure, and intelligent analytics to support energy efficient and low-carbon operation. However, many existing buildings and heating systems remain constrained by limited sensor coverage, heterogeneous infrastructure, poor data quality, fragmented building management systems, and insufficient digital readiness. This paper presents a practice-informed conceptual and deployment-oriented study of how long-life Internet of Things sensing platforms, local-first edge processing, and AI-supported analytics can support future smart district heating applications. Drawing on lessons from the LoLiPoP-IoT project, the paper proposes a layered architecture that connects energy-harvesting sensors, data acquisition, local preprocessing, AI-assisted interpretation, and human-centred decision support. The study does not claim a fully validated autonomous control system; instead, it identifies the technical and organisational conditions required before AI-enabled district heating optimisation can be reliably implemented. A key finding is that AI deployment in older buildings or buildings with multiple heating arrangements may be limited or delayed when reliable baseline data, interoperable interfaces, and validated datasets are unavailable. The paper concludes by outlining future research needs related to quantitative validation, digital twins, trustworthy AI, heating-and-cooling integration, cost-benefit assessment, and replication of the LoLiPoP-IoT architecture in operational energy environments.