By Nidhi DhullReviewed by Susha Cheriyedath, M.Sc.Aug 8 2024
A recent article published in Data in Brief presented a comprehensive dataset on the University of Sharjah's (UoS’s) innovative initiative to transform into a smart campus by 2030. Internet-of-Things (IoT) sensors were used to precisely monitor energy consumption and environmental conditions from 1st January 2024 to 20th June 2024.
![Monitoring Building Energy Consumption Using IoT Sensors](https://d251cvb8f7e7p0.cloudfront.net/images/news/ImageForNews_23583_17231893516287080.jpg)
Image Credit: Aree_S/Shutterstock.com
Background
The construction sector increasingly relies on advanced computational models to accurately analyze energy consumption patterns in campus buildings and achieve energy efficiency. The application of artificial intelligence (AI) in building energy management, particularly for anomaly detection, holds significant promise for optimizing energy use and identifying inefficiencies. AI systems can detect deviations from typical energy consumption patterns, enabling timely interventions to reduce waste and enhance operational efficiency. As a result, institutions can significantly lower operational costs and promote sustainable energy practices through AI-driven analysis and response to energy data.
In this study, the researchers compiled a comprehensive dataset on energy consumption and environmental conditions from a university campus to support the development and refinement of machine learning models for energy classification, anomaly detection, and recommendation systems.
Data Description
The dataset was organized into four primary containers, each reflecting different aspects of the data collected from the UoS. These containers included:
-
Raw Data on Power Consumption: This container comprised multiple CSV files, each representing power consumption data for different campus appliances, such as coffee machines, microwaves, fridges, kettles, printers, desktops, and water dispensers. The files included time stamps, voltage, and current measurements.
-
Raw Data on Environmental Conditions: Similar to the power consumption data, this container contained CSV files recording indoor conditions like temperature, humidity, and occupancy. These measurements were taken from sensors placed in the interdisciplinary lab, kitchen, and mailroom on campus.
-
Processed Summary Data: This container held summarized data on power consumption and environmental conditions, presented either as additional CSV files or visual charts, depending on the processing methods used. This summary data facilitated quick, high-level analysis.
-
Documentation and Metadata: The fourth container included comprehensive documentation detailing the dataset's structure, the nature of each file, and metadata about the sensors used, their locations, and the data collection protocols.
All data was combined and stored in InfluxDB, which supported high-resolution time series data acquisition. This allowed for precise calculations of mean temperature and humidity values at predefined intervals, ensuring data integrity and granularity for a thorough analysis of building dynamics.
Experimental Design, Materials, and Methods
The data collection at the UoS campus targeted specific areas including the lab, kitchen, and mailroom. A central data management and edge computing hub was established using the Odroid N+2, which provided a robust platform for handling extensive data workflows. This hub was integrated into a local network running a customized smart building management system to facilitate dynamic data collection.
Smart plugs and sensors were installed on appliances such as fridges, microwaves, and kettles to monitor power usage. Temperature, humidity, and occupancy sensors were also strategically placed in key locations to maximize coverage and complement the power consumption measurements. Passive infrared sensors were then used to track occupancy, and their detection range and accuracy were thoroughly evaluated.
To ensure data accuracy, each environmental sensor was calibrated using standard instruments, and extensive post-installation tests were conducted to verify data integrity and robustness of wireless connectivity. Energy consumption data was recorded every minute, while environmental conditions were logged every five minutes.
Database management strategies were implemented to efficiently handle, compress, and export the large volume of data. Additionally, the Markov Transition Field (MTF) method was applied to capture temporal dynamics. This technique transformed one-dimensional time series data into two-dimensional matrices, facilitating visual analysis of temporal trends.
Conclusion
Overall, this dataset offers a detailed perspective on energy consumption and environmental parameters within a campus setting. It is a valuable resource for researchers studying energy efficiency, IoT applications in smart buildings, and environmental monitoring. Additionally, students and academics can gain meaningful insights into the intersection of technology and sustainability.
However, there are some limitations to consider. The dataset's five- to six-month duration may not capture long-term trends or seasonal variations in energy use and environmental conditions. Its focus on a university campus also means the findings may not be directly applicable to other building types or climatic contexts.
Moreover, the dataset's size may be insufficient for training complex deep-learning models, potentially limiting their accuracy in predicting energy consumption. Additionally, while sensor accuracy can affect IoT-based data collection, this issue can be managed through regular calibration and validation with standard instruments.
Journal Reference
Oulefki, A., Amira, A., Kurugollu, F., & Soudan, B. (2024). Dataset of IoT-Based Energy and Environmental Parameters in a Smart Building Infrastructure. Data in Brief, 110769. DOI: 10.1016/j.dib.2024.110769, https://www.sciencedirect.com/science/article/pii/S2352340924007352
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.