Search Results (1,377)

Substation-level carbon-emission factors (CEFs) are operationally relevant because substations concentrate transformer losses, auxiliary consumption, and sulfur hexafluoride (SF6) leakage at the interface between transmission and distribution. However, static or annual emission-factor methods average over heterogeneous operating regimes and cannot capture the pronounced non-stationarity of substation CEFs driven by seasonal loads, stochastic maintenance events, cooling-system switching, and extreme weather. To support high-frequency dynamic carbon tracing, dispatch optimization, and audit compliance, this study proposes a physics-guided fusion framework integrating a temporal foundation model, MOIRAI-2, with a Uniform Physics-Informed Neural Network (UPINN). A 15-dimensional physically constrained feature vector is constructed from IEEE C57.91 thermal-circuit equations and ideal-gas state equations, including transformer top-oil/hot-spot temperature, SF6 pressure/density estimation, and oil-forced/air-forced (OFAF) or oil-directed/water-forced (ODWF) cooling status. MOIRAI-2 uses Any-Variate Attention with binary attention bias to model intra-variate temporal dependencies and cross-variate physical couplings, whereas UPINN embeds thermal-balance, SF6 leakage-kinetics, and CEF conservation residuals as soft constraints. An adaptive gating network balances data-driven pattern recognition and physics-driven smoothness across steady-state, extreme-event, and maintenance regimes. Validation on a 220 kV substation dataset achieves a mean absolute error (MAE) of 0.352 gCO₂e/kWh, outperforming random forest (RF), gradient boosting machine (GBM), long short-term memory (LSTM), and a Pure Transformer by 24.0%, 19.1%, 23.0%, and 14.4%, respectively. Ablation studies show that the 15-dimensional physical-feature expansion improves accuracy by 8.8%, whereas physics-loss regularization reduces prediction variance by 37%. UPINN decomposition further indicates that transformer total loss, ambient temperature, and load factor dominate CEF dynamics, and rainfall cooling reduces CEF by 0.04 gCO₂e/kWh per 20 mm increment. The framework provides a physically consistent and intrinsically interpretable basis for dynamic substation carbon accounting and low-carbon operation. Full article

The Boussinesq equations are widely used to model incompressible thermally driven flows, but their numerical simulation becomes more challenging in domains containing fixed or moving rigid obstacles, particularly when the prescribed obstacle motion is uncertain. In this work, we develop a Brinkman-type fictitious-domain formulation for the incompressible Boussinesq system on a fixed computational domain. Brinkman penalization is used to impose the prescribed solid velocity, while a thermal penalty term enforces the temperature inside the immersed bodies. The resulting problem is discretized by a finite element method using skew-symmetric convective forms and grad-div stabilization. The uncertainty in the obstacle motion is introduced through a random oscillation amplitude and treated by a non-intrusive stochastic collocation method based on Gauss–Legendre quadrature. Stability and convergence of the fully discrete scheme are established. Numerical experiments for moving heated and hot–cold obstacles demonstrate the influence of the uncertain oscillation amplitude on the Nusselt number, kinetic energy, mean vorticity, and the spatial distributions of the mean and standard deviation of the solution fields. The results indicate that the proposed method provides a stable numerical framework for simulating incompressible thermally coupled flows with moving immersed obstacles under uncertainty in the prescribed obstacle motion. Full article

Extreme rainfall events significantly affect groundwater systems by altering hydrogeological conditions and accelerating the migration of pollutants, exacerbating the ecological vulnerability of over-exploited areas. However, there is a lack of systematic research on the ecological response mechanisms of microbial communities under extreme rainfall. This study focuses on the over-exploited area in the western North China Plain. Based on 16S rRNA sequencing analysis of 49 groundwater samples collected from the typical alluvial–diluvial plain in the study area impacted by extreme precipitation induced by Typhoon Doksuri, we revealed the effects of rainfall on the taxonomic composition, assembly mechanisms, and environmental drivers of groundwater microbial communities. We found that extreme rainfall influenced the distribution of dominant genera in groundwater and enriched tolerant bacterial communities, further causing a decline in groundwater microbial diversity. Our study confirmed that abundant subcommunities displayed strong microbial collaborations, while rare subcommunities exhibited higher sensitivity and evolved toward opportunistic groups after extreme rainfall. Rainfall influenced the composition of rare taxa, while enhancing the influence of stochastic processes to the assembly of subcommunities. Functionally, confined water favored anaerobic carbon fixation, methanogenesis and dissimilatory sulfate reduction, whereas the Calvin–Benson–Bassham pathway, nitrification, and denitrification were enriched in phreatic water. Additionally, abundant taxa exhibited significantly responses to carbon nutrients in shallow groundwater, whereas rare taxa were governed by mineral ions in deep confined aquifers after rainfall. This study expanded our knowledge of the groundwater microbiome driven by extreme rainfall events and was of great significance to the assessment of ecological stability in the western North China Plain. Full article

This paper proposes a two-stage, reliability-driven multi-objective planning framework for fast electric vehicle charging stations (FCSs) integrated with solar photovoltaic (PV) generation and battery energy storage systems (BESS) in coupled power–transportation networks. The framework simultaneously addresses electrical network constraints, transportation-driven charging demand, and techno-economic trade-offs in high EV penetration scenarios. A benchmark IEEE 69-bus radial distribution system is co-simulated with a 25-node transportation network to realistically capture spatial and temporal interactions between EV mobility and grid operation. To quantify the combined impacts of voltage stability, service continuity, and charging uncertainty, a novel average voltage deviation reliability index (AVDRI) is introduced. Spatially and temporally varying EV charging demand is modeled using a hybrid approach that integrates queuing theory with gravity-based traffic interaction models, enabling a realistic representation of stochastic arrival patterns and route-dependent charging behavior. In the first stage, a multi-objective optimization problem is formulated to determine the optimal locations and charging capacities of FCSs, minimizing system power losses, reliability degradation, and total system cost while maximizing EV serviceability. Multi-objective particle swarm optimization (MOPSO), multi-objective grey wolf optimization (MOGWO), and a proposed hybrid GWOPSO algorithm are comparatively evaluated. In the second stage, a bisection-based sizing strategy is employed to determine the optimal PV and BESS capacities required to mitigate solar intermittency and peak power generation mismatches. Results demonstrate that the proposed hybrid GWOPSO based framework achieves superior convergence characteristics and delivers significant improvements in voltage profile, reliability indices, power loss reduction, and overall techno-economic performance compared to conventional approaches. Full article

We develop a self-contained variational calculus for the coordination of synchronised agents. An agent is modelled as an overdamped (Onsager–Machlup) dynamical system on a five-dimensional manifold $\mathcal{M}=[0,1]\times [0,2{\pi }^2]\times {[0,1]}^3$ whose coordinates are an internal Kuramoto order parameter R, a phase variance ${\sigma }^2$, and three S-entropy coordinates $(S_k,S_t,S_e)$. Crucially, and in contrast to a preliminary version of this work, every construct used here—the S-entropy coordinates, the partition potential, and the governing Lagrangian—is derived within the paper from three stated axioms (bounded phase space, categorical exclusion, finite resolution); the manuscript depends on no external or unpublished result. We prove (i) a triple-equivalence identity establishing the S-entropy coordinates from maximum-entropy counting; (ii) the explicit Euler–Lagrange equations of the agent action, given in full rather than asserted; (iii) the Kuramoto synchronisation bifurcation, with the corrected critical coupling $K_c=2/\left[\pi g\left(0\right)\right]$ (for a Gaussian frequency law, $K_c=\sqrt{2\pi }{\sigma }_{\omega }(2/\pi )\approx 1.596{\sigma }_{\omega }$), replacing an erroneous coefficient in the preliminary version; (iv) a formal, analogy-free definition of partition extinction and an observable-commutation theorem that separates functional from ontological indistinguishability; and (v) a coordination calculus for ensembles. We are explicit that the five coordination bands on R are operational classification cut-offs, not phase transitions: the only genuine transition in the model is the Kuramoto bifurcation. Rather than the self-referential numerical checks of the preliminary version, we report nine external experiments against public data (atomic shell structure, the Kuramoto onset by direct simulation, the Wiedemann–Franz constant across eleven metals, sleep-EEG order parameters from PhysioNet, enzyme kinetics, antidepressant meta-analysis, superconductor transport, and representation-disjoint classifiers on a benchmark dataset). Two of these external tests forced corrections to the theory and are reported as such. We are deliberately narrow about applicability: the S-entropy coordinates measure the shape of an agent’s state distribution, not the content or ecological salience of its information; the model is stochastic (Langevin) but single-scale and on a fixed manifold, so it does not represent the multiscale fluctuation and metastability of real biological collectives; and we exhibit a concrete interpretable Lagrangian and a real-data ensemble instance rather than claim to capture any such collective in full. A Scope and Limitations section delimits the defensible claims. Full article

The large-scale integration of electric vehicles (EVs) can increase load fluctuations, operating costs, and security risks in active distribution networks (ADNs). To address these challenges, this study proposes a multi-objective optimal scheduling strategy based on a Multi-Objective Dung Beetle Optimization (MODBO) algorithm. A Monte Carlo simulation is first used to model stochastic EV charging behavior, followed by representative scenario selection using a minimum-distance criterion. A multi-objective scheduling model is then established considering photovoltaic utilization, ADN operating cost, system net-load variance, and voltage deviation. Case studies on a modified IEEE 33-bus system with 200 EVs show that uncoordinated charging increases the maximum load from 5672 kW to 6321 kW and raises the net-load variance to 4.15. With coordinated scheduling, the proposed method reduces the maximum load to 5672 kW and the variance to 1.29, while achieving 96.69% photovoltaic utilization and an operating cost of CNY 17,573. The results demonstrate that the proposed strategy effectively coordinates EV charging with distributed energy resources, mitigates load fluctuations, and improves the operational performance of ADNs. Full article

OCR for low-resource languages is still held back by the same small number of issues: too little labeled image-text data, too few benchmarks, and thin language-specific tooling. Kurdish Sorani is a particularly awkward case. It is written in a modified Arabic script, runs right to left, and has orthographic habits that standard Arabic OCR engines handle poorly. This paper describes a transformer OCR system for Sorani trained almost entirely on synthetic data, meaning line images rendered on the fly from a text corpus rather than manually transcribed scans. The pipeline has three parts: corpus-driven line synthesis, a deterministic script-aware normalization step based on character-level transliteration, and a TrOCR encoder–decoder recognizer. Text lines are rendered with randomly sampled fonts and sizes, then passed through stochastic augmentation to mimic realistic distortions. The system is evaluated twice. On an in-distribution synthetic set of 200 rendered lines, the best model reaches a character error rate of 0.0434, a word error rate of 0.1246, and 64.0% exact matches. More importantly, on a real-world test set of 19 scanned Kurdish documents (467 lines, 28,468 characters) processed end-to-end through detection and recognition, it reaches a character error rate of 0.0305 and a word error rate of 0.1770, beating both Arabic and Kurdish Tesseract baselines and an existing Kurdish TrOCR model while being considerably smaller than the latter. A controlled ablation, in which eight variants are trained under one shared budget and scored on identical images, then isolates what each design choice contributes. The label space is the largest design effect, and the reason is concrete: the decoder’s pre-trained tokenizer has no representation for seven common Sorani graphemes, which cover 14.7% of the corpus and place a floor under any model trained on native-script labels. Corpus size dominates overall and behaves as a threshold, font diversity helps with diminishing returns, and stochastic augmentation buys robustness at a small cost in in-distribution accuracy. Aligning detected lines against the transcribed ones further shows that line detection contributes under 1% of the reported character error on this material. The broader point, at least for Sorani, is that the synthetic training data and the label space in which the model predicts have to be designed together: a compact recognizer built that way outperforms a substantially larger released Kurdish model on genuine document images. Full article

Aiming at the problems of high energy consumption, high carbon emissions from data centers and the difficulty of renewable energy accommodation in distribution networks driven by rapid growth in computing tasks, this paper proposes a collaborative trading method for data center–power grid–energy storage systems to improve spatiotemporal flexibility. Firstly, an integrated mechanism model including IT equipment, HVAC cooling systems, delay-tolerant batch tasks and UPS energy storage is established to quantify multi-dimensional internal flexible regulation potential. Secondly, an improved k-means algorithm is adopted for scenario reduction of wind–PV outputs, and a stochastic-robust collaborative trading optimization model considering carbon emission cost is constructed. Multiple practical constraints are incorporated, including power balance, power flow limits, nodal voltage bounds, task service latency and state of charge limits of energy storage. An improved particle swarm optimization with premature-convergence indicator is developed to solve this nonlinear, non-convex, mixed-variable problem. Simulations are carried out on a modified IEEE 33-node test system over a 24 h scheduling horizon. Numerical results demonstrate that compared with the conventional demand-response strategy, the proposed method reduces total operational cost by 10.7%, curtails wind–PV abandoned power, and achieves 28.6% peak-shaving ratio for data center load. Monte Carlo repeated experiments indicate that the improved Particle Swarm Optimization (PSO) reaches a 95% feasible solution rate with an average computation time of 26.8 s for day-ahead dispatch, which satisfies practical engineering requirements. Full article

The efficient use of existing road infrastructure has become increasingly important in densely developed urban areas where large-scale roadway expansion is constrained. Variable directional lanes (VDLs) and traffic signal control can reallocate roadway capacity and improve network performance. However, most existing studies optimize lane configurations and signal timing under a fixed route-flow distribution and therefore do not capture the feedback between control decisions and travelers’ route choices. To address this limitation, this study proposes a bi-level framework for coordinating VDLs and traffic signals at multiple intersections. The upper-level model determines the VDL functions and signal-control parameters to minimize total system travel time, while the lower-level static Logit-based stochastic user equilibrium model endogenously redistributes fixed origin–destination (OD) demand among candidate paths. Thus, OD demand remains fixed within each analysis period, whereas the route-flow distribution responds endogenously to the interaction between traffic control and aggregate route-choice responses. A hybrid solution procedure combining the Non-dominated Sorting Genetic Algorithm II and the Method of Successive Averages is used to solve the coupled control–assignment problem. Numerical experiments on a hypothetical network showed that incorporating route-choice feedback improved coordinated VDL–signal control under the tested conditions. In a supplementary comparison with the pre-optimization BPR-based reference scenario, the average route travel time decreased by 7.67–12.84% across the five representative demand periods, including reductions of 12.52% and 12.84% during the morning and evening peak periods, respectively. Microscopic simulation provided an additional numerical consistency check, with average discrepancies of 4.66% before optimization and 4.38% after optimization between the analytical and simulation results. These findings indicate that incorporating aggregate route-choice feedback can support sustainable urban traffic management by reducing travel time and congestion and improving the utilization of existing transportation infrastructure. However, further validation using real-world data and larger-scale networks is required, and environmental benefits should be evaluated explicitly using energy-consumption and emission indicators. Full article

Synthetic load profiles are essential for distribution network planning, protection sizing, and demand-side management studies. However, most generation methods are validated on a single load typology, and their transferability remains unexamined. Two broad paradigms dominate the generation literature: data-driven approaches that learn the demand distribution directly from historical records (generative adversarial networks, diffusion models, Markov-chain generators) and bottom-up, physically motivated approaches that reconstruct demand from the superposition of discrete appliance ON/OFF events. Same-data, same-metric comparisons across these two families for structurally distinct load typologies remain absent from the literature, and this gap is the one this paper addresses. This paper presents a systematic comparison of three stochastic models (a first-order autoregressive (AR(1)) profile, a physically constrained ON/OFF event model, and a nonlinear-least-squares (NLS) calibrated variant) applied to two fundamentally different load typologies measured with a Class A power-quality recorder at 10-min resolution: a single-family residential dwelling (13 days, 1860 samples) and an institutional commercial building (9 days, 1333 samples). Evaluation spans six distributional statistics (mean, standard deviation, and the percentiles $p_{50}$, $p_{90}$, $p_{95}$ and $p_{99}$), the Kolmogorov–Smirnov (KS) statistic, root mean square error (RMSE), and hourly variance profiles. In this two-site study, load typology, not model sophistication, emerges as the dominant factor shaping fit quality. The simple AR(1) reproduces all percentiles within 7% for the near-Gaussian commercial load (skewness 3.47). By contrast, no model reproduces the centre, the dispersion and the upper tail of the highly skewed residential load (skewness 6.56) simultaneously to within 10%. On the residential tail the AR(1) ensemble is the least biased ($p_{99}$: −5.8%) but by far the most dispersed across realizations (CV = 15.3%), whereas the event-based models are more stable but biased, so model choice on skewed loads is a bias–stability trade-off rather than an accuracy ranking. The NLS-calibrated model attains near-exact residential $p_{95}$ reproduction (−0.3%) and commercial upper-tail errors below 2%, at a computational cost roughly three orders of magnitude above the AR(1). The dynamic characterization of demand obtained from these models, including the magnitude and frequency of the detected events, provides elements that may be of interest for the sizing and operation of photovoltaic systems in the context considered. Based on these two cases, a preliminary recommendation matrix mapping models to engineering applications is proposed, motivating typology-specific model selection rather than universal approaches; broader validation on additional sites is identified as future work. Full article

The transition to electric vehicles (EVs) is widely promoted as a strategy to reduce greenhouse gas emissions from transportation. However, the environmental benefits of EVs depend strongly on the electricity mix used for charging. If charging is predominantly supplied by fossil-fuel-based grid electricity, the resulting carbon dioxide (CO₂) emissions may remain substantial. Renewable-powered charging stations offer a solution, yet their spatial distribution creates a trade-off: if they are located further from the driver than their grid counterparts, then more CO₂ might be emitted along the way. This study developed and validated a framework for quantifying this trade-off. A mathematical model was first constructed, in which charging stations were spatially distributed following a Poisson process, and renewable availability was described by a Gaussian distribution. Emissions were measured in terms of mCO₂, the mass of CO₂ emitted by driving and charging. The model was then tested by comparing it with a MATLAB-based computer simulation incorporating stochastic station distributions and vehicle energy states. Both approaches identified a distinct minimum in the emission–distance curve. The mathematical model located the optimum at 3.737 km, while the computer simulation confirmed the result within an interval of 2.844–4.266 km. These findings prove the existence of an optimal charging distance. The results highlight the value of considering various parameters during EV infrastructure planning and offer practical guidance to reduce life-cycle emissions in sustainable mobility systems. Full article

Long-term distribution planning requires weather sequences that capture not only changes in temperature and precipitation, but also the dependence and persistence among heat, humidity, wind, solar radiation, and rainfall. This paper develops multivariate weather scenario generation for Shaanxi Province, China, using daily data from six NEX-GDDP-CMIP6 models for 1985–2014 and 2031–2060 under SSP2-4.5 and SSP5-8.5 at four locations. Two generators are compared: a vector autoregressive model with seven-day residual blocks, and a season-conditioned multivariate nearest-neighbor analog. The comparison leaves out complete five-year periods within 24 climate-model–location units and tests whether the future-minus-historical changes in 16 weather and compound-event indices are preserved. The two methods each obtain the lower overall loss in 12 units. The vector autoregressive method better preserves most marginal and correlation signals, whereas the nearest-neighbor method better preserves RX3day, hot–dry–low-wind days, dry-spell duration, and three-day compound stress. Both are retained to generate 960 30-year paths, which drive a fixed IEEE 33-node resilience example showing that generator differences propagate to demand, renewable availability, repair duration, and unserved energy. The resulting conditional scenario sets provide an auditable weather basis for climate-resilient and sustainable long-term planning of distribution systems, which is a prerequisite for a sustainable energy transition under a changing climate. Full article

Large-scale public events generate concentrated passenger demand during egress periods, often overwhelming urban transit systems. This paper proposes a multi-modal evacuation framework that coordinates in-service buses temporarily diverted from existing lines and dedicated shuttle vehicles pre-positioned at depots. The problem is formulated as a two-layer stochastic optimization under travel time uncertainty: the upper layer determines pre-event shuttle fleet sizing, while the lower layer makes real-time dispatching decisions for both modes. We propose an Uncertainty-Aware Reinforcement Learning framework with Categorical DQN (UARL-CD) that learns a robust dispatching policy through a reward function aligned with the lower-level objective, explicitly accounting for travel time uncertainty via distributional value representation and stochastic training, with an action masking mechanism enforcing operational constraints. Simulation experiments based on a realistic stadium evacuation scenario demonstrate that the proposed framework significantly outperforms deterministic optimization and rule-based strategies, achieving a 31.6% reduction in evacuation completion time and a 48.4% reduction in average passenger waiting time compared to shuttles alone, while maintaining robustness to travel time uncertainty with only 4.0% performance degradation and online decisions executed within the 2-min decision interval. Full article

Multi-product production systems require effective coordination among production, intermediate storage, downstream processing, and customer demand, particularly under stochastic operating conditions and finite capacity. Unlike existing analytical studies that mainly examine either single-product push–pull systems or multi-product manufacturing systems separately, this work develops an exact CTMC framework that jointly captures product variety, shared buffering, downstream parallelization, sequence-dependent setups, and machine unreliability. The system comprises an unreliable upstream machine with sequence-dependent setup changes, a finite intermediate buffer, a distribution center modeled as a pooled processing resource with (M) identical reliable channels, and dedicated finished-goods buffers serving product-specific demand. A high-dimensional continuous-time Markov chain is formulated, and a systematic algorithm is developed to construct the infinitesimal generator matrix and compute steady-state performance measures. Numerical experiments examine intermediate buffer capacity, downstream processing capacity, priority rules, and upstream machine reliability. Increasing buffer capacity from 0 to 20 increases total throughput from 0.5937 to 1.0988, whereas further expansion to 100 yields only 1.1606, while average work-in-process reaches 19.8524. Downstream capacity exhibits similar diminishing performance gains, while priority rules and machine reliability affect product-level and overall throughput. These findings highlight throughput–inventory trade-offs and demonstrate the framework’s applicability for evaluating alternative configurations. Full article

The frequent occurrence of extreme high-temperature events has significantly affected the operating characteristics and distributed photovoltaic (PV) hosting capacity of distribution networks. However, existing hosting capacity assessment methods rarely consider the accumulated heat effect caused by sustained high temperatures. To address this issue, this paper proposes a coordinated planning method for enhancing distributed PV hosting capacity under extreme high-temperature scenarios. First, an accumulated heat load model is developed to characterize the temporal cumulative influence of sustained high temperatures on temperature-sensitive loads. Meanwhile, the uncertainties associated with PV output fluctuations and load demand variations are considered to represent the stochastic characteristics of source-side generation and load-side consumption. Subsequently, a multi-objective source–network–load coordinated planning model is established to maximize distributed PV hosting capacity while minimizing the hosting capacity enhancement cost. A multi-objective artificial bee colony (MO-ABC) algorithm incorporating Sobol sequence-based quasi-Monte Carlo sampling (Sobol-MC) and a constraint domination-based constraint handling strategy are further developed to solve the proposed model efficiently. Simulation results on the modified IEEE 33-bus distribution system show that the proposed method increases distributed PV hosting capacity by 69.52% under extreme high-temperature scenarios through coordinated optimization of PV inverter reactive power control, VAR compensation, and Incentive-based Demand Response (IDR). Full article