Search Results (284)

Accurate air conditioning (AC) load prediction under extreme weather conditions is critical for power grid stability and energy management. While tree-based ensemble methods such as XGBoost, LightGBM, and Gradient Boosting have become the dominant paradigm in short-term load forecasting, their effectiveness for weather-sensitive AC load prediction—particularly during extreme weather events—remains insufficiently examined. This study presents a systematic comparison of four machine learning models—Support Vector Regression with RBF kernel (SVR-RBF), LightGBM, XGBoost, and Gradient Boosting—for daily AC load estimation conditioned on measured same-day meteorological and calendar features. Based on five years of processed load and weather data from a major city in Southwest China, we construct 31 features and evaluate model performance across four scenarios: normal days, weather-extreme days, high-load P85, and high-load P90. Hyperparameters are selected within the first four years by expanding-window validation, and the fifth year is held out for testing. Our results reveal that SVR-RBF achieves an overall $R^2$ of 0.9772, substantially outperforming LightGBM (0.9091), XGBoost (0.8972), and Gradient Boosting (0.9094); paired moving-block bootstrap intervals for the tree-minus-SVR MAE differences exclude zero. The advantage of SVR-RBF is most pronounced under extreme conditions: on weather-extreme days, SVR-RBF attains $R^2=0.8316$ versus $R^2=0.39$ for the best tree model. This pattern is consistent with the smooth U-shaped temperature–load relationship captured by the RBF kernel. Additional trend-sensitivity analysis shows that annual load growth and target-level extrapolation also explain a substantial part of the tree-model degradation. Furthermore, we quantify the performance limitation of the restricted feature set: while normal-day estimation achieves $R^2=0.984$, high-load P90 days reach $R^2=0.843$. These findings support SVR-RBF as a strong baseline for measured-weather conditional AC load estimation while emphasizing that temporal shift and training-domain coverage must be considered when interpreting model differences. For energy-system applications, the lower errors under high-load and weather-extreme conditions are relevant to peak-demand assessment, reserve planning, and demand-side management. Full article

Industrial tariffs and demand-response (DR) programs pay for two quantities—daily energy and daily peak demand—yet short-term forecasting research addresses mostly the first, and rarely under the constraints of small industrial sites: strongly bimodal operation and only one to two years of records. We propose the Two-Stage Adaptive Framework (TSAF), a metering-only method that detects bimodality, classifies each day as active or inactive from its partial-day consumption, and fits a regression model to active days only; the same 21 features serve both targets. On 15 min data from ten plating factories of the Ansan Plating Industrial Complex (19 months), TSAF reaches 16.2% mean MAPE on daily energy against 58.7% for a day-ahead reference, and no deep-learning model beats the 22-parameter Ridge regressor. On daily peak, evaluated on active days, a joint multi-factory Transformer reaches 8.4% MAPE against a 14.0% constant-predictor floor, and a training-free tabular foundation model (TabPFN) reaches a comparable 7.0% without cross-factory data; intraday peak timing (≈2 h mean error) marks the limit of the meter-only design. External validation on 36 stratified UCI clients delimits the framework’s scope, and a weather ablation, specified in advance of estimation, finds no significant gain (p = 0.23). After adjusting for Stage 1 misclassification and intraday dispatch feasibility, the ten factories gain about 60 million KRW (US$46,000) per year and avoid 15.4 tCO₂ under Korean market conditions. Because TSAF needs only the smart-meter feed, power suppliers and DR aggregators can deploy it without access to customer-facility internals. Full article

Data centers can provide demand-side flexibility by coordinating computing workloads, cooling systems, and on-site energy resources. However, facility demand varies with information technology (IT) load and cooling operation, making a fixed power usage effectiveness (PUE) or a forecast independent of planned controls inconsistent with dispatch. This paper proposes a control-aware, multi-horizon PUE forecasting framework for coordinated data center demand-side management (DSM) and microgrid dispatch. The key idea of this control-aware approach is to forecast PUE using planned workload and cooling schedules as inputs. A power-consistent Temporal Fusion Transformer (PC-TFT) predicts quantiles of non-IT overhead power and rack inlet temperature from telemetry, weather forecasts, admitted requests, and candidate workload and cooling schedules. Facility power and PUE are derived from the algebraic power balance, ensuring consistency among IT, overhead, and facility power and PUE values no lower than 1. Empirical split conformal calibration and temporally dependent scenarios characterize forecast uncertainty. A trajectory-conditioned piecewise-affine control response map with a recursive thermal state links the forecasts to a risk-informed model predictive controller that coordinates workloads, cooling, photovoltaic generation, battery storage, and grid exchange. The proposed framework is validated through EnergyPlus simulations of a Shenzhen data center, coupled with workload and microgrid simulations. Forecasting performance is compared with persistence and matched-input neural baselines, while dispatch is benchmarked against deterministic and oracle controllers. The results demonstrate improved multi-horizon PUE forecasting accuracy and empirical interval calibration, lower operating cost and peak grid demand, higher renewable energy utilization, and fewer service quality violations. These findings indicate that control-aware, power-balance-constrained probabilistic PUE forecasts can provide a reliable basis for coordinated data center DSM and microgrid dispatch. Full article

Underprediction of high electrical demand can be more operationally consequential than an equally sized overprediction, yet standard point forecasting models are optimized primarily for average error. This case study evaluates a multivariate quantile Transformer as a safety-oriented next observation forecasting layer for a university laboratory. A timestamp-level audit identified 33,374 native measurements collected from 22 April to 12 December 2024 at a median interval of approximately 10 min. The final leakage-free pipeline uses only real observations, performs the chronological split before sequence generation, fits all scalers on training data only, and rejects windows containing gaps greater than 30 min. Persistence, fixed-order SARIMA, LSTM, GRU, CNN–LSTM, and an MSE-trained Transformer were evaluated on the same 6595-sample test period. GRU achieved the best deterministic accuracy (MAE 0.017744 kW; RMSE 0.023278 kW), whereas the proposed τ = 0.90 Transformer intentionally traded point accuracy (MAE 0.033830 ± 0.001133 kW) for asymmetric risk control. Across five independent runs, it achieved a pinball loss of 0.004453 ± 0.000069 kW, empirical coverage of 87.95 ± 1.01%, and a peak underprediction rate of 26.64 ± 5.32%, compared with 72.94–100% for the conventional benchmark outputs. Additional τ = 0.75 and τ = 0.95 experiments demonstrate the expected accuracy–safety trade-off. MAPE is not used as a primary metric because near-zero loads make percentage errors unstable. The results support the proposed model as a complementary upper quantile forecasting layer for this small, dynamic facility; they do not establish general performance at feeder or system scale. Full article

Residential short-term load forecasting (STLF) plays a critical role in maintaining grid stability and enabling effective demand response in smart grid environments. However, residential electricity consumption exhibits strong stochastic variability, non-linear patterns, and abrupt changes caused by occupant behavior and environmental factors, making direct forecasting of raw power signals highly susceptible to high-frequency noise and temporal prediction lag. To address these challenges, this study proposes a Robust Residual Attention Network (Robust-RAN) for trend-oriented residential STLF. The proposed framework integrates localized causal denoising, residual multi-head attention mechanisms, and cyclic temporal encoding to mitigate observable phase lag and stabilize predictions against stochastic fluctuations, capture both short- and long-term temporal dependencies, and preserve daily periodicity. The proposed model was evaluated on the public UCI Appliances Energy Prediction dataset using comprehensive regression, classification, and residual analysis metrics. The experimental results demonstrate that Robust-RAN achieves a Coefficient of Determination (R²) of 0.9063 and a Mean Absolute Error (MAE) of 13.00 W. Under high-load conditions, the proposed approach further attains a peak detection accuracy of 96.63% with an F1-score of 0.8529, indicating reliable prediction of critical demand events. The results demonstrate that the proposed framework provides a stable and accurate estimation of the underlying residential load trend, offering a stable forecasting framework evaluated on the UCI benchmark dataset for demand response, peak- load management, and advanced residential energy management applications. Full article

In this technological era, electrical energy is the bloodstream for the economic and social development of any country. It is the need of the time that developing countries like Pakistan have strategic planning for efficient generation and utilisation of electricity and have as much cheap electricity as possible at their disposal while respecting environmental constraints. The overloaded and ageing infrastructure of an electrical power network can impact system reliability and the sustainability of power generation, transmission and distribution mechanisms. The initiation of the planning process depends upon accurate load estimation to optimally fulfil consumers’ power needs. This paper compares statistical, hybrid and deep learning (DL) mechanisms, including SARIMAX, SARIMA with gradient boosting (SARIMA-GB), long short-term memory (LSTM) network, STL decomposition with LSTM, and CWT-LeNet-5-LSTM, for the prediction of residential electricity demand in the LESCO region of central Pakistan. The study uses 7670 daily feeder observations recorded between 1 January 2002 and 31 December 2022. The series is modelled at its native daily resolution and partitioned chronologically into a fitting span of 5216 days, a validation span of 920 days and a test span of 1534 days beginning 20 October 2018. All models receive the same block of 14 exogenous calendar variables, four annual Fourier harmonic pairs, day-of-week and month sine and cosine terms, a weekend indicator and a linear trend, and all neural models are trained with a validation split, early stopping and restoration of the best weights rather than for a fixed number of epochs. Accuracy is assessed with MAE, RMSE, MAPE and peak normalized RMSE under one recursive protocol at forecast leads of 1, 7, 14 and 30 days, because the ranking of the frameworks depends on the lead. Averaged over three random initialisations, the proposed CWT-LeNet-5-LSTM attains the lowest error at every multi-step lead, reaching an MAPE of 2.35 ± 0.28% at lead 30 against 3.32% for the multivariate LSTM, 3.89% for SARIMA-GB and 4.47% for SARIMAX. At lead 1, SARIMA-GB is the more accurate model (0.76% against 1.20 ± 0.17%) because the previous day’s observed load dominates one-step prediction for a series whose lag-one autocorrelation is 0.978. An architecture ablation isolates the contribution of the wavelet stage, the convolutional stage, the anisotropic pooling and the calendar fusion. A stratified analysis across seasons, weekdays, weekends and high-, medium- and low-load days shows where the advantage is concentrated. Additionally, Diebold–Mariano tests identify the statistical significance of the differences. Full article

As urban rail transit systems enter a stage of high-quality development, accurate short-term passenger flow forecasting has become essential for efficient operation and management. To improve the accuracy and robustness of multi-step short-term passenger flow forecasting under fluctuating demand conditions, this study develops a feature-enhanced Informer framework that embeds Complex Network Representation (CNR) into the Informer architecture to jointly capture both the topological characteristics of the subway network and spatiotemporal passenger flow dynamics. External factors, including subway schedules and land use around stations, are further integrated to enrich the input features. In addition, the ProbSparse self-attention mechanism is adopted to improve long-sequence dependency modeling, thereby enabling efficient multi-step passenger flow forecasting. Experiments were conducted on the Beijing metro passenger flow dataset from January to October 2024 to evaluate the proposed model. The dataset covered 264 stations and was aggregated at 15 min intervals. Based on historical passenger flow and multi-source features, the model predicts passenger flow over multiple future time steps. The overall evaluation metrics were calculated on the test set and averaged over all test samples and observed stations. The experimental results show that, compared with the standard Transformer model, the proposed model reduces the average prediction error by 16.59% on weekdays and 20.48% on weekends while maintaining stable predictive performance during peak hours. Sensitivity analysis and ablation studies are further conducted to evaluate the model performance across different station types and forecasting horizons. The results demonstrate that the proposed model can provide reliable decision support for intelligent urban rail transit operations, including transport capacity scheduling, passenger service improvement, and operating cost reduction. Full article

The rapid growth of Electric Vehicles (EVs) has accelerated the deployment of Electric Vehicle Charging Stations (EVCSs), making their integration into modern power systems increasingly important. While EVCSs support transportation electrification and global decarbonization goals, large-scale integration introduces technical challenges that affect power system operation, reliability, and planning. This review provides a comprehensive assessment of the impact of EVCS integration on power system performance by examining charging technologies, charging stations, charging modes, and the principal components of EVCSs. The review discusses the effects of EV charging on load demand, peak load, voltage profile, voltage stability, active and reactive power losses, transformer loading, and overall grid performance. It further evaluates mitigation strategies, including smart charging, coordinated charging, Demand Response (DR), Renewable Energy Sources (RESs), Battery Energy Storage Systems (BESSs), Vehicle-to-Grid (V2G) technology, and Artificial Intelligence (AI)-based energy management. The application of Machine Learning (ML), Deep Learning (DL), Reinforcement Learning (RL), and advanced optimization algorithms for charging coordination and demand forecasting is also reviewed. Finally, the paper identifies current research challenges and future directions related to charging uncertainty, renewable energy integration, cybersecurity, interoperability, and infrastructure development. The findings demonstrate that intelligent charging strategies combined with renewable energy integration, energy storage, V2G, and AI significantly improve the reliability, efficiency, resilience, and sustainability of future EV-integrated power systems. Full article

Rapid urbanization and growing passenger demand are placing increasing pressure on metro systems to provide reliable, sustainable, and passenger-centered urban mobility services. This paper presents Masar, an AI-powered digital twin prototype for Riyadh Metro that integrates synthetic passenger-flow simulation, short-term congestion forecasting, real-time monitoring, an Arabic conversational assistant, and an AI-assisted Lost & Found service. The system combines a passenger mobile application with a staff operations dashboard to support both passenger guidance and operational decision-making. Because granular Riyadh Metro operational data are not publicly available, a reproducible synthetic data generation and preparation pipeline was developed. The pipeline uses a dual-Gaussian demand model to represent morning and evening demand peaks, scales station-level occupancy using capacity, day-type, and event multipliers, and extends the simulation to trip, stop-time, and carriage-level passenger-flow records. Machine learning models were trained and evaluated on a dataset of 194,580 min-level records generated across six representative stations and organized into monthly CSV files; the same simulation engine was subsequently deployed in production to synchronize live station data with Firebase Firestore, supporting real-time digital-twin visualization and inference. Four machine learning models were evaluated for 30 min congestion forecasting. XGBoost achieved the strongest overall performance, with an RMSE of 363.42, an R² of 0.9229, and a severe congestion detection rate of 0.9944. The findings demonstrate the technical feasibility of the proposed architecture within a synthetic evaluation environment; the reported metrics reflect internal consistency rather than validated accuracy on real operational data, and should be interpreted as proof-of-concept evidence pending field deployment. Full article

To address the high dependence on diesel generation, renewable energy variability, and limited demand-side flexibility of remote islanded microgrids, this study develops a low-carbon economic dispatch framework for an islanded photovoltaic–wind–diesel–battery energy storage system with coordinated demand response. The proposed model minimizes the operating cost, pollutant treatment cost, and load-loss penalty cost while satisfying generation-output, battery state-of-charge, charging and discharging, demand-response, and islanded power-balance constraints. To solve the resulting high-dimensional, nonlinear, and strongly constrained optimization problem, a rotation quantum particle swarm optimization algorithm (RQPSO) is proposed. In contrast to the conventional velocity–position update, RQPSO independently encodes each decision variable using a full-dimensional quantum phase representation and performs the search through a shortest-path rotation-guided phase-updating mechanism. Adaptive angular mutation, elite local refinement, and stagnation-aware restart are further incorporated to balance global exploration, local exploitation, and convergence stability. The algorithm is evaluated using nine 30-dimensional benchmark functions and representative 24 h forecasted load and renewable-generation profiles for Island data. Under the reliability-priority scheduling scheme, RQPSO achieves a total scheduling cost of 69,017.69 CNY, diesel fuel consumption of 6636.20 kg, and estimated CO₂ emissions of 18,332.49 kg. Compared with conventional PSO, these three indicators are reduced by 9.34%, 12.25%, and 12.25%, respectively. RQPSO also reduces the total cost by 6.16–27.36% relative to six comparison algorithms. The results demonstrate that the coordination of demand response and battery storage can improve peak–valley regulation, reduce diesel dependence and emissions, and maintain feasible and economical operation under different renewable-generation conditions. Full article

The growing demand for sustainable urban infrastructure requires reliable and energy-efficient lighting solutions, particularly for urban parks characterized by dynamic energy consumption patterns. This study proposes an integrated analytical framework combining deterministic energy system modeling, a novel interaction-oriented metric, and machine learning techniques to evaluate hybrid photovoltaic (PV)–wind systems for urban park lighting applications. Multi-year hourly meteorological and lighting-demand data from millet gardens in Balıkesir and Çanakkale, Türkiye, were analyzed. A Hybrid Energy Synergy Index (HESI) was introduced to quantify the degree of concurrent contribution and coordination between PV and wind resources relative to lighting demand. The results show substantial demand variability, with an average daily demand of approximately 1609 kWh and peak values exceeding 25,600 kWh. Photovoltaic generation dominated total renewable energy production (92.95%), whereas wind energy contributed 7.05%. The average HESI value (≈0.113) indicated weak source coordination, accompanied by persistent energy deficits that occasionally exceeded −20,000 kWh. Reliability analysis revealed severe system inadequacy, with a reliability rate of only 0.000547 (0.055%). Within the evaluated design space, the highest-performing configuration consisted of 49 PV panels and 19 wind turbines; however, reliability improvements remained limited, indicating that capacity expansion alone is insufficient to ensure satisfactory performance. Machine learning models achieved high predictive accuracy for HESI forecasting (R² = 0.9902; MAE = 0.0027), while explainable artificial intelligence identified solar radiation and wind speed as the dominant environmental drivers. The results highlight the importance of integrating renewable energy capacity planning with energy storage support, improved source coordination, and adaptive energy management strategies to enhance the reliability of sustainable urban lighting systems. Full article

Data centers are facing rapidly increasing electricity demand and carbon emissions, while the intermittency of renewable energy creates a significant temporal mismatch between renewable generation and data center load demand. To bridge this temporal mismatch, we propose a coordinated optimization strategy that integrates shared hydrogen energy storage facilities with load scheduling mechanisms. A multi-objective MILP model is formulated to minimize annualized cost, renewable energy curtailment, and carbon emissions. Simulation results show that, compared with the no-shared-station case, the proposed electricity–hydrogen coordination strategy with load shifting yields significant benefits: the annualized total cost decreases from 13.65 to 4.74 million yuan; annual carbon emissions are reduced from 6297 to 1432 tons; and peak-period electricity purchases are reduced from 5373 to 906 MWh. Under the representative daily forecast condition, Scenario S4 achieves zero renewable curtailment when grid export is permitted; therefore, the renewable-electricity utilization rate reaches 100.00% within the model boundary. When grid export is prohibited, the utilization rate decreases to 98.59%, with 179,100 kWh of annualized renewable curtailment. The research findings indicate that integrating shared hydrogen energy storage with the load flexibility of data centers can effectively reduce the system’s overall operating costs, promote the integration of renewable energy, and achieve low-carbon operation. Full article

Urbanization often enhances residents’ material living standards but fails to fully satisfy their sports and spiritual needs. Using the Beijing–Tianjin–Hebei (BTH) region of China as a case study, this study performed statistical, spatial hot/cold spot, and landscape patch analyses to characterize the quantity, spatial configuration, and supply–demand matching characteristics of sport/cultural land (SCL) and park/greenspace land (PGL). This paper proposes a decision support approach for balancing supply and demand of the two land-use types through reasonable population forecasting. The main results were as follows: (1) SCL and PGL varied markedly across BTH. Hebei had the largest SCL and PGL areas, while Tianjin had the smallest; Beijing had the highest area proportion and strongest clustering. (2) Hebei shaped the BTH landscape pattern, with highly fragmented but regular, well-connected, and clustered SCL, while its PGL was intact and strongly connected yet dispersed. Beijing and Tianjin showed opposite features. (3) The per capita SCL and PGL area generally met demands, but spatial mismatches were prominent, peaking in the 1.19 km buffer and concentrated in urban cores. (4) The BTH population will grow fastest in Beijing and slowest in Tianjin. Beijing has the most obvious SCL provision–demand imbalances, and Hebei has the most severe PGL conflicts, mainly in the 1.19 km buffer. These findings unveil the distinct mismatch patterns of recreational land across multi-level urban agglomerations, thereby addressing a research gap regarding spatial mismatch governance for cross-provincial mega-regions in developing nations. The proposed optimization strategies offer practical references for achieving a balanced allocation of recreational land, advancing the Healthy China 2030 initiative and contributing to the attainment of SDG 3. Full article

Accurate daily tourist flow prediction is crucial for optimizing tourism management, allocating public resources, and ensuring sustainable ecological development in scenic areas. However, forecasting visitor volume at the Jiuzhaigou Scenic Area poses significant challenges for conventional deep learning architectures due to the data’s high volatility, strong seasonality, and complex non-stationarity. To address these limitations, this study proposes a novel integrated forecasting framework, SD-ConvLSTM-Attn (Seasonal Trend Decomposition-Enhanced Convolutional Long Short-Term Memory with Attention). Using daily visitor data spanning 30 March 2015 to 10 April 2026, collected from the official website of the Jiuzhaigou Scenic Area on 11 April 2026, the model employs a “divide-and-conquer” strategy. A Seasonal Trend decomposition layer first decouples the non-stationary raw sequence into a deterministic trend and stochastic seasonal components. Subsequently, a ConvLSTM module is utilized to extract local features, while a Multi-Head Attention mechanism is integrated to capture long-range temporal dependencies, including recurring Golden Week demand spikes. The experimental results demonstrate that the SD-ConvLSTM-Attn architecture achieves competitive predictive performance against six benchmark architectures, including LSTM, CNN-LSTM, ConvLSTM, TimeMixer, DLinear, and PatchTST, exhibiting superior accuracy in peak flow capture. Furthermore, a hierarchical capacity warning system and predictive resource scheduling protocol are proposed to support dynamic scenic area management. Full article

As passenger travel and freight transport demand continue to rise, transportation remains a major source of global carbon emissions. Predicting transportation carbon emissions helps support sustainable transportation planning and the development of scientifically sound emission reduction policies. However, there are currently no unified standards for accounting for transportation carbon emissions, and different approaches yield widely varying predictions, creating challenges for formulating carbon emission policies. In this study, a top–down approach based on China’s energy statistical data is adopted to calculate regional transportation carbon emissions (TCE). Twenty factors influencing carbon emissions are selected, and Spearman correlation analysis is used to examine the correlations among these factors. Lasso regression and the STIRPAT model are employed to quantitatively analyze the influence of each factor. Subsequently, an improved hybrid deep learning model, GA–CNN–LSTM–Attention, is proposed to predict carbon emissions under baseline, restriction, and control scenarios. Finally, an empirical analysis is grounded in a case study of Fujian Province, China. The quantitative analysis results show that, of all the factors, the urbanization rate has the greatest impact on transportation carbon emissions in the region, and the GA–CNN–LSTM–Attention prediction model achieves higher forecasting accuracy. Under the first two scenarios, carbon emissions in the region show a year-by-year increasing trend from 2022 to 2035, with no emissions peak observed. Under the controlled scenario, carbon emissions are projected to peak in 2030, reaching approximately 70.28 million tons of CO₂. The prediction results provide a scientific basis for sustainable transportation development and government policymaking on carbon reduction. Full article