Sustainability and innovation: a new approach for predictive maintenance of IoT systems

A NEW MODEL FOR PREDICTING THE LIFETIME OF RECHARGEABLE BATTERIES POWERING WIRELESS SENSORS

SCAI Lab’s R&D team conducted and published the study “LSTM-based Battery Life Prediction in IoT Systems: a proof of concept”^{^[1]}, realising an innovative model for predicting the remaining life of rechargeable batteries.

A Proof of Concept is a demonstration to verify that specific ideas or theories can be used in the real world. In the study conducted by SCAI Lab’s R&D team, the Proof of Concept involved the creation of a model based on LSTM (Long Short-Term Memory) neural networks for predicting the remaining life of rechargeable batteries in IoT systems. Objectives of the PoC Demonstrate feasibility: The PoC was to demonstrate that the LSTM model could accurately predict the remaining life of lithium-ion batteries (SOH – State of Health) and their RUL (Remaining Useful Life), expressed in number of remaining recharge cycles. Assessing accuracy: Using public datasets, such as those made available by the Hawaii Natural Energy Institute, the PoC aimed to verify the accuracy of the model’s predictions. Experimental results showed an accuracy of 97%, indicating that the model is highly accurate. Practical integration: Another objective of the PoC was to demonstrate that the model could be successfully integrated into the IBIS ECO Decision Support System (DSS), a research project for intelligent monitoring and management of energy performance and comfort in buildings.

aims to realise an innovative system for the intelligent monitoring and management of the energy performance and comfort of existing buildings^[2]. In particular, while working on IoT networks of wireless sensors, several critical issues emerged related to the reliability of battery-powered devices. This led to research aimed at implementing a tool for the predictive maintenance of wireless IoT sensors capable of estimating the remaining battery life.

The machine learning model developed and presented at the conference reduces the risk of abrupt operational interruptions and/or data loss by IoT sensors and contributes to the design of increasingly energy-efficient devices. It uses a LSTM (Long-Short-Term Memory) network to monitor the health of lithium-ion (SOH) batteries and predict their remaining useful life (RUL) expressed in the number of remaining charging cycles.

The experimental results of this methodology, tested using public battery datasets made available by the Hawaii Natural Energy Institute repository^[3], demonstrated 97% accuracy. From this study, a specific PoC was developed and integrated into the DSS IBIS ECO – a decision support system designed for intelligent monitoring and management of energy performance and comfort in buildings. The experimentation conducted within the project’s demonstrator institutes (University of Basilicata and Montemurro Primary School) will further validate the approach in real life.

The work, presented at the third International Conference on Statistics and Data Science held in Palermo last April, was published in the volume “Proceedings of the Statistics and Data Science 2024 Conference – New perspectives on Statistics and Data Science”, which collects the conference proceedings.

^[1] Vanessa Verrina; Andrea Vennera; Annarita Renda; LSTM-based Battery Life Prediction in IoT Systems: a proof of concept; “Proceedings of the Statistics and Data Science 2024 Conference – New perspectives on Statistics and Data Science”; anno 2024; pag. 322-327; https://unipapress.com/book/proceedings-of-the-statistics-and-data-science-2024-conference/

^[2] Project funded under the ERDF Operational Programme 2014-2020 – Action 1B.1.2.2. Public Notice: “Complex research and development projects ‘CORES’” – Thematic areas: “Energy and Bioeconomy” – CUP G49J19001400004.

^[3] https://www.batteryarchive.org/list.html

expertise

Research & Innovation

Business Intelligence & Big Data Analysis

Internet of Things

Developed by SCAI LAB

Learn more about SCAI Lab

SCAI Lab TECHNOLOGY: IDEAS

SCAI Lab provides consulting services and effective system integration and application management solutions, in addition to contributing to key App R&D developments.

Contatta SCAI Lab