Historical Load Research Articles

Research on Artificial Intelligence Methods for Feature Extraction and Prediction of Current Fluctuations

Accurate prediction of current fluctuations in power systems is crucial for ensuring the safety and stability of grid operations. Traditional prediction methods face significant limitations when dealing with data complexity and dynamic changes. The introduction of artificial intelligence technologies offers a new perspective for current fluctuation predictions. Based on Long Short-Term Memory networks (LSTM), this study proposes an intelligent prediction method combining deep fusion and feature extraction of multi-source data. This method processes multi-dimensional information such as historical current data, weather conditions, and load demands to achieve high-accuracy and real-time predictions of current fluctuations. The results show that this method significantly improves prediction accuracy, enhances model adaptability, and provides reliable support for power dispatch and resource optimization, offering important reference value for the development of smart grids.

Association Between Dynamic Viral Rebound and Longitudinal Measures of Viral Load/CD4 Counts Among People with HIV in South Carolina.

Monitoring HIV viral rebound (VR) is crucial, as it indicates an increased risk of infection, transmission, disease progression, and drug resistance. This study aims to identify the association between dynamic VR and historical viral load (VL)/CD4 count measures. A 15-year South Carolina population-based electronic health record data were used for the study. VR was defined as the return of detectable levels of VL (>200 copies/mL) after stable viral suppression (VS) (two consecutive VS, i.e., VL ≤200 copies/mL). A generalized linear mixed model was used to evaluate the association between dynamic VR and historical time-dependent predictors, such as nadir CD4 count and comorbidities, within a year prior to each VR. Subgroup analysis for men who have sex with men (MSM) was also conducted. Among 8,185 people with HIV (PWH), 1,173 (14.3%) had a history of VR. Lower nadir CD4 count (≥500 vs. <200 cells/μL; adjusted odds ratio [aOR]: 0.51, 95% confidence interval [CI]: [0.43, 0.60]), younger age (>60 years old vs. 18-30 years old; aOR: 0.43, 95% CI: [0.29, 0.63]), and being Black (Black vs. White; aOR: 1.58, 95% CI: [1.34, 1.85]) were associated with a higher risk of VR, while MSM (MSM vs. heterosexual; aOR: 0.81, 95% CI: [0.67, 0.96]) were associated with decreased VR risk. The rate of VR among PWH in South Carolina is significant. Within-1-year VL/CD4 test is critical for identifying PWH at risk for VR. Tailored interventions are needed for PWH at risk for VR to achieve sustained suppression and better health outcomes.

Testing and implementation of LEU+ 6% enriched UO2 fuel at The PULSTAR reactor

The PULSTAR reactor is a light water moderated and cooled, UO2 fueled reactor featuring both 4% and 6% enriched fuel assemblies. The PULSTAR was operated for over 40 years with 4% enriched assemblies. Originally designed to demonstrate key reactor concepts in the 1970s, the PULSTAR now serves as a hub of research opportunities. Recently, the PULSTAR has successfully demonstrated methods for licensing and operations with higher enriched fuel in a typical light water reactor configuration. Using MCNP, a model was developed and benchmarked against historical core loadings of the PULSTAR. Key reactivity contributions as well as kinetic effects were simulated. From this model, predictive calculations were completed to demonstrate the compliance of operations, with 6% enriched assemblies, to all required technical specifications. Based on the results, the PULSTAR reactor license was successfully amended to include both 4% and LEU+ 6% enriched assemblies. Results from three core loadings from 2017 to the present demonstrate the adherence of the PULSTAR to its technical specifications and the related US NRC regulations, while operating with LEU+ 6% enriched fuel. Operational experience with these new mixed enrichment cores continues to show excellent agreement between the predicted and observed reactor behavior. The precedent set by the simulations and operations with the PULSTAR’s mixed enrichment cores can be beneficial to the planned introduction of higher enriched (LEU+) fuel in the light water commercial reactor (LWR) fleet in the US and worldwide.

Load forecasting method based on CNN and extended LSTM

With the continuous development of power systems and the growing demand, accurate power load prediction has become essential to ensure stable system operation and optimize energy allocation. The collection and analysis of historical load data have become increasingly important, and the factors affecting load changes have grown more complex. To address this challenge, this paper proposes a short-term load prediction method based on convolutional neural networks (CNN) and multilayer extended long short-term memory (LSTM) neural networks. The method first extracts local features from historical data using 1D convolution operations in the CNN. Then, a multilayer extended LSTM network captures long-term dependencies in the data from multiple dimensions for more accurate load prediction. The prediction experiments use real load and related factor data from a city in eastern Mongolia. Results demonstrate that the proposed method outperforms other deep learning-based load prediction methods in direct single-step, four-step, and 32-step load predictions in terms of MAPE, RMSE, and MAE indexes. Additionally, the proposed model can be extended to more prediction application scenarios, including intelligent transportation and financial analysis.

TimeGPT in load forecasting: A large time series model perspective

Machine learning models have made significant progress in load forecasting, but their forecast accuracy is limited in cases where historical load data is scarce. Inspired by the outstanding performance of large language models (LLMs) in computer vision and natural language processing, this paper aims to discuss the potential of large time series models in load forecasting with scarce historical data. Specifically, the large time series model is constructed as a time series generative pre-trained transformer (TimeGPT), which is trained on massive and diverse time series datasets consisting of 100 billion data points (e.g., finance, transportation, banking, web traffic, weather, energy, healthcare, etc.). Then, the scarce historical load data is used to fine-tune the TimeGPT, which helps it to adapt to the data distribution and characteristics associated with load forecasting. Simulation results show that TimeGPT outperforms the popular benchmarks for load forecasting on several real datasets with scarce training samples, particularly for short look-ahead times. However, it cannot be guaranteed that TimeGPT is always superior to benchmarks for load forecasting with scarce data, since the performance of TimeGPT may be affected by the distribution differences between the load data and the training data. In practical applications, operators can divide the historical data into a training set and a validation set, and then use the validation set loss to decide whether TimeGPT is the best choice for a specific dataset.

Long-term load forecasting for smart grid

Abstract The load forecasting problem is a complicated non-linear problem connected with the weather, economy, and other complex factors. For electrical power systems, long-term load forecasting provides valuable information for scheduling maintenance, evaluating adequacy, and managing limited energy supplies. A future generating, transmission, and distribution facility's development and planning process begins with long-term demand forecasting. The development of advanced metering infrastructure (AMI) has greatly expanded the amount of real-time data collection on large-scale electricity consumption. The load forecasting techniques have changed significantly as a result of the real-time utilization of this vast amount of smart meter data. This study suggests numerous approaches for long-term load forecasting using smart-metered data from an actual distribution system on the NIT Patna campus. Data pre-processing is the process of converting unprocessed data into a suitable format by eliminating possible errors caused by lost or interrupted communications, the presence of noise or outliers, duplicate or incorrect data, etc. The load forecasting model is trained using historical load data and significant climatic variables discovered through correlation analysis. With a minimum MAPE and RMSE for every testing scenario, the proposed artificial neural network model yields the greatest forecasting performance for the used system data. The efficacy of the proposed technique has been through a comparison of the acquired results with various alternative load forecasting methods.

Advanced Short-Term Load Forecasting with XGBoost-RF Feature Selection and CNN-GRU

Accurate and efficient short-term load forecasting (STLF) is essential for optimizing power system operations. This study proposes a novel hybrid forecasting model that integrates XGBoost-RF feature selection with a CNN-GRU neural network to enhance prediction performance while reducing model complexity. The XGBoost-RF approach is first applied to select the most predictive features from historical load data, weather conditions, and time-based variables. A convolutional neural network (CNN) is then employed to extract spatial features, while a gated recurrent unit (GRU) captures temporal dependencies for load forecasting. By leveraging a dual-channel structure that combines long- and short-term historical load trends, the proposed model significantly mitigates cumulative errors from recursive predictions. Experimental results demonstrate that the model achieves superior performance with an average root mean square error (RMSE) of 53.29 and mean absolute percentage error (MAPE) of 3.56% on the test set. Compared to traditional models, the prediction accuracy improves by 28.140% to 110.146%. Additionally, the model exhibits strong robustness across different climatic conditions. This research validates the efficacy of integrating XGBoost-RF feature selection with CNN-GRU for STLF, offering reliable decision support for power system management.

An energy management method for combined load forecasting of fuel cell ships

Abstract The article proposes an energy management method (EMM) for fuel cell ships, which consists of two parts: load forecasting (LF) and real-time optimized scheduling (RTOS). LF is composed of Long Short-Term Memory neural networks, which are capable of being trained to predict future loads based on historical ship load information. RTOS consists of model predictive control and can allocate energy in real time. Initially, shiploads and battery status are collected in real time, and then LF passes the predicted load sequence to RTOS. RTOS optimizes real-time energy allocation based on the predicted load and the current operating status of two sets of batteries, while meeting system constraints and maintaining the battery SOC at a healthy level, to minimize operational cost consumption. The simulations demonstrate that the EMM can consistently maintain SOC at a healthy level and reduce fuel cell ship operating costs to some extent, thereby improving economic efficiency.

A secure and efficient data aggregation scheme for cloud–edge collaborative smart meters

Smart meters are part of the Advanced Measurement Infrastructure (AMI) system in the smart grid. It facilitates data transfer between consumers and electricity suppliers (ES). However, the mass deployment of smart meters (SM) brings heavy overhead to grid operation and poses serious privacy threats. To this end, this paper proposes a secure and efficient data aggregation scheme of cloud–edge collaboration smart meters. At first, we standardize the users’ historical electricity load features and use the improved K-Means clustering algorithm to calculate the Euclidean distance between feature vectors to obtain the classification results of users’ load features. On this basis, ES generates relevant parameters to encrypt meter data and protect users’ data privacy based on classification results. The aggregator (Ag) performs the data aggregation, generates the overall signature using the Schnorr aggregation signature method, and sends it to the cloud server (CS). The ES queries the CS to obtain data and parses it to realize the customer billing service. Meanwhile, this paper executes a series of experiments, and the results show that the proposed scheme exhibits significant advantages in privacy protection and system operation efficiency.

Mechanical characterization of large diameter UHMWPE ropes under quasi-static tensile loading: An experimental study

The Large diameter UHMWPEFR (Ultra High Molecular Weight Polyethylene Fiber Rope) are ideal for heavy lifting and slinging, with their mechanical properties and behaviour characterisation are key to the application.Therefore, this study conducts uniaxial tensile tests on a 115 mm diameter rope (FR-115) to investigate its mechanical properties under quasi-static tensile conditions. considering the anisotropic structural characteristics of UHMWPEFR, a method and criteria for eliminating inter-strand gaps and uneven force through pre-tensioning were proposed. Subsequently, tensile tests were carried out to examine the mechanical properties and constitutive relationship of the rope. The experimental results indicate that pre-tensioning can effectively eliminate the influence of non-elastic deformation of the rope, meeting the formal tensioning requirements based on the stability criteria of the rope structure constructed from mechanical and physical parameters. The stress-strain curve of FR-115 shows a linear bilinear shape without a yield segment during the tensioning process, presenting a typical brittle failure mode. The mechanical behavior is closely related to the historical load effects and loading speed, with the rope transitioning from soft and tough to hard and brittle, exhibiting hysteresis. Based on the above analysis, the graded load and time-varying modulus parameter are fixed, and the internal force is more sensitive to the quasi-static tensile velocity; a two-stage principal model is established based on the analysis of curve characteristics and time-dependent correlation, with 50 % MBS(Minimum Break Strength) as the characterization point, and the UHMWPEFR principal relationship in the pre-period is linear-elastic, and the strain-rate effect is taken into account for the dense section, and the rate correlation and nonlinear characteristics of the dense section of the rope can be accurately portrayed with the introduction of Cowper-Symonds.

تطبيق شبكات العصب الاصطناعي وأساليب المنطق الضبابي لتوقع الحمل على المدى القصير لدى شبكة كهرباء غرب ليبيا

This research paper investigates the application of artificial neural networks (ANNs) and fuzzy logic methods for short-term load forecasting in the Western Libyan Electric Network. Historical hourly load data collected during the year 2023 by the General Electricity Company of Libya (GECOL). The study aims to develop accurate and reliable load forecasting models to support efficient grid operation and resource planning. Historical load data and meteorological variables are utilized to train and evaluate the forecasting models. The performance of the ANNs and fuzzy logic methods is compared, and the suitability of each approach for load forecasting in the Western Libyan Electric Network is assessed. The results demonstrate the effectiveness of both ANNs and fuzzy logic methods in short-term load forecasting, providing valuable insights for electricity providers and policymakers in the region. Keywords: neural' networks, fuzzy logic, load forecasting

Short term load forecasting analysis using machine learning method: An SVM based study

Short-Term Load Forecasting (STLF) is crucial for effective energy management, enabling utilities to optimize electricity generation, distribution, and pricing strategies. This study explores the application of three machine learning models—Linear Regression (LR), Artificial Neural Networks (ANN), and Support Vector Machines (SVM)—to predict short-term electrical load demand. Each model was trained using historical load data enriched with temporal features to capture daily, seasonal, and other variations in electricity consumption. The SVM model demonstrated strong predictive capability, achieving a Mean Absolute Error (MAE) of 1887.41 MW, Mean Squared Error (MSE) of 6942.36 GW, Root Mean Squared Error (RMSE) of 2634.83 MW, and an R2 score of 91.95%. In comparison, the ANN model showed slightly higher errors, while the LR model had the highest error rates, indicating its limitations in capturing non-linear relationships. The results suggest that SVM and ANN models are more effective than LR for STLF due to their ability to handle non-linear dependencies and high-dimensional data. This study highlights the potential of machine learning techniques in enhancing the accuracy and reliability of load forecasting, ultimately supporting better decision-making in energy management.

Application of artificial neural network for peak load forecasting in 150 kV Semarang power system

Accurate load forecasting is essential for reliable and efficient operation of power systems. Traditional forecasting methods often struggle with capturing complex nonlinear patterns in load data. Artificial neural networks (ANNs) have emerged as a promising alternative due to their ability to learn complex relationships from historical data (Syed et al. in IEEEA 9:54992–55008, 2021. https://doi.org/10.1109/ACCESS.2021.3071654). This study investigates the potential of ANNs for short-term peak load forecasting in a 150 kV power system in Semarang, Indonesia. The study examines the impact of different input variables, including historical peak load, minimum load, population, and energy production, on forecasting accuracy. Several ANN architectures are trained and evaluated using mean absolute percentage error (MAPE) and mean squared error (MSE) metrics as reported by Demuth and De Jesús (neural network design). The results indicate that ANNs can achieve high accuracy in predicting peak load, with MAPE values below 10%. The study also demonstrates the importance of carefully selecting input variables and training parameters for optimal model performance. The findings highlight the potential of ANNs for improving load forecasting accuracy in power systems, contributing to enhanced grid reliability and operational efficiency. The findings of this study contribute to a deeper understanding of the application of ANNs in power system load forecasting. They demonstrate the potential of ANNs to achieve high accuracy and provide valuable insights into the factors influencing model performance. The findings are relevant for power system operators, researchers, and policymakers working to improve grid reliability and efficiency as reported by Prabha Kundur and Malik (Power System Stability and Control, McGraw-Hill Education, New York, 2022. https://www.accessengineeringlibrary.com/content/book/9781260473544).

Hydrogen storage planning robust to year‐round net load fluctuation

AbstractLong‐term hydrogen storage systems are considered a solution to the long‐term supply imbalance caused by different seasonal characteristics in renewable energy output and load. However, most existing planning models assume the whole year‐round net load series is known in advance and the optimal operation sequence of long‐term storage is derived accordingly. This assumption underestimates the necessary long‐term storage capacity because of the difficulty of accurate fine‐grained long‐term forecasting. This work proposes a long‐term hydrogen storage planning framework that is robust to year‐round net load fluctuation. The daily average component from the historical net load series is first derived to formulate the long‐term operation scenario set. On this basis, a robust planning model is developed where the trend of long‐term storage operation is forced to be identical for all long‐term operation scenarios and not to respond to one specific scenario with full information of the net load sequence. The results show that the planning results given by the proposed framework are more robust in real‐time operation with less load shedding.

Ultra-short-term Load Forecasting Based on XGBoost-BiGRU

High-precision load forecasting serves as the foundation for power grid scheduling planning and safe economic operation. In scenarios where only historical power load data is available without other external information, fully exploiting meaningful features from the temporal load sequence is crucial for improving the accuracy of load forecasting. Therefore, an ultra-short-term load forecasting method that combines eXtreme gradient boosting (XGBoost) and bidirectional gated recurrent unit (BiGRU) is proposed in this paper. Considering various factors that affect loads, a candidate feature set is established, which includes temporal information and historical loads. XGBoost is used to select the features that contribute significantly to load forecasting, forming an optimal feature set. These optimal features are then used as inputs to the BiGRU, and the bayesian optimization algorithm is applied to optimize the network hyperparameters. Then the load forecasting model for the next 15 minutes based on BiGRU is generated by training iteratively. The proposed XGBoost-BiGRU method is validated on real load data from a province in China. Experimental results demonstrate that the method can effectively avoid the impact of redundant features, improving both prediction accuracy and efficiency. The research has significant importance for guiding real-time supply-demand balance calculations and scheduling in power grids.

Efficient grid management: smart forecasting of short-term power load using PSO-LSTM

Recent load forecasting techniques combining machine learning models and hyperparameter optimization algorithms have shown success for short-term load forecasting (STLF) task, but they often require complex programming, higher computational costs, and greater parameter tuning. In this paper, we introduce an improved STLF model that combines Long Short-Term Memory (LSTM) neural network with Particle Swarm Optimization (PSO) for enhanced performance. In the proposed approach, the number of hidden neurons in different LSTM layers, learning rate and the number of iterations for training are optimized using the PSO algorithm. To validate the effectiveness of this method, meteorological data and historical load data from a real-world power grid are used as input. The experimental results reveal PSO significantly enhances hyperparameter tuning for LSTM neural networks, leading to improved predictive modelling. The PSO-LSTM model performed better than the LSTM model by more than 20% (in terms of Mean Absolute Error), and showed low sensitivity to hyperparameters. Comparative analysis with alternative approaches from the literature further validates the PSO-LSTM’s effectiveness in STLF. Additionally, the model achieved stable multi-step prediction capabilities, with average errors of 3.6445 for MAE, 4.6509 for RMSE, and 4.6519 for MAPE over a 1–4 day ahead lead times. This study highlights PSO-LSTM’s enhanced robustness and accuracy in power load prediction while addressing hyperparameter tuning challenges through self-optimization.

GCN-Transformer-Based Spatio-Temporal Load Forecasting for EV Battery Swapping Stations under Differential Couplings

To address the challenge of power absorption in grids with high renewable energy integration, electric vehicle battery swapping stations (EVBSSs) serve as critically important flexible resources. Current research on load forecasting for EVBSSs primarily employs Transformer models, which have increasingly shown a lack of adaptability to the rapid growth in scale and complexity. This paper proposes a novel data-driven forecasting model that combines the geographical feature extraction capability of graph convolutional networks (GCNs) with the multitask learning capability of Transformers. The GCN-Transformer model first leverages Spearman’s rank correlation to create a multinode feature set encompassing date, weather, and historical load data. It then employs data-adaptive graph generation for dynamic spatio-temporal graph construction and graph convolutional layers for spatial aggregation tailored to each node. Unique swapping patterns are identified through node-adaptive parameter learning, while the temporal dynamics of multidimensional features are managed by the Transformer’s components. Numerical results demonstrate enhanced accuracy and efficiency in load forecasting for multiple and widely distributed EVBSSs.

Damage inference of statically deterministic structural trusses is carried out by Bayesian updating

In the face of harsh environments, aging construction materials, and increased traffic, the performance of Bridges in service continues to deteriorate. To avoid the serious consequences caused by structural damage, it is of great significance to infer the damage of the truss structure and evaluate the reliability of the bridge. In this paper, a random sampling process is used to characterize the resistance degradation, and the historical load test information and Bayesian method are used to update the resistance of serving Bridges. The modeling of structural mechanics solver determines the maximum axial force on the damaged structural member by loading different nodes. Secondly, node displacement is calculated. Then, the displacement results are taken as new information as the parameters of prior distribution for Bayesian update, and finally, the maximum probability of structural damage is obtained. The results show that the probability of EA values between 1.34 and 1.38 is the greatest. The uncertainty and variability in the process of resistance degradation of reinforced concrete bridge members can be described more reasonably by the Bayesian updating process. By integrating the real resistance degradation information detected during bridge service into practical application, and using the research method in this paper to analyze and calculate, Bridges' future service status and remaining service life can be predicted more accurately.

Investigation of a model predictive control (MPC) strategy for seasonal thermochemical energy storage systems in district heating networks

This research investigates the integration of model predictive control (MPC) with seasonal thermochemical energy storage systems (STES) within district heating networks, focusing on the Nottingham district heating as a case study. The primary aim is to develop an MPC strategy that utilizes machine learning models for accurate heat load forecasting. This strategy optimizes the charging and discharging cycles of thermochemical energy storage systems to mitigate the mismatch between heating energy supply and demand by storing surplus heat during summer and utilizing it during winter. We employed and validated machine learning models, including support vector regression (SVR), regression trees, and long short-term memory (LSTM) networks, using historical heat load and meteorological data. A validated numerical model of the thermochemical energy storage system (TCES) was integrated into the MPC framework, formulated as a mixed-integer linear program to optimize the STES's operations. The performance of the MPC strategy was benchmarked against a rule-based control approach under varying supply capacities to evaluate scalability and robustness. Our findings reveal that each machine learning model achieved comparable performance, with CVRMSE values within the 9–11% range. The LSTM model, in particular, provided accurate multi-step forecasts essential for the MPC framework. Incorporating these models into the MPC strategy allowed for precise heat demand predictions, enhancing the management of energy storage and distribution. Results confirmed that MPC effectively shifts energy seasonally, reduces reliance on auxiliary heating during winter, and minimizes waste heat. The MPC strategy outperformed the rule-based control by storing a significantly higher percentage of waste heat and meeting a greater portion of the additional heat demand that was not covered by the auxiliary heat supply. The system demonstrated effective performance under varying supply capacities, with the MPC strategy efficiently utilizing stored heat to meet demand at 80% supply capacity, achieving a waste heat reduction to 4% and meeting most of the heat demand. However, performance declined at 60% capacity, indicating the need for careful consideration of supply capacities in system design. This study highlights the potential of integrating machine learning models with MPC to enhance the performance and adaptability of district heating systems with STES, minimizing waste heat and efficiently meeting energy demands.

Long-term deformation of rock salt under creep–fatigue stress loading paths: Modeling and prediction

Rock salt, due to its water solubility, low permeability, high plasticity, and damage self-healing ability, is one of the best candidate rock types for underground energy storage. Utilizing salt caves to construct compressed air energy storage (CAES) facilities can effectively enhance the utilization of renewable energy. Due to the need for peak shaving, the surrounding rock of the salt cavern will undergo discontinuous cyclic loading with varying gas injection rates and pressures, namely, alternating creep–fatigue loading. Considering the actual peak-shaving cycle of a CAES plant, long-term creep–fatigue tests of rock salt with different loading cycles and stress levels were conducted. The results indicate that in long-term creep–fatigue tests for rock salt, the lower the loading stress rate is, the greater the deformation of the rock salt. The variation in the stress limit has a greater effect on creep than on fatigue loading and unloading. The deformation rate of rock salt is influenced by alterations in the stress state. Based on the test results, according to the Norton creep model, a new creep–fatigue constitutive model for rock salt was established by defining a state variable that characterizes the level of rock hardening and introducing unloading as well as crack factors. This model can accurately describe the impact of historical loading and unloading processes on the viscoplastic mechanical characteristics of rock salt. The rock salt creep–fatigue test results were used to verify the constitutive model. A comparison of the fitting curve of the different stress loading paths with the test curve reveals good consistency, indicating that the model comprehensively considers the effects of time, load, and state on rock salt creep–fatigue, effectively describing the viscoplastic deformation characteristics of rock salt under different stress paths. These research findings provide important guidance for ensuring the stability of salt caverns used for CAES.