Search Results (216)

To address the deepening electro-carbon coupling and flexibility shortages in active distribution networks with high renewable energy penetration, this paper proposes a bi-level collaborative planning strategy considering dynamic electro-carbon factors. First, considering the spatial–temporal correlation of wind and solar outputs, typical renewable energy scenarios are generated using the Frank-Copula function and clustering algorithms. Second, a bi-level planning model for the Soft Open Point integrated with an Energy Storage System (E-SOP) is established: the upper level optimizes the siting and sizing of E-SOPs to minimize the annualized comprehensive cost; the lower level incorporates a dynamic stepped carbon trading mechanism and a continuous price-based demand response (PBDR) mechanism to achieve optimal operational economy. For model solving, a hybrid bi-level decomposition strategy combining the Dhole Optimization Algorithm (DOA) and second-order cone programming (SOCP) is adopted, utilizing a coordinated dual-level solution interaction to favorably support numerical stability. Case studies on a modified IEEE 33-node system demonstrate that the proposed scheme reduces the annualized comprehensive cost by 12.3% and transforms the carbon trading expenditure into a net revenue, thereby significantly enhancing the low-carbon economic efficiency and operational flexibility of the distribution network. Full article

Accurate forecasting of urban electric vehicle (EV) charging demand is important for power system operation, sustainable transport electrification, and charging infrastructure planning. However, this task remains challenging because EV charging demand is shaped by temporal usage patterns as well as changing relationships among weather conditions, operational factors, and historical charging behavior. Many existing forecasting models treat these explanatory variables mainly as parallel inputs, while their mutual relationships are often predefined, simplified, or left implicit in the temporal learning process. To support AI-driven charging demand management, this study proposes an adaptive graph-based recurrent network (A-GRN) for city-level aggregated EV charging-load forecasting. In the proposed framework, key explanatory variables are represented as feature nodes, and their connections are learned through an adaptive adjacency matrix rather than a fixed spatial topology. The adaptive graph neural network (AGN) module captures feature-level interactions, while a dual-path gated recurrent unit module (DG-GRU) extracts temporal representations from the charging-load sequence. Experiments on a city-level EV charging dataset show that A-GRN outperforms several baseline models, including naive persistence forecasting, GRU, LSTM, BiGRU, TCN, and GCN. Compared with the BiGRU baseline, A-GRN reduces MAE, MSE, and RMSE by 31.36%, 34.65%, and 20.48%, respectively. In the original physical unit, the MAE is reduced from 187.43 kWh to 128.64 kWh, and the RMSE is reduced from 222.69 kWh to 177.08 kWh. The results indicate that feature-level graph learning can improve short-term EV charging-load forecasting, especially when the target is an aggregated urban load rather than the load of a single charging station. The proposed model provides a data-driven forecasting tool for sustainable urban charging demand management, low-carbon transport operation, and charging infrastructure planning. Full article

This study investigates how the hydrogen supply chain should be designed under alternative carbon tax scenarios to decarbonize heavy-duty freight transportation. A bi-objective, multi-period optimization model is developed to minimize the total daily system cost while constraining CO₂ emissions using the Augmented ε-constraint approach, thereby revealing the trade-off between economic and environmental objectives. The model was applied to Türkiye’s heavy-duty transportation sector and solved under zero, moderate, and aggressive carbon tax scenarios. The results show that the levelized cost of hydrogen (LCOH) ranges from 2.06 to 14.06 $/kg H₂. High carbon pricing increases the LCOH by 29.06% in hybrid designs, while raising the renewable energy share from 2.04% to 46.97% in centralized supply chains. Sensitivity analysis reveals that a ±20% variation in electrolyzer-based production costs does not alter the network topology but shifts the LCOH between 13.10 and 15.02 $/kg H₂ in emission-focused solutions. The findings indicate that in renewable-energy-based decentralized structures, higher carbon tax policies primarily increase the LCOH. Still, the overall technology mix and network topology remain largely unchanged compared to the no-tax case. However, in centralized supply chains, carbon pricing affects both the energy sources and selected technologies. By integrating Türkiye’s 2030–2053 policy milestones into a multi-period framework, this study distinguishes itself by providing a comprehensive, multi-period planning framework tailored to the economic and logistical realities of developing countries. Unlike existing models, our approach quantifies how evolving carbon tax trajectories decisively drive infrastructure investment by analyzing the direct impact of different tax levels on the operational and strategic decisions of heavy-duty transport. This research represents the first joint assessment of carbon tax policy instruments and the evolution of long-term hydrogen supply chains, offering a decision-making framework for policy-driven energy transitions in similar emerging economies. Full article

Under the goals of carbon peaking and carbon neutrality and the development of zero-carbon parks, the continuous expansion of distributed photovoltaic (PV) installations has made grid-export constraints increasingly prominent. To investigate their influence on energy storage configuration and system operation, this paper incorporates the grid-export ratio constraint into the planning and scheduling process of a park-level PV-storage-charging system. A bilevel optimization model is established, in which the upper level minimizes the annual total cost (ATC), while the lower level minimizes the annual operating cost (AOC), considering time-of-use electricity prices, PV curtailment penalty, power shortage penalty, and battery degradation cost. The model is solved by a genetic algorithm (GA) and CPLEX. The results show that, for the studied industrial park, the 20% grid-export ratio is an important case-specific turning point under the given PV capacity, load level, electricity price, storage cost, and grid-connection conditions. Compared with the scheme without energy storage, the scheme with energy storage achieves lower PV curtailment and better economic performance. Sensitivity analyses further show that the PV curtailment penalty coefficient, energy storage investment cost, and PV installed capacity affect the optimal storage configuration and system economics. This study can provide a reference for energy storage planning and operation optimization of park-level PV-storage-charging systems under grid-export constraints. Full article

Background: Predicting rehabilitation outcomes at admission supports tailored therapy plans and efficient use of resources for patients undergoing intensive inpatient rehabilitation, including those with stroke, orthopedic, and other neurological conditions. Nonetheless, current machine learning (ML) methods face limitations, including the ceiling effect in absolute functional gain measures, the uniform treatment of diverse patient groups, and reliance on black-box models that lack clinical transparency. Methods: This retrospective observational study analyzed a fully anonymized, publicly available dataset of 3419 patients admitted to the intensive rehabilitation unit at IRCCS San Raffaele Hospital, Rome, Italy, from 2015 to 2018. To mitigate the ceiling effect, a normalized Barthel Index gain metric was developed. K-means clustering (K = 2, trained solely on the training set) identified patient admission profiles based on functionality, which were then used as predictive features. Eight machine learning classifiers were tested across three groups (All Patients, Orthopedic, Neurological). SHAP-based explainability was employed at four levels: global, diagnostic group, patient functional profile, and individual. Finally, clinical decision rules and bedside stratification profiles were derived and validated with an internal held-out test set (n = 684). Results: Normalization significantly increased the correlation between admission BI and gain (r = 0.130 to r = 0.520), supporting the presence of a ceiling-related limitation in absolute gain metrics. Two distinct functional admission profiles with statistically significant group differences were identified—High-Burden (38% below-median recovery) and Moderate-Burden (21%)—with cluster membership the third most important predictor (13.9% SHAP importance). The highest AUC-ROC values were 0.831 for all patients (XGBoost), 0.864 for neurological patients (Gradient Boosting), and 0.839 for orthopedic patients (Gradient Boosting). Multilevel SHAP analysis showed age as the primary predictor for neurological patients (mean |SHAP| = 0.360) but the third for orthopedic patients (0.350), highlighting clinical relevance. Validation using SHAP values from the Gradient Boosting model showed a Spearman correlation of ρ = 0.925 (p = 1.13 × 10⁻³⁰), with eight of the top ten features overlapping, indicating that these patterns are not model-specific but reflect the underlying data. Risk zone stratification found 80.7% of patients in high-confidence zones (accuracy > 80%). The clinical decision rules achieved 70.8% accuracy with full transparency, and the elderly (≥75 years) combined with a low BI (<25) profile showed an 89.6% model accuracy with only 10.4% recovery above the median. Conclusions: This explainable, profile-informed ML pipeline addresses key methodological limitations in predicting rehabilitation outcomes. It also provides a foundation for integrating models into clinical practice, pending prospective, external validation of the results. Before clinical implementation, validation across multicenter cohorts is essential. Full article

To address the challenges posed by the diversification of electric vehicle charging demand and the low-carbon economic operation of distribution networks, this paper proposes a bi-level low-carbon distribution network expansion planning method considering the integration of multi-type EV charging facilities. The planning layer of the model aims to minimize the annual total system cost and performs coordinated decision-making for multi-type charging facilities, new line construction, and distributed generation. By introducing a coordinated configuration mechanism for multi-type charging facilities, the model effectively matches diverse user charging demands. In the operation layer, the Aumann–Shapley value method is employed to fairly and accurately quantify carbon emission responsibilities, based on which system carbon allowances are determined. An integrated green certificate-tiered carbon trading mechanism is then established. Meanwhile, a low-carbon demand response model considering dynamic carbon emission factors is introduced to enable low-carbon optimal operation of the distribution network. Finally, simulations are conducted on a modified IEEE 33-bus system. The results demonstrate that the proposed method can effectively reduce total system cost and carbon emissions while satisfying diverse charging demands. Full article

For power systems with a high penetration of renewable energy, energy storage allocation is important for enhancing system flexibility and supporting renewable energy integration. Existing planning methods cannot simultaneously reflect source-load uncertainty and the stage-wise evolution of system development. To address this issue, this paper proposes a stage-wise energy storage planning framework based on bilevel optimization. The proposed method employs an LSTM model to construct representative wind power, photovoltaic power, and load time series for the subsequent optimization analysis, and applies K-means clustering to extract representative operating scenarios. The Qinghai power system is selected as a case study for validation. The results show that the proposed method can reasonably capture the stage-wise characteristics of storage demand, with deviation rates of 4.6% for storage power and 3.2% for storage capacity. Under low-, medium-, and high-growth scenarios, storage demand increases significantly with renewable development scale. In the high-growth scenario, the required storage capacity increases from 277,836 MWh in 2030 to 926,120 MWh in 2035. Meanwhile, the role of storage shifts from short-term power balancing to peak shaving and inter-temporal energy shifting, while the optimal storage duration remains stable at 3–4 h. The proposed framework provides a basis for long-term energy storage planning in power systems with high renewable penetration. Full article

Traffic flow prediction is important for route planning, signal control, and traffic guidance. However, traffic-state sequences usually exhibit non-stationarity, periodicity, and complex temporal dependencies, which makes it difficult for traditional statistical methods and single deep learning models to simultaneously capture short-term local fluctuations and long-term evolutionary trends. To address this issue, this paper proposes a dual-stream latefusion CNN-LSTM with an adaptive gated shortcut for traffic flow prediction, denoted as AGS-CNN-LSTM. The proposed method does not aim at explicit spatial-topology modeling; instead, it focuses on improving the fusion mechanism of CNN-LSTM-based models under settings without graph-structure constraints. Based on two public datasets, PeMS-BAY and PeMSD8, this study constructs multi-step prediction tasks with horizons of 15 min, 30 min, 60 min, 90 min, and 120 min and compares the proposed model with MLP, SimpleRNN, 1DCNN, LSTM, Serial CNN-LSTM, CNN-LSTM-Attention, BiLSTM-Attention, TCN-LSTM, Transformer Encoder, DLinear, and DS-CNN-LSTM (w/o Gate). The experimental results show that AGS-CNN-LSTM does not consistently achieve the best performance across all datasets, prediction horizons, and evaluation metrics. Nevertheless, it performs close to the best baseline models on the 30 min and 60 min tasks of PeMS-BAY and achieves competitive RMSE and R² results on the 15 min, 30 min, and 60 min tasks of PeMSD8. Further ablation experiments indicate that the adaptive gated shortcut can enhance the predictive capability of the dual-stream late-fusion structure in some scenarios, although its benefits are dependent on the dataset and prediction horizon. Overall, the proposed model is more appropriately regarded as a lightweight fusion-mechanism improvement for CNN-LSTM-based models under settings without explicit graph-structure constraints, rather than a comprehensive replacement for complex graph neural networks, Transformerbased models, or models incorporating multiple external factors. Therefore, the findings should be interpreted as proof-of-concept evidence for a lightweight CNN-LSTM fusion enhancement under constrained non-graph-input settings, rather than as evidence of broad generalizability in complete road-network-level traffic forecasting. Full article

Multimodal transport has emerged as an effective solution for improving freight efficiency and promoting sustainable logistics, reducing environmental impacts; however, route choice remains challenging under uncertain demand and dynamic transshipment time. This study addresses this problem by developing a bi-objective route-choice model that minimises total transport cost and total transport time while explicitly capturing the correlation between freight demand and transshipment time. The model is transformed into a deterministic equivalent using chance-constrained programming, enabling rigorous optimisation under predefined confidence levels and solved by a simulated annealing-based genetic algorithm (SAGA), which combines the global exploration capability of genetic algorithms with the local search efficiency of simulated annealing to improve convergence and solution quality. By incorporating carbon emission costs into the objective functions, the model supports environmentally and economically sustainable transport strategies. A numerical case study is conducted to validate the proposed approach. The results show that when freight demand is significantly below the capacity threshold, the optimal solution tends to adopt a single-mode transport scheme with stable route structure, whereas higher demand necessitates multimodal strategies, with cost–time trade-offs clearly observed. Sensitivity analysis further reveals a clear trade-off between cost and time: a time-oriented strategy dominated by rail transport reduces total transport time by approximately 20%, whereas a cost-oriented strategy relying on waterway transport decreases total cost by about 73%. These findings demonstrate the effectiveness of the proposed model and provide decision support for efficient and sustainable multimodal transport planning under demand uncertainty. Full article

Background/Objectives: This paper presents a bi-objective mathematical programming model for sizing hybrid renewable energy systems (HRESs) in isolated mini-grids and compares the optimized solutions with the first-year operation of a real system deployed in La Miel, Panama. Methods: The model minimizes the levelized cost of energy (LCOE) and the expected energy not served (EENS), using an $\epsilon$-constraint approach over a one-year time series (8760 h) of measured demand. For La Miel, the annual demand is 132,578 kWh with a peak load of 28.4 kW. Four configurations are evaluated: (A) diesel-only, (B) photovoltaic (PV)+diesel, (C) PV+batteries, and (D) PV+diesel+batteries. The results are compared with the installed plant (E) including 107 kWp PV, a 40 kVA diesel generator, and lead-acid battery banks (4560 Ah nominal capacity). Results: The optimized hybrid configuration (D) achieves near-zero EENS with an LCOE of 41.4–41.8 cts-USD/kWh, compared to 56.6 cts-USD/kWh for diesel-only. The real system achieves EENS = 0% with LCOE = 48.3 cts-USD/kWh and an annual renewable penetration of 53.2% (up to 68.4% in March 2020), while the optimized case reaches 79.6% on average (up to 95.3% in March). Conclusions: The distinctive contribution of the study is the direct ex ante versus ex post comparison between optimized planning outcomes and the documented first-year operation of the installed system. Operational constraints observed on site (e.g., minimum battery SoC of 60% to comply with voltage quality limits) and demand growth explain part of the LCOE gap between optimized and real performance. Full article

Scenario generation and reduction based on a single variable (e.g., photovoltaic power or load forecasting) is a mainstream approach in current power system planning. However, such methods often overlook the temporal correlation between source and load, which can compromise the credibility of the generated scenarios and lead to suboptimal planning outcomes. To address this issue, this paper proposes an integrated model for microgrid energy storage planning and operation that explicitly considers the joint distribution of source–load scenarios. First, a comprehensive similarity metric is developed by combining dynamic time warping (DTW) distance, slope distance, and source–load correlation distance. An improved K-medoids clustering algorithm is then employed to cluster the joint source–load time series, generating a set of typical scenarios that effectively preserve the coupling characteristics between photovoltaic generation and load demand. Subsequently, a bi-level optimization model is formulated, with energy storage capacity as the primary decision variable. The upper-level planning problem aims to maximize the return on investment (ROI) under energy storage investment constraints, determining the optimal capacity configuration. The lower-level operational problem maximizes the daily net revenue by optimizing the charging and discharging strategies of the energy storage system. Through iterative interaction between the two levels, the model achieves optimal coordination between investment decisions and economic dispatch. Case studies on a campus microgrid demonstrate that the proposed joint scenario generation method effectively captures the temporal correlation between source and load, enhancing both the credibility of the scenarios and the economic rationality of the integrated planning and operation framework. Full article

Accurate energy consumption prediction is fundamental for enhancing range estimation and trip planning in battery electric vehicles (BEVs) under real-world conditions. This study develops a route-level benchmark utilizing 1 Hz data acquired via ECU/OBD-II interfaces (CAN 500 kbps) across ten diverse real-world driving routes. The input feature set comprises vehicle speed, longitudinal acceleration, estimated motor torque, road altitude, and accelerator pedal position. Ground truth energy consumption was derived from battery voltage and current, integrated via the trapezoidal rule. We performed a comparative analysis between five memoryless regressors (FNN, SVR, GPR, QRNN, and Bagged Trees) and three sequence models (LSTM, GRU, and BiLSTM) trained on 20-second temporal windows. The results indicate that the GRU model achieved the highest overall performance (mean RMSE = 0.1142 kWh, R² = 0.9545 and MAE = 0.072 kWh), while Bagged Trees emerged as the most robust static model (mean RMSE = 0.1587 kWh). Temporal models outperformed static ones on routes with high dynamic variability, whereas Bagged Trees excelled in five specific scenarios. These findings provide a controlled within-route benchmark for time-resolved cumulative energy estimation and highlight the need for chronological and cross-route validation before drawing deployment-oriented generalization claims. Full article

Achieving Great Britain’s 2050 net-zero target requires strategic integration of hydrogen into the national energy system. This study evaluates the system-wide impacts of hydrogen blending (0–100%) using a bi-level optimisation framework that combines long-term cooperative investment planning with short-term operational Optimal Power and Gas Flow (OPGF) simulation. The strategic layer models infrastructure investment decisions under a cooperative game-theoretic structure, where system value is allocated among electricity, hydrogen production, and storage technologies using the Shapley-value payoff mechanism. Contrary to traditional centralised cost-minimisation models, our findings demonstrate that a cooperative planning structure identifies superior transition pathways. Comparative results reveal that at 100% hydrogen penetration, the cooperative framework reduces total system CO₂ emissions by 31%, lowers operational costs by 26%, and decreases total electricity supply requirements by 8% relative to centralised planning. Furthermore, the cooperative approach significantly enhances economic resilience, yielding a more robust Net Present Value (NPV) across all blending levels compared to centralised planning, while ensuring project profitability at lower blending thresholds (20%) where traditional models remain loss-making. Simulation results indicate that hydrogen blending up to 20% maintains operational stability with manageable increases in operational cost. Full hydrogen conversion (100%) increases peak electricity supply requirements by approximately 30% relative to low-blending scenarios due to electrolysis-driven load expansion and conversion losses. The findings demonstrate that hydrogen blending represents a viable transitional pathway when supported by integrated infrastructure development and cooperative stakeholder coordination, enabling a more efficient and economically sustainable phased progression towards Great Britain’s 2050 net-zero target. Full article

To deal with the high standards, strong restrictions, and high repeatability that are inside State Grid small-scale infrastructure projects, this research puts forward a norm-driven generative design method, which conquers the low efficiency, compliance dangers, and semantic breakage that are usual in manual modeling. Taking standards such as Q/GDW 11382.3-2015 as the knowledge origin, we construct an ALBERT-BiLSTM-CRF semantic parsing model and change natural-language clauses into executable design restrictions via normative text pre-processing, BIO sequence marking, and rule triplet mapping. Therefore, model training and assessment produce Accuracy, Precision, Recall, and F1 of 98.05%, 95.49%, 95.88%, and 95.59% separately, with 100% precision for logical comparison and conjunction labels; thus, this provides a steady semantic base for the rule base. At the component level, a three-part coding plan and unit module collection are built based on OmniClass and GB/T 51269, which makes semantic consistency and traceability between components and space functions possible. At the system level, a continuous work process is carried out through the Revit API, which covers scheme making, automatic arrangement, and deliverable output. Hence, validation on a real case in a digital operation center for the power system shows that the design time for the third-floor administrative office area was cut from about 20 h to around 4 h, and the first-time solution met all code restrictions, which improves efficiency and compliance in a significant way. The results point out that norm-driven generative design can supply deployable automation and high-quality outputs for small-scale power infrastructure, which provides a sustainable database for digital twins and smart O&M. Full article

This study addresses an integrated airport gate assignment and departure scheduling problem under capacity constraints while explicitly accounting for the operational role of apron resources. A bi-objective mixed integer linear programming model is developed to jointly determine gate or apron assignments and departure times by considering passenger transfer times, taxi operations, runway separation, and schedule deviations. The first objective minimizes a normalized composite measure of passenger transfer burden, taxi penalties, and departure schedule deviation, whereas the second objective minimizes apron usage. The epsilon constraint method is used to generate exact Pareto-efficient solutions. Computational experiments on synthetically generated congested hub airport instances with 20 flights show that increasing physical gate capacity from 3 to 5 improves the average value of Objective 1 from 1.37 to 0.92 and reduces average apron usage from 10.00 to 4.00 flights. In the highlighted 20-flight and 5-gate scenario, increasing apron usage from 3 to 5 assignments reduces the standard deviation of departure time deviations from 8.0 to 7.6 min. The results show that selective apron usage improves system-level schedule stability and that gate capacity and apron flexibility should be evaluated jointly in tactical airport planning. Full article