Short-term load forecasting with deep learning: Improving performance with post-training specialization

Short-term power load prediction is not only a critical part of power system dispatching but also an essential task for power marketing, grid planning, and other management departments; deep learning is an artificial intelligence method that has gained extensive attention in recent years; this paper selected three models with the most typical signs of deep learning recurrent neural networks and studied their model performance for short-term power load predictions, including power load data preprocessing, feature selection of prediction models, determination of model parameters, and comparison of prediction characteristics between shallow and deep networks, to explore and prove the applicability of deep learning neural networks in short-term power load predictions, ultimately improving the accuracy of short-term load predictions.KeywordsPower loadDeep learningPrediction

In traditional short-term load dispatching methods for distributed power systems, complex data processing, and low recognition accuracy hinder effective system optimization.To address this issue, an intelligent short-term load dispatching method for distributed power systems based on deep learning was proposed, aiming to improve voltage quality and reduce network losses.The proposed method first collected historical load data from the system, including active power and reactive power, and utilized a long short-term memory neural network (LSTM) for short-term load forecasting.In the forecasting process, the input gate, output gate, and forget gate operations were employed to accurately handle load variations.Based on the forecasting results, a short-term load dispatching model was constructed to minimize network losses, voltage deviations, and power abandonment rates, while incorporating constraints such as power flow, node voltage, and load regulation coefficients.The bee colony Quantum-behaved Particle Swarm Optimization (QPSO) algorithm was used to achieve intelligent load allocation in the distributed power system.Experimental results show that, the proposed method effectively controls voltage fluctuations within 0.4 p.u. while reducing network losses to below 0.14 MW, significantly improving the overall system performance.The conclusion provides scientific technical supports for the optimization and dispatching of distributed power systems and validates the applicability and effectiveness of intelligent methods in complex power system scenarios.

SAC-ConvLSTM: A novel spatio-temporal deep learning-based approach for a short term power load forecasting

Accurate short-term load forecasting (STLF) is critical to the operation and control of the power grid. In the literature, deep learning techniques have shown its superiority for accurate load forecasting over other methods due to its strong regression ability. However, the training process of the deep learning systems is time-consuming providing that they contain a large number of parameters due to the complex structure. This paper presents a novel broad learning system (BLS) approach for load forecasting. The BLS, which is established as a flat network, is an alternative to the deep learning system by expanding the structure in a wide sense. The expanded broad structure of BLS ensures its strong approximation capability of nonlinear mapping. Moreover, the BLS has a much faster training process compared to deep learning due to its simple structure. The proposed method is tested with an open dataset at low-voltage level from Australia and the testing results demonstrate the effectiveness of the proposed method.

Short-term building energy consumption prediction is of great significance for the optimized operation of building energy management systems and energy conservation. Due to the high-dimensional nonlinear characteristics of building heat loads, traditional single machine-learning models cannot extract the features well. Therefore, in this paper, a combined model based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), four deep learning (DL), and the autoregressive integrated moving average (ARIMA) models is proposed. The DL models include a convolution neural network, long- and short-term memory (LSTM), bi-directional LSTM (bi-LSTM), and the gated recurrent unit. The CEEMDAN decomposed the heating load into different components to extract the different features, while the DL and ARIMA models were used for the prediction of heating load features with high and low complexity, respectively. The single-DL models and the CEEMDAN-DL combinations were also implemented for comparison purposes. The results show that the combined models achieved much higher accuracy compared to the single-DL models and the CEEMDAN-DL combinations. Compared to the single-DL models, the average coefficient of determination (R2), root mean square error (RMSE), and coefficient of variation of the RMSE (CV-RMSE) were improved by 2.91%, 47.93%, and 47.92%, respectively. Furthermore, CEEMDAN-bi-LSTM-ARIMA performed the best of all the combined models, achieving values of R2 = 0.983, RMSE = 70.25 kWh, and CV-RMSE = 1.47%. This study provides a new guide for developing combined models for building energy consumption prediction.

Accurate short-term electric load forecasting is significant for the smart grid. It can reduce electric power consumption and ensure the balance between power supply and demand. In this paper, the Stacked Denoising Auto-Encoder (SDAE) is adopted for short-term load forecasting using four factors: historical loads, somatosensory temperature, relative humidity, and daily average loads. The daily average loads act as the baseline in final forecasting tasks. Firstly, the Denoising Auto-Encoder (DAE) is pre-trained. In the symmetric DAE, there are three layers: the input layer, the hidden layer, and the output layer where the hidden layer is the symmetric axis. The input layer and the hidden layer construct the encoding part while the hidden layer and the output layer construct the decoding part. After that, all DAEs are stacked together for fine-tuning. In addition, in the encoding part of each DAE, the weight values and hidden layer values are combined with the original input layer values to establish an SDAE network for load forecasting. Compared with the traditional Back Propagation (BP) neural network and Auto-Encoder, the prediction error decreases from 3.66% and 6.16% to 2.88%. Therefore, the SDAE-based model performs well compared with traditional methods as a new method for short-term electric load forecasting in Chinese cities.

Short-term load forecasting is essential for the power company's operation and grid operators because it is necessary to ensure adequate capacity and proper power generation arrangement; this will affect operating efficiency and short-term decisions. Meanwhile, the Covid-19 epidemic as a nonlinear factor will be effective in short-term load forecasting and based on previous solutions, electrical load forecasting may not be accurate. A nonlinear and complex relationship between the factors affecting the load forecasting problem explains the need to use intelligent methods such as machine learning. This paper analyses the effect of Covid-19 epidemic countermeasures on short-term electric load forecasting in Iran. To forecast the short term electrical load, a deep neural network with a hybrid architecture and peak power consumption data, average temperature, and Covid-19 epidemic countermeasure data over 15 months during the Covid-19 epidemic was used. The results indicate an increase in forecasting accuracy considering the countermeasure's data. Also, the proposed model validation with data related to the fourth wave of the Covid-19 epidemic and the data of countermeasures modeling in Iran show the effectiveness and reasonable accuracy of the proposed model during the Covid19 epidemic.

The electricity load forecasting handles the process of determining how much electricity will be available at a given time while maintaining the balance and stability of the power grid. The accuracy of electricity load forecasting plays an important role in ensuring safe operation and improving the reliability of power systems and is a key component in the operational planning and efficient market. For many years, a conventional method has been used by using historical data as input parameters. With swift progress and improvement in technology, which shows more potential due to its accuracy, different methods can be applied depending on the identified model. To enhance the forecast of load, this paper introduces and proposes a framework developed on graph database technology to archive large amounts of data, which collects measured data from electrical substations in Pristina, Kosovo. The data includes electrical and weather parameters collected over a four-year timeframe. The proposed framework is designed to handle short-term load forecasting. Machine learning Linear Regression and deep learning Long Short-Term Memory algorithms are applied to multiple datasets and mean absolute error and root mean square error are calculated. The results show the promising performance and effectiveness of the proposed model, with high accuracy in load forecasting.

Short-Term Load Forecast at residential house level plays a critical role in home energy management system. While a variety of machine learning based load forecasting methods have been proposed, their prediction performance have not been assessed against cyber threats on smart meters which have been increasingly reported. This paper investigates the vulnerability of four extensively used machine learning algorithms for residential short-term load forecast against cyberattacks, including Nonlinear Auto Regression with external input (NARX) neural network, support vector machine (SVM), decision tree (DT), and long-short-term memory (LSTM) deep learning. We use the REFIT dataset which collected whole-house aggregated loads at 8-second intervals continuously from 20 houses over a two-year period in the U.K. The results were determined and show the predictions using NARX and LSTM. Four cyberattack models are investigated, including pulse, scale, ramp, and random. The vulnerability assessment results indicate the LSTM provides the most accurate prediction without cyberattacks. However, the prediction accuracy of the LSTM fluctuates when there are cyber-attacks. Among the four cyberattacks, the random attack triggered the larges variations on the predication results.

Short-term load forecasting plays a significant role in the operation of power systems. Recently, deep learning has been generally employed in short-term load forecasting, primarily in the extraction of the characteristics of digital information in a single dimension without taking into account of the impact of external variables, particularly non-digital elements on load characteristics. In this paper, we propose a joint MFF-SAM-GCN to realize short-term load forecasting. First, we utilize a Bi-directional Long Short-Term Memory (Bi-LSTM) network and One-Dimensional Convolutional Neural Network (1D-CNN) in parallel connection to form a multi-feature fusion (MFF) framework, which can extract spatiotemporal correlation features of the load data. In addition, we introduce a Self-Attention Mechanism (SAM) to further enhance the feature extraction capability of the 1D-CNN network. Then with the deployment of a Graph Convolutional Network (GCN), the external non-digital features such as weather, strength, and direction of wind, etc., are extracted. Moreover, the generated weight matrices are incorporated into the load features to enhance feature recognition ability. Finally, we exploit Bayesian Optimization (BO) to find the optimal hyperparameters of the model to further improve the prediction accuracy. The simulation is taken from our proposed model and six benchmark schemes by using the bus load dataset of the Shandong Open Data Network, China. The results show that the RMSE of our proposed MFF-SAM-GCN model is 0.0284, while the SMAPE is 9.453%,the MBE is 0.025, and R-squared is 0.989, which is better than the selected three traditional machine learning methods and the three deep learning models.

Short-term residential load forecasting plays a crucial role in smart grids, ensuring an optimal match between energy demands and generation. With the inherent volatility of residential load patterns, deep learning has gained attention due to its ability to capture complex nonlinear relationships within hidden layers. However, most existing studies have relied on default loss functions such as mean squared error (MSE) or mean absolute error (MAE) for neural networks. These loss functions, while effective in overall prediction accuracy, lack specialized focus on accurately predicting load peaks. This article presents a comparative analysis of soft-DTW loss function, a smoothed formulation of Dynamic Time Wrapping (DTW), compared to other commonly used loss functions, in order to assess its effectiveness in improving peak prediction accuracy. To evaluate peak performance, we introduce a novel evaluation methodology using confusion matrix and propose new errors for peak position and peak load, tailored specifically for assessing peak performance in short-term load forecasting. Our results demonstrate the superiority of soft-DTW in capturing and predicting load peaks, surpassing other commonly used loss functions. Furthermore, the combination of soft-DTW with other loss functions, such as soft-DTW + MSE, soft-DTW + MAE, and soft-DTW + TDI (Time Distortion Index), also enhances peak prediction. However, the differences between these combined soft-DTW loss functions are not substantial. These findings highlight the significance of utilizing specialized loss functions, like soft-DTW, to improve peak prediction accuracy in short-term load forecasting.

Short-term power load forecasting and management based on quadratic decomposition and deep learning

Accurate cooling load prediction is essential for energy-efficient HVAC system operation. However, the stochastic and nonlinear nature of load data challenges conventional neural networks, causing prediction delays and errors. To address this, a novel hybrid model is developed. The approach first applies a two-stage decomposition (CEEMDAN with K-means and VMD) to process complex cooling load data. Then, a CNN-BiLSTM network optimized by the Crested Porcupine Optimizer and integrated with an attention mechanism is constructed for prediction. Experimental results demonstrate the model’s high performance, achieving a 96.75% prediction accuracy with a MAPE of 3.25% and an R2 of 0.9929. The proposed model shows strong robustness and generalization, providing a reliable reference for intelligent building energy management.

The forecasting of short term electricity load plays a vital role in power system. It is essential for the power system's reliable, secure, and cost-effective functioning. This paper contributes significantly for enhancing the accuracy of short term electricity load forecasting. It presents a hybrid forecasting model called Gated Recurrent Unit with Ensemble Empirical Mode Decomposition and Boruta feature selection (EBGRU). It is a hybrid model that addresses the non-stationary, non-linearity and noisy issues of the time series input by using Ensemble Empirical Mode Decomposition (EEMD). It also addresses overfitting and curse of dimensionality issues of load forecasting by identifying the pertinent features using Boruta wrapper feature selection. It effectively handles the uncertainty and temporal dependency characteristics of load and forecasts the future load using deep learning based Gated Recurrent Unit (GRU). The proposed EBGRU model is experimented by using European and Australian Electricity load datasets. The temperature has high correlation with load demand. In this study, both load and temperature features are considered for the accurate short term load forecasting. The experimental outcome demonstrates that the proposed EBGRU model outperforms other deep learning models such as RNN, LSTM, GRU, RNN with EEMD and Boruta (EBRNN) and LSTM with EEMD and Boruta (EBLSTM).

Smart grid being the major component of our power system is progressing fast to become more flexible and interactional. For the proper operation and planning in these systems, the load forecasting is of prime concern. In spite of the fact that many traditional methods are present for load forecasting but still a prediction of electrical load is needed to be explored further as the variation from time and surrounding weather conditions make it too complex. In this paper, we present deep learning based recurrent neural network approach for the short term load prediction using data captured from smart meters. This framework helps us to handle the over-fitting problem and uncertainty issues. The Tensor flow as a deep learning platform is used for our implementation. We compared our proposed model with the autoregressive integrated moving average model (ARIMA) and Fbprophet model and it is manifested that the proposed model outperforms in terms of RMSE, MSE, and MAE.