Results in Engineering 19 (2023) 101296 Available online 17 July 2023 2590-1230/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). Contents lists available at ScienceDirect Results in Engineering journal homepage: www.sciencedirect.com/journal/results-in-engineering Research paper A new approach to seasonal energy consumption forecasting using temporal convolutional networks Abdul Khalique Shaikh a,∗ , Amril Nazir b , Nadia Khalique a , Abdul Salam Shah c , Naresh Adhikari d a Department of Information Systems, Sultan Qaboos University, Muscat, 123, Oman b Department of Information Systems and Technology Management, Zayed University, Abu Dhabi, 144534, United Arab Emirates c Department of Computer Engineering, University of Kuala Lumpur (UniKl-MIIT), Kuala Lumpur, 50250, Malaysia d Department of Computer Science, Slippery Rock University, 1 Morrow Way, Slippery Rock, PA 16057, USA A R T I C L E I N F O A B S T R A C T Keywords: Energy forecasting Seasonal energy Smart grids Temporal convolutional networks There has been a significant increase in the attention paid to resource management in smart grids, and several energy forecasting models have been published in the literature. It is well known that energy forecasting plays a crucial role in several applications in smart grids, including demand-side management, optimum dispatch, and load shedding. A significant challenge in smart grid models is managing forecasts efficiently while ensuring the slightest feasible prediction error. A type of artificial neural networks such as recurrent neural networks, are frequently used to forecast time series data. However, due to certain limitations like vanishing gradients and lack of memory retention of recurrent neural networks, sequential data should be modeled using convolutional networks. The reason is that they have strong capabilities to solve complex problems better than recurrent neural networks. In this research, a temporal convolutional network is proposed to handle seasonal short- term energy forecasting. The proposed temporal convolutional network computes outputs in parallel, reducing the computation time compared to the recurrent neural networks. Further performance comparison with the traditional long short-term memory in terms of MAD and sMAPE has proved that the proposed model has outperformed the recurrent neural network. 1. Introduction Different smart city initiatives by government and private organiza- tions have incorporated information and communication technologies (ICTs) to meet cities’ growing challenges. International policies and sci- entific literature have widely embraced the smart homes, smart grids, and overall intelligent city concept. This concept makes cities smarter for citizens by utilizing many ICT innovations hitting us alarmingly. According to scientific evidence, smart cities are based on the follow- ing foundational theories: ICTs, urban planning, environmental con- siderations, living labs, and creative industries [16]. In addition, the associated concepts illustrate how ICT can assist in addressing almost every urban challenge. The latest ICT trends identified in the literature analysis are smart grids, IoT, big data, open data, and e-government [20,32,45]. The topic of consideration for this study is smart grids. The world’s urban population makes up about half of the total population of the entire planet [59]. Cities are increasingly crowded, resulting in de- * Corresponding author. E-mail address: shaikh@squ.edu.om (A.K. Shaikh). clining quality and quantity of services for their residents. The increased population has caused challenges for the energy sector to produce en- ergy as per future demand [12]. The prediction of energy consumption is essential for effective demand-side management. The models of energy consumption prediction mostly use statistical and machine learning models [1,46]. The previously recurrent neural network-based models have been applied for short-term energy predic- tion, while the seasonal factor has not been considered [34]. Traditional statistical models have performed better with the small amount of data, but the energy consumption data has drastically increased with the in- crease in the urban population [41]. The statistical models have certain limitations; hence the deep learning models have been widely adopted to handle the time series energy consumption data. Recurrent neural networks are essential algorithms for forecasting energy consumption [40]. The other most prominent algorithms are temporal convolutional neural networks (TCNs). The governments focus on providing effective https://doi.org/10.1016/j.rineng.2023.101296 Received 31 March 2023; Received in revised form 10 July 2023; Accepted 11 July 2023