Search

Focusing on the issue of significant randomness and low latitude of short-term household electrical load data, this paper proposes a novel short-term load multi-step forecasting method based on long-and short-term time series network (LSTNet). Firstly, the time sliding window method is used to sample massive historical load data to construct feature maps as input. Secondly, convolutional neural network (CNN) and long short-term memory (LSTM) are used to capture temporal short-term local information and long-term related information respectively, and autoregressive (AR) models are used as linear components. Then, the models will be evaluated using a scheme called walk-forward validation, and the average absolute percentage error (MAPE) and root mean square error (RMSE) are used as accuracy evaluation indicators. Finally, the four-year electric load data of a family in Paris, France is used to verify the proposed method and comprehensively compare the proposed method with the three most popular load forecasting algorithms. The experimental results show that in the prediction results for the next week, the MAPE and RMSE of the prediction method proposed in the paper are smaller than those of other algorithms, which can more effectively express the time series relationship of household short-term load and have higher prediction accuracy.

To fully mine the relationship between temporal features in load data, improve the accuracy and efficiency of short-term load forecasting and overcome the difficulties caused by load nonlinearity and volatility in accurate load forecasting. In this paper, a hybrid neural network short-term load forecasting model based on temporal convolutional network (TCN) and gated recurrent unit (GRU) is proposed. Firstly, the correlation between meteorological features and load is measured with the distance correlation coefficient, and the fixed-length sliding time window method is used to reconstruct the features. Next, temporal convolutional network is adopted to extract the hidden historical information and time relationship including meteorological features, electricity price, etc., and a better-performing gated recurrent unit is utilized for perdition. Furthermore, the state-of-the-art AdaBelief optimizer and Attention mechanism are utilized to enhance the prediction accuracy and efficiency. The effectiveness and superiority of the proposed model are verified by load and weather data from Spain and PJM power system data. Short-term load forecasting results in different periods and comprehensive comparisons with the performance of different models show that the proposed model can provide accurate load forecasting results rather quickly. The highlights of this paper are that temporal convolutional network and gated recurrent unit are combined for load forecasting for the first time, and the forecasting performance is improved by the novel optimizer AdaBelief and feature selection based on distance correlation coefficient.

Short-term load forecasting (STLF) with excellent precision and prominent efficiency plays a significant role in the stable operation of power grid and the improvement of economic benefits. In this paper, a novel model based on data mining and deep learning is proposed. Firstly, the preprocessing of data includes normalization of historical load, and fuzzification of influencing factors (meteorological factors, date types and economy) based on Pearson correlation coefficient (PCC). Secondly, kernel fuzzy c-means (KFCM) modified by particle swarm optimization (PSO-KFCM) algorithm clusters the daily load curve. In the clustering experiments, the within-cluster sum of squared error (SSE) index is presented to determine the number of clusters and the clustering validity has a 31.9% enhancement compared with the traditional FCM algorithm. Thirdly, the cosine similarity establishes the resemblance between the prediction date and each cluster, and the similar cluster is determined according to the principle of maximum similarity. Finally, a multivariate and multi-step hybrid model MMCNN-LSTM based on convolution neural network (CNN) and long short-term memory (LSTM) neural network is proposed to forecast the load in following 24 hours, in which similar cluster data is applied to training set. To demonstrate the effectiveness of proposed integrated technique, the accuracy has been verified in three predictive experiments. The fruitful results indicated that the average mean absolute percent error (MAPE) in the entire test set was only 1.34%, a 3.02% reduction compared to a single LSTM.

International audience; Since electricity plays a crucial role in countries' industrial infrastructures, power companies are trying to monitor and control infrastructures to improve energy management and scheduling. Accurate forecasting is a critical task for a stable and efficient energy supply, where load and supply are matched. This article discusses various algorithms and a new hybrid deep learning model which combines long short-term memory networks (LSTM) and convolutional neural network (CNN) model to analyze their performance for short-term load forecasting. The proposed model is called parallel LSTM-CNN Network or PLCNet. Two real-world data sets, namely ''hourly load consumption of Malaysia '' as well as ''daily power electric consumption of Germany'', are used to test and compare the presented models. To evaluate the tested models' performance, root mean squared error (RMSE), mean absolute percentage error (MAPE), and R-squared were used. In total, this article is divided into two parts. In the first part, different machine learning models, including the PLCNet, predict the next time step load. In the second part, the model's performance, which has shown the most accurate results in the first part, is discussed in different time horizons. The results show that deep neural networks models, especially PLCNet, are good candidates for being used as short-term prediction tools. PLCNet improved the accuracy from 83.17% to 91.18% for the German data and achieved 98.23% accuracy in Malaysian data, which is an excellent result in load forecasting.

In this study, a new technique is proposed to forecast short-term electrical load. Load forecasting is an integral part of power system planning and operation. Precise forecasting of load is essential for unit commitment, capacity planning, network augmentation and demand side management. Load forecasting can be generally categorized into three classes such as short-term, midterm and long-term. Short-term forecasting is usually done to predict load for next few hours to few weeks. In the literature, various methodologies such as regression analysis, machine learning approaches, deep learning methods and artificial intelligence systems have been used for short-term load forecasting. However, existing techniques may not always provide higher accuracy in short-term load forecasting. To overcome this challenge, a new approach is proposed in this paper for short-term load forecasting. The developed method is based on the integration of convolutional neural network (CNN) and long short-term memory (LSTM) network. The method is applied to Bangladesh power system to provide short-term forecasting of electrical load. Also, the effectiveness of the proposed technique is validated by comparing the forecasting errors with that of some existing approaches such as long short-term memory network, radial basis function network and extreme gradient boosting algorithm. It is found that the proposed strategy results in higher precision and accuracy in short-term load forecasting.

While the general belief is that the best way to predict electric load is through individualized models, the existing studies have focused on one-for-all models because the individual models are difficult to train and require a significantly larger data accumulation time per individual. In recent years, applying deep learning for forecasting electric load has become an important research topic but still one-for-all has been the main approach. In this work, we adopt transfer learning and meta learning that can be smoothly integrated into deep neural networks, and show how a high-performance individualized model can be formed using the individual's data collected over just several days. This is made possible by extracting the common patterns of many individuals using a sufficiently large dataset, and then customizing each individual model using the specific individual's small dataset. The proposed methods are evaluated over residential and non-residential datasets. When compared to the conventional methods, the meta learning model shows 7.84% and 15.07% RMSE improvements over the residential and non-residential datasets, respectively. Our results suggest that the individualized models can be used as effective tools for many short-term load forecasting tasks.

This paper presents a novel procedure for short-term load forecasting in distribution management systems. The load is forecasted for feeders that can be of a primarily residential, commercial, industrial or combined type. Each feeder has various amounts of distributed energy resources installed, which accounts for multiple different load patterns. Hence, the distribution management systems cannot use a single short-term load forecasting model for all forecasts. The proposed procedure addresses the specificity of each particular feeder type, by creating customized short-term load forecasting models. It uses a genetic algorithm to select the best inputs for different multiple linear regression models. The genetic algorithm chooses variables from a dataset constructed using load and temperature measurements. The dataset is extended by adding non-linear transformations and mutual interaction effects of the measurements, as well as calendar variables. This extension enables for the modelling of non-linear influences and extracts the non-linearity to the domain of input variables. The models? performance is assessed by the mean absolute percentage error. The proposed procedure is applied to a set of measurements collected from an US electric power utility, which operates in the city of Burbank, Cal., USA. The obtained multiple linear regression model is compared with a previously proposed na?ve benchmark, and a special comparison model, developed by correlation analysis. The proposed method is extendable to suit distribution management systems with different types of electricity consumers.

In an increasingly open electricity market environment, short-term load forecasting (STLF) can ensure the power grid to operate safely and stably, reduce resource waste, power dispatching, and provide technical support for demand-side response. Recently, with the rapid development of demand side response, accurate load forecasting can better provide demand side incentive for regional load of prosumer groups. Traditional machine learning prediction and time series prediction based on statistics failed to consider the non-linear relationship between various input features, resulting in the inability to accurately predict load changes. Recently, with the rapid development of deep learning, extensive research has been carried out in the field of load forecasting. On this basis, a feature selection algorithm based on random forest is first used in this paper to provide a basis for the selection of the input features of the load forecasting model. After the input features are selected, a hybrid neural network STLF algorithm based on multi-model fusion is proposed, of which the main structure of the hybrid neural network is composed of convolutional neural network and bidirectional gated recurrent unit (CNN-BiGRU). The input data is obtained by using long sliding time windows of different steps, then multiple CNN-BiGRU models are trained respectively. The forecasting results of multiple models are averaged to get the final forecasting load value. The load datasets come from a region in New Zealand and a region in Zhejiang, China, are used as load forecast examples. Finally, a variety of load forecasting algorithms are introduced for comparison. The experimental results show that our method has a higher accuracy than comparison models.

Load forecasting has always been an essential part of power system planning and operation. In recent decades, the competition of the market and the requirements of renewable integration lead more attention to probabilistic load forecasting methods, which can reflect forecasting uncertainties through prediction intervals and hence benefit decision-making activities in system operation. Moreover, with the development of smart grid and power metering techniques, power companies have collected enormous load data about electricity customers and substations. The abundant load data allow us to utilize medium voltage measurement data to achieve better accuracy in high voltage transmission substation load forecasting. In this paper, a bottom-up probabilistic forecasting method is proposed for high voltage transmission substation short-term load forecasting, in which the probability distributions of medium voltage day-ahead load forecasting values are estimated and added up to form high voltage load predictions. Two bottom-up frameworks based on load patterns collected from medium voltage outgoing lines and substations are proposed respectively, in which mismatches between load data at different levels are estimated for correcting high voltage predictions. The comparison of predictions obtained by traditional and bottom-up methods demonstrates that the proposed method obtains high voltage load forecasting more accurately and give narrower prediction intervals.

With the innovation of power market and the development of energy intelligent technology, load forecasting technology as an important direction of power system development plays an important role in power system planning. Aiming at the problem of insufficient feature extraction and low prediction accuracy, a short-term load forecasting model of multi-scale CNN-LSTM hybrid neural network considering the real-time electricity price is proposed in this paper. Firstly, the maximum information coefficient method is used to analyze the correlation between electricity price and load. The historical load, real-time electricity price, weather and other factors are constructed in the form of continuous feature maps as input. Secondly, the Convolutional Neural Network (CNN) is used to cascade the shallower and deeper feature from four different scales. Feature vectors of different scales are fused as the input of Long Short Term Memory (LSTM) network , and LSTM network is used for short-term load forecasting. Finally, the proposed method is used to predict the real load data of a city in Liaoning Province. The experimental results show that the proposed method has higher prediction accuracy than the standard LSTM model, Support Vector Machine (SVM) model, Random Forest (RF) model and Auto Regressive Integrated Moving Average (ARIMA) model. Besides, the prediction results show that this study has high application value and provides a new way for the development of power load forecasting.

Residential demand response is vital for the efficiency of power system. It has attracted much attention from both academic and industry in recent years. Accurate short-term load forecasting is a fundamental task for demand response. While short-term forecasting for aggregated load data has been extensively studied, load forecasting for individual residential users is still challenging due to the dynamic and stochastic characteristic of single users’ electricity consumption behaviors, i.e., the variability of the residential activities. To address this challenge, this paper presents a short-term residential load forecasting framework, which makes use of the spatio-temporal correlation existing in appliances’ load data through deep learning. Multiple time series are conducted in the framework to describe electricity consumption behaviors and their internal spatio-temporal relationship. And a method based on deep neural network and iterative ResBlock is proposed to learn the correlation among different electricity consumption behaviors for short-term load forecasting. Experiments based on real world measurements have been conducted to evaluate the performance of the proposed forecasting approach. The results show that both the appliances’ load data and iterative ResBlocks can help to improve the forecasting performance. Compared with existing methods, measurements on Root Mean Squared Error, Mean Absolute Error and Mean Absolute Percentage Error for the proposed approach are reduced by 3.89%-20.00%, 2.18%-22.58% and 0.69%-32.78%. In addition, further experiments are conducted to evaluate the impact of using appliances’ load data, iterative ResBlocks as well as other factors for the proposed approach.

Short-term load forecasting is a critical task in the smart grid, which can be used to optimize power deployment and reduce power losses. Recurrent neural networks (RNNs) are the most popular deep learning models for short-term load forecasting. However, despite of achieving better forecasting accuracy than the traditional models, the performance of the existing RNN-based load forecasting approaches is still unsatisfactory for practical usage. Therefore, in this work, we have proposed input attention mechanism (IAM) and hidden connection mechanism (HCM) to greatly enhance the accuracy and efficiency of RNN-based load forecasting models. Specifically, we use IAM to assign the importance weights on input layers, which have better performances in both efficiency and accuracy than traditional attention mechanisms. To further enhance the models’ efficiency, HCM is then applied to utilize residual connections to enhance the model’s converging speed. We have applied both IAM and HCM on four state-of-the-art RNN implementations, and then conducted extensive experimental studies on two public datasets. Experimental results show that the proposed RNNs with IAM and HCM models achieve much better performances than the state-of-the-art baselines in both accuracy and efficiency. Ablation studies show that both IAM and HCM are essential to achieve such superior performances.

In most deep learning-based load forecasting, an intact dataset is required. Since many real-world datasets contain missing values for various reasons, missing imputation using deep learning is actively studied. However, missing imputation and load forecasting have been considered independently so far. In this article, we provide a deep learning framework that jointly considers missing imputation and load forecasting. We consider a family of autoencoder/long short-term memory (LSTM) combined models for missing-insensitive load forecasting. Specifically, autoencoder (AE), denoising autoencoder (DAE), convolutional autoencoder (CAE), and denoising convolutional autoencoder (DCAE) are considered for extracting features, of which the encoded outputs are fed into the input of LSTM. Our experiments show that the proposed DCAE/LSTM combined model significantly improves forecasting accuracy no matter what missing rate or type (random missing, consecutive block missing) occurs compared to the baseline LSTM.

Modeling an accurate forecasting model for short-term load is still challenging due to the diverse causes of load changing and lack of information on many of these causes. In this paper, error trend is used to reveal the trend effect caused by unknown load affecting factors and proposed adaptive second learning of error trend (A-SLET) to self-adapt the trend effect. Furthermore, the training set is classified based on balance point temperature and then parallelly trained and tested adaptive forecaster for hot days and adaptive forecaster for cold days with proper data. Combining A-SLET with parallel forecasting and training set classification, Adaptive and Parallel forecasting strategy based on Second Learning of Error Trend (AP-SLET) is proposed. The work studied two distinct load patterns, one in the USA and the other in Australia. Considering the yearly forecasting horizon, MAPE of the adaptive and parallel forecasting strategy is 1.87%-4.04% for ME-Maine of New England and 2.81%-4.41% for New South Wales. Compared to the state-of-art forecasting methods, MAPE of the adaptive and parallel forecasting strategy is reduced by 17.03%-33.33%, RMSE and MAE are reduced by 34.05% and 35.38% respectively. The experimental results demonstrate the proposed strategy can transform unknown and unavailable load affecting factors into known forecasting features and then adapt it to improve forecasting performance. The proposed strategy is also forecaster independent and equally applicable to almost all load scenarios regardless of geographical and seasonal differences.

Load forecasting is one of the critical tasks for enhancing the energy efficiency of smart grids. Even though recent deep learning-based load forecasting models have shown excellent forecasting performance, one of the common problems they faced was that their forecasting accuracy was highly dependent on the data quality and quantity available for the model training. Collecting a sufficient amount of high-quality data is expensive and time-consuming. Recently, a generative adversarial network (GAN) has shown its potential as a solution to the data shortage problem by generating virtual data based on a small amount of real data, and several studies have used GAN to generate electric load data for training forecasting models. However, due to the noise data problem of GANs, their predictive performance also deteriorated. To solve this problem, in this study, we propose a two-stage data generation scheme that more effectively generates input and output variables for short-term load forecasting. In the first stage, we generate virtual calendar and temperature data used as input variables using a conditional tabular GAN (CTGAN). In the second stage, we generate electric load data corresponding to the input variables using a deep learning-based regression model. Lastly, we construct our forecasting model by training another regression model using a mixture of generated data and real data. To verify the effectiveness of our scheme, we conducted extensive experiments using various datasets and data generation models. We report some of the results.

As the development of smart grids and electricity markets around the world, short-term load forecasting (STLF) plays an increasingly important role in safe and economical operations of power systems. Facing massive multivariate and heterogeneous data from smart grid environment, traditional STLF methods no longer meet the requirements of comprehensive prediction performance in big data era. Therefore, a novel STLF model named EGA-STLF is proposed in this paper. The core idea of the proposed model is mainly based on distributed representation, bidirectional gated recurrent unit (Bi-GRU), and attention mechanism. EGA-STLF encodes non-numerical variables in input data into sparse vectors by distributed representation during preprocessing, which can more reasonably reflect the correlations between different variables across the temporal sequence. Moreover, the Bi-GRU layer in EGA-STLF processes the past and the future information simultaneously to fully extract temporal and nonlinear features from input data for the improvement of forecasting accuracy. In addition, the introduction of attention mechanism highlights the role of key features in load forecasting, which is beneficial to generate more accurate forecasting results for EGA-STLF. The validity and superiority of EGA-STLF is verified by taking the actual data from Victoria state and Queensland state in Australia for experimental research in which mean absolute percentage error (MAPE) and root mean square error (RMSE) are selected as evaluation metrics. The experimental results indicate that EGA-STLF outperforms the state-of-the-art models based on SVR and MLP in term of comprehensive forecasting accuracy. Hence, EGA-STLF is a promising method to create economic benefits in power industry.

Deep neural networks of deep learning algorithms can be applied into regressions and classifications. While the regression performances and classification performances of the deep neural networks are depending on the hyper-parameters of the deep neural networks. To mitigate the adverse effect of the hyper-parameters for the deep learning algorithms, this paper proposes deep forest regression for the short-term load forecasting of power systems. Deep forest regression includes two procedures, i.e., multi-grained scanning procedure and cascade forest procedure. These two procedures can be effectively trained by two completely random forests and two random forests with the default configuration. Then, the deep forest regression is applied into the short-term load forecasting of power systems. The forecasting performances of deep forest regression are compared with that of numerous intelligent algorithms and conventional regression algorithms under the model with the data of previous 7-day, 21-day, and 40-day. Besides, the forecasting performances of deep forest regression with different parameters are compared. The numerical results show that the deep forest regression with default configured parameters can increase the accuracy of the short-term forecasting and mitigate the influences of the experiences for the configuration of the hyper-parameters of deep learning model.

Short-term load forecasting (STLF) plays a vital role in the reliable, secure, and efficient operation of power systems. Since electric load variation results from diverse factors, accurate and stable load forecasting remains a challenging task. To increase the forecasting accuracy and stability, in this paper, we newly propose a short-term load forecasting method based on the cross multi-model and second decision mechanism. First, we combine horizontal and longitudinal training set selection method to construct the cross training sets, which acquire both the horizontal and longitudinal characteristics of the load variation. Second, to improve the generalization ability and extend the application scope, we construct forecasting multi-models by training multiple forecasting algorithms with cross training sets. Finally, to aggregate the forecasting outputs obtained by the forecasting multi-models, we propose a second decision mechanism based on a decision multi-model and adaptive weight allocation strategy, which overcomes the limited learning ability shortcoming of single decision models and further improves the forecasting accuracy. Case studies based on electrical load data from the state of Maine, the region of New England, Singapore, and New South Wales of Australia show that both the accuracy and the stability of the proposed method are superior to the compared models.

Electricity load forecasting is an important part of power system dispatching. Accurately forecasting electricity load have great impact on a number of departments in power systems. Compared to electricity load simulation (white-box model), electricity load forecasting (black-box model) does not require expertise in building construction. The development cycle of the electricity load forecasting model is much shorter than the design cycle of the electricity load simulation. Recent developments in machine learning have lead to the creation of models with strong fitting and accuracy to deal with nonlinear characteristics. Based on the real load dataset, this paper evaluates and compares the two mainstream short-term load forecasting techniques. Before the experiment, this paper first enumerates the common methods of short-term load forecasting and explains the principles of Long Short-term Memory Networks (LSTMs) and Support Vector Machines (SVM) used in this paper. Secondly, based on the characteristics of the electricity load dataset, data pre-processing and feature selection takes place. This paper describes the results of a controlled experiment to study the importance of feature selection. The LSTMs model and SVM model are applied to one-hour ahead load forecasting and one-day ahead peak and valley load forecasting. The predictive accuracy of these models are calculated based on the error between the actual and predicted loads, and the runtime of the model is recorded. The results show that the LSTMs model have a higher prediction accuracy when the load data is sufficient. However, the overall performance of the SVM model is better when the load data used to train the model is insufficient and the time cost is prioritized.

Short-Term Load Forecasting is critical for reliable power system operation, and the search for enhanced methodologies has been a constant field of investigation, particularly in an increasingly competitive environment where the market operator and its participants need to better inform their decisions. Hence, it is important to continue advancing in terms of forecasting accuracy and consistency. This paper presents a new deep learning-based ensemble methodology for 24 h ahead load forecasting, where an automatic framework is proposed to select the best Box-Jenkins models (ARIMA Forecasters), from a wide-range of combinations. The method is distinct in its parameters but more importantly in considering different batches of historical (training) data, thus benefiting from prediction models focused on recent and longer load trends. Afterwards, these accurate predictions, mainly the linear components of the load time-series, are fed to the ensemble Deep Forward Neural Network. This flexible type of network architecture not only functions as a combiner but also receives additional historical and auxiliary data to further its generalization capabilities. Numerical testing using New England market data validated the proposed ensemble approach with diverse base forecasters, achieving promising results in comparison with other state-of-the-art methods.

Due to the strong volatility of the electrical load and the defect of a time-consuming problem, in addition to overfitting existing in published forecasting methods, short-term electrical demand is difficult to forecast accurately and robustly. Given the excellent capability of weight sharing and feature extraction for convolution, a novel hybrid method based on ensemble GoogLeNet and modified deep residual networks for short-term load forecasting (STLF) is proposed to address these issues. Specifically, an ensemble GoogLeNet with dense block structure is used to strengthen feature extraction ability and generalization capability. Meanwhile, a group normalization technique is used to normalize outputs of the previous layer. Furthermore, a modified deep residual network is introduced to alleviate a vanishing gradient problem in order to improve the forecasting results. The proposed model is also adopted to conduct probabilistic load forecasting with Monte Carlo Dropout. Two acknowledged public datasets are used to evaluate the performance of the proposed methodology. Multiple experiments and comparisons with existing state-of-the-art models show that this method achieves accurate prediction results, strong generalization capability, and satisfactory coverages for different prediction intervals, along with reducing operation times.

International audience; Electric grid, as we nowadays know it, is undergoing a significant transformation. What we are now witnessing is an undoubted change of trend towards a decentralized and decarbonized electric grid, where the electric generation based on local resources will take on special relevance. In this context, the encouragement of collective selfconsumption becomes one of the key issues when it comes to taking steps forward to this end. One of the aspects that will contribute to this aim is the development of a consumptionforecasting tool. Hence, a load-forecasting model based on NARX Neural Network is proposed in the following paper. The prediction of the next day (24h) load profile of an individual building is carried out aiming an optimal management of the flexible loads so to achieve the maximum self-consumption. To ensure a consistent behavior of the NARX Neural Network model, identification and removing of outliers, together with data normalization and fixing common time interval has been carried out. The first results of the research are promising, being obtained a 17,6% MAPE in NARX and 25,19% with LSTM model, both evaluated during a regular week on winter in adverse conditions .

Power system planning in numerous electric utilities merely relies on the conventional statistical methodologies, such as ARIMA for short-term electrical load forecasting, which is incapable of determining the non-linearities induced by the non-linear seasonal data, which affect the electrical load. This research work presents a comprehensive overview of modern linear and non-linear parametric modeling techniques for short-term electrical load forecasting to ensure stable and reliable power system operations by mitigating non-linearities in electrical load data. Based on the findings of exploratory data analysis, the temporal and climatic factors are identified as the potential input features in these modeling techniques. The real-time electrical load and meteorological data of the city of Lahore in Pakistan are considered to analyze the reliability of different state-of-the-art linear and non-linear parametric methodologies. Based on performance indices, such as Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MAE), the qualitative and quantitative comparisons have been conferred among these scientific rationales. The experimental results reveal that the ANN–LM with a single hidden layer performs relatively better in terms of performance indices compared to OE, ARX, ARMAX, SVM, ANN–PSO, KNN, ANN–LM with two hidden layers and bootstrap aggregation models.

Short-term load forecasting (STLF) is fundamental for the proper operation of power systems, as it finds its use in various basic processes. Therefore, advanced calculation techniques are needed to obtain accurate results of the consumption prediction, taking into account the numerous exogenous factors that influence the results’ precision. The purpose of this study is to integrate, additionally to the conventional factors (weather, holidays, etc.), the current aspects regarding the global COVID-19 pandemic in solving the STLF problem, using a convolutional neural network (CNN)-based model. To evaluate and validate the impact of the new variables considered in the model, the simulations are conducted using publicly available data from the Romanian power system. A comparison study is further carried out to assess the performance of the proposed model, using the multiple linear regression method and load forecasting results provided by the Romanian Transmission System Operator (TSO). In this regard, the Mean Squared Error (MSE), the Mean Absolute Error (MAE), the Mean Absolute Percentage Error (MAPE), and the Root Mean Square Error (RMSE) are used as evaluation indexes. The proposed methodology shows great potential, as the results reveal better error values compared to the TSO results, despite the limited historical data.

International audience; Short-term electricity load forecasting is key to the safe, reliable, and economical operation of power systems. An important challenge that arises with high-frequency load series, e.g., hourly load, is how to deal with the complex seasonal patterns that are present. Standard approaches suggest either removing seasonality prior to modeling or applying time series decomposition. This work proposes a hybrid approach that combines Singular Spectrum Analysis (SSA)-based decomposition and Artificial Neural Networks (ANNs) for day-ahead hourly load forecasting. First, the trajectory matrix of the time series is constructed and decomposed into trend, oscillating, and noise components. Next, the extracted components are employed as exogenous regressors in a global forecasting model, comprising either a Multilayer Perceptron (MLP) or a Long Short-Term Memory (LSTM) predictive layer. The model is further extended to include exogenous features, e.g., weather forecasts, transformed via parallel dense layers. The predictive performance is evaluated on two real-world datasets, controlling for the effect of exogenous features on predictive accuracy. The results showcase that the decomposition step improves the relative performance for ANN models, with the combination of LSTM and SAA providing the best overall performance.

The optimal operation planning (OOP) of electrical distribution systems (EDS) is very sensible to the quality of the short-term load forecasts. Assuming aggregated demands in EDS as univariate non-stationary seasonal time series, and based on historical measurements gathered by smart meters, this paper presents a parsimonious short-term load forecasting method to estimate the expected outcomes of future demands, and the standard deviations of forecast errors. The chosen short-term load forecasting method is an adaptation of the multiplicative autoregressive integrated moving average (ARIMA) models. Seasonal ARIMA models are parsimonious forecasting techniques because they require very few parameters and low computational resources to provide an adequate representation of stochastic time series. Two approaches are used in this paper to estimate the parameters that constitute the proposed multiplicative ARIMA model: a frequentist and a Bayesian approach. Advantages and disadvantages of both methods are compared by simulating a centralized self-healing scheme of a real EDS that uses the forecasts to deploy a robust restoration plan. Results shown that the proposed seasonal ARIMA model is a fast, precise, straightforward and adaptable load forecasting method, suitable for OOP of highly supervised EDS.

Electric load forecasting has always been a key component of power grids. Many countries have opened up electricity markets and facilitated the participation of multiple agents, which create a competitive environment and reduce costs to consumers. In the electricity market, multi-step short-term load forecasting becomes increasingly significant for electricity market bidding and spot price calculation, but the performances of traditional algorithms are not robust and unacceptable enough. In recent years, the rise of deep learning gives us the opportunity to improve the accuracy of multi-step forecasting further. In this paper, we propose a novel model multi-scale convolutional neural network with time-cognition (TCMS-CNN). At first, a deep convolutional neural network model based on multi-scale convolutions (MS-CNN) extracts different level features that are fused into our network. In addition, we design an innovative time coding strategy called the periodic coding strengthening the ability of the sequential model for time cognition effectively. At last, we integrate MS-CNN and periodic coding into the proposed TCMS-CNN model with an end-to-end training and inference process. With ablation experiments, the MS-CNN and periodic coding methods had better performances obviously than the most popular methods at present. Specifically, for 48-step point load forecasting, the TCMS-CNN had been improved by 34.73%, 14.22%, and 19.05% on MAPE than the state-of-the-art methods recursive multi-step LSTM (RM-LSTM), direct multi-step MS-CNN (DM-MS-CNN), and the direct multi-step GCNN (DM-GCNN), respectively. For 48-step probabilistic load forecasting, the TCMS-CNN had been improved by 3.54% and 6.77% on average pinball score than the DM-MS-CNN and the DM-GCNN. These results show a great promising potential applied in practice.

The load of transformers shows higher volatility and uncertainty than do the system-level and substation-level loads. This paper proposes a two-stage short-term load forecasting (STLF) model for power transformers. 1) Three state-of-the-art technologies are applied to predict the aggregated substation-level load by taking the historical load, weather, and calendar data as inputs. In this stage, no specific STLF model needs to be developed, which allows the forecasters to select the most accurate prediction results for transformer-level load forecasting. 2) The load distribution factor (LDF) is defined as the ratio of the transformer load to the substation load. The relationship between LDF and substation load is captured by nonlinear regression functions under different substation operating conditions, and the load of each parallel transformer is predicted using these nonlinear regression functions. Each nonlinear function can be accurately established even if the historical load data are scarce under some irregular operating conditions. Three application examples show the effectiveness and rationality of the proposed method. The third example demonstrates that STLF of transformers is necessary because it provides important information for optimizing substation operating schemes and equipment maintenance plans.

Short-Term Load Forecasting (STLF) for End-User Transformer Level (EUTL) is challenging due to the high penetration of Electric Heating Loads (EHLs), which exhibit significant uncertainty, nonlinearity, and variability. In this paper, a STLF model is proposed based on the Stacked Auto-Encoder Extreme Learning Machine (SAE-ELM) deep learning framework, which can be used to extract hidden features from the time series load data. In order to improve the capability of extracting deep and diverse features from the data and generate a useful knowledge representation structure, a novel specialized feature indices set is proposed to construct the training sample set. The sliding trend, fluctuation rate, grade of change, and smoothness of the time series are considered and quantified as elements of the training sample set. Then, deep nonlinear features are extracted by using the SAE-ELM with no iterative parameter tuning needed. To illustrate the validity of the proposed model, five numerical cases are conducted. Comparison of results shows that the proposed model improves the capability and sensitivity of dealing with load volatility and forecasting accuracy.

Nowadays electricity load forecasting is important to further minimize the cost of day-ahead energy market. Load forecasting can help utility operators for the efficient management of a demand response program. Forecasting of electricity load demand with higher accuracy and efficiency can help utility operators to design reasonable operational planning of generation units. But solving the problem of load forecasting is a challenging task since electricity load is affected by previous history load, several exogenous external factors (i.e., weather variables, social variables, working day or holiday), time of day, and season of the year. To solve the problem of short-term load forecasting (STLF) and further improve the forecasting accuracy, in this paper we have proposed a novel hybrid STLF model with a new signal decomposition and correlation analysis technique. To this end, load demand time series is decomposed into some regular low frequency components using improved empirical mode decomposition (IEMD). To compensate for the information loss during signal decomposition, we have incorporated the effect of exogenous variables by performing correlation analysis using T-Copula. From the T-Copula analysis, peak load indicative binary variable is derived from value at risk (VaR) to improve the load forecasting accuracy during peak time. The data obtained from IEMD and T-Copula is applied to deep belief network for predicting the future load demand of specific time. The proposed data driven method is validated on real time data from the Australia and the United States of America. The performance of proposed load forecasting model is evaluated in terms of mean absolute percentage error (MAPE) & root mean square error (RMSE). Simulation results verify that, the proposed model provides a significant decrease in MAPE and RMSE values compared to traditional empirical mode decomposition based electricity load forecasting.

WOS: 000499332300003 The constant increase in consumption of electricity has become one of the biggest problems today. The evaluation of energy resources has also made it worthwhile to consume it. In this respect, the transmission of electric energy and the operation of power systems have become important issues. As a result, reliable, high quality and affordable energy supply has become the most important task of operators. Realizing these elements can certainly be accomplished with good planning. One of the most important elements of this planning is undoubtedly consumption estimates. Therefore, knowing when consumers will consume energy is of great importance for operators as well as energy producers. Consumption estimates or, in other words, load estimates are also important in terms of the price balance that will occur in the market. In this study, the short-term load estimation of Duzce, Turkey is performed with Artificial Neural Networks (ANN). In the study, the April values were taken as reference and the estimates were obtained according to the input results of this month. As a result of this study, it is seen that the load consumption with nonlinear data can be successfully forecasted by ANN. Duzce UniversityDuzce University This study was carried out in cooperation with SEDAS and Duzce University within the scope of "Influence of Optimization Method on Uncertainty Costs of Profile Coefficients Methodology and Optimization Project" supported by EPDK on May 28.

The introduction of a new pricing mechanism, the increasing-block tariff (IBT), will not only affect electricity bills for residents, but also lead to a change in residential electricity consumption behaviours. Understanding these consumption patterns will help create more accurate load forecasting and increase the efficiency of the IBT. This study proposes an innovative clustering-based approach for short-term load forecasting under the IBT in China. The new approach initially partitions households into homogeneous groups each of which has distinctive consumption patterns under the IBT, each consumer segment can then select the most appropriate model for load forecasting, and the predicted load demands of different clusters are aggregated to derive the total usage. In particular, the IBT-related attributes are newly introduced into the clustering analysis. The utility and effectiveness of the proposed model is confirmed through a realistic dataset that contains the daily household-level consumption data of 533 households from April 2014 to February 2015. Consequently, the households are classified into five clusters with distinctive consumption patterns, including low-demand and insensitivity to high temperature ( C l u s t e r 1 ), ordinary users and sensitivity to high temperature ( C l u s t e r 2 ), ordinary users and sensitivity to the IBT ( C l u s t e r 3 ), high-demand consumers and sensitivity to high temperature ( C l u s t e r 4 ), and luxury consumers ( C l u s t e r 5 ). In addition, the obtained experimental results demonstrate that the proposed approach can not only achieve better prediction accuracy (e.g., the mean absolute percentage error (MAPE) improves from 3.82% to 2.28% by using autoregressive integrated moving average (ARIMA)), but also provide better flexibility for hybrid modelling. From the practical implication point of view, the proposed forecasting model can help power companies to provide a reliable and high-quality electricity supply as well as to establish appropriate schedules of operations and maintenance within a certain area. Moreover, the identified consumption behaviours can be analysed and used to improve the design and promote awareness/acceptance of the IBT.

Forecasting time series with multiple seasonal cycles such as short-term load forecasting is a challenging problem due to the complicated relationship between input and output data. In this work, we use a pattern representation of the time series to simplify this relationship. A neural network trained on patterns is an easier task to solve. Thus, its architecture does not have to be either complex and deep or equipped with mechanisms to deal with various time-series components. To improve the learning performance, we propose weighting individual errors of training samples in the loss function. The error weights correspond to the similarity between the training pattern and the test query pattern. This approach makes the learning process more sensitive to the neighborhood of the test pattern. This means that more distant patterns have less impact on the learned function around the test pattern and lead to improved forecasting accuracy. The proposed framework is useful for a wide range of complex time-series forecasting problems. Its performance is illustrated in several short-term load-forecasting empirical studies in this work. In most cases, error weighting leads to a significant improvement in accuracy.

With economic growth, the demand for power systems is increasingly large. Short-term load forecasting (STLF) becomes an indispensable factor to enhance the application of a smart grid (SG). Other than forecasting aggregated residential loads in a large scale, it is still an urgent problem to improve the accuracy of power load forecasting for individual energy users due to high volatility and uncertainty. However, as an important variable that affects the power consumption pattern, the influence of weather factors on residential load prediction is rarely studied. In this paper, we review the related research of power load forecasting and introduce a short-term residential load forecasting model based on a long short-term memory (LSTM) recurrent neural network with weather features as an input.

Nowadays, supplying demand load and maintaining sustainable energy are important issues that have created many challenges in power systems. In these types of problems, short-term load forecasting has been proposed as one of the management and energy supply modes in power systems. In this paper, after reviewing various load forecasting techniques, a deep learning method called bidirectional long short-term memory (Bi-LSTM) is presented for short-term load forecasting in a microgrid. By collecting relevant features available in the input data at the training stage, it is shown that the proposed procedure enjoys important properties, such as its great ability to process time series data. A microgrid in rural Sub-Saharan Africa, including household and commercial loads, was selected as the case study. The parameters affecting the formation of household and commercial load profiles are considered as input variables, and the total household and commercial load profiles of the microgrid are considered as the target. The Bi-LSTM network is trained by input variables to forecast the microgrid load on an hourly basis by recognizing the consumption pattern. Various performance evaluation indicators such as the correlation coefficient (R), mean squared error (MSE), and root mean squared error (RMSE) are utilized to analyze the forecast results. In addition, in a comparative approach, the performance of the proposed method is compared and evaluated with other methods used in similar studies. The results presented for the training phase show an accuracy of R = 99.81% for the Bi-LSTM network. The test and load forecasting stage are performed by the Bi-STLM network, with an accuracy of R = 99.34% and forecasting errors of MSE = 0.1042 and RMSE = 0.3243. The results confirm the high performance of the proposed Bi-LSTM technique, with a high correlation coefficient when compared to other methods used for short-term load forecasting.

Recently, multistep-ahead prediction has attracted much attention in electric load forecasting because it can deal with sudden changes in power consumption caused by various events such as fire and heat wave for a day from the present time. On the other hand, recurrent neural networks (RNNs), including long short-term memory and gated recurrent unit (GRU) networks, can reflect the previous point well to predict the current point. Due to this property, they have been widely used for multistep-ahead prediction. The GRU model is simple and easy to implement, however, its prediction performance is limited because it considers all input variables equally. In this paper, we propose a short-term load forecasting model using an attention based GRU to focus more on the crucial variables and demonstrate that this can achieve significant performance improvements, especially when the input sequence of RNN is long. Through extensive experiments, we show that the proposed model outperforms other recent multistep-ahead prediction models in the building-level power consumption forecasting.

Forecasting energy demand within a distribution network is essential for developing strategies to manage and optimize available energy resources and the associated infrastructure. In this study, we consider remote communities in the Arctic located at the end of the radial distribution network without alternative energy supply. Therefore, it is crucial to develop an accurate forecasting model to manage and optimize the limited energy resources available. We first compare the accuracy of several models that perform short-and medium-term load forecasts in rural areas, where a single industrial customer dominates the electricity consumption. We consider both statistical methods and machine learning models to predict energy demand. Then, we evaluate the transferability of each method to a geographical rural area different from the one considered for training. Our results indicate that statistical models achieve higher accuracy on longer forecast horizons relative to neural networks, while the machine-learning approaches perform better in predicting load at shorter time intervals. The machine learning models also exhibit good transferability, as they manage to predict well the load at new locations that were not accounted for during training. Our work will serve as a guide for selecting the appropriate prediction model and apply it to perform energy load forecasting in rural areas and in locations where historical consumption data may be limited or even not available.

Forecasting of daily loads is crucial for the Distribution System Operators (DSO). Contemporary short-term load forecasting models (STLF) are very well recognized and described in numerous articles. One of such models is the Adaptive Neuro-Fuzzy Inference System (ANFIS), which requires a large set of historical data. A well-recognized issue both for the ANFIS and other daily load forecasting models is the selection of exogenous variables. This article attempts to verify the statement that an appropriate selection of exogenous variables of the ANFIS model affects the accuracy of the forecasts obtained ex post. This proposal seems to be a return to the roots of the Polish econometrics school and the use of the Hellwig method to select exogenous variables of the ANFIS model. In this context, it is also worth asking whether the use of the Hellwig method in conjunction with the ANFIS model makes it possible to investigate the significance of weather variables on the profile of the daily load in an energy company. The functioning of the ANFIS model was tested for some consumers exhibiting high load randomness located within the area under supervision of the examined power company. The load curves featuring seasonal variability and weekly similarity are suitable for forecasting with the ANFIS model. The Hellwig method has been used to select exogenous variables in the ANFIS model. The optimal set of variables has been determined on the basis of integral indicators of information capacity H. Including an additional variable, i.e., air temperature, has also been taken into consideration. Some results of ex post daily load forecast are presented.

An accurate short-term load forecasting (STLF) is one of the most critical inputs for power plant units&rsquo, planning commitment. STLF reduces the overall planning uncertainty added by the intermittent production of renewable sources, thus, it helps to minimize the hydrothermal electricity production costs in a power grid. Although there is some research in the field and even several research applications, there is a continual need to improve forecasts. This research proposes a set of machine learning (ML) models to improve the accuracy of 168 h forecasts. The developed models employ features from multiple sources, such as historical load, weather, and holidays. Of the five ML models developed and tested in various load profile contexts, the Extreme Gradient Boosting Regressor (XGBoost) algorithm showed the best results, surpassing previous historical weekly predictions based on neural networks. Additionally, because XGBoost models are based on an ensemble of decision trees, it facilitated the model&rsquo, s interpretation, which provided a relevant additional result, the features&rsquo, importance in the forecasting.

Short-Term Load Forecasting (STLF) is a fundamental instrument in the efficient operational management and planning of electric utilities. Emerging smart grid technologies pose new challenges and opportunities. Although load forecasting at the aggregate level has been extensively studied, electrical load forecasting at fine-grained geographical scales of households is more challenging. Among existing approaches, semi-parametric generalized additive models (GAM) have been increasingly popular due to their accuracy, flexibility, and interpretability. Their applicability is justified when forecasting is addressed at higher levels of aggregation, since the aggregated load pattern contains relatively smooth additive components. High resolution data are highly volatile, forecasting the average load using GAM models with smooth components does not provide meaningful information about the future demand. Instead, we need to incorporate irregular and volatile effects to enhance the forecast accuracy. We focus on the analysis of such hybrid additive models applied on smart meters data and show that it leads to improvement of the forecasting performances of classical additive models at low aggregation levels. (C) 2020 International Institute of Forecasters. Published by Elsevier B.V. All rights reserved.

The most important thing to operate a power system is that the power supply should be close to the power demand. In order to predict the amount of electric power transaction (EPT), it is important to choose and decide the variable and its starting date. In this paper, variables that could be acquired one the starting day of prediction were chosen. This paper designated date, temperature and special day as variables to predict the amount of EPT of the Korea Electric Power company. This paper also used temperature data from a year ago to predict the next year. To do this, we proposed single deep learning algorithms and hybrid deep learning algorithms. The former included multi-layer perceptron (MLP), convolution neural network (CNN), long short-term memory (LSTM), gated recurrent unit (GRU), support vector machine regression (SVR), and adaptive network-based fuzzy inference system (ANFIS). The latter included LSTM + CNN and CNN + LSTM. We then confirmed the improvement of accuracy for prediction using pre-processed variables compared to original variables We also assigned two years of test data during 2017–2018 as variable data to measure high prediction accuracy. We then selected a high-accuracy algorithm after measuring root mean square error (RMSE) and mean absolute percent error (MAPE). Finally, we predicted the amount of EPT in 2018 and then measured the error for each proposed algorithm. With these acquired error data, we obtained a model for predicting the amount of EPT with a high accuracy.

The load forecasting of a coal mining enterprise is a complicated problem due to the irregular technological process of mining. It is necessary to apply models that can distinguish both cyclic components and complex rules in the energy consumption data that reflect the highly volatile technological process. For such tasks, Artificial Neural Networks demonstrate advanced performance. In recent years, the effectiveness of Artificial Neural Networks has been significantly improved thanks to new state-of-the-art architectures, training methods and approaches to reduce overfitting. In this paper, the Recurrent Neural Network architecture with a small-size model was applied to the short-term load forecasting of a coal mining enterprise. A single recurrent model was developed and trained for the entire four-year operational period of the enterprise, with significant changes in the energy consumption pattern during the period. This task was challenging since it required high-level generalization performance from the model. It was shown that the accuracy and generalization properties of small-size recurrent models can be significantly improved by the proper selection of the hyper-parameters and training method. The effectiveness of the proposed approach was validated using a real-case dataset. © 2020 by the authors. Licensee MDPI, Basel, Switzerland.

Nowadays, electricity demand forecasting is critical for electric utility companies. Accurate residential load forecasting plays an essential role as an individual component for integrated areas such as neighborhood load consumption. Short-term load forecasting can help electric utility companies reduce waste because electric power is expensive to store. This paper proposes a novel method to evolve deep neural networks for time series forecasting applied to residential load forecasting. The approach centres its efforts on the neural network architecture during the evolution. Then, the model weights are adjusted using an evolutionary optimization technique to tune the model performance automatically. Experimental results on a large dataset containing hourly load consumption of a residence in London, Ontario shows that the performance of unadjusted weights architecture is comparable to other state-of-the-art approaches. Furthermore, when the architecture weights are adjusted the model accuracy surpassed the state-of-the-art method called LSTM one shot by 3.0%.

To minimise environmental impact, to avoid regulatory penalties, and to improve competitiveness, energy-intensive manufacturing firms require accurate forecasts of their energy consumption so that precautionary and mitigation measures can be taken. Deep learning is widely touted as a superior analytical technique to traditional artificial neural networks, machine learning, and other classical time-series models due to its high dimensionality and problem-solving capabilities. Despite this, research on its application in demand-side energy forecasting is limited. We compare two benchmarks (Autoregressive Integrated Moving Average (ARIMA) and an existing manual technique used at the case site) against three deep-learning models (simple Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU)) and two machine-learning models (Support Vector Regression (SVR) and Random Forest) for short-term load forecasting (STLF) using data from a Brazilian thermoplastic resin manufacturing plant. We use the grid search method to identify the best configurations for each model and then use Diebold&ndash, Mariano testing to confirm the results. The results suggests that the legacy approach used at the case site is the worst performing and that the GRU model outperformed all other models tested.

Load forecasting provides essential information for engineers and operators of an electric system. Using the forecast information, an electric utility company’s engineers make informed decisions in critical scenarios. The deregulation of energy industries makes load forecasting even more critical. In this article, the work we present, called Nearest Neighbors Load Forecasting (NNLF), was applied to very short-term load forecasting of electricity consumption at the national level in Mexico. The Energy Control National Center (CENACE—Spanish acronym) manages the National Interconnected System, working in a Real-Time Market system. The forecasting methodology we propose provides the information needed to solve the problem known as Economic Dispatch with Security Constraints for Multiple Intervals (MISCED). NNLF produces forecasts with a 15-min horizon to support decisions in the following four electric dispatch intervals. The hyperparameters used by Nearest Neighbors are tuned using Differential Evolution (DE), and the forecaster model inputs are determined using phase-space reconstruction. The developed models also use exogenous variables; we append a timestamp to each input (i.e., delay vector). The article presents a comparison between NNLF and other Machine Learning techniques: Artificial Neural Networks and Support Vector Regressors. NNLF outperformed those other techniques and the forecasting system they currently use.

Short-term load forecasting (STLF) for industrial customers has been an essential task to reduce the cost of energy transaction and promote the stable operation of smart grid throughout the development of the modern power system. Traditional STLF methods commonly focus on establishing the non-linear relationship between loads and features, but ignore the temporal relationship between them. In this paper, an STLF method based on ensemble hidden Markov model (e-HMM) is proposed to track and learn the dynamic characteristics of industrial customer's consumption patterns in correlated multivariate time series, thereby improving the prediction accuracy. Specifically, a novel similarity measurement strategy of log-likelihood space is designed to calculate the log-likelihood value of the multivariate time series in sliding time windows, which can effectively help the hidden Markov model (HMM) to capture the dynamic temporal characteristics from multiple historical sequences in similar patterns, so that the prediction accuracy is greatly improved. In order to improve the generalization ability and stability of a single HMM, we further adopt the framework of Bagging ensemble learning algorithm to reduce the prediction errors of a single model. The experimental study is implemented on a real dataset from a company in Hunan Province, China. We test the model in different forecasting periods. The results of multiple experiments and comparison with several state-of-the-art models show that the proposed approach has higher prediction accuracy.

Short-term load forecasting (STLF) plays an important role in the economic dispatch of power systems. Obtaining accurate short-term load can greatly improve the safety and economy of a power grid operation. In recent years, a large number of short-term load forecasting methods have been proposed. However, how to select the optimal feature set and accurately predict multi-step ahead short-term load still faces huge challenges. In this paper, a hybrid feature selection method is proposed, an Improved Long Short-Term Memory network (ILSTM) is applied to predict multi-step ahead load. This method firstly takes the influence of temperature, humidity, dew point, and date type on the load into consideration. Furthermore, the maximum information coefficient is used for the preliminary screening of historical load, and Max-Relevance and Min-Redundancy (mRMR) is employed for further feature selection. Finally, the selected feature set is considered as input of the model to perform multi-step ahead short-term load prediction by the Improved Long Short-Term Memory network. In order to verify the performance of the proposed model, two categories of contrast methods are applied: (1) comparing the model with hybrid feature selection and the model which does not adopt hybrid feature selection; (2) comparing different models including Long Short-Term Memory network (LSTM), Gated Recurrent Unit (GRU), and Support Vector Regression (SVR) using hybrid feature selection. The result of the experiments, which were developed during four periods in the Hubei Province, China, show that hybrid feature selection can improve the prediction accuracy of the model, and the proposed model can accurately predict the multi-step ahead load.

Load forecasting impacts directly financial returns and information in electrical systems planning. A promising approach to load forecasting is the Echo State Network (ESN), a recurrent neural network for the processing of temporal dependencies. The low computational cost and powerful performance of ESN make it widely used in a range of applications including forecasting tasks and nonlinear modeling. This paper presents a Bayesian optimization algorithm (BOA) of ESN hyperparameters in load forecasting with its main contributions including helping the selection of optimization algorithms for tuning ESN to solve real-world forecasting problems, as well as the evaluation of the performance of Bayesian optimization with different acquisition function settings. For this purpose, the ESN hyperparameters were set as variables to be optimized. Then, the adopted BOA employs a probabilist model using Gaussian process to find the best set of ESN hyperparameters using three different options of acquisition function and a surrogate utility function. Finally, the optimized hyperparameters are used by the ESN for predictions. Two datasets have been used to test the effectiveness of the proposed forecasting ESN model using BOA approaches, one from Poland and another from Brazil. The results of optimization statistics, convergence curves, execution time profile, and the hyperparameters’ best solution frequencies indicate that each problem requires a different setting for the BOA. Simulation results are promising in terms of short-term load forecasting quality and low error predictions may be achieved, given the correct options settings are used. Furthermore, since there is not an optimal global optimization solution known for real-world problems, correlations among certain values of hyperparameters are useful to guide the selection of such a solution.

Short-term load forecasting is integral to the energy planning sector. Various techniques have been employed to achieve effective operation of power systems and efficient market management. We present a scalable system for day-ahead household electrical energy consumption forecasting, named HousEEC. The proposed forecasting method is based on a deep residual neural network, and integrates multiple sources of information by extracting features from (i) contextual data (weather, calendar), and (ii) the historical load of the particular household and all households present in the dataset. Additionally, we compute novel domain-specific time-series features that allow the system to better model the pattern of energy consumption of the household. The experimental analysis and evaluation were performed on one of the most extensive datasets for household electrical energy consumption, Pecan Street, containing almost four years of data. Multiple test cases show that the proposed model provides accurate load forecasting results, achieving a root-mean-square error score of 0.44 kWh and mean absolute error score of 0.23 kWh, for short-term load forecasting for 300 households. The analysis showed that, for hourly forecasting, our model had 8% error (22 kWh), which is 4 percentage points better than the benchmark model. The daily analysis showed that our model had 2% error (131 kWh), which is significantly less compared to the benchmark model, with 6% error (360 kWh).

Short-term load forecasting ensures the efficient operation of power systems besides affording continuous power supply for energy consumers. Smart meters that are capable of providing detailed information on buildings energy consumption, open several doors of opportunity to short-term load forecasting at the individual building level. In the current paper, four machine learning methods have been employed to forecast the daily peak and hourly energy consumption of domestic buildings. The utilized models depend merely on buildings historical energy consumption and are evaluated on the profiles that were not previously trained on. It is evident that developing data-driven models lacking external information such as weather and building data are of great importance under the situations that the access to such information is limited or the computational procedures are costly. Moreover, the performance evaluation of the models on separated house profiles determines their generalization ability for unseen consumption profiles. The conducted experiments on the smart meter data of several UK houses demonstrated that if the models are fed with sufficient historical data, they can be generalized to a satisfactory level and produce quite accurate results even if they only use past consumption values as the predictor variables. Furthermore, among the four applied models, the ones based on deep learning and ensemble techniques, display better performance in predicting daily peak load consumption than those of others.