Temporal Convolutional Networks Applied to Energy-Related Time Series Forecasting

Abstract

Modern energy systems collect high volumes of data that can provide valuable information about energy consumption. Electric companies can now use historical data to make informed decisions on energy production by forecasting the expected demand. Many deep learning models have been proposed to deal with these types of time series forecasting problems. Deep neural networks, such as recurrent or convolutional, can automatically capture complex patterns in time series data and provide accurate predictions. In particular, Temporal Convolutional Networks (TCN) are a specialised architecture that has advantages over recurrent networks for forecasting tasks. TCNs are able to extract long-term patterns using dilated causal convolutions and residual blocks, and can also be more efficient in terms of computation time. In this work, we propose a TCN-based deep learning model to improve the predictive performance in energy demand forecasting. Two energy-related time series with data from Spain have been studied: the national electric demand and the power demand at charging stations for electric vehicles. An extensive experimental study has been conducted, involving more than 1900 models with different architectures and parametrisations. The TCN proposal outperforms the forecasting accuracy of Long Short-Term Memory (LSTM) recurrent networks, which are considered the state-of-the-art in the field.

Optimal Day-Ahead Self-Scheduling and Operation of Prosumer Microgrids Using Hybrid Machine Learning-Based Weather and Load Forecasting

“…The advantages of machine learning algorithms have been utilized for forecasting short-term load or weather data, separately [22,23]. The time-series forecasting methods for prediction of load demand and weather conditions have received a great deal of attention [13,14,24,25]. However, most of them merely focused on one or two parameters (e.g.…”

“…Forecasting results of the utilized machine learning algorithms are compared to analyze the performance of each method. This approach has been used in many recent studies where specific parameters are forecasted, and results are discussed accordingly [24,25]. However, in day-ahead scheduling and operation of PMGs, monitoring daily predicted load and weather data should be considered as one of the important tasks of the PMG EMS.…”

“…Therefore, (21)(22)(23)(24) are considered for prohibiting the BSS to be fully charged and discharged. Since it is not possible to charge and discharge the BSS at the same operational time interval, it is necessary to apply (25), which shows the binary value for charging and discharging of BSS.…”

“… Proposing a novel optimal operation of PMGs using simultaneous forecasting of demand and supply side's parameters; Although several papers in the literature, such as [9][10][11][22][23][24][25] [19,20]. However, they considered a "pre-defined" load pattern without any particular load forecasting methodology.…”

Context-Aware Traffic Prediction Framework Based on Series Decomposition

“…Among other methods, we can consider like TCN that is deemed applicable to specific issues, and simply relies on the dilation convolution capturing longer temporal information with a growing reception field. TCN ignores the local periodic characteristics of convolution features [25]. Other hybrid TCN approaches, e.g., Multi-Stage TCN (MS-TCN), Ensemble Empirical Mode Decomposition-Temporal Convolutional Network (EEMD-TCN) and Temporal Graph Convolutional Network (T-GCN) are integrating external information to help improving the forecasting ability [26], [27].…”