Energy Pattern Classification and Prediction in an Educational Institution using Deep Learning Framework

Abstract

Building Energy Management is the most promising, trending, and essential way to enhance Building energy performance but it varies from one system to another. It can be achieved only by extracting the entire knowledge of the system. For this, a deep learning framework is proposed. The framework visualizes the Energy profiles of various workspaces in a building and also accounts for the Generating sources feeding them. It imparts awareness among consumers on effective handling of available Energy sources. kRNN-LSTM is the proposed framework applied to real-time Smart energy meter data observed in the Two-storeyed Electrical department building block of Thiagarajar College of Engineering, Madurai. k-means clustering is done to achieve the underlying pattern uniformities while RNN based LSTM gives a one-month ahead prediction. The LSTM model is compared with the Machine learning-based ARIMA model and basic Naïve time series model by employing quality metrics computed from each model. The proposed model gives 94% accuracy and the reason behind the superiority of the model is due to temporal analytics dependencies and application of sliding window concept to Data for updating and learning the patterns of energy usage. In addition to this, the framework also points out some sort of information such as workspaces with higher energy profiles, Annual peaks in consumption, duration, and range of existence of peak demand, preferable Generating source on a cost basis, etc., Moreover, a suggestion for reducing the cost involved in Energy utilization is also depicted.

Keywords:

• Building Energy Management (BEM)
• building energy performance
• deep learning framework
• energy demand prediction
• pattern defining
• smart meter technology

Additional information

Funding

The authors extend their sincere thanks to the management of the respective college for their constant encouragement and flexibility in doing this research work and the financial support from Science and Engineering Research Board (SERB), India under the Early Career Research (ECR) Scheme (F. No. ECR/2017/000827) is gratefully acknowledged.

Notes on contributors

A. C. Vishnu Dharssini

A. C. Vishnu Dharssini received her Bachelor’s Degree in Electrical and Electronics Engineering, and Masters in Power Systems Engineering with distinction from Anna University, Chennai in 2019, and 2021 respectively. She is the recipient of the “Best Outgoing Student” Award during her Master’s degree in Power Systems Engineering from Thiagarajar College of Engineering in 2021. She is currently pursuing her Ph.D. at the Department of Electrical and Electronics Engineering, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India. Her topics of interest include power system analysis, smart grid and Machine learning. E-mail: [email protected]

S. Charles Raja

S. Charles Raja, is received the Young Scientist Fellowship - 2009–2010 from the TNSCST, Career Award for Young Teachers 2014-18 from AICTE, and Early Career Research Award 2017-20 from DST-SERB. He is a co-author of a book entitled ‘Electrical Power Systems: Analysis, Security and Deregulation’, PHI Publication, 2017. He has successfully completed a sponsored research project from DST - SERB with the cost of Rs. 30 Lakh. He has published 25 papers in International/National conferences and 45 research papers in International/National Journals. He is a member of PES Society and a Life member of ISTE. He is presently working as an Associate Professor in the Department of EEE, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India. His topics of interest include smart grid, IoT, Machine learning, and power system restructuring. E-mail: [email protected]

T. Karthick

T. Karthick, received his Bachelor’s Degree in Electrical Engineering from Anna University Chennai, India in 2008 and a Master’s degree in Power Management from Anna University, Chennai, India in 2012. He has completed his Ph.D. at the Department of Electrical and Electronics Engineering, Thiagarajar College of Engineering, India. He is currently the Managing Director of Quantanics Techserv Pvt.Ltd., Madurai, Tamil Nadu, India. His areas of interest include IoT, IIoT, smart grid, Industry 4.0, machine learning, deep learning, artificial intelligence, Block Chain and IOT Leanardo, Utilities, Smart Grids/Meters, Smart Cities, AMI, and Analytics. E-mail: [email protected]

P. Venkatesh

P. Venkatesh has received the BOYSCAST Fellowship award in the year 2006 from the Department of Science and Technology, India for carrying out Post-Doctoral Research Work at the Pennsylvania State University, U.S.A. He is an author of a book entitled ‘Electrical Power Systems: Analysis, Security and Deregulation’, PHI Publication, 2017. He has published 4 papers in International/National conferences and 63 research papers in International/National Journals. He is a Fellow in the Institution of Engineers (India) (IE), and a life member in Indian Society for Technical Education (ISTE). Currently, he is working as a Professor in the Department of Electrical and Electronics Engineering, Thiagarajar College of Engineering, Madurai, India. His area of research interest includes renewable energy system, power system restructuring, and the application of evolutionary computation techniques to power system problems. E-mail: [email protected]