Electrical Load Pattern Research Articles

Short-Term Load Forecasting for Regional Smart Energy Systems Based on Two-Stage Feature Extraction and Hybrid Inverted Transformer

Accurate short-term load forecasting is critical for enhancing the reliability and stability of regional smart energy systems. However, the inherent challenges posed by the substantial fluctuations and volatility in electricity load patterns necessitate the development of advanced forecasting techniques. In this study, a novel short-term load forecasting approach based on a two-stage feature extraction process and a hybrid inverted Transformer model is proposed. Initially, the Prophet method is employed to extract essential features such as trends, seasonality and holiday patterns from the original load dataset. Subsequently, variational mode decomposition (VMD) optimized by the IVY algorithm is utilized to extract significant periodic features from the residual component obtained by Prophet. The extracted features from both stages are then integrated to construct a comprehensive data matrix. This matrix is then inputted into a hybrid deep learning model that combines an inverted Transformer (iTransformer), temporal convolutional networks (TCNs) and a multilayer perceptron (MLP) for accurate short-term load forecasting. A thorough evaluation of the proposed method is conducted through four sets of comparative experiments using data collected from the Elia grid in Belgium. Experimental results illustrate the superior performance of the proposed approach, demonstrating high forecasting accuracy and robustness, highlighting its potential in ensuring the stable operation of regional smart energy systems.

Measurement of Dynamic Electric Consumption Trend in a Broiler House in Arkansas

Highlights We quantified the diurnal and seasonal electric load pattern during four consecutive growouts of heavy broilers over an 11-month period on a commercial farm. Consumption was measured separately for ventilation fans, lights, feed motors, well pump motors and radiant tube heater blowers. Ventilation fans were the largest consumer of energy and the largest contributor to peak demand on the farm. Average monthly electric demand ranged from 31% to 77% of the corresponding monthly peak demand, calculated from the largest 15-min period. The electricity cost of raising heavier-weight broiler chickens is higher in the summer ($22 per 1,000 kg live weight) than in the winter ($5 per 1,000 kg live weight). Abstract. Characterizing electric energy use in a commercial broiler house—partitioned by bird age, season, and electric load type—can provide engineers with insight for designing energy conservation practices, as well as help to analyze and design on-site renewable energy production sub-systems. This study monitored electric energy consumption in a commercial broiler house in Arkansas, raising heavy broilers (with a capacity of 17,000 birds) during an 11-month period that included four, 8-week flocks. Average electricity use (RMS voltage and current) was measured at a sampling interval of 5 min for all of the major load types, including ventilation, lighting, feed motors, and well-water pump motor. Results indicated that electric consumption for ventilation fans was the greatest load, accounting for over 88% of the total. The endwall fans (for tunnel ventilation) contributed to the peak electric demand (VA) in nine of the 11 months. Over the 11-month monitoring period, broiler chicken production had an average electric energy consumption (per unit live-weight mass produced) of 43 kWh per metric ton of live weight (in a winter flock) to 184 kWh per metric ton (in a summer flock). The corresponding electric costs were $5 per metric ton and $22 per metric ton in winter and summer flocks, respectively (based upon an electricity unit cost of $0.12/kWh). Keywords: Electricity demand, Energy efficiency, Poultry housing.

Structured stochastic models based on multi-source heterogeneous data for predicting internal electricity load of non-residential buildings

Electricity load patterns in buildings are increasingly diverse and inherently present building-by-building and temporal variations. This study develops structured stochastic models to capture individual and temporal variations in non-residential building groups (e.g., education, office, culture, and retail buildings). Hourly electricity load data of non-residential buildings, categorized by building use, are sourced from two open datasets (Ministry of Land, Infrastructure and Transport and Korea Electric Power Corporation). First, heterogeneous public datasets are processed and utilized to provide both detailed building information and high-resolution electricity load data. Second, a structured set of four stochastic models are formulated using linear regression and multivariate normal distribution models. Model 1 utilizes the beta distribution to derive daily average electricity load per unit area for each building use. Model 2 utilizes the burr distribution to determine the daily variation of electricity load from the daily average. Models 3 and 4 utilize a multivariate normal distribution to represent the hourly average electricity load profile and its time variation, respectively. The coefficient of variation of the root mean square error (CVRMSE) for the structured probabilistic models ranges from 5.2 % to 14.8 % regarding five key load characterization variables. CVRMSEs for near-peak load, near-base load, rise time, high-load duration, and fall time are 8.5–13.9, 9.5–14.7, 7.3–14.8, 6.4–13.9, and 5.2–11.4 %, respectively. The validation results demonstrate that the proposed probabilistic models capture actual variations in measurements. Structured statistical models improve prediction performance by distinctly reflecting individual and temporal variations in electricity loads.

High-resolution electric power load data of an industrial park with multiple types of buildings in China

Considering the growing demand for electricity in industrial parks, understanding their electric power load patterns is critical for improving energy efficiency and ensuring the rational utilization of energy resources. However, the detailed electric power load data of various buildings in industrial parks are rarely available and accessible, which hinders the related studies. In this context, we present the electric power load data of 6 years (from January 1, 2016 to December 31, 2021) for various types of buildings in an industrial park in Suzhou, China. The data are obtained from smart meters and have various time resolutions (i.e., 5 minutes, 30 minutes, and 1 hour). This work describes the data collection, processing process, and different imputation methods. The high-resolution electric power load data can be used for various research tasks, including load prediction, load pattern recognition, anomaly detection, and demand response strategy development.

Enhanced short-term load forecasting with hybrid machine learning models: CatBoost and XGBoost approaches

The focus of this paper is to improve short-term load forecasting for electric power. To achieve this goal, the study explores and evaluates hybrid models, specifically using the CatBoost and XGBoost algorithms, which are optimized with different optimizers. The study incorporates hourly electricity load data and also includes temperature data to enhance the precision of the forecasting models. Statistical metrics are then used to assess the performance of these models. The study evaluates the performance of the hybrid models on both training and testing datasets. It finds that the CatBoost-Arithmetic Optimization Algorithm hybrid model outperforms the other models in the training dataset. However, in the testing dataset, the XGBoost- Arithmetic Optimization Algorithm hybrid model demonstrates superior performance compared to the CatBoost models. The study conducts an importance and sensitivity analysis to understand which variables have the most significant impact on the target variable, which is likely electricity load. The results of this analysis reveal that temperature is the most influential variable affecting the target variable. Additionally, the month variable is identified as having a notable impact on load forecasting. These findings suggest that employing hybrid models, particularly those optimized with appropriate algorithms, can significantly improve the accuracy of short-term load forecasting. Moreover, the study highlights the importance of incorporating temperature data into these models, as temperature is a key driver of electricity load patterns.

Data driven machine learning models for short‐term load forecasting considering electrical vehicle load

AbstractElectric vehicles (EVs) are gaining popularity due to their fuel efficiency and ability to reduce greenhouse gas emissions. Significant penetration of EVs with unregulated charging, which can have a substantial impact on power networks. Accurate load predictions, including the charging of EVs are crucial for ensuring the cost‐effective and dependable operation of power systems. In order to estimate short‐term load in the presence of EV load, the XG (extreme gradient) boost algorithm is proposed and the effectiveness of the performances are checked against other models. A variety of distinct meteorological parameters and the electrical load pattern for the years 2017 and 2018 for Northeast India are used to train the machine learning classifier models. R‐squared value analyses are also performed to identify the most correlated input parameters that influence the results of various models. Analysis shows that temperature, cloud cover, heat index, dew point, wind chill, and perceived temperature are substantially connected with electricity consumption. The performance of XG boost is outperformed by comparing prediction results with decision tree, random forest, K nearest neighbors, and logistic regression. Three separate case studies were employed for verification using the precision, F1 score, sensitivity, specificity, and accuracy metrics. Our findings demonstrate that, in comparison to other forecasting models, XG Boost exhibits higher accuracy (83.84%, 81.51%, and 85.97%) and robust forecasts.

Identification of natural disaster impacted electricity load profiles with k means clustering algorithm

Natural disasters threat the resilience of the electricity system. However, little literature has investigated the electricity system’s recovering process and progress after natural disasters’ hit which strongly influence the system operators’ planning and quality of the security of supply for the electricity customers. To fill the research gap, this paper applies an unsupervised machine learning method, the k means clustering algorithm, to investigate the normal/abnormal electricity load profiles, identify natural disaster- and electrical fault-impacted electricity load profiles with a case study of the Lombok electricity system, Indonesia, and ½-hourly electricity load data from 2015 until 2021. The results show that electricity consumption in Lombok has increased over the years, which match the installed production capacity of Lombok. The results prove that the disturbance-induced electricity load patterns and especially natural disaster-impacted load profiles can be identified by the k means clustering algorithm. Especially, the pre-, during, and post-natural disaster impacted load patterns can be portrayed. Furthermore, the investigation results regarding the impacts of natural disasters and electrical faults on the performance of the electricity system, show that the natural disaster-induced load reductions and electrical fault-induced load reductions differ from the short and long-term perspectives. Moreover, the results can facilitate the electricity system operators to better understand the load patterns, predict ND strikes’ impact on the electricity system and conduct better long-term energy management strategies.

Industrial park electric power load pattern recognition: An ensemble clustering-based framework

Electric power load pattern recognition from various accumulated load data is performed for energy efficiency improvement, power system operation support, and demand side management. Industrial park is a typical energy consumption scenario, in which there are many different types of energy users. In this study, a framework based on ensemble clustering is proposed for electric power load pattern recognition in industrial parks. The proposed framework can not only accelerate the ensemble clustering process but also improve the clustering performance. Specifically, we focus on the electric power load pattern recognition of business office industrial parks with air conditioners as the main energy-consuming device. The hourly air conditioning load data and total electric power load data of eight buildings in an industrial park in Suzhou, China, from January 1, 2019, to December 31, 2019, were used in the experiment. The results showed the effectiveness of the proposed framework in recognizing the electric power load patterns in industrial parks. The identified electric power load patterns could be potentially used to support more efficient and targeted industrial park energy management decision-making.

A novel short receptive field based dilated causal convolutional network integrated with Bidirectional LSTM for short-term load forecasting

The Short-Term Load Forecasting (STLF) is a pre-eminent task for reliable power generation and electrical load dispatching in the power system. Numerous machine-learning and deep-learning forecasting algorithms have been presented in literature for performing an accurate electrical load forecast. However, the complicated structure of machine-learning and deep-learning multi-layer and with increased filter size architectures provoke the overfitting issue, which degrades the performance of STLF engines in the presence of highly diversified weather and temporal variations. This paper proposes a novel two-stage Encoder-Decoder (ED) network with improved generalization capability and forecasting accuracy. The proposed architecture is based on Short Receptive field based Dilated Causal Convolutional (SRDCC) network in the first stage and Bi-directional Long Short-Term Memory (BiLSTM) network in the second stage. Using real valued data, the proposed ED architecture is quantitatively and qualitatively analyzed in comparison with state-of-the-art machine-learning and hybrid deep-learning STLF models. The evaluation matrix used for the comparison consists of six evaluation parameters. The extensive experimentation for multi-step ahead STLF validates the efficiency of the proposed technique in terms of accuracy in comparison with other employed models. The CNN-LSTM revealed to have best performance among all other implemented parametric and non-parametric forecasting models; however, the proposed ED architecture proves to be 35% more accurate compared to CNN-LSTM and have the tendency to capture the local trends in an electrical load pattern more accurately. Moreover, a detailed comparative analysis on the computational complexity of the proposed ED architecture is also conducted to show the real implementation prospect.

Interpretable Short-Term Electrical Load Forecasting Scheme Using Cubist

Daily peak load forecasting (DPLF) and total daily load forecasting (TDLF) are essential for optimal power system operation from one day to one week later. This study develops a Cubist-based incremental learning model to perform accurate and interpretable DPLF and TDLF. To this end, we employ time-series cross-validation to effectively reflect recent electrical load trends and patterns when constructing the model. We also analyze variable importance to identify the most crucial factors in the Cubist model. In the experiments, we used two publicly available building datasets and three educational building cluster datasets. The results showed that the proposed model yielded averages of 7.77 and 10.06 in mean absolute percentage error and coefficient of variation of the root mean square error, respectively. We also confirmed that temperature and holiday information are significant external factors, and electrical loads one day and one week ago are significant internal factors.

A Novel Temporal Feature Selection Based LSTM Model for Electrical Short-Term Load Forecasting

An accurate electrical Short-term Load Forecasting (STLF) is an eminent factor in the power generation, electrical load dispatching and energy planning for the power supply companies, specifically in developing countries. This paper proposes a novel temporal feature selection-based Long Short-term Memory (LSTM) model developed by the combination of standard Artificial Neural Network (ANN) layer and LSTM for electrical short term load forecasting. The LSTM model has excellent capability of predicting the stochastic nature of an hour ahead electrical loads. The standard ANN layer consisting 11 neurons is used as an input to LSTM cells. Such a combination of ANN layer with LSTM was never proposed before. The proposed model accommodates variations in weather as well as temporal inputs like humidity, holidays, and date-time features in the hourly load data of the power supply company situated in Johor, Malaysia. This paper gives the insights of hyper parameter tuning to capture the more generalized electrical load patterns in the dataset without compromising the time complexity of the proposed model. The proposed approach was compared with five existing approaches, namely: ANN, LSTM model 1, LSTM model 2, LSTM model 3 and Convolutional Neural Network-LSTM (CNN-LSTM) using hourly load dataset of Johor. The experimental results demonstrate that the proposed approach outperformed the existing approaches in terms of root mean square error, mean absolute percentage error and Diebold-Mariano statistical inference test within 95% confidence interval.

Demand charges and user flexibility – Exploring differences in electricity consumer types and load patterns within the Swedish commercial sector

Demand-based charges have been employed as a tool intended to reduce electricity users’ maximum demand but there is a lack of consensus regarding their efficacy. One reason for this may be the diversity in the flexibility potential of different types of users. This study explores the flexibility potential of different types of electricity consumers in the small to medium-sized commercial sector (35-63A) in response to a compulsory demand charge. The objective is to characterize varying levels of flexibility with respect to different types of commercial users with different load patterns. A multivariate clustering technique was used to group commercial users with comparable load patterns based on a year of hourly data before the tariff change was introduced. This method was used to: (1) match users from the intervention area and reference area with similar load patterns, without losing any user data, and (2) compare how users with different load patterns react differently to the tariff change. We found clear distinctions in the types of commercial users in each cluster and their response to the tariff, demonstrating the extent to which demand flexibility may be dependent on the nature of an organization’s activities and its respective load patterns. The highest demand flexibility was found in clusters which had a large share of users in the IT sector, commerce and public administration. The lowest demand flexibility was found in the real estate and education sectors. Future research should further investigate these variations and explore the possibilities of tailoring interventions to the specific types of users.

A study on the transferability of computational models of building electricity load patterns across climatic zones

Significant reduction in energy demand from non-domestic buildings is required if greenhouse emission reduction targets are to be met worldwide. Increasing monitoring of electricity consumption generates a real opportunity for gaining an in-depth understanding of the nature of occupant-related internal loads and the connection between activity and demand. The stochastic nature of the demand is well-known but as yet there is no accepted methodology for generating stochastic loads for building energy simulation. This paper presents evidence that it is feasible to generate stochastic models of activity-related electricity demand based on monitored data. Two machine learning approaches are used to develop stochastic models of plug loads; an autoencoder (AE) and a Functional Data Analysis (FDA) model. Using data from two office buildings located in different countries, the transferability of models is explored by training the models on data from one building and using the trained models to predict demand for the other building. The results show that both models predict plug loads satisfactorily, with a good agreement with the mean demand and quantification of the uncertainty.

A data mining-based framework for the identification of daily electricity usage patterns and anomaly detection in building electricity consumption data

With the development of advanced information techniques, smart energy meters have made a considerable amount of real-time electricity consumption data available. These data provide a promising way to understand energy usage patterns and improve building energy management. However, previous studies have paid more attention to methodologies for the identification of energy usage patterns and are limited in the interpretability and applications of the patterns. In this context, this paper proposes a general data mining-based framework that can extract typical electricity load patterns (TELPs) and discover insightful information hidden in the patterns. The framework integrates multiple data mining techniques and mainly consists of three phases: data preparation, identification of TELPs and knowledge discovery in the patterns. A new clustering method with a two-step clustering analysis is proposed to identify the TELPs at the individual building level. Before clustering, five statistical features that represent the shapes of electricity load profiles are first defined to reduce the dimensions of daily electricity load profiles. The first clustering step aims at detecting outliers of daily electricity load profiles (DELPs) by using the density-based spatial clustering application with noise (DBSCAN) algorithm clustering technique, which addresses the data quality issues for electricity consumption data derived from energy consumption monitoring platforms (ECMPs). The second clustering step aims at grouping similar DELPs by means of the k-means algorithm to extract TELPs. The effectiveness of the proposed clustering method is demonstrated by a comparison with two single-step clustering techniques. Furthermore, a classification and regression tree (CART) algorithm is employed to discover insightful knowledge on TELPs and improve the interpretability of clustering results, namely, to explain the relations between dynamic influencing factors related to electricity consumption and TELPs. The proposed framework is applied to analyze the time-series electricity consumption data of three practical office buildings in Chongqing, and its effectiveness has been confirmed. A potential application of discovered knowledge is presented: early fault detection of anomalous electricity load profiles. The proposed framework can provide building managers with an efficient way to understand the characteristics of building electricity usage patterns and detect anomalies therein.

Heterogeneity of Electricity Consumption Patterns in Vulnerable Households

A key aspect of the design of specific tariff structures is to identify and characterize homogeneous electricity consumption profiles. Recent research in residential electricity demand has explored load profile segmentation via cluster analysis combined with descriptive data from the dwelling and occupants, which has partly explained electricity load patterns and their underlying drivers but has failed to investigate any consumption heterogeneity among similar households. Thus, the aim of this paper is to reverse this approach and investigate the extent that households with similar characteristics have different electricity consumption patterns. This study combines population-based register data with hourly electricity consumption data for a sample of 67 Danish households. First, a homogenous household group is selected based on several indicators that signal vulnerability. The specific group under investigation is single-person, older, low-income households in detached housing. Second, K-means clustering is used to identify similarities and differences in consumption patterns. The results indicate four distinct vulnerable household profiles characterized by different start and end times of peak and off-peak times, peak intensities, and overall consumption, which vary across seasons. These profiles are discussed concerning the performance of everyday practices and the design of demand-side management strategies targeted at vulnerable households.

Optimal participation of demand response aggregators in reconfigurable distribution system considering photovoltaic and storage units

Feeder reconfiguration is an important operational process in power distribution grids, which is implemented to enhance the system’s performance by managing the switches. Given variations of the electricity price and load pattern in smart distribution networks, the operational problems of the distribution system are largely time-dependent and very complex. To deal with these temporal dependencies, it is important to extend the problem across different time intervals. Toward this end, a multi-objective optimization model is presented in this study for dynamic feeder reconfiguration problem in the distribution system over multiple time intervals considering distributed generators, energy storage systems, and solar photovoltaic units. The demand response program including interruptible/curtailable service is proposed to enable energy consumers to rethink their energy consumption patterns based on incentive and punitive policies. The common objectives considered in the feeder reconfiguration problem are power loss and voltage deviation which are important objectives for traditional distribution systems, but less attention has been paid to distribution network voltage security. This study considers operational cost, energy loss and voltage stability index as objective functions that can meet operational and voltage security expectations. Dynamic feeder reconfiguration problem is a complex integer nonlinear program problem and hence it is difficult to solve which requires the use of appropriate optimization algorithms to converge to global optima or find close to global optima. In this paper, a modified honey bee matting optimization algorithm based on the new mating mechanism is presented to solve the multi-objective dynamic feeder reconfiguration problem. The proposed approach uses an eliminating zone concept to finds a set of non-dominated solutions during the search process. Furthermore, a fuzzy decision-maker is adopted to select the best compromise solution among the non-dominated solutions. The suggested approach is tested on the IEEE 33-node and 136-node test systems and its superiorities are shown through comparison with other evolutionary algorithms.

국내 전력 부하패턴 불균형 완화를 위한 수소 기반 에너지 저장 시스템 용량 산정 연구

국내 전력 부하패턴 불균형 완화를 위한 수소 기반 에너지 저장 시스템 용량 산정 연구

An empirical analysis of domestic electricity load profiles: Who consumes how much and when?

With the increased share of renewables in power generation, end users play a key role in keeping the demand at levels that better match variable supply, maintaining lower overall system costs, and reducing carbon dioxide emissions. To increase the potential for demand-side flexibility, a deeper understanding of domestic electricity load profiles is needed. Informed by customer grouping based on similar consumption patterns and drivers, targeted interventions can be better designed to time-shift peak loads and reduce overall demand. Thus, the objectives of this study are (i) to segment households in relation to their electricity load patterns using K-means clustering and (ii) to investigate household characteristics that have an influence on electricity load patterns by employing multinomial probit regression. This study uses hourly electricity consumption for 2017, combined with population-based register data for a large sample of Danish households. The results indicate that four distinct Danish household groups are characterized by different timing and magnitudes of electricity consumption, which are influenced by specific sociodemographics and dwelling characteristics. Similarities between the groups emerge with respect to the evening peak consumption, seasonal variation in electricity demand, and weekend morning demand ramp-up. Challenges and opportunities for domestic load profiling in the power industry and policymaking are discussed.

PERAMALAN BEBAN LISTRIK JANGKA PENDEK MENGGUNAKAN METODE ADAPTIVE NEURO FUZZY INFERENCE SYSTEM (ANFIS) DI GARDU INDUK NUSA DUA BALI

Electric load in Nusa Dua Bali has increased from 2013-2017 by an average of 11.83%. The increase in electric load requires the electrical energy service provider to be able to adjust the electricity demand and be able to increase its reliability, The effort that can be done is to predict the electric load. Electric load forecasting can be done by various methods, ANFISo (Adaptiveo Neuroo Fuzzyo Inferenceo Systemo) is one method that is often used in forecasting electrical loads. ANFIS is able to explain the reasoning process and do data learning. The data used are the electric load, temperature, humidity and time, the data was chosen because changes in temperature and humidity affect people's habitual patterns in using air conditioners (electric load patterns). The electric load pattern is trained 100 times using ANFIS with the type of membership function is trimf, and [3 3 3 3] is the number of membership function. The indicator to determining the accuracy of the electrical load forecasting pattern results with the real electric load pattern used the MAPE (Mean Absolute Percentage Error) value, which the MAPE standard value that good is less than 10%. The test results from this study produced a MAPE value of 6.98%.

Solving the Cold-Start Problem in Short-Term Load Forecasting Using Tree-Based Methods

An energy-management system requires accurate prediction of the electric load for optimal energy management. However, if the amount of electric load data is insufficient, it is challenging to perform an accurate prediction. To address this issue, we propose a novel electric load forecasting scheme using the electric load data of diverse buildings. We first divide the electric energy consumption data into training and test sets. Then, we construct multivariate random forest (MRF)-based forecasting models according to each building except the target building in the training set and a random forest (RF)-based forecasting model using the limited electric load data of the target building in the test set. In the test set, we compare the electric load of the target building with that of other buildings to select the MRF model that is the most similar to the target building. Then, we predict the electric load of the target building using its input variables via the selected MRF model. We combine the MRF and RF models by considering the different electric load patterns on weekdays and holidays. Experimental results demonstrate that combining the two models can achieve satisfactory prediction performance even if the electric data of only one day are available for the target building.