Energies doi: 10.3390/en19225496

Authors: Xin Lei Weixin Pang Qiang Fu Yuhua Ma Xiaohan Zhang Huiyun Wen Yang Ge Lu Liu

In recent years, the co-production of natural gas hydrate and shallow gas has been regarded as a promising technical solution to tackle the aforementioned challenges; however, the influence of well-type adaptability of single-wellbore in co-production has been insufficiently studied, hindering the optimization of co-production schemes. In this paper, typical symbiotic reservoirs containing natural gas hydrate and shallow gas in the Qiongdongnan Basin in the South China Sea were selected as research targets. Large-scale thermo-hydro-chemical multi-field coupled numerical models were established to investigate the effects of well types and wellbore gas flow characteristics on the production performance of vertical, horizontal, and multilateral wells under single-wellbore and separate-wellbore co-production modes. Gas production performance under different well configurations and various scales of underlying gas accumulations was investigated, along with the synergistic pressure evolution mechanism and interlayer gas channeling patterns during co-production. As revealed by simulation results, hydrate gas productivity reaches 5.56 times that of vertical wells with horizontal well deployment, while the volume of dissociated gas exceeds 140 times the baseline value. This pronounced dissociation-production mismatch is likely caused by pore gas trapping, which is in turn governed by the adopted relative permeability. Wellbore configuration exerts a significant influence on interlayer gas channeling during co-production. The pressure drop required for hydrate dissociation is far greater than that for shallow gas recovery in co-production via a single wellbore; reservoir pressure differential within the first day triggers shallow gas channeling into the hydrate layer through the wellbore and induces temporary productivity loss, whereas separate wellbores for hydrate and shallow gas extraction can effectively inhibit such wellbore gas channeling. Well locations and perforation intervals are suggested to be arranged in a differentiated manner according to the scale ratio of hydrate reservoirs to underlying gas formations and physical properties of interlayers, so that coordinated production enhancement can be realized and safe, efficient, and unified exploitation can be achieved. This study provides theoretical support for well configuration design of hydrate-shallow gas co-production.

Energies doi: 10.3390/en19225495

Authors: Maanasa Bhat Linh Le Khanh Hoang Sili Deng

Growing renewable energy deployment requires large-scale, long-duration energy storage and transport. Chemical energy carriers can meet this demand by enabling easy storage and transport of renewable electricity. While liquid-phase energy carriers such as liquid hydrogen are more commonly used, solid-state carriers offer advantages with regard to safety and density. Here, we evaluate aluminum as a solid-state hydrogen carrier in a closed-loop Al/Al2O3 system (Al). This is compared to liquid hydrogen (LH2) and liquid organic hydrogen carrier (LOHC) cases through techno-economic analysis (TEA) and environmental assessment. In the Al/Al2O3 closed-loop system, metallic Al is transported to the point of use and reacted with water to generate hydrogen. The resultant Al2O3 is returned to the point of production for regeneration. Under baseline assumptions of carrier production in the United States and hydrogen delivered to Japan, the levelized costs of hydrogen (LCOH) are $7.44, $8.32, and $9.09/kg-H2 for LH2, LOHC, and Al, respectively. Despite similar levelized costs, the pathways exhibit distinctly different cost structures. Sensitivity analysis shows that reducing the energy demand, capital cost, and operating cost of Al regeneration could lower its LCOH to $7.65/kg-H2. CO2 emissions are estimated as 5.04, 14.91, and 12.9 kg-CO2/kg-H2 for LH2, LOHC, and Al, respectively, with Al achieving 15% lower baseline emissions than LOHC when low-carbon electricity is used for aluminum regeneration. Including direct CO2 emissions from the carbon anode in the Al regeneration process increases overall emissions to 23.94 kg-CO2/kg-H2. These results demonstrate that the solid-state carrier Al shifts the primary burden of hydrogen transportation from downstream storage and logistics to a concentrated, improvable upstream regeneration process, providing a potentially competitive pathway for large-scale, long-distance hydrogen transportation.

Energies doi: 10.3390/en19225493

Authors: Liu Li Wang Chen

Typhoon disasters expand single-section faults into multiple simultaneous faults. Conventional switching functions construct expected nodal states under unidirectional main-source dominance, neglecting the blocking of downstream supply paths and leaving the shared-branch amplitude relation undecided. An adaptive fault section localization method is proposed. An effective contribution coefficient of the main source accounts for supply-path blocking, and a direction discriminant built from the upstream path impedance and distributed-generation capacity converts the shared-branch expected state into an amplitude comparison, forming an adaptive dynamic switching function. Three measurement types are fused into a model solved by enumeration over a pruned candidate set. On the 123-node feeder, 500 scenarios per multiplicity with one fixed seed give accuracies of 98.0%, 85.0%, 70.6% and 50.0% for single to quadruple faults, against 99.4%, 84.4%, 61.6% and 39.6% for the fixed-value function, an advantage of 9.0 and 10.4 percentage points for triple and quadruple faults. Accuracy presumes dispersed damage and reliable outage reporting, falling to 30.0% under clustered triple faults. The blocking-relation share rises from 0% to 57.6%, and under non-metallic faults the direction discriminant contributes independently of the blocking term.

Energies doi: 10.3390/en19225494

Authors: Zhe Xu Baohang Zhang Lin Yang Fenggang Yuan Han Zhang

Previous studies have applied particle swarm optimization to coupled offshore wind farm layout and cable design and used geometry-guided genetic search to preserve spatial configurations during crossover. Transferring a successful relocation between layouts requires reassessing its wake and cable effects in the receiving layout. We propose cross-swarm recombination particle swarm optimization (CSRPSO) for offshore layouts with embedded capacity-constrained routing under a levelized cost of energy (LCOE) objective. Two swarms use wake interactions and cable connections to guide turbine relocations. Successful relocations are reconstructed in the other swarm’s personal-best layouts and compared with candidates generated from current layouts using complete LCOE evaluations. Six algorithms were compared across eight wind-farm cases over 30 repetitions per case, using matched seeds and initial populations, populations of 30 and 15,000 complete evaluations including initialization. CSRPSO achieved the lowest mean and median LCOE in every case, reducing mean LCOE by 0.041–0.697% versus adaptive replacement strategy-incorporated genetic learning particle swarm optimization (AGPSO), the closest baseline. At 10,000 evaluations, one-third below budget, its mean LCOE was below AGPSO’s final mean in every case. Mean annual energy production exceeded AGPSO’s in all eight cases, and mean cable expenditure was lower in six.

Energies doi: 10.3390/en19225490

Authors: Mingfang Lu Junbin Gao

When pumped storage participates in a joint energy and frequency regulation market, competitors’ cost parameters and bidding strategies are not fully observable. The plant must also coordinate pumping and generation schedules across periods and allocate limited capacity between the two markets. The resulting high-dimensional continuous action space and strong physical constraints make bidding optimization under incomplete information sample-inefficient and difficult to converge. This paper therefore proposes a pumped storage bidding strategy that incorporates prior information from forecast electricity prices. First, a directed graph neural network price forecasting model driven by multi-scenario clustering is developed. Historical price data are clustered to decouple operating scenarios, and the directed dependencies between price and its drivers, together with multi-scale temporal features, are learned separately for each scenario to improve day-ahead price forecasting accuracy. Second, the forecast prices are converted into a pumping/generation state prior and a bidding price benchmark prior, thereby constraining the bidding action space. Finally, Bayesian multi-agent reinforcement learning is used to solve the bidding and clearing model. The case study results show that the proposed forecasting model improves the coefficient of determination (R2) from 0.787 to 0.868. Compared with bidding strategies that consider only a single prior, the proposed strategy increases the net revenue of pumped hydro storage by 18.3%.

Energies doi: 10.3390/en19225491

Authors: Bartłomiej Gaweł Andrzej Paliński

Long-term energy forecasting in a dynamically changing environment requires combining information contained in historical data with expert knowledge about structural, technological and policy changes. However, expert scenarios often disagree, and their relative credibility is difficult to assess. We propose a framework for combining a statistical forecast with expert scenarios while explicitly retaining uncertainty about both scenario outcomes and their relative importance. We examine this problem using gas demand in Poland up to 2040. A statistical forecast based on historical data provides a transparent reference, while expert scenarios developed by three organizations represent alternative structural developments that may not be captured by historical extrapolation alone. Uncertainty around the statistical forecast is represented by a probability distribution, while expert deviations from it are represented by triangular fuzzy numbers. We fix the total weight assigned to expert alternatives but leave its allocation among them unspecified. The combined result is a probability box, which gives lower and upper probability bounds. In the current case study, all expert scenarios imply higher demand than the statistical reference in 2030, while their demand levels differ strongly in 2040. In 2035, the midpoint of the combined expected-value interval remains close to the statistical reference despite disagreement among scenarios. Giving more weight to expert alternatives raises this midpoint in 2030 but lowers it in 2040. The results show that combining statistical forecasts with expert knowledge can reveal the dependence of long-term conclusions on structural assumptions and scenario credibility, particularly when historical relationships may not remain stable. The proposed framework therefore supports a transparent synthesis of statistical evidence and expert judgement without requiring a single precise weighting of alternative future scenarios.

Energies doi: 10.3390/en19225492

Authors: Penghan Li Ying Wang Wei Wang Yunpeng Zhou Xiaojie Shi Dan Wu

To strengthen voltage and frequency stability in islanded renewable hubs without traditional generator support, grid-forming (GFM) converters and synchronous condensers (SCs) can be configured jointly to provide support capabilities. However, their underlying physical principles and costs are different. A joint allocation of these two equipments can achieve complementary dynamic supports with a minimum cost. This study formulates a lifecycle-cost minimization problem over continuous GFM and SC capacities for an islanded renewable hub connected to the grid via an LCC-HVDC link. A differential-algebraic equation (DAE) model captures the entire dynamic response from the pre-fault stable condition to the post-fault transients. To solve the continuous trajectory-constraint planning problem efficiently, a trust-region-based sequential linear programming (SLP) algorithm is proposed, restricting local steps through trajectory-constraint approximations. Meanwhile, a nonlinear dynamic simulation validates trial designs. Independent coarse, refined grid scans and dynamic responses across different scenarios confirm that the continuous search accurately identifies the dynamic feasibility boundary and achieves the minimum annual cost. The framework provides an efficient methodology for the joint GFM–SC planning problem, respecting both the dynamical and the steady state constraints.

Energies doi: 10.3390/en19225489

Authors: Qiang Gao Yingbo Wang Zhen Yin Shaoliang Xu Xiaogang Lin Jinghao Li Jinhai Liu

Accurate day-ahead electricity consumption forecasting is important for electricity market trading and cost management. In this paper, a decomposition-driven dual-branch network is proposed for this task. First, the original consumption sequence is decomposed into trend and residual components, which are modeled by a linear forecasting branch and an improved ModernTCN branch, respectively. In the residual branch, gated multi-scale blocks are employed to extract temporal patterns at different scales, and the resulting features are adaptively fused using input-dependent weights. The proposed method is evaluated on real-world electricity consumption data from Ansteel Mining. It achieves the highest mean number of qualified test days among the compared methods. Across five independent runs, an average of 178.8 qualified days out of 182 test days is obtained, with a standard deviation of 1.1 days. In ablation experiments, the increase in qualified days is mainly attributed to trend decomposition, while average errors and their variability across runs are reduced by the gated multi-scale block. On a public industrial load dataset, the proposed method achieves the lowest mean absolute error among the compared methods. The effectiveness of the proposed method for day-ahead industrial electricity consumption forecasting is demonstrated by these results.

Energies doi: 10.3390/en19225488

Authors: Priyanshu Kumar Purnima Verma Anubhav Kumar Pandey Gopisetti Manikanta Priya Kumari Renukaprasad Gangaraju Akshay Kumar Saha

This paper investigates the influence of modelling approaches on the control performance of a DC–DC boost converter using state-space averaging (SSAT) and closed-loop identification (CLIT) within a common two-degree-of-freedom internal model control (T-DOF IMC) framework. The SSAT model is analytically derived using nominal converter parameters, whereas CLIT models are obtained by fitting transfer functions to experimentally measured closed-loop reference-to-output responses. Among the five identified models, the one providing the best response fit and closest transient characteristics to the desired response is selected for controller design. Both controllers are tuned using a frequency-domain procedure based on gain and phase margin specifications. The proposed designs are experimentally evaluated under set-point changes, input-voltage variations, duty-cycle disturbances, and load-resistance changes. The results show that the CLIT-based design generally provides faster transient recovery and a higher multiplicative gain-uncertainty bound (approximately 35%) than the SSAT-based design (approximately 5%) under the adopted small-gain criterion. However, its performance is not uniformly superior across all experimental conditions and integral performance indices. These findings demonstrate the influence of model selection on closed-loop performance and highlight CLIT as a practical alternative to conventional analytical modelling for DC–DC boost converter control.

Energies doi: 10.3390/en19225487

Authors: Andreas T. van Nimwegen Jos M. C. van ‘t Westende Hajo P. Pereboom Hadi Dashtaki Hesari Pejman Shoeibi Omrani

Drag reducing agents (DRAs) can reduce the pressure losses in turbulent pipeflow in district heating and cooling networks (DHCNs), which can reduce the operational costs in current networks and allow for reduced investment, installation, and construction costs for future DHCNs. However, their hydraulic effectiveness, performance and stability at DHCN operating conditions remain insufficiently understood. In this work, measurements are performed at relevant conditions for DHCNs in a dedicated flow loop for different DRAs at temperatures of 20–115 °C. The results demonstrate that the DRAs are able to reduce the pressure drop under representative DHCNs operating temperatures by 50–80%. Furthermore, the tested DRAs exhibit stable drag reduction performance during week-long experiments and maintained chemical stability over month-long testing periods. However, the governing mechanisms for drag reduction also result in decreased heat transfer performance in heat exchanger experiments, highlighting a trade-off between hydraulic efficiency and thermal performance. These findings provide experimental evidence supporting the feasibility of DRAs for DHCN applications and form an important step towards future field trials and demonstrations.

Energies doi: 10.3390/en19225486

Authors: Yu Fang Jinlou Peng Yan Ping Xiaoyan Liu Jianping Yan Kai Zhou

The overall performance, pollutant emissions, and operational reliability of gas turbines are strongly influenced by combustor design. However, combustor design remains challenging because of the strong coupling among turbulent flow, fuel–air mixing, chemical reactions, heat transfer and cooling, and combustion instabilities. Conventional design approaches are often constrained by long development cycles, high computational costs, and difficulties in coordinated multi-objective optimization. In recent years, artificial intelligence (AI) techniques have emerged as promising tools for addressing these limitations. This review establishes a process-oriented framework based on the design requirements of gas turbine combustors and the methodological characteristics of AI. Recent advances are systematically reviewed in six directions: fast performance prediction and sensitivity analysis, combustion flow-field reconstruction and multiphysics representation, intelligent turbulent combustion simulation and chemical-reaction reduction, surrogate-model-driven multi-objective optimization design, combustion-state recognition and instability warning, and online state estimation, digital twins, and closed-loop development. The review indicates that AI is moving beyond task-specific prediction, reconstruction, diagnosis, and optimization tools toward a more integrated role in combustor design, simulation, monitoring, and decision support. However, engineering deployment remains limited by scarce high-quality experimental data, insufficient cross-condition and cross-configuration generalization, limited interpretability, weak enforcement of physical constraints, and incomplete uncertainty and trust-boundary assessment. Future research should prioritize high-quality multi-source data systems, physics-constrained and uncertainty-aware learning, interpretable and verifiable optimization, and closed-loop digital-twin frameworks, so that AI-assisted combustor design can evolve from local algorithmic support into a reliable capability for engineering research and development.

Energies doi: 10.3390/en19225485

Authors: Rafael A. Garcia Tinoco Maximiliano Martinez Jonathan Muñoz Tabora Hermen Octavio Alvarenga Pérez

Large battery energy storage systems (BESSs) are often sized assuming that their operation does not affect electricity prices. For utility-scale assets this assumption fails: their bids change which generators are committed and, therefore, the prices from which the battery earns its revenue. This paper proposes a bilevel framework that jointly determines the power rating, energy capacity and hourly bids of a price-maker BESS that maximize the project net present value (NPV). The upper level is solved with particle swarm optimization (PSO), while the lower level reproduces the day-ahead market clearing as a security-constrained unit commitment (SCUC) that retains generator commitment decisions, startup costs and network constraints, instead of the convex approximations required by existing bilevel sizing models. The battery participates in energy, Reg-Up and Reg-Down markets, and its degradation is evaluated with a non-linear Rainflow-based model. Simulations on the IEEE 3-bus and 57-bus systems with CAISO-based profiles yield optimal energy-to-power ratios of 1.9–2.45 h, below the common 4-h design, because reserve services provide most of the revenue. A sensitivity analysis on renewable penetration reveals a non-monotonic relationship between renewable share and the optimal power rating, driven by changes in reserve scarcity.

Energies doi: 10.3390/en19225484

Authors: Long Yu Xiaohu Yan Qiang Lu Jichao Xue Zhongshun Teng Shixiang Yang Jietuo Luo Ye Bai Zhonghao Wang

Thermal inertia in district heating networks can expand the electrical flexibility of combined heat and power plants. Realizing this value requires matching sustainable modulation with its auxiliary electricity costs over the requested service duration. This study couples a dynamic heat network model and flow control with system resistance, manufacturer pump curves and minimum frequency constraints. A service-to-electricity price ratio identifies actions that cover the incremental pumping costs. The case study shows that disturbance propagation and compensating flow change mean pumping demand across durations at the same modulation amplitude. Under ideal variable-speed operation, a heat-to-power coefficient of 0.20 MWe/MWth and a 20–26 ∘C indoor-temperature filter, 94 of 164 eligible amplitude–baseline groups change coverage classification across durations at a price ratio of 0.20. Reusing the 4 h mean pumping power incorrectly classifies 26.8% of matched 8–24 h windows as cost-covered. Building parameter and controller response sensitivities quantify shifts in cost thresholds and thermal admission. The largest feasible amplitude and highest unit pumping cost can also occur at different operating points, depending on the temperature limits. These findings support joint duration–capacity–cost screening of individual flexibility actions under specified operating and service assumptions.

Energies doi: 10.3390/en19225483

Authors: Cem Haydaroğlu

PEM fuel cells and electrolyzers can support frequency in low-inertia microgrids, but a switched-off unit needs time to start, every start–stop cycle degrades the stack, and a hydrogen chain with about 30% round-trip efficiency drains its tank even at zero net energy exchange. This paper proposes a degradation- and inventory-aware hierarchical frequency controller: a rate-aware tube model predictive control (MPC) power layer; a commitment layer that applies the classical ski-rental rule with break-even times derived from stack degradation and standby costs (commitment cost bounded by twice the offline optimum plus one start, excluding frequency performance); and a hydrogen-inventory layer that shifts the conventional-unit schedule to hold the tank level. On an electrochemical plant with a realistic information structure and 48 unseen test cases, the controller matches the frequency performance of a perfect-foresight commitment at 5% higher cost and reduces the integral absolute frequency error by 26% versus fuel-cell-only standby, 19% versus a tuned hold-time rule and 28–36% versus tuned PI, PID, fractional-order PID, fuzzy and MPC controllers. The gain grows with the electrolyzer start-up time and persists under random plant–model mismatch. Against a hold-time rule tuned on hour-long profiles, however, the frequency advantage is minimal; the remaining benefits are lower cost and the absence of cold electrolyzer onsets. Over ten days, the inventory layer keeps the tank within 1.09–1.29 kg; without it, the tank empties after 17–20 h.

Energies doi: 10.3390/en19225482

Authors: Inga Zicmane Aleksandrs Sahnovskis Roman Petrichenko Sergejs Kovalenko Gatis Junghans

The paper presents an integrated approach to Optimal Power Flow (OPF) in transmission systems with a high share of renewable energy sources (RESs) supported by energy storage systems (ESSs). Unlike conventional approaches, the proposed integrated evaluation framework minimizes operating cost subject to technical (network and ESS) constraints, while separately reporting RES utilization/curtailment and CO2 outcomes as post-solve indicators. The model was implemented as a browser-based JavaScript application using a custom Big-M Simplex linear programming solver and tested on a 24 h horizon under multiple scenarios representing conditions informed by Baltic/Latvian load, generation, and price profiles, applied to a modified IEEE 9-bus test network, including standard operation, high price volatility, high RES penetration, and a transmission line outage (N-1 criterion). The results demonstrate that ESS integration together with RES usage maximization reduces total operating costs by up to 10.2% compared to the baseline strategy without ESS and optimization (about 4.7 percentage points of this saving come from the lower minimum output of the conventional unit alone), increases RES share by about 5 percentage points, and decreases RES curtailment to 0%, under a terminal state of charge condition that prevents the ESS from ending the horizon more depleted than it started (imposed in HiGHS runs of the same linear program). Under high RES scenarios, ESSs become essential for balancing variability, while in volatile markets they support stable dispatching within the model. A sensitivity analysis highlights the dominant influence of market price fluctuations, with smaller effects of ESS capacity and RES share on total operating cost. The findings indicate that, within the transportation-type network model used, the proposed approach improves economic efficiency and environmental performance, offering a practical decision-support tool for transmission system operators and energy planners.

Energies doi: 10.3390/en19225481

Authors: Xiaofan Xie Duo Xu Bing Zeng Hui Liao Yunmin Xie Mengyao Liang Zhen Yao Tianjiao Cao Quanyun Zheng

The area of the corona discharge region on an insulator is reported as a diagnostic indicator, yet it is almost always obtained by thresholding a segmentation probability map at the conventional 0.5 level and is reported without any uncertainty statement. Using 1600 solar-blind ultraviolet frames, we show that this practice is the dominant error source. Selecting the operating point on a validation split reduces the median relative area error from 20.8% to 3.3% and, unexpectedly, increases Dice from 0.874 to 0.903; a 7.76 M-parameter U-Net then reaches 3.26 ± 0.39% area error, within 0.07 percentage points of an oracle threshold. Dice is not a valid proxy for area: a two-pixel registration shift changes the area by 0.03% but Dice by 0.159, because the annotations are fragmented into about 31 sub-ten-pixel blobs. Split conformal prediction yields intervals whose empirical coverage (93.7–97.8%) meets or exceeds the nominal level even when checkpoint selection, threshold selection and calibration use disjoint data, with a domain of validity of roughly −5% to +10% gain deviation. A GUM budget attributes 15.7% relative standard uncertainty to annotation-boundary ambiguity, against 2.1% for the model and acquisition combined, so the measurement is annotation-limited. The protocol makes the reported image-domain area fraction an interpretable measurand with a stated uncertainty.

Energies doi: 10.3390/en19225480

Authors: Xuan Wei Cunshi Tang Huihui Qiao Yongchao Wang Mingda Xu

Existing deep learning-based fault diagnosis models for wind turbine planetary gearboxes cannot cope with the problem of poor domain generalization ability caused by differences in data distribution under different operating conditions. To address this limitation, a domain generalization fault diagnosis method based on multi-level feature alignment (DGFD-MLFA) for planetary gearbox is proposed. Firstly, the multi-level feature extraction (MLFE) module is designed, which includes a convolutional denoising autoencoder, a multi-scale separable convolution, and a transformer encoder. The MLFE module can extract both local and global features while denoising. Secondly, the multi-level feature alignment (MLFA) module includes the Local Maximum Mean Discrepancy (LMMD) and the Maximum Mean Discrepancy (MMD) is designed to align local and global features respectively, allowing the model to ultimately extract domain-invariant features and enhancing its fault diagnosis capability under unknown operating conditions. Experiment results on nine domain generalization fault diagnosis tasks based on the wind turbine planetary gearbox dataset, including ablation studies and comparisons with different feature alignment methods, show that the proposed model significantly improves fault diagnosis accuracy under both interpolation and extrapolation domain generalization tasks.

Energies doi: 10.3390/en19225479

Authors: Sadi Serdar Güneli Muhammet Ali Arserim

Grid disconnection in converter-dominated systems is difficult to distinguish from faults and load changes with comparable or larger transients. This paper investigates an observer–energy event-discrimination framework built on a role-separated architecture for a grid-connected virtual synchronous generator (VSG): the VSG keeps the voltage-forming role, so its power balance is available internally, while a phase-locked loop (PLL) acts only as an external, communication-free observer at the point of common coupling. Three signatures are examined: PLL-frequency deviation, VSG–PLL phase deviation, and the power-balance signature. Phase deviation settles at a nonzero value after a permanent grid-connected load step, and is excluded from the decision rule. In steady state, the swing equation and damping approximately relate the two thresholds (0.10 p.u. ≈ 0.1006 Hz); both must be exceeded for 300 ms. For a 30 kW local load, simulated disconnection is detected 0.553 s after the grid opens (threshold crossing plus dwell), and the fault and grid-connected load step raise no false alarm; no speed advantage is claimed. The deficit-side non-detection-zone edge lies near 26.6 kW under reference conditions and varies with reactive-power mismatch, load model, and power reference. Offline one-at-a-time variations of parameters, grid strength, X/R, and voltage noise preserved 30 kW detection without false alarms.

Energies doi: 10.3390/en19225475

Authors: Gergo Varhegyi Mutasim Nour

Techno-economic planning identifies renewable generation, storage, and flexibility portfolios but cannot establish their AC-network feasibility or dynamic security. This study develops an automated Python framework coupling PLEXOS planning with steady-state and RMS dynamic assessment in PSS/E. Critical hourly operating conditions are transferred into network cases, tested against disturbances, and evaluated using voltage, frequency nadir, and rate of change of frequency (RoCoF). The framework is demonstrated across four systems (23–505 nodes) and successive renewable energy sources (RES) penetration stages. Across 144–256 screened PSS/E cases per system, the largest individual violation category affected 12.24–41.67% of cases, establishing a conservative minimum affected-case share and revealing technical limitations not represented in the upstream planning assessment, including AC-feasibility, voltage-recovery, and frequency-security limitations. Both benchmark systems exhibited their highest RoCoF near 75% RES rather than at maximum penetration; the IEEE 39-node case reached 1.87 Hz/s, while its lowest nadir occurred at 50% RES. Halving available fast frequency response increased peak RoCoF by 0.05–0.11 Hz/s and lowered nadir by 0.05–0.50 Hz. These findings show that critical transition stages depend on synchronous inertia, network flows, operating conditions, and flexibility deployment. The framework also identifies actionable requirements, including regional voltage support and transmission reinforcement, for subsequent long-term planning.

Energies doi: 10.3390/en19225476

Authors: Meng Zhu Yang Liu Di An Xin Hui Kexin Li Mo Zhou Yibo Zhou Xiaotian Zhang Chuanliang Cao Ziyun Wang

With the large-scale integration of renewable-energy generating units, power-system inertia continues to decline and becomes unevenly distributed in space. Consequently, frequency responses differ markedly among regions following a disturbance, and assessment based on a single center-of-inertia (COI) frequency may conceal local frequency-security risks. This paper proposes a regional inertia assessment and minimum-inertia calculation method based on linearized analysis. First, a generalized relative correlation gain matrix is constructed from a linearized multi-machine model to quantify the dynamic correlation of generator frequency responses. Spectral clustering and the Calinski–Harabasz (CH) index are then used to determine the system partition. Next, synchronous-machine inertia, load-voltage characteristics, virtual inertia, and primary frequency control are considered in stages. The minimum regional inertia requirements are calculated under rate-of-change-of-frequency (RoCoF) and frequency-nadir constraints, and the larger value is taken as the regional inertia threshold. Simulations on the IEEE 10-machine 39-bus system show that the proposed partitions capture regional differences in frequency response and that the estimated values agree closely with the actual aggregated inertia after synchronous machines are replaced. When inertia is configured according to the calculated thresholds, the RoCoF and frequency nadir in every region satisfy the security constraints. Further analysis shows that the minimum inertia requirements under both constraints increase with disturbance magnitude. At a stringent RoCoF limit, the requirement is dominated by the RoCoF constraint; as the limit is relaxed, the frequency-nadir constraint progressively becomes dominant. The proposed method provides a reference for regional inertia assessment and frequency-security control in power systems with high renewable penetration.

Energies doi: 10.3390/en19225478

Authors: Damian Frej Marcin Ślęzak Marek Jaśkiewicz Ewa Dębicka Miloš Poliak Rafał Chaba Przemysław Kubiak Dariusz Tarnapowicz Tymoteusz Zaleski

Urban digitalization requires site-specific evidence to determine where intelligent sensing, edge processing, warning actuation and energy management functions are justified. This repeated cross-sectional field study quantified changes in smartphone-related behaviour and electrified micromobility at an unsignalized roundabout in Kielce, Poland, and translated them into technical requirements and an energy budget for a privacy-preserving, energy-aware crossing infrastructure. Two observers recorded 27,487 crossing-area users in 1728 arm-specific five-minute intervals during 18 standardized two-hour sessions in 2023–2025. Any phone use increased from 41.87% to 48.63% (difference 6.76 percentage points, 95% CI 5.31–8.20; Cohen’s h = 0.14), and a combined electric micromobility exposure indicator increased from 11.84% to 22.39% (h = 0.28), while bicycle prevalence remained near 8%. The phone-use trend was robust to the clustering of observations: session-clustered logistic regression, a random-intercept model and an exact stratified permutation test using the 18 sessions as independent units (p < 0.001) gave consistent results, and a composition analysis showed that the trend was not explained by the growing share of micromobility users. The effects are small in standardized terms but correspond to approximately 120 additional phone-engaged users per hour at peak volumes. The framework specifies homography-based trajectory prediction, time-to-collision conflict detection, a logistic risk score with data-driven threshold selection, k-anonymous aggregate data release and a staged validation plan with key performance indicators. A field-calibrated, state-based energy budget estimated 0.63–0.73 kWh/day for event-based operation of a four-approach edge node, 28–42% less than presence-activated and 60–65% less than continuously active operation. Transport-energy savings cannot be inferred because trip distances and displaced modes were not observed.

Energies doi: 10.3390/en19225477

Authors: Kornelia Szmit Stanisław Bielski Michał Bilczak

From the end of 2020, every new building in the European Union had to be a nearly zero-energy building (nZEB). Afterwards, new-dwelling prices grew faster than existing-dwelling prices in most countries. The European Central Bank linked part of this growth differential to energy costs and other factors. Its energy explanation is tested against a phase shift of new-dwelling prices (later recording or slower adjustment) and the cyclical sensitivity of existing dwellings, with country, cross-country, panel and event-study estimates for 26 Member States (without Greece). After correction for multiple testing, the differential shows no association with national energy indicators: energy prices, heating fuels, nZEB values and carbon intensity of electricity, which alone nears significance before correction. Much of the measured rise is a matter of timing. Between 2019–2021 and 2022–2025 the differential rose more where building permits fell more, consistent with a cyclical account, but the rise and this link largely disappear when new-dwelling prices are compared with existing-dwelling prices two or four quarters earlier. The contemporaneous differential thus overstates the lasting change in relative prices. The asymmetry itself is not new: in 2007–2014, before any nZEB requirement, existing dwellings fell further than new ones in most countries. The timing of national rate increases does not identify a credit channel. Energy premia are better estimated within the existing stock, and segment comparisons should allow for the phase shift.

Energies doi: 10.3390/en19225473

Authors: Qi Zhu Yu Sun Lie Li Yifan Zhao

Islanded microgrid topology and converter current limiting can alter transformer protection currents. We formulate primary and backup coordination as a topology-indexed feasible domain. A bridge test identifies source contributions that cross the protected transformer, and a joint model tests differential sensitivity and stability, overcurrent pickup and timing, load security, and device limits. At the selected inputs, four supported bridge states in an eight-entry reference library retain non-empty domains. The weak-GFL case has a narrow backup–pickup interval (0.984–1.042 pu), and its feasibility classification changes across the tested lower-current factors; the high-load case remains empty (1.860 pu load-security lower bound versus 1.326 pu sensitivity upper bound). A seven-bus reference EMT study includes 112 baseline cases, a 1232-case direction/location/resistance scan, and a 288-case coupled-controller scan. The CT stress screen reduces the all-source reference-domain volume from 0.08391 to 0.07059. All numerical results use the declared reference-network and device inputs.

Energies doi: 10.3390/en19225474

Authors: Menaouar Berrehil El Kattel Robson Mayer Lincoln Moura de Oliveira Reuber Saraiva de Santiago Cristina do Carmo Lucio Berrehil El Kattel Benameur Berrehil El Kattel Telles Brunelli Lazzarin Luiz Henrique Silva Colado Barreto Braz de Jesus Cardoso Filho

This paper presents a three-port unidirectional isolated DC–DC converter for high-power applications, featuring two isolated output ports supplied by a single input source. The proposed topology integrates a full-bridge switching stage, a high-frequency transformer with two secondary windings, and independent rectification stages to provide two output voltages. The converter operating principle is analyzed through its main switching stages, and closed-form expressions are derived for the DC voltage gains, inductor design, capacitor design, and voltage and current stresses of the main semiconductor and passive components. In addition, a parameterized expression for the input inductor current ripple is obtained, showing complete ripple cancellation at a duty cycle of 0.5 and identifying the operating points associated with the maximum ripple. A 7.5 kW laboratory prototype with a 260 V input voltage and 800 V and 400 V output ports was developed to experimentally validate the proposed analysis. The experimental results show good agreement with the theoretical prediction. The converter achieves a maximum measured efficiency of 96.81% at approximately 4.5 kW and maintains an efficiency above 96.5% from approximately 3.5 to 7 kW. These results demonstrate the feasibility of the proposed three-port isolated converter for high-power applications.

Energies doi: 10.3390/en19225472

Authors: Martin García-Vaquero Antonio Sánchez-Bayón Ana Belén Miquel-Burgos

European Union energy transition policy has developed through successive layers of climate, energy security, industrial and electricity market measures. This structured critical review examines identifiable cross-policy mechanisms within three interaction families: energy security, decarbonisation and strategic/industrial autonomy; carbon pricing and industrial competitiveness; and electrification, system capacity and implementation. An evidence-contingent Structured Cross-Policy Interaction Analysis (SCPIA) organises 18 policy pairs and 28 outcome-specific records, distinguishing policy design, implementation evidence, observed effects, modelled effects and prospective risks. Classification is outcome-specific and permits abstention when evidence is insufficient. Of the 28 admitted outcomes, 21 receive a class (11 reinforcing, eight trade-offs and two direct conflicts) and seven remain justified abstentions; 22 rest mainly on legal design rather than observed effects. The same policy pair can reinforce one outcome while implying a prospective trade-off for another, as in the coordination of the EU carbon border adjustment with the phase-out of free emission allowances. Two human assessors examined a purposively selected subset of 15 outcome records and agreed on the initial reported interaction and class assignments in all 15; subsequent source-threshold assessment is recorded separately. The review contributes a traceable, evidence-sensitive procedure for comparing cross-policy interactions and integrates substantive findings across three EU energy transition policy families.

Energies doi: 10.3390/en19225471

Authors: Hong Yang Binghao Zhang Zixuan Xiao Zhuozhan Han Jiasheng Huang Yue Xie

Thermal aging alters the voltage-dependent dielectric response of cross-linked polyethylene (XLPE) cable insulation. Characterizing the temperature dependence of this response is necessary for comparing insulation measurements obtained under different thermal conditions. A high-voltage variable-frequency measurement circuit was established to investigate 10 kV XLPE cables thermally aged at 150 °C for 0–42 d. Voltage–current measurements covered peak total applied voltages of 100–2000 V and frequencies of 0.01–1000 Hz at 30 and 90 °C. Polarization-derived resistance, differential scanning calorimetry, and capacitance measurements at 1 V provided complementary references for insulation aging. Resistance decreased with aging, whereas the capacitance response depended on temperature and provided limited separation between some aging stages. The voltage dependence of admittance was most pronounced at 0.01–0.1 Hz. At 0.01 Hz, the normalized complex-admittance increment increased and then decreased with aging, reaching maxima of 5.5 and 4.0 at 30 and 90 °C, respectively. The characteristic boundary frequency and transition voltage generally increased with aging and were lower at 90 °C than at 30 °C. These shifts differed from the nonmonotonic evolution of nonlinear strength. At 0.01 Hz, conductance exhibited larger normalized voltage-dependent deviations than susceptance. The results distinguish changes in nonlinear strength from shifts in the frequency extent and voltage transition of the response. Joint analysis of these characteristics and the admittance components provides complementary descriptors of aging-related dielectric changes at different test temperatures.

Energies doi: 10.3390/en19225470

Authors: Dimitris Al. Katsaprakakis Nikos Papadakis Nikos Kallitsounakis George M. Stavrakakis Marina Leontidi Ioannis Papadakis George Viskadouros

This study evaluates photovoltaic (PV) self-production with and without battery energy storage for five contrasting consumer profiles under the meteorological conditions considered for Crete, Greece: a university campus, an aquarium, an industrial facility, a residential group and a travel agency. Annual chronological simulations at hourly resolution assess demand coverage, surplus generation and economic performance over a 25-year evaluation period, with no revenue assigned to surplus electricity. The model includes battery charging and discharging power limits, annual degradation rates of 0.5% for PV power capacity and 2% for battery energy capacity, and equipment replacement assumptions. Normalised generation and storage parameters facilitate comparison across cases. A one-at-a-time sensitivity analysis examines six technical and economic parameters for one selected configuration per consumer. Without storage, the tested configurations supply 36.1–45.8% of annual demand at the university, 36.5% at the aquarium, 47.7–67.3% at the industrial facility, 24.2–29.2% in the residential group and 50.2–55.1% at the travel agency. Among the reported configurations, those achieving the highest mean annual demand coverage in each case supply 89.5–94.6% of demand, while retaining surplus fractions of 20.1–77.1%. In the residential case, increasing PV capacity from 500 to 1000 kW at a fixed storage capacity of 1000 kWh raises coverage by only 2.6 percentage points, while surplus increases from 55.3% to 77.1%. Degradation affects useful supply differently from nominal asset performance: surplus generation buffers the decline in direct PV supply, whereas battery degradation generally drives the reduction in total coverage. Economic outcomes are particularly sensitive to electricity purchase price and demand level, with the residential and travel agency configurations showing greater dependence on favorable economic assumptions. The findings support joint PV–battery sizing based on demand timing and marginal useful energy gains, rather than annual energy matching or maximum coverage alone.

Energies doi: 10.3390/en19225469

Authors: Hassan Qudrat-Ullah

Large-scale nuclear energy projects depend on credible governance arrangements to support international cooperation and address institutional and societal concerns surrounding nuclear energy development. This study applies the Qudrat-Ullah Nuclear Peace and Trust (Q-NPT) framework to examine the governance of the Egypt–Russia El Dabaa Nuclear Power Plant (ERENPP). The framework comprises four pillars—Trust, Equity, Transparency, and Inclusion—with institutional alignment treated as a cross-cutting dimension. Using a qualitative single-case study design and deductive content analysis, the study examines documentary evidence from intergovernmental agreements, financing arrangements, regulatory and international review documents, official project statements, media reports, civil-society publications, and relevant academic literature. The analysis distinguishes formal commitments and reported activities from implementation, stakeholder perceptions, and demonstrated outcomes. The findings identify formal bilateral and institutional arrangements relevant to trust and coordination, alongside governance questions concerning the disclosure of project information, the distribution of costs and benefits, and the scope and influence of stakeholder participation. Although documented consultation activities and regulatory milestones provide evidence of governance processes, they do not independently establish public confidence, equitable outcomes, or the effectiveness of stakeholder engagement. The study demonstrates how Q-NPT can structure documentary evidence across the four pillars, identify cross-pillar relationships, and make evidence gaps explicit without assigning an unsupported overall governance rating or inferring causal effects. It contributes a context-sensitive application of Q-NPT to cross-border nuclear energy governance and identifies priorities for further assessment, including accessible information disclosure, meaningful stakeholder engagement, institutional coordination, and transparent evaluation of project costs and benefits. The framework offers a complementary analytical approach for examining governance in nuclear energy partnerships and informing comparative research on energy-system development.

Energies doi: 10.3390/en19225468

Authors: Matteo Buffo Federico Mucciarelli Andrea Moser Matteo Varesco Grigoris Protopsaltis Bill Gialelis Diego Viesi Francesco Locatelli

The integration of intermittent renewable energy sources and the electrification of end uses are placing unprecedented pressure on European electricity systems, increasing the need for flexible resources capable of balancing supply and demand at multiple voltage levels. Within this context, industrial Demand-Side Flexibility (iDSF) has emerged as a promising resource, offering both system-level benefits and economic opportunities for industries. However, its implementation remains constrained by a complex interplay of technical, organizational, regulatory, and market barriers that vary across countries. The present study delivers an exploratory cross-country assessment for Italy, Austria, Greece, and Bulgaria, combining desk research on regulatory and market frameworks with a consultation involving forty-five relevant stakeholders. Usable flexibility emerged as considerably narrower than technical potential due to continuous process constraints and organizational limitations, and a systematic tension appeared between users’ need for operational control and the automated activation required by system operators. Availability-based remuneration proved to be the most influential factor driving participation, whereas national differences persist regarding market maturity and the spectrum of iDSF opportunity at the transmission and distribution levels. The findings provide actionable implications for flexibility services in the explored regions, focusing on stable remuneration and simplified procedures to foster broader participation.

Energies doi: 10.3390/en19225467

Authors: Lijian Li Qingchun Meng Zhenjie Wang Shengliang Wang Yu Sun Wenhao Li

Multi-layer combined development is widely adopted in low-to-medium-permeability sandstone reservoirs to balance economic efficiency and development performance during supercritical CO2 miscible flooding. However, permeability heterogeneity often induces severe inter-layer interference and gas channeling. To quantitatively clarify the dynamic influence of permeability heterogeneity on commingled injection and production systems, this study conducted single-core baseline tests (ranging from 19.17 mD to 205.60 mD) and three-core parallel physical simulation experiments at 150 °C and 50 MPa with varying permeability contrasts (nominal ratios of 5, 7, and 10; actual ratios of 5.15, 7.22, and 10.73, by fixing the medium- and low-permeability layers). The experimental results demonstrate that recovery increased with permeability across the tested assemblies, with single-core ultimate recovery increasing monotonically from 64.69% to 78.68% as permeability rises, accompanied by earlier CO2 breakthrough and lower peak pressure drops. In parallel combined development systems, permeability contrast serves as the dominant factor associated with inter-layer flow competition and dynamic shielding. At a high permeability contrast of 10, the high-permeability layer in the combination experiences premature gas breakthrough, with instantaneous split ratios generally between 60% and 85% (mean 71.63%) and consistent with substantial pressure relief. The observed zero recovery in the low-permeability layer is consistent with a dynamic shielding mechanism, in which early pressure relief in the high-permeability layer is interpreted as limiting the driving pressure drop available for low-permeability mobilization. As the permeability contrast decreases to 7 and 5, the flow resistance imbalance is significantly mitigated, allowing the low-permeability layer’s recovery factor to increase to 19.06% and 38.49%, respectively. At nominal ΔK = 5, this corresponds to a 26.20 percentage-point deficit and a 40.50% relative suppression compared with its single-core baseline (64.69%). Furthermore, system overall recovery improves substantially from 45.27% to 59.91%, and cumulative gas–oil ratios drop noticeably. The smallest tested permeability contrast gave the highest combined recovery under the experimental conditions. This is an experimentally observed case, not a general or calibrated threshold. The novelty of this work lies in providing a controlled-variable three-core parallel dataset under supercritical miscible conditions, providing macroscopic evidence interpreted as dynamic shielding of low-permeability layers, and identifying the smallest tested permeability contrast as a reference condition for optimized sweep efficiency. This study provides crucial physical simulation support and technical references for rational layer-combination division, stratified gas injection strategies, and subsequent mobility control optimization in heterogeneous reservoirs.

Energies doi: 10.3390/en19225466

Authors: Min Zhang Yiwei Zhang Xuelin Zheng Weiyao Zhu Yubao Gao

Multi-well fracturing serves as a common technology to increase reservoir permeability for shale gas and oil recovery. The swarm-like fractures from multi-well fracturing are controlled by the stress shadows between fractures. By employing optical fiber to monitor stress shadow characteristics among multiple horizontal wells in real time during fracturing, this study investigates the morphology and underlying causes of swarm-like fractures under various fracturing sequences. Experimental results indicate that Zipper fracturing enhances fracture propagation uniformity, while Alternate fracturing increases the breakdown pressure. The hydraulic fracture exhibits a staged propagation pattern, inducing a rapid transition from tension to compression strain around the fracture associated with strain-energy release. The maximum values within both the compressive and tensile zones of the stress shadow emerge at fracture initiation. Upon the completion of fracture propagation, the region within 40 mm of the fracture surface constitutes the tensile zone, while the area spanning 100 mm from the fracture represents the primary compressive zone of the stress shadow. Compared to Zipper fracturing, fractures in Sequential fracturing induce pronounced tensile strain concentration, yielding a larger tensile strain zone and causing the stress shadow of subsequent fracturing wells to affect the already completed wells. These findings delineate how stress-shadow interference governs the propagation of fracture swarms across multiple horizontal wells under controlled laboratory conditions. The results indicate that sequencing protocols can effectively modulate localized stress interference to promote more balanced fracture network development, thereby offering mechanistic insights into fracturing sequence optimization for unconventional reservoirs.

Energies doi: 10.3390/en19225465

Authors: Atef Chibani Ayache Boultif Tawfiq Chekifi Farhan Lafta Rashid Arman Ameen Selman Djeffal Mohamed Razi Morakchi Samir Gouga Kamel Mesmoudi Oussama Benbouaziz Mohamed Kezzar Moustafa Boukraa

Greenhouses operating in hot, arid climates are exposed to large solar gains during the day and significant heat losses at night, which makes passive thermal buffering attractive. In this work, a north-wall storage layer made of the paraffin RT25 is embedded in aluminum, copper or steel foam, and its effect on the greenhouse microclimate is analyzed with a transient two-dimensional CFD model that couples porous-medium resistance, discrete-ordinates solar radiation, turbulent airflow and enthalpy–porosity phase change. Each composite is evaluated against an otherwise identical pure-RT25 wall. The simulated temperature, liquid-fraction, enthalpy and velocity fields show that the foam material modifies conductive heat spreading inside the wall, the timing of latent storage and the buoyancy-driven circulation in the enclosure. The thermal and phase-change results correspond to a 45 °C exterior/15 °C nighttime scenario, while the airflow profiles correspond to a separate 45 °C/10 °C scenario. A single foam pore density (20 PPI) is examined, so no pore-density optimum is claimed. Against a building-roof PCM experiment, the model reproduces the measured inner roof-surface temperature history with a root-mean-square error of 0.21 °C; this benchmark supports the thermal response of the PCM layer but not the complete greenhouse–foam system. Because no HVAC load balance is solved, the findings are expressed as comparative physical responses rather than as quantified energy savings.

Energies doi: 10.3390/en19225462

Authors: Weidong Chen Qingfeng Hou Gaohua Liu Zhuoyan Zhu Yuanyuan Wang Hongbin Ma Yanfang Yin Xiaojia Li Xiong Yan Fei Guo Xiaoling Yi Weidong Liu

Graphene oxide-based nanomaterials are widely used in different areas due to their distinguished interfacial activity. Leveraging the interfacial behavior of graphene oxide (GO) at the oil–water interface, this study developed a hybrid heavy oil viscosity reducer (HHOVR) comprising GO nanosheets and a macro-molecular emulsion stabilizer (MES). A key advantage of the HHOVR lies in its effectiveness under mild shear conditions, where it forms stable oil-in-water (O/W) emulsions due to the complementary effect of GO and the MES. These emulsions resist coalescence, ensuring sustained emulsification capability and viscosity reduction. The HHOVR demonstrated exceptional heavy oil viscosity reduction performance, achieving rates exceeding 85% under multiple test conditions, with some cases surpassing 95%, indicating strong potential for large-scale application. Core flooding experiments and core-scale numerical simulation indicated a significant enhancement in displacement efficiency, with an enhancement of about 25%. The HHOVR’s high viscosity reduction rate, practical applicability, and remarkable emulsion stability highlight its promising prospects for field deployment and provide robust support for advancing chemical-enhanced oil recovery (EOR) technologies.

Energies doi: 10.3390/en19225464

Authors: Wenyu Jiang Anqi Dai Miankun Li Chutong Xu

Workload flexibility and battery storage can support data-center energy management, but evaluating their operating value requires explicit checks of execution feasibility, numerical solution quality, and economic accounting. This study assesses these dimensions in a six-region benchmark with 50,000 tasks and 2407 hourly intervals. Event-time checks reveal resource violations hidden by hourly accounting; regional reassignment restores modeled feasibility but can worsen service and cost outcomes. Same-objective dispatch comparisons and numerical sensitivity analyses improve export revenue and grid-exchange smoothness relative to the historical low-carbon result, without further reducing already-zero purchased-grid emissions. Matched controls and conditional cycle-wear tests show that balanced-policy cash savings can reverse, whereas tested cost-priority adaptations retain positive improvements. Six restricted joint mixed-integer linear programs yield no additional incumbent gain. A seven-point ε-constraint analysis of cost–grid-variation tradeoffs, followed by post hoc checks, identifies one numerically non-dominated candidate and tolerance-dominating alternatives for the other six under the stated tolerances. Auxiliary forecasting remains separate and does not establish a statistically significant advantage of the tested nonlinear predictor. The findings concern deterministic, full-information operating comparisons and do not establish full-scale joint optimality, physical charging provenance, or lifecycle investment profitability.

Energies doi: 10.3390/en19225463

Authors: Kezheng Jiang Yuqi Ao Chang Ye Xun Xu Menghan Xiao

In islanded microgrids, a sudden drop in photovoltaic (PV) power causes severe frequency dips, whereas grid-forming energy storage relying on droop control alone exhibits a steady-state deviation that cannot be eliminated; increasing the damping coefficient converges only slowly and is constrained by the droop regulation ratio and the frequency deadband. To address this, this paper proposes a two-timescale frequency control method based on meteorological warnings. On the slow timescale, meteorological warnings are used to apply active-power feedforward against the PV drop caused by cloud passage, compressing the disturbance that droop must compensate from the entire power deficit to the prediction residual; on the fast timescale, virtual inertia and damping are pre-scheduled according to the bound of the prediction residual, raised before the disturbance and released afterwards. An input-to-state stability (ISS) theorem is established for the closed-loop system, proving that the frequency deviation is input-to-state stable with respect to the bounded prediction residual, from which the inertia/damping scheduling law and the feasible residual region are derived. Using a random forest predictor trained on three years of hourly measured meteorological data from a region in Hubei, together with a four-method benchmark and end-to-end ablation experiments, the results show that under an 85% PV drop the proposed method restores the steady-state frequency from 48.33 Hz to 50.00 Hz and reduces the maximum rate of change of frequency (RoCoF) in the fast transient from 0.81 Hz/s to 0.17 Hz/s with feedforward alone and to 0.06 Hz/s with the complete two-timescale controller; with a real event-level prediction error of 0.064, the steady-state compliance rate (not lower than 49.8 Hz) reaches 91.7% using the empirical event-level error distribution (94.4% under the Gaussian approximation), far above that of conventional droop control.

Energies doi: 10.3390/en19225461

Authors: Bo Yi Yi Guo Yi Huang Ping Zhang Kunming Lu Xingtao Guo Yunhe Liu

Developing cascade water energy storage potential is an effective way to enhance hydro–wind–solar flexibility and promote the accommodation of variable renewable energy. This study proposes an optimization method for developing cascade water energy storage potential that considers reservoir dredging, energy storage pump station, and reversible pumped-storage unit. Firstly, the characteristics of various measures for developing water energy storage potential are analyzed, and the operating boundaries and constraints of the cascade water energy storage system are determined. Secondly, considering the volatility of grid-connected large-scale green energy, a short-term multi-objective cascade water energy storage system optimization scheduling model is established with the goal of renewable energy generation and peak-shaving benefits. Finally, a multi-objective fuzzy evaluation method is adopted to quantitatively analyze the impact of different measures on the green energy generation and peak-shaving of the cascade water energy storage system. The cascade hydropower system of Gansu section of the Yellow River basin is selected as a case study, and the results indicate that (1) at matched average daily cascade hydropower output, reservoir 2 dredging reduces the residual-load standard deviation by 27.8% on average, whereas the composite evaluation of generation and peak shaving improves it by 18.0%; (2) deployment of energy storage pump station and reversible pumped-storage unit can improve the overall operational performance of the cascade water energy storage system by 11.3% and 41.0%, respectively; (3) Among the schemes evaluated at the specified capacities, reservoir dredging combined with reversible-unit expansion performs best for the modeled generation and peak-shaving objectives.

Energies doi: 10.3390/en19225459

Authors: Ziliang Zhao Zhiguang Ge Fanyu Meng Zhanfeng Wang Bin Guo Man Zhang

The construction of typical driving cycles is essential for evaluating vehicle energy consumption, yet existing methods struggle to balance search coverage, computational efficiency, and disturbance control. This paper proposes a novel method based on the Markov Chain Resampling Reachable Domain with Reinforcement Learning (RL-MC-RRD). Reinforcement learning jointly determines the sub-segments to be refined and their corresponding bridging-path search budgets, followed by state lookup table-based resampling to further correct the residual statistical discrepancies. Validated using one million seconds of real driving data from Qingdao, the final driving cycle constructed by RL-MC-RRD achieved an average deviation of only 0.206% across 16 characteristic parameters. Compared to traditional MC and MC-GA methods, it improves construction accuracy by 50-fold and 8-fold. In addition, the RL-MC-RRD-generated driving cycle exhibited deviations of 1.309% in electric-energy consumption and 3.075% in fuel consumption relative to the real-world driving cycle and offers an effective approach to accurate driving cycle construction, vehicle energy management, and performance evaluation.

Energies doi: 10.3390/en19225460

Authors: Haris Ur Rehman Horani Rafique Ur Rehman Memon Muhammad Amin Štefan Bojnec

Climate change can jointly alter electricity demand and generation availability, yet these effects are often assessed separately. This study develops an audit-controlled climate-conditioned electricity adequacy framework for Pakistan by linking an FY2025 hourly climate–demand model, ERA5-Land and GFDL-ESM4 SSP3-7.0 climate information, a geolocated generation inventory, and monthly supply–demand accounting for 2031–2050. The selected physical demand model, H0_no_heating, uses a 20 ∘C cooling threshold, hour-specific cooling sensitivity, humidity, calendar controls, and six-hour thermal memory. Its leave-one-month-out diagnostics are RMSE = 1392.4 MW, MAE = 1057.9 MW, and R2=0.8725. Under the 19,986 MW central-ready quantified supply basis and a climate-only demand counterfactual, 41 of 240 months exhibit negative monthly mean-balance margins, all during June–August. July 2044 is most constrained, with supply of 18,385.5 MW, mean demand of 19,678.8 MW, and a −1293.2 MW margin. A separate 31,179 MW sensitivity asset set remains non-deficit in monthly mean balance across 12 deterministic hydro/non-hydro states. These states have no probability weights. The framework therefore provides a reproducible monthly climate stress test, not an hourly probabilistic reliability assessment, and identifies strong seasonal concentration of climate-conditioned adequacy stress.

Energies doi: 10.3390/en19225453

Authors: Jie Wu Li Chen Zhihong Tang Lide Zhang Jianwei Li

Liquid metals with a low Prandtl number (Pr) show a strong separation between momentum and heat transport, so Reynolds averaged Navier–Stokes (RANS) closures based on a prescribed turbulent Prandtl number may lose accuracy for local temperature and wall heat flux predictions. This work evaluates a four-parameter RANS closure that obtains the turbulent viscosity and turbulent thermal diffusivity from separate mechanical and thermal transport scales. The closure couples the shear stress transport k-ω momentum model with kθ-ϵθ thermal transport equations. The model is tested for forced convection under imposed wall heat flux in periodic channel flow, heated vertical lead–bismuth eutectic pipe flow, and the KALLA 19-rod wire-wrapped bundle. In the channel, direct benchmark comparisons at Reτ=180, 395, and 640 assess mean temperature and turbulent transport, while additional cases examine near-wall scaling and model trends up to Reτ=4400. In the pipe, the model follows the fully developed variation of the Nusselt number with the Péclet number (Pe) over 201≤Pe≤4890 and gives a lower Nusselt number error than the constant Prt calculations considered here. In the bundle, the model predicts local temperatures with mean absolute errors of 2.40, 3.36, and 4.51 K at the three measurement levels (ML1, ML2, and ML3) while also predicting the main pressure drop and thermal nonuniformity trends. The results support the independent thermal scale closure as a useful RANS approach for forced low Pr liquid metal heat transfer.

Energies doi: 10.3390/en19225458

Authors: Han Zhang Fuxiang Li Yang Wei Hongli Liu Wei Tang Lingyu Chen

Provincial carbon footprint factors of electricity underpin product carbon footprints, green certificate trading, and carbon border adjustment, yet their uncertainty is usually reported as one aggregate interval that does not show which part better measurement could reduce and which part an accounting convention fixes. This case study examines eight plants in one Chinese province with original 2025 accounting workbooks (four coal-fired, four hydropower). Each is rebuilt into a normalized parametric model reproducing the original total. Measurable quantities are propagated by Monte Carlo sampling and apportioned with variance-based sensitivity indices; database choices and accounting conventions are enumerated as discrete scenarios and compared with, rather than added to, the probabilistic interval. For the coal-fired plants, the measurement-driven 95% interval is about −11% to +12%, dominated by coal input; the background database shifts the factor by a comparable amount, whereas every accounting convention shifts it by less than 30% of the interval half-width. For the hydropower plants, the interval is wider, yet harmonizing the amortization horizon of civil works alone moves the factor by 1.5 to 2.8 times that half-width. The findings are specific to the sample and suggest differentiating data collection and factor publication by generation type.

Energies doi: 10.3390/en19225457

Authors: Sami Youssef Jose Marie Saad Joseph Kesserwani Christy Lahoud Joy Najem Chawki Lahoud

Lebanon’s prolonged electricity crisis has renewed interest in renewable energy; yet the wave energy resource of its 220 km Mediterranean coastline has not been reassessed since 2013 and has never been evaluated within an integrated techno-economic framework. Model-derived daily significant wave heights for 2022 and 2023 at three coastal sites (Tripoli, Beirut and Saida), obtained from an operational wave forecast and compared with three other wave datasets, were analyzed with linear wave theory, consistently with the offshore origin of the data. Three wave energy converters (Pelamis, Wave Dragon and AquaBuoy) were evaluated from their power matrices, followed by techno-economic and CO2 emission assessments. The annual mean wave power flux was 2.6–4.3 kW/m (other wave datasets: 1.4–3.2 kW/m), about 41% lower than the preliminary shallow-water estimate, and a 31-year reanalysis shows that 2022 was an average year and 2023 about 10% below average. All converters operated far below their rated power: Wave Dragon had the highest capacity factor (about 4.9%), three to four times lower than for the same device in the Madeira Archipelago, but with interpolated power matrices its output was halved and Pelamis ranks first. With 2026 costs and electricity valued at the weighted avoided cost of the current utility–generator supply (0.419 USD/kWh, an upper-bound revenue case), the 20-year Net Present Value of Wave Dragon is +1.3 million USD under ideal availability but −1.0 million USD with 90% availability and 5% losses; viability therefore requires an electricity value of about 0.45 USD/kWh or more and confirmation of the production estimate. Pelamis and AquaBuoy are not viable. Wave Dragon would avoid about 2400 t of CO2 per year in gross operational terms.

Energies doi: 10.3390/en19225456

Authors: Runzhi Li Peng Luo Shixuan Lu Hang Wen Na Jia

Optimizing the attachment point, azimuth, and length of each fishbone branch independently lets the well design adapt to reservoir heterogeneity, but many candidate designs then violate geometric constraints because branches intersect or come too close. This study combines repair-based constraint handling with a two-stage workflow: a genetic algorithm (GA) selects the branch attachment points, and particle swarm optimization (PSO) then optimizes the azimuth and length of each branch. Out-of-bounds coordinates are reflected back into their ranges, and infeasible candidates are repaired by Gaussian mutation or regenerated by Latin hypercube sampling with opposition-based learning (hybrid regeneration). The operators are established; the contribution is their combination for clearance constraints among the branches of one well and its systematic evaluation. The workflow was coupled with CMG STARS and tested on eight synthetic models based on the Lloydminster heavy oil and Bakken tight oil formations. Relative to the best of twelve fixed-pattern designs, which use the full 2000 m branch-length limit, the optimized five-branch designs produced 13.2–29.4% (Lloydminster) and 5.8–20.8% (Bakken) more oil over 15 years. Across 16 combinations of reservoir model and branch number, the proposed workflow outperformed GA–PSO without regeneration in all 16 cases (by 0.5–18.8%) and PSO-only with evenly spaced attachment points in 15 of the 16 cases (by up to 18.4%). In five independent runs on each of four selected models, its mean production exceeded that of GA–PSO without regeneration by 1.4–9.6% and that of PSO-only by 5.8–16.4%. In the tested models, fewer, longer branches (five versus ten attachment points) outperformed more, shorter branches of equal total length.

Energies doi: 10.3390/en19225455

Authors: Remigiusz Jasiński Aleksander Ludwiczak Paula Kurzawska-Pietrowicz Andrzej Ziółkowski Jakub Niklas Maciej Mazur Aleksander Rajch

The decarbonization of aviation represents one of the most complex challenges in the global energy transition due to stringent safety requirements, long aircraft lifecycles, and dependence on high-energy-density fuels. This study examines the role of hydrogen-based solutions in enabling long-term emission reductions across both airborne propulsion systems and airport energy infrastructure. The analysis integrates multiple technological pathways, including hydrogen fuel cells, direct hydrogen combustion in gas turbines, hydrogen-derived synthetic aviation fuels, and the development of hydrogen-enabled airport energy systems. A narrative literature review examines technological readiness, infrastructure requirements, storage strategies, economic constraints, and safety considerations across the selected hydrogen-based aviation pathways. The results indicate that hydrogen fuel cells are most suitable for short- to medium-range applications, while liquid hydrogen combustion may become viable for long-range aircraft, subject to advancements in cryogenic storage and aircraft integration. Hydrogen-derived e-SAF provides compatibility with existing fleets but depends heavily on renewable electricity availability and production scalability. Airport infrastructure emerges as a critical enabler, requiring coordinated development of on-site and off-site hydrogen supply chains, storage systems, and safety frameworks. The findings suggest that hydrogen should not be considered a single-technology substitute for Jet A-1 but rather a systemic energy platform requiring phased implementation, cross-sector integration, and policy alignment. The reviewed evidence suggests that hydrogen propulsion, hydrogen-derived synthetic fuels, and supporting airport infrastructure may make complementary contributions to aviation decarbonization. Their deployment depends on low-carbon energy availability, aircraft integration, economic feasibility, and certification; the review does not establish a preferred transition scenario or demonstrate achievement of net-zero aviation by 2050.

Energies doi: 10.3390/en19225454

Authors: Dong Yao Xiaoxue Liu Zheng Cao Yu Gong Yang Zhao Yinchi Shao

Wind and photovoltaic (PV) power output exhibits substantial temporal, spatial, and terrain-related variability, making sample partitioning critical for constructing representative output profiles. This study develops a unified framework for evaluating sample partitioning rules before clustering, rather than treating them as predetermined preprocessing conditions. The framework constructs 18 candidate schemes by combining different temporal, spatial, and terrain-based partitioning rules, and evaluates their suitability using six qualitative criteria and nine quantitative consistency measures through an integrated weighting and ranking approach. The selected scheme is subsequently used to organize daily output samples for clustering, from which hourly P50 characteristic profiles and P40–P60 empirical bands are extracted. For empirical validation, the framework is applied to wind and PV generation data from the Australian National Electricity Market over 2020–2025. The scheme combining monthly, county-level, and terrain-class partitioning, designated as A3, achieves the highest overall ranking among the 18 candidates. The resulting profiles characterize the intraday fluctuations in wind generation and the single-peaked daytime pattern of PV generation, while the quantile bands describe output dispersion within the dominant clusters. The methodological contribution lies in establishing a systematic pre-evaluation approach for sample partitioning rules and integrating it with subsequent output characterization, providing a structured basis for constructing condition-specific wind and PV output benchmarks.

Energies doi: 10.3390/en19225452

Authors: Aamer Abbas Shah Khubab Ahmed Aftab Ahmed Almani Shuchen Wang Liwen Wang

This paper presents a novel photovoltaic (PV) power prediction method based on a radial basis function neural network (RBF-NN) and randomness-operator particle swarm optimization (ROPSO). The ROPSO method uses the range of meteorological variables to construct a search space to optimize the RBF-NN. First, ROPSO analyzes the possible range of unknown parameters defined by the meteorological conditions of the plant site. Then, based on the range of unknown parameters, ROPSO constructs a squeezed search space for the RBF-NN, thus speeding up the parameter identification process and making the identification process more efficient. A real case study of a 20 MW grid-connected photovoltaic (PV) system in Jining City, China, is used to demonstrate the effectiveness of the proposed forecasting technique. Various evaluation metrics, including the mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R2), are employed to assess the effectiveness of the prediction scheme. The simulation and numerical results indicate that the radial basis function neural network model optimized by ROPSO can achieve higher efficiency compared to other alternatives.

Energies doi: 10.3390/en19225451

Authors: Myunghun Lee Hyun Sik Yoon

This study investigates heat-transfer and drag control in tandem circular cylinders using proximal policy optimization (PPO)-based deep reinforcement learning for closed-loop synthetic-jet actuation. Two-dimensional unsteady laminar computational fluid dynamics simulations were coupled with PPO controllers. Separate policies were trained for six spacing ratios (L/D = 5–10) at Reynolds numbers Re = 100, 150, and 200, using equally weighted heat-transfer and drag objectives. The sum of the percentage drag reductions and heat-transfer changes of both cylinders was positive at all 18 training conditions (5.2–40.3). Upstream heat transfer increased by 2.3–4.4% across all training conditions, while downstream drag reduction reached 49.3%. At Re = 100, downstream heat transfer was largely maintained, whereas higher Reynolds numbers exhibited reduced downstream heat transfer accompanying greater drag reduction, revealing a thermal–aerodynamic trade-off. Representative flow-field analyses associated downstream responses with changes in mean velocity and temperature, and upstream drag responses with distinct vorticity and temperature distributions. This sum remained positive at all 12 untrained conditions at Re = 125 and 175 but decreased markedly at L/D = 9 and 10. These findings highlight the potential and limitations of PPO-based synthetic-jet control for balancing heat-transfer enhancement and drag reduction in tandem-cylinder configurations.

Energies doi: 10.3390/en19225450

Authors: Peng Zou Dahui Hao Wei Yan Qiangzhong Xiao Jing Yan Ruihua Wang Huanbin Chen

Deepwater, low-permeability, medium- to high-strength sandstone reservoirs are susceptible to pressure depletion and fluid–rock interactions during production. These processes can induce plastic instability and particle migration in the near-wellbore region, ultimately causing sand production. Taking the deepwater L gas field as the study area, this work combines laboratory experiments with numerical simulation to systematically investigate reservoir mechanical properties and the evolution of sand production. The target interval is a low-permeability, fine-grained sandstone reservoir. Quartz is the dominant whole-rock mineral, illite is the dominant clay mineral, and the smectite content is low. The median grain size d50 is mainly 58–139 μm, and the sediments are generally poorly sorted. The uniaxial compressive strength is approximately 43 MPa, indicating a medium- to high-strength sandstone, whereas water immersion causes a pronounced reduction in rock strength. A hydro-mechanically coupled sand-production model was employed in ABAQUS 2026 to incorporate porous-medium fluid seepage, rock mechanical response, and an empirical sand-erosion criterion. The model was used to quantify how reservoir parameters affect near-wellbore plastic deformation and sand production. With progressive reservoir depletion, the equivalent plastic strain increases and sand production remains concentrated in the near-wellbore region. The normalized sensitivity of the maximum equivalent plastic strain decreases in the following order: cohesion, Young’s modulus, ratio of maximum horizontal principal stress to overburden stress, internal friction angle, Poisson’s ratio, and production drawdown. Production forecasts for a development well further show that sand production evolves in three distinct stages: an initial sand-production stage, a rapid sand-production growth stage, and a fluctuating and declining sand-production stage. Numerical simulation results indicate that sand-production risk persists throughout the predicted 15-year production period, highlighting the need to incorporate appropriate sand-control measures into the well completion design and maintain sand-production management throughout the production life cycle. The results provide a basis for sand-production risk assessment and completion and development optimization in comparable deepwater, low-permeability sandstone gas reservoirs.

Energies doi: 10.3390/en19225449

Authors: Vishnu Suresh Francesco Montana Eleonora Riva Sanseverino

This article presents a framework combining deep quantile regression (DQR) and two-stage stochastic optimization for ancillary service provision by an aggregate of prosumers. The purpose of the study is to support bidding decisions by considering uncertainties in electricity demand and photovoltaic generation together with the operation of electricity and hydrogen equipment. The aggregate includes battery storage, hydrogen storage, and an electrolyzer. DQR predicts the 5th, 50th, and 95th demand and generation quantiles, which are used to define three weighted operating scenarios. The service bid is determined in the first stage, while equipment schedules and grid exchanges are adjusted for each scenario in the second stage. Three optimization policies were investigated, which include deterministic median-forecast, expected-cost, and conditional value at risk (CVaR) optimization. These were evaluated across 24 operating conditions and different service prices. It was observed that, at service prices of 125, 130, and 135 €/MWh, the expected-cost approach reduced expected operating costs in ten upward-service conditions per price, with savings of up to €15 per one-hour case compared with median-only bidding. Moreover, at the baseline price of 200 €/MWh, CVaR reduced worst retained-scenario costs by declining ten upward-service offers, although expected operating costs increased. The results demonstrate how considering forecast uncertainty can help avoid economically unfavorable commitments and support the selection of bids according to expected returns and adverse-scenario costs.

Energies doi: 10.3390/en19225446

Authors: Yankai Xing Tianhang Yang Haopeng An Minglei Zhu Dongsheng Cai

High photovoltaic penetration complicates islanded microgrid scheduling because source–load uncertainty interacts with limited storage energy and distribution network constraints. This study develops a coordinated scheduling framework based on the point estimation method (PEM) and multiresolution surrogate-assisted success-history adaptive differential evolution (MSA-SHADE). A 24 h horizon is divided into 96 intervals of 15 min. Five weighted operating conditions represent baseline operation and positive and negative deviations in photovoltaic generation and load. The decision variables comprise storage power allocation coefficients and the droop parameters of two virtual synchronous generators. The objective minimizes expected storage degradation, network loss, photovoltaic curtailment, and load shedding costs. Exact radial power-flow evaluations enforce voltage, frequency, branch current, converter capacity, and state-of-charge constraints. Tests on modified IEEE 33-bus and 69-bus systems yield the lowest expected objective values among the evaluated algorithms. The respective costs are CNY 1028.669 and CNY 879.313, with runtimes of 22.6 and 43.9 min. Both tests require 652 exact evaluations, a 32.9% reduction from the 972 evaluations required by the comparison algorithms. The results demonstrate that joint scheduling and control parameter tuning can improve operating economy while surrogate screening reduces computational effort.

Energies doi: 10.3390/en19225447

Authors: Dacheng Wang Minghui Duan Zhe Wang Xuchen Wang Yaoxiang Zhang Guoqiang Zu Gaozhong Lv Xin Liu Fengzhang Luo Shihan Chen

This paper proposes a multi-objective distributed energy storage planning method that couples distribution-level electricity market simulation to improve the coordination between storage investment, market participation, and distribution network operation. The method incorporates technical–market-based candidate-bus screening, joint energy and frequency regulation market clearing with endogenous price formation, and life-cycle economic assessment. An improved NSGA-II algorithm is employed to obtain the Pareto-optimal solutions, while the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) is used to select a compromise solution among annualized life-cycle net cost, nodal voltage deviations, and network losses. The proposed method is evaluated on an extended IEEE 33-bus distribution system using 24-h typical-day load and renewable generation profiles. Compared with the baseline scenario, the selected storage configuration reduces average network losses from 0.20135 MW to 0.08235 MW and average nodal voltage deviations from 0.028 p.u. to 0.019 p.u., while increasing the renewable energy accommodation rate from 81.73% to 95.63%. The resulting storage investment achieves a life-cycle net present value of CNY 0.8745 million with a simple payback period of 7.33 years. In addition, compared with the PSO-based energy storage planning method, the method proposed in this paper achieves greater economic benefits while effectively improving the system operating state. These results demonstrate the effectiveness of integrating electricity market simulation and life-cycle economic evaluation into distributed energy storage planning.

Energies doi: 10.3390/en19225448

Authors: Zhen Liu Hongyue Chen Jun Mao Xianyang Liu Qi Chen Dongxu Guo Deyuan Tong

Linear compressors are commonly used in household appliances owing to their high efficiency. Previous studies on linear compressor dynamics neglected voltage waveform impact. In this paper, a moving magnet linear compressor is proposed and its numerical model incorporating stator assembly vibration is established. Dynamic properties of the linear compressor are compared in time and frequency domains driven by symmetrical and asymmetrical voltage waveforms. Simulation results show that the piston and stator assemblies exhibit a relative motion irrespective of waveforms. Under symmetrical voltage waveforms, greater oscillation amplitudes and velocities are exhibited, driven by the cosine waveform. The maximum compressor efficiency is achieved under the cosine waveform. Among asymmetrical voltage waveforms, the cosine waveform without troughs results in larger oscillation amplitudes. However, the maximum compressor efficiency is observed under the cosine waveform without crests. The maximum COP is achieved under the trapezoid waveform among all the symmetrical and asymmetrical waveforms. The smoother current curve is obtained driven by the voltage waveform with larger trough heights. Experimental results validate the analytical model.

Energies doi: 10.3390/en19225445

Authors: Jeongju Park Hyeongjun Jo Yeonbin Choi Soobae Kim

Port electrification under the green-port transition sharpens peaks and raises demand charges, motivating photovoltaic–battery energy storage (PV–BESS). Yet PV–BESS value depends on forecast error, operating policy and tariff peak memory, a ratchet carrying one peak-setting error across months; evaluations under perfect foresight or monthly-only billing leave their interaction unquantified. We quantify it in a controlled one-year Korean container-terminal simulation (synthetic load, public PV data, rolling dispatch). Three peak-memory structures and forecast-error levels are tested for four published policies; re-billing, an execution-layer ablation and 20 PV-forecast realizations separate mechanical from behavioral effects. At calibrated error, the ratchet raised mean point-forecast regret by 6–8 million KRW/yr (bootstrap interval excluding zero for one policy; no sign test significant), reduced scenario-policy regret by 3 million (nominally significant), multiplied annual-tariff dispersion two- to ninefold and made its sign event-dependent. In the reference realization, threshold-induced early depletion under monthly-only billing interacted with 12-month re-billing, raising point-forecast regret to 2.4–3.3 times the monthly-only value; with that response off, four of five policies matched the re-billing benchmark. Scenario information was worth 12.5 million KRW/yr under the ratchet, 2.7 under monthly-only billing. PV–BESS value thus depends jointly on forecast accuracy, tariff peak memory, policy information use and event-level dispatch.

Energies doi: 10.3390/en19225444

Authors: Selim Gulhan Seda Kul Seyit Alperen Celtek

Urban wind energy applications are characterized by low wind speeds, variable wind directions, and turbulent flow conditions that influence small-scale wind turbine performance. Although machine learning methods have been widely investigated for wind speed forecasting, the implications of prediction errors for turbine energy production and carbon emission reduction remain less extensively examined. This study evaluates the effects of machine-learning-based wind-speed predictions on the estimated energy production and carbon emission reduction potential of two commercial 10 kW horizontal-axis (HAWT) and vertical-axis (VAWT) wind turbines under inland urban wind conditions. Five prediction models, namely Linear Regression, Support Vector Machine, Random Forest, XGBoost, and Long Short-Term Memory, were evaluated using hourly meteorological measurements collected in Karaman, Türkiye, over one year. Turbine power output was estimated from measured and predicted wind speeds using power curves with corrections for local air density. Among the evaluated models, XGBoost achieved the lowest root mean square error (RMSE) of 0.5957 m/s and the highest coefficient of determination (R2) of 0.7961 on the chronological test dataset. However, the predicted wind speeds showed 15.1% lower variability than the measurements, underrepresenting higher-wind-speed events. Based on measured wind speeds, the equivalent annual energy production was estimated at 0.6559 MWh for the HAWT and 0.9622 MWh for the VAWT, corresponding to capacity factors of 0.75% and 1.10%, respectively. Using XGBoost predictions reduced these estimates by 30.7% and 24.3%, while the corresponding net annual CO2 emission reductions decreased from 0.4041 to 0.2801 tCO2 for the HAWT and from 0.5928 to 0.4487 tCO2 for the VAWT. Variance scaling and quantile mapping reduced the magnitude of the energy estimation errors, although the wind-speed prediction errors increased. Under the assumed cost parameters, the estimated levelized costs of energy exceeded 6 USD/kWh, indicating limited economic feasibility for both turbines at the investigated site. The results indicate that wind-speed prediction accuracy alone does not fully characterize the reliability of subsequent energy and carbon assessments, particularly under low-wind conditions.

Energies doi: 10.3390/en19225443

Authors: Santiago Romero Solana Carlos Cohen Manrique Eberto Porto Mass Alejandro Guerrero Hernández Luis Quiroz Herazo Andrea Burbano Bustos

Campus electrical load is increasingly recorded at minute resolution, but analog signal conditioning adds noise, and whether removing it improves load forecasting is rarely measured against a reference. This paper documents a four-layer IoT architecture at the Corporación Universitaria del Caribe (CECAR, Colombia), in which an ADE9000 analog front-end measures the electrical parameters and a reference signal (Sref), i.e., the electrical data before the sensor, supports a supervised evaluation. Savitzky–Golay, Kalman and Whittaker–Eilers smoothers were compared against Sref over wide hyperparameter grids, with temporal validation, in two buildings and two academic years; Whittaker–Eilers ranked first in every case and reduced the 1 min RMSE by 29–52%. Long short-term memory (LSTM) networks were then trained on smoothed and unsmoothed current with ten seeds in four building–year domains. At the 5 min forecasting resolution, acquisition noise accounted for only about 1–6% of the 24 h mean absolute error, and smoothing did not improve the forecasts consistently: its pooled effect was equivalent to zero within ±2% at 1, 6 and 24 h. A ridge autoregressive baseline outperformed the LSTM in most cases, and weather inputs used as history did not help. Cleaner signals improve measurement fidelity, not forecast accuracy, at this resolution.

Energies doi: 10.3390/en19225441

Authors: Chunning Gao Zhiwei Song Weiliang Xiong Congbo Gou Jie Zhang Zhuangzhuang Zhang Pengxu An Hailong Zhang Junhong Jia Qiuyang Zhao

Mature domestic continental low-permeability oil reservoirs in China have generally entered the ultra-high water-cut and high-recovery stage. Complex reservoir heterogeneity, development characteristics prone to gas channeling, and unclear seepage mechanisms of air foam flooding present engineering challenges to efficient development. To address these issues, this study investigates oxygen-reduced air foam flooding in the Chang-6 ultra-low-permeability sandstone reservoir in the WY block of Changqing Oilfield. The microscopic pore network was characterized using high-pressure mercury intrusion, BET (Brunauer-Emmett-Teller) nitrogen adsorption, and scanning electron microscopy. Changes in reservoir and fluid physicochemical characteristics before and after development were evaluated using the sessile drop method, pendant drop method, and Wilhelmy plate method. Core-flooding experiments were performed to confirm preferred key injection parameters, quantitatively decompose individual enhanced oil recovery mechanisms, and verify pore-scale seepage mechanisms via online nuclear magnetic resonance. The results reveal multi-scale pore network characteristics of the target reservoir and confirm the significant reduction in oil–water interfacial tension during foam flooding. Parameter sensitivity analysis identifies preferred operating conditions within the investigated experimental range, including an injection velocity of 0.2–0.3 mL/min, a foaming-agent concentration of 0.4–0.5%, a gas-to-liquid ratio of 2:1, and a 5-cycle slug schedule. Oil recovery decomposition analysis indicates that ultimate oil recovery comprises a relative share of 77.63% from baseline waterflooding, 13.44% from surfactant-assisted oil mobilization, 6.45% from gas assistance, and 2.48% from physical foam plugging. Furthermore, online nuclear magnetic resonance T2 (Transverse relaxation time) spectrum analysis reveals distinct permeability-dependent air foam flow regimes: as permeability decreases, flow behavior transitions from smooth foam propagation with concurrent oil washing and plugging, to the formation of dynamic foam plugging, and finally to severe shear-induced foam destruction in narrow throats that causes air foam flooding to fail. This study provides preferred injection parameters and theoretical grounds for mechanism decomposition in air foam enhanced oil recovery for ultra-low-permeability reservoirs, offering key insights into air foam seepage behavior.

Energies doi: 10.3390/en19225442

Authors: Liang Shen Yukun Bao

Measuring electricity demand response requires the electricity load that an activated customer would have exhibited without activation. Selective activation and shared disturbances complicate estimation of this unobserved baseline. This study develops Quantile Generalized Synthetic Control (QGSC), combining a historical conditional quantile predictor with residual information from contemporaneous non-activated customers. A shrinkage-controlled common location shift updates the predicted quantiles, followed by empirical interval calibration using held-out historical event controls. Evaluation combines measured commercial and industrial electricity load with simulated activation and response. Five overlapping 60-customer cohorts are drawn from a single-city pool of 102 customers. Relative to a CalTRACK-style regression, QGSC reduces the 19-quantile grid approximation to the continuous ranked probability score from 170.440 to 164.458 kW, or 3.51%. Nominal 90% interval coverage increases from 77.89% to 85.16%, accompanied by a 9.79% increase in normalized width. Point-measurement mean absolute error decreases by 3.33%. Matched ablations attribute most of the grid-score reduction to contemporaneous correction, while donor and subgroup diagnostics reveal residual bias and undercoverage. These descriptive results support a modest incremental benefit for ex-post measurement under the specified retrospective conditions. They do not establish nominal coverage, statistically significant main-experiment superiority, or performance under actual demand response operations.

Energies doi: 10.3390/en19225439

Authors: Marco Vivero Tenorio René Cortijo Leyva Henry Díaz-Iza

The electricity supply in rural communities of Ecuador faces limitations due to the strong dependence on hydropower sources and the lack of alternative backup systems. Photovoltaic systems with automated solar tracking mechanisms emerge as a viable alternative to optimize energy harvesting and enhance energy security. This study aims to design, implement, and validate a 200 Wp photovoltaic generator equipped with a 2000 W inverter with 100 Ah battery backup, integrating a two-axis automated solar tracking mechanism and a remote control interface accessible through mobile devices. The methodology included system sizing using the Photovoltaic Geographical Information System (PVGIS), structural modeling in Autodesk Inventor Professional, implementation of a control system based on an Arduino Mega 2560 with current, voltage, and temperature sensors, and the development of a remote monitoring interface using the RemoteXY platform. Experimental results showed that the tracking mechanism maintained the perpendicular orientation of the solar panels relative to incident radiation, achieving ranges of 137° in elevation and 47° in declination. Automatic switching through the ATS device was successfully verified without observable interruptions of the connected load. According to the manufacturer specifications, the ATS transfer time is below 50 ms, although this value was not independently measured during the experimental validation. Stable access to the remote interface was also achieved, with an external network connection success rate of 93.33%. The failure rate under low-radiation conditions was approximately 17%. The system represents a significant technological contribution for rural areas with limited access to reliable electricity, and establishes the basis for future improvements in control algorithms and sensor sensitivity for solar tracking systems.

Energies doi: 10.3390/en19225440

Authors: Likhabiso Lempe Akshay Kumar Saha

The increasing integration of distributed energy resources (DERs) has transformed conventional radial distribution networks into active distribution systems characterized by bidirectional power flow, variable fault currents, and dynamic operating conditions. While DERs improve system reliability, efficiency, and sustainability, they introduce significant challenges to conventional protection and fault detection schemes. This paper presents a comprehensive review of fault detection methods for DER-integrated distribution networks, including Support Vector Machine (SVM) approaches, neural network-based methods, harmonic analysis techniques, and adaptive protection strategies. The impact of DER penetration on protection coordination, nuisance tripping, protection blindness, high-impedance fault (HIF) detection, and islanding is critically examined. In addition, passive, active, and artificial intelligence (AI)-based islanding detection methods are reviewed and compared in terms of detection accuracy, non-detection zone (NDZ), and detection speed. This review shows that AI-based approaches, particularly deep learning and hybrid machine learning techniques, consistently achieve superior fault and islanding detection performance. However, most reported methods are validated using simulation-generated datasets under predefined operating conditions, limiting confidence in their practical deployment. This study highlights the need for large-scale field validation, real-world datasets, hardware implementation, and robust protection frameworks capable of accommodating high DER penetration levels and evolving grid operating conditions.

Energies doi: 10.3390/en19225438

Authors: Olena Stryhunivska Rafał Rumin Robert Giel Sylwia Werbinska-Wojciechowska

Hydrogen-based propulsion is a promising option for low-emission transport, including emerging high-speed concepts such as Hyperloop. This study evaluates the operational performance of a two-stage hydrogen refueling station concept for hydrogen-powered Hyperloop-type vehicles. The proposed configuration combines direct hydrogen supply from a transport container with final refueling from a high-pressure buffer. A discrete-event simulation model developed in FlexSim represents hydrogen deliveries, container replacement, vehicle arrivals, refueling, buffer replenishment and station occupancy under different arrival patterns and hydrogen supply conditions. Full refueling to 100% is considered alongside a 70% target as a fallback option when sufficient high-pressure hydrogen is unavailable. The assessment combines station-level indicators with departure hydrogen inventory, theoretical vehicle range and route feasibility. In the scheduled scenario, the station served all 16 vehicles, with an average vehicle time in the system of 35.4 min. The results identify high-pressure hydrogen availability and buffer replenishment as key operational constraints. Partial refueling may facilitate departures during temporary high-pressure shortages, provided that the onboard hydrogen inventory is sufficient for the subsequent route. The proposed approach links station performance with vehicle hydrogen requirements and provides a basis for assessing hydrogen refueling infrastructure under variable transport demand.

Energies doi: 10.3390/en19225437

Authors: Julian Patiño Ortíz Adán Acosta-Banda Verónica Aguilar-Esteva Eduardo Campos-Mercado Ricardo Carreño-Aguilera Liliana Hechavarría Difur Miguel Patiño Ortíz

Techno-economic optimization of residential photovoltaic (PV) systems often yields continuous capacities that are not directly buildable. This study develops an integrated framework linking techno-economic sizing, discrete component selection, rooftop geometry, low-voltage grid interconnection, and screening-level decarbonization. Three residential targets previously obtained with HOMER Pro (0.460, 1.310, and 3.890 kW) at 4.45 MXN/kWh were translated into commercial configurations. Evaluation of 430–460 W modules and compatible microinverters selected a 435 W module and IQ8A architecture, producing 1-, 3-, and 9-module systems of 0.435, 1.305, and 3.915 kWp with a mean absolute sizing deviation of 2.15%. At a 16° tilt, the minimum inter-row spacing was 0.817 m, with reserved rooftop areas of 6.754, 13.433, and 36.770 m2. Hourly PVGIS-ERA5 simulations (2019–2023), accounting for temperature, electrical losses, inter-row shading, and microinverter clipping, yielded mean annual AC generation of 855.32, 2565.95, and 7688.61 kWh, respectively. The corresponding annual clipping losses were 4.50, 13.50, and 40.51 kWh. Using the official 2024 Mexican grid emission factor of 0.444 tCO2e/MWh, the estimated avoided operational emissions were 0.380, 1.139, and 3.414 tCO2e/year, respectively. The framework links optimization to implementable residential PV designs, complemented by independent economic screening and illustrative point-of-connection voltage-rise analysis, both subject to site-specific validation.

Energies doi: 10.3390/en19225436

Authors: Mohamed Bousla Thabet Mohammed Alzahrani Saud Hameed Alotaibi Ali Haddi Youness El Mourabit Hilal Essaouini Hassna Salime Badre Bossoufi

Utility-scale wind farms must provide at the point of connection (PoC) a reactive power range that is often wider than the aggregated capability of their turbines. A previous assessment of a 120 MW doubly fed induction generator wind farm connected to the Moroccan 225 kV network identified an absorption deficit at low active power. The present paper does not revisit that diagnosis; it contributes a reproducible and cross-validated procedure for sizing and verifying the remedial compensation. Seven passive and hybrid (STATCOM-based) configurations are evaluated against a +48/−36 Mvar requirement for PoC voltages of 0.9–1.087 pu, first on 616 PowerFactory load-flow scenarios and then with an independent Newton–Raphson model built from the complete collector data, which reproduces the PowerFactory results with a mean absolute error of 0.07 Mvar when the on-load tap changers (OLTCs) are active. The independent model extends the verification to 23,331 operating points (1% active-power and 0.01 pu voltage steps), non-uniform turbine dispatch, equipment tolerances, transformer loading limits, the continuity of the reactive power range across switching states and harmonic resonance screening. Covering the 39.0 Mvar deficit at zero output requires at least 37.8 Mvar of 33 kV reactors with OLTC action and 40.0 Mvar without it, because the reactor output falls with the square of the collector voltage. An 8 Mvar STATCOM with 3 × 10 Mvar reactors falls 0.7 Mvar short, whereas a 10 Mvar STATCOM with 3 × 10.5 Mvar reactors, or 3 × 11 Mvar reactors, satisfies the requirement at all evaluated steady-state operating points, with worst-case margins of 2.2–2.5 Mvar that shrink to 0.6–0.9 Mvar for a −5% reactor tolerance. At the voltage extremes the OLTCs reach their end positions, but all 33 kV and turbine voltages remain within their continuous operating ranges, and limiting the turbine transformers to their rating still leaves an absorption margin of at least 45 Mvar at high output. Relative to the existing capacitor-bank states, the added reactors and STATCOM introduce no new low-order resonance in the screening model. The procedure is transferable to other plants subject to wide-range reactive power requirements.

Energies doi: 10.3390/en19225434

Authors: Thanh Hai Nguyen Quoc Nam Trinh Ton Duc Do Asif Nawaz

The high penetration of renewable energy sources (RES), particularly power-electronic-interfaced wind and solar generation, is transforming conventional power systems into converter-dominated smart grids. While this transition supports decarbonization goals, it introduces major control and stability challenges due to reduced inertia, weakened grid strength, and complex interactions among converters and network dynamics. This paper presents a comprehensive review of control strategies and stability enhancement techniques for smart grids with high renewable energy penetration. Grid-forming and grid-following control strategies, including virtual inertia and damping emulation, droop-based, adaptive, and distributed control schemes, are reviewed, with emphasis on their operating principles, advantages, limitations, and suitable applications. Stability issues under weak grid conditions, high converter density, and hybrid AC/DC configurations are discussed, including frequency and voltage instability, converter interactions, synchronization challenges, and system response to disturbances. The roles of energy storage systems, flexible AC transmission system (FACTS) devices, and high-voltage direct current (HVDC) technologies in enhancing power system stability are also examined. Finally, the paper identifies key research gaps, practical implementation challenges, and emerging trends, supported by a comparative assessment of representative existing review studies, highlighting future directions toward resilient, flexible, and stable smart grids capable of accommodating very high shares of renewable energy.

Energies doi: 10.3390/en19225435

Authors: Zijia Tie Yanbin Du Yang Liu Xuewen Lei Yuxuan Li Yujia Kang

With the continuous growth of electric power load, transformers act as the core equipment in power grids, and control of their operating temperature control is directly related to insulation service life and system safety. Parameter tuning of traditional PID control relies on engineering, and it is difficult to achieve systematic balance among energy efficiency, response speed and stability. This paper proposes a parameter optimization method for an intelligent transformer refrigeration system based on Latin Hypercube Sampling (LHS), Kriging surrogate model and Multi-Objective Particle-Swarm Optimization (MOPSO). The results show that the optimal parameter combination is Kp = 679.60, Ki = 0.12 and Kd = 39.89. Compared with parameters tuned by traditional empirical methods, the optimized COP increases by approximately 15%, and the overshoot decreases by approximately 96.3%. These results verify the effectiveness of the proposed method in balancing energy efficiency and dynamic performance.

Energies doi: 10.3390/en19225433

Authors: Elnur Alizade

Magnetic treatment of injection water has been proposed as a low-cost, chemical-free means of improving waterflooding, but laboratory evidence that links field intensity, pressure, electrokinetic behaviour and oil displacement remains limited. This study investigates the effect of a constant transverse magnetic field (40–150 mT) applied to injection water upstream of a linear sand–clay (75:25) porous-medium model, 120.4 cm long and 4.2 cm in internal diameter. First, single-phase water flow was examined at 1.6–30.4 atm, recording the potential difference, electrical resistance and outflow rate. Second, three oils with kinematic viscosities of 24.9 mm2/s (paraffin oil), 47.49 mm2/s (machine oil) and 198 mm2/s (refined transformer oil) were displaced by untreated and magnetically treated water at pressures up to 35.2 atm. Magnetic treatment lowered the potential difference and resistance of the water-saturated medium and increased the water outflow rate. In oil displacement, the final oil displacement coefficient increased from 47.38% to 60.64% for machine oil, from 42.30% to 54.32% for paraffin oil and from 37.08% to 50.36% for transformer oil at the most effective intensity of 125 mT, while the water displacement coefficient decreased. Magnetic treatment also stabilised the electrokinetic response during displacement. A generalised filtration law that includes electromagnetic forces is presented; with the streaming field taken from the measured potential difference, it reproduces the single-phase data up to 20.8 atm at all intensities with one parameter set (R2 = 0.967) and shows that treatment mainly lowered the threshold pressure difference, from 4.1 to 0.9 atm at 125 mT. Untreated control runs in the same setup, the non-monotonic dependence on field intensity and heat-balance estimates indicate that heating, vibration and degassing do not explain the effects. The between-run uncertainty of the oil displacement coefficient is estimated at 1–2 percentage points.

Energies doi: 10.3390/en19225432

Authors: Ovis D. Irefu Godslove I. Ebiega Breana G. Smithers Miguel F. Acevedo

Microgrids that combine renewable generation with battery storage are being deployed in power systems worldwide to improve supply continuity and to integrate variable renewable energy, yet most evidence on their reliability comes from simulation, and field studies rarely report measured reliability indices together with the load, storage, and weather period under which they were obtained. This paper reports on the design and field evaluation of a 24 V direct-current microgrid at the University of North Texas Discovery Park, Denton, Texas, USA (33.2° N) that combines photovoltaic (PV) and wind generation with lead-acid storage and a parabolic-dish thermoelectric branch on a shared bus. Reliability was assessed from a 33-day metered campaign (696,595 records) and a design-basis suite using loss of power supply probability (LPSP), loss of load expectation (LOLE), expected energy not supplied (EENS), and two storage-independent indices. At the metered load, wind reduced battery-supplied energy by 17.9% (13.5% at night) and generation-adequacy LPSP from 47.9% to 39.4% by generating at night and under low irradiance. Under a 140 W design-basis load, wind reduced LPSP from 4.18% to 0.93% (EENS from 0.140 to 0.031 kWh/day), but the same load evaluated on the record, which covers a different period of measured weather, yielded up to 24.1%, showing that reliability indices depend jointly on load, storage, and the weather period over which they are computed. For the thermoelectric branch, characterized in seven campaigns over 32 days (19 August–19 September 2026), diagnostic elimination showed that manual-tracking dwell time and optical-interface coupling govern output; moving the thermoelectric module bank closer to the dish focal point, with the dish aiming procedure unchanged, raised delivered power from 13.11 W to 50.96 W.

Energies doi: 10.3390/en19225431

Authors: Cemal Zehir Tuğçe Ekiz Yılmaz Mehmet Reha Özder Mesut Akdere

Effective renewable energy policy requires benchmarks that reflect countries’ structural energy characteristics rather than generic performance rankings. This study develops an explainable machine-learning framework that identifies structurally comparable peers and derives peer-relative priorities, using annual World Bank data for 37 Organisation for Economic Co-operation and Development (OECD) countries over 2010–2021. Thirteen energy, economic and demographic indicators are standardized and analysed with principal component analysis, Ward hierarchical clustering and cosine similarity, supported by bootstrap, density-based, temporal, scaling and leave-one-country-out robustness checks. Random Forest with Gini, permutation and SHAP importance is used descriptively to characterize the clusters, and directionally filtered standardized gaps between Türkiye and its closest peer form a Strategic Gap Priority Index (SGPI). Three weakly separated regimes emerge: renewable-intensive economies, larger urbanized fossil-based economies, and less urbanized lower-income economies. Under historical averages, Türkiye and the United Kingdom are mutually closest peers, and coal-based electricity generation and net energy import dependence are the two leading SGPI priorities under all weighting schemes tested. These results are conditional on the profile definition: adding annual trends makes Hungary the closest peer, and excluding hydropower makes non-hydro renewable generation the second priority.

Energies doi: 10.3390/en19225430

Authors: Xutao Chen Xin Cai Shifa Lin Zhimeng Luan Saixian Bian

Trailing-edge adhesive voids are typical manufacturing defects that affect wind turbine blade stability. This study investigates the influence of the spanwise location and length of missing-adhesive voids on the buckling reliability of a 90 m composite wind turbine blade. BUCKLE eigenvalue analysis was used to determine the initial geometric imperfection mode, and the modified Riks method was employed to evaluate geometrically nonlinear buckling. For a void centered at 75 m with a length of 400 mm, the BUCKLE and Riks factors for the medium mesh were 0.4386 and 0.3822, respectively. For the prescribed defect cases, eight uncertain material and geometric inputs were considered. A full quadratic response surface including pairwise interactions was developed from an 81-point central composite design, and Monte Carlo simulation and total-effect Sobol sensitivity analysis were used to assess failure probability. For this blade and loading pattern, the sampled void centers at 75 and 80 m had relatively low capacities, while increasing void length reduced capacity and shifted conditional failure-probability curves toward lower load factors. These location-specific results support defect prioritization, not a universal critical region or a certified reliability level.

Energies doi: 10.3390/en19225427

Authors: Yifei Xu Xiaomei Huang Binji Wang

Hydrogen-enriched natural gas (HENG) is an important pathway for the low-carbon transition of urban energy systems. Understanding its leakage dispersion characteristics and hazardous gas cloud evolution is essential for residential safety management. In this study, a full-scale residential leakage experimental platform was developed, and a validated computational fluid dynamics (CFD) model was established. The effects of kitchen layout, hydrogen blending ratio, and leakage orifice diameter on indoor gas accumulation and hazardous gas cloud evolution were systematically investigated, with hazardous gas cloud volume adopted as the risk indicator. The results show that enclosed kitchens restrict gas migration and promote local accumulation near the leakage source, whereas open kitchens accelerate gas transport into adjacent rooms. Increasing the hydrogen blending ratio from 0% to 20% increased the hazardous gas cloud volume from 52.1 to 59.2 m3 at 3600 s, mainly due to the reduced flammability threshold and enhanced volumetric release characteristics of HENG. Leakage orifice diameter exhibited a more significant influence on hazardous gas cloud expansion; the hazardous gas cloud volumes were 37.4, 30.5, and 6.36 m3 for 10, 6, and 2 mm orifices, respectively, with the 10 mm orifice producing a hazardous region approximately 5.9 times larger than that of the 2 mm orifice. These findings provide quantitative insights for HENG leakage risk assessment, gas sensor placement, and emergency response strategies in residential buildings.

Energies doi: 10.3390/en19225429

Authors: Yuxin Wu Xueqian Fu

This PRISMA-informed structured review examines trustworthy deep learning for low- and medium-voltage distribution feeders and coordinated multi-feeder systems, from standard test feeders to utility networks; it does not assume one fixed bus count. Renewable generation, electric vehicles, flexible demand, storage, and digital monitoring create nonlinear, temporal, spatial, and probabilistic decision problems. We organize the evidence through a task–data–model–risk–validation pathway that links each operational task to its data structure, defensible model family, principal trustworthiness risk, and minimum deployment evidence. Common mathematical formulations provide shared evaluation criteria for forecasting, state estimation, reliability, resilience, control, and deployment. Reported results are interpreted only within their stated datasets and test conditions. Compact temporal models are defensible first-line baselines for moderate-horizon forecasting; topology-aware and physics-informed models are suited to tasks where network structure and feasibility matter; and deep reinforcement learning is a candidate for sequential control only when constraints and fallback behavior are tested. Exact database search dates and retrieval counts were not retained, so the review does not claim exhaustive systematic coverage or a quantitative meta-analysis.

Energies doi: 10.3390/en19225428

Authors: Bora Alboyacı Mehmet Aytaç Çınar Yunus Berat Demirol Rukiye B.Aymaz

The selection and sizing of grid components for energy distribution are essential for both the reliable operation of the network and its sustainability. Increasing harmonic components, driven especially by electric vehicle charging stations, energy storage systems, and microgrid structures, affect the operating conditions of the distribution grid. This study examines the behavior of cable and busbar systems commonly used for energy distribution, under high harmonic operating conditions. For this purpose, electromagnetic analyses were carried out using the finite element method. In this study, one busbar system and three different cable systems, each consisting of three parallel cables with approximately equivalent current-carrying capacity, were examined. In these analyses, the current distributions in the conductors of the cable and busbar systems, the magnetic field distributions around the conductors, and the losses in the systems were examined under current waveforms with different total harmonic distortion values. The results of the analysis show that busbar systems can be more advantageous than cable systems with respect to conductor current distribution and magnetic field levels. Furthermore, when examined with respect to losses, busbar systems can be more advantageous, particularly as harmonic levels increase.

Energies doi: 10.3390/en19225426

Authors: Yongqin Liang Xiuxiu Sun Haodong Zhang Teng Ma

The diesel ignition ammonia combustion model has become an effective method to solve the difficulty of ammonia combustion. However, the high NOx emissions of ammonia/diesel engine need to be further resolved due to the addition of diesel fuel. The simulation model of ammonia/diesel marine engine was established and validated in this paper. The effects of ammonia injection pressure, ammonia injection-rate profile, and pilot diesel injection timing on thermal and fuel NOx emissions were investigated. The results showed that increasing ammonia injection pressure can improve combustion characteristics but also increase NOx emissions, especially thermal NOx. The 150 MPa injection pressure can achieve a high score based on the results of multiple-criteria decision-making. The rectangular injection rate has great potential for improving thermal efficiency and reducing emissions, and it achieves the maximum score among all the proposed injection strategies. The appropriate injection timing of diesel achieves better combustion performance improvement and lower NOx emission, compared to the results of the injection pressure method. The results can provide theoretical guidance and methodological support for ammonia combustion in marine engines. Increasing the ammonia injection pressure from 50 to 250 MPa increased ITE by 7.6% and NOx emissions by 101.9%, while the thermal NOx fraction rose from 43.9% to 56.9%. The rectangular injection-rate profile achieved the highest MCDM score of 0.291.

Energies doi: 10.3390/en19225425

Authors: Grace Kiarie Oludolapo Olanrewaju

The demand for fresh water in coastal regions has increased the energy demand for desalination systems. This study models a hybrid solar PV, battery, and grid energy system for reverse osmosis desalination under realistic solar conditions in Mombasa, Kenya. An adaptive system that responds to real-time irradiance data is modelled in MATLAB R2025a using a five-year hourly solar irradiance dataset from NASA POWER, an internet database with renewable energy datasets. The analysis done determines the control and economic efficiency of using fuzzy logic control that prioritizes solar PV power, battery, and grid and ensures seamless transitions between energy sources. A well-tuned PID controller is compared with the fuzzy controller. The results show that the fuzzy logic hybrid system outperformed the PID controller, achieving near-autonomous reliance on solar PV and battery energy. Solar PV power provided most of the energy to power reverse osmosis, while the stored charge was effectively used during periods of low solar irradiance. The attained specific energy consumption for the reverse osmosis desalination process was 5.82 kWh/m3, which substantially minimized grid energy use, hence reducing operational costs. The findings highlight the importance of intelligent control in optimizing hybrid energy in desalination applications, offering a viable pathway to low-carbon, sustainable freshwater production in regions with high solar irradiance.

Energies doi: 10.3390/en19225423

Authors: Lei Zhang Hu Zhou An Tao Sen Hu Deping Ke Chong Han

Representative-day models can underestimate long-duration storage requirements because independent daily cycles fragment multi-day renewable shortages. This study proposes an event-enriched temporal aggregation method that retains load-weighted low-availability events as contiguous windows while keeping the modeled horizon fixed at 2160 h. Using German hourly wind, photovoltaic, and load profiles, capacity decisions are optimized on 2017–2018 data and evaluated chronologically on unseen 2019 data. Compared with conventional representative days, event enrichment increases the planned storage energy from 7.474 to 36.023 MWh and reduces its deviation from the 42.416 MWh full-chronology solution from 82.38% to 15.07%. In out-of-sample operation, energy not served decreases from 19.164 to 1.773 MWh, representing a 90.75% reduction, while annualized evaluation cost falls by 22.83%. Lower energy not served is obtained for all six clustering seeds and across the tested storage-cost range. Sensitivity analyses indicate that the benefit depends on the planning shortage allowance and on whether the historical record contains events relevant to the evaluation year. Therefore, retaining selected multi-day shortages improves storage-capacity fidelity and out-of-sample ENS performance without increasing the temporal budget.

Energies doi: 10.3390/en19225424

Authors: Huabo Shi Liang Wang Zanao Hu Hanghang He Pengyu Pan Jianyong Song Ning Jiao Shunliang Wang

Full-size converter variable-speed pumped storage units (FSC-VSPSUs) feature bidirectional power flow. They serve as critical equipment for enhancing regulation margins and promoting renewable energy integration. However, bidirectional energy flow elevates oscillations and instability risks. To address this issue, this paper establishes a complex-vector model of the FSC-VSPSU based on dc-bus decoupling and reveals the underlying mechanism of the internal frequency coupling effect. Comparing the coupling admittance of the FSC-VSPSU with that of a traditional grid-following converter (GFL) reveals that a pair of coupling peaks, absent in traditional converters, appear near the fundamental frequency of the FSC-VSPSU, with the admittance magnitude reaching twice that of the grid-following converter. The influencing factors of system stability under generating and pumping modes are comparatively analyzed using the impedance criterion. This indicates that the machine-side parameters exert a relatively minor impact on grid-connected stability. But the system presents a greater instability risk in pumping mode than in generating mode. Finally, the simulation results validate the correctness of the theoretical analysis presented in this paper.

Energies doi: 10.3390/en19225413

Authors: Li Ji Hoa Thanh Diep Li Yang

Artificial intelligence (AI) is increasingly reshaping the new energy industry, yet its relationship with enterprise resilience remains insufficiently understood. Using 54 Chinese A-share listed new energy enterprises from 2018 to 2023, this study evaluates resilience efficiency across production, storage, and application segments. A parallel two-stage dynamic directional distance function data envelopment analysis (DDF-DEA) model is developed to assess operational, innovative, and sustainable capabilities, with AI incorporated as an exogenous environmental variable. Econometric models are further used to examine the association between AI and resilience efficiency. The results show that overall resilience efficiency remains relatively low, with the production segment performing best and the storage segment worst. Sustainable capability is close to the efficiency frontier, whereas innovative capability constitutes the main constraint. AI-conditioned efficiency scores are generally higher than unconditional scores, with the largest difference observed in the application segment and primarily in operational capability. Regression results provide modest statistical support for a positive association between AI and overall resilience efficiency. After controlling for firm characteristics and firm and year fixed effects, the association remains statistically significant at the 10% level. These findings highlight the importance of differentiated AI strategies for strengthening resilience across new energy industry segments.

Energies doi: 10.3390/en19225422

Authors: Ke Liu Shuaishuai Nie Xiuping Zhong Dongbin Pan

The permeability of boundary layers in gas hydrate deposits is a critical factor influencing production performance, yet its impact has not been fully evaluated from a techno-economic perspective. This study developed a “reservoir simulation + cash flow analysis” evaluation framework to address this gap. Specifically, field-scale production simulations were performed to obtain gas-water production data under open- and closed-boundary conditions. A net present value (NPV) model incorporating a comprehensive range of costs, including drilling, production, operation, and maintenance, was then developed to calculate key economic indicators. The results demonstrate that development under open-boundary conditions is economically unviable, even at a high gas price of 5.4 RMB/m3, as the low gas revenues could not offset the substantial lifting energy expenditures, which represent the primary cost component. In contrast, development under closed-boundary conditions yields positive NPVs (21.8–124.1 million RMB), with a payback period of 3–7 years, an internal rate of return of 17.3–53.9%, and an allowable investment in boundary sealing of 20–85 million RMB. Additionally, increasing well spacing and reducing production pressure are identified as effective in enhancing economic performance; the optimal strategy for field development is the combination of well spacing > 150 m and production pressure < 3 MPa.

Energies doi: 10.3390/en19225421

Authors: Iván D. Hernández Jorge W. González Oscar Vasco-Echeverri Idi A. Isaac Hugo A. Cardona Gabriel J. López

Geomorphological stratification of the territory is rarely integrated explicitly into sizing models for hybrid solar–wind energy matrices, even though relief modulates both resources in opposite ways. This work proposes a four-stage model that combines geomorphological stratification by elevation ranges, resource-specific solar and wind characterization, a decision-tree surrogate model that reproduces the coupled physical simulation (PVWatts and PyWake) with high fidelity (R2 > 0.99), and optimization through mixed-integer linear programming (MILP). MILP is subject to a hybridization constraint that prevents a single technology from occupying more than 80% of the installed area. The model was applied to a first location with four altitudinal strata and contrasted with a second one. For the first site, the wind resource was non-viable across all strata, and the model consistently selected exclusively solar configurations. At the second location, it selected a hybrid configuration with 73% wind and 27% solar energy, whose proportions are governed by the hybridization threshold. Uncertainty and sensitivity analyses showed that the technology selected is robust to the choice of representative point and to the out-of-sample behavior of the surrogate, whereas the wind non-viability of the first site depends on the reference height and shear assumptions of the reanalysis data.

Energies doi: 10.3390/en19225420

Authors: Ammar Alammar

Phase change material (PCM) envelopes store latent heat only if the material melts during the day and solidifies at night, and elevated urban night-time temperatures can prevent this regeneration. Adaptive thermal coupling has been proposed as a regeneration strategy, but it remains unclear when it delivers genuine latent-storage benefits rather than acting as dynamic insulation. This study identifies the thermal bounds of that operating window for an office building in four Saudi cities, representing hot-arid (Riyadh), hot-humid coastal (Jeddah and Dammam), and highland (Abha) climates, and three urban morphologies of low, medium, and high density. Morphology-specific weather files from the Urban Weather Generator drove 1524 EnergyPlus simulations of static PCM envelopes. These were followed by phase-state diagnostics, adaptive-coupling scenarios compared against inert-layer controls (identical layers without latent heat), and a cooling-setpoint sensitivity analysis. The annual recharge margin (the PCM solidification temperature minus the mean summer nightly minimum of the urban air temperature) alone did not predict adaptive performance. Instead, useful operation was bounded by a regeneration floor set by night-time conditions and an activation ceiling tied to the cooling setpoint. In the cases tested, the upper limit of useful melting temperatures was about 2 K above the setpoint, and melting temperatures below it were also effective. Three regimes emerged. Under sink-limited conditions, coupling increased PCM cycling without meaningful savings. Within the latent adaptive window, coupling reduced the monthly cooling supplied by a median of 3.5–3.8% relative to the static PCM envelope at margins of 0 to +1.5 K and by less than 1% at −1.5 to 0 K. Of this reduction, 82–87% was attributable to the PCM and its interaction with the control. In Abha with a 24 °C layer, the annual reduction was 3.4–6.0% relative to the static envelope and 4.9–10.0% relative to the wall without PCM. At higher recharge margins, the PCM contribution diminished, and dynamic insulation dominated. In the summers of the three hot cities (Riyadh, Jeddah and Dammam), the regeneration floor lay above the activation ceiling, so no tested melting temperature was effective; cooler months opened a seasonal window. The resulting empirical framework indicates, for the conditions examined, when adaptive PCM envelopes provide useful latent storage under given urban microclimates and operating conditions.

Energies doi: 10.3390/en19225419

Authors: Muhammad Waseem Asghar Amina Shoukat Muhammad Safyan

The increasing integration of photovoltaic (PV) generation and electric vehicles (EVs) into distribution networks creates new challenges related to energy management, voltage regulation and network losses. Battery energy storage systems (BESSs) can provide additional flexibility by storing surplus renewable energy and supporting EV charging demand. However, the simultaneous sizing and placement of PV systems, BESSs and electric vehicle charging stations (EVCSs) constitutes a nonlinear mixed optimization problem involving discrete and continuous decision variables. This paper develops a hybrid swarm-intelligence framework for the coordinated planning of PV, BESS and EVCS infrastructure in a radial distribution network. The proposed Multi-Agent Adaptive-Acceleration Particle Swarm Optimization without Velocity and Grey Wolf Optimizer (MA-AA-PSWV-GWO) combines multi-agent search, adaptive acceleration, velocity-free particle swarm optimization and grey-wolf search mechanisms. The Baran–Wu 33-bus distribution network is considered using Saudi-oriented solar generation, electricity tariff and EV charging parameters. Four planning scenarios are investigated, ranging from fixed-location sizing to joint placement, sizing and 24 h BESS energy management. The proposed method is compared with PSO, MAPSO, GWO and AA-PSWV using a common objective function and network constraints. Over 10 independent runs, the proposed method achieves the lowest mean objective values and the standard deviations. The results show that the proposed hybrid framework provides competitive solution quality, consistency and feasibility for coordinated PV, BESS and EVCS planning.

Energies doi: 10.3390/en19225417

Authors: Mouna Nassim Mohamed Bezza

Data centers consume vast amounts of electricity, and cooling still accounts for a large share of it. Artificial intelligence is increasingly used to reduce this demand, but computing resources and cooling are usually optimized as two separate problems, a separation we call the IT–cooling divide. This review examines how AI is applied to cooling, to Information and Communication Technology (ICT) resources, and to the data center as a whole (Full-DC). We searched Scopus and Web of Science (last searched on 3 August 2026) for journal articles published in 2021–2025 that report quantitative results. We supplemented the database search with a targeted search of ScienceDirect and reference checking. We report the process following PRISMA 2020. As no formal risk-of-bias tool exists for simulation-based engineering studies, the validation level and baseline of each study were recorded, and results were synthesized narratively. The first author selected 45 studies (16 cooling, 21 ICT, 8 Full-DC) on the basis of titles and abstracts and read all of them in full for data extraction. Because selection was made by a single reviewer, the reported shares describe the reviewed sample rather than the whole literature; a reproducible re-screening of all 933 database records, checked double-blind by both authors, indicates that reinforcement learning is over-represented, and metaheuristics under-represented, in this sample. Reinforcement learning is the most frequent primary approach for control and deep learning for prediction. Energy results are heterogeneous: they use different baselines (PID control, heuristics, competing algorithms), are reported at different levels (cooling plant, IT servers, network, or facility), and come almost entirely from simulation, trace-driven, or offline evaluations; no included study reports closed-loop deployment in production. Among the studies reporting an energy reduction in percent, the median value is about 9% for cooling (n = 7) and 13% for ICT (n = 7), and the two Full-DC studies report 16–17%. Only two included studies explicitly co-optimize computing and cooling decisions. Joint IT–cooling optimization therefore remains a promising but thinly tested direction, and progress depends on facility-level reporting, real-world validation, and open datasets. The review was not registered and received no external funding.

Energies doi: 10.3390/en19225418

Authors: Yan Wu Hui Yu Zi-Cheng Wang Bao-Shen Xu Yong-Hu Zhang Xue-Li Ma Ying-Ying Zhang Cheng-Yao Chen Nan Li

Traditional chemical absorption-based CO2 capture processes suffer from high consumption of regeneration energy, uneven temperature distribution inside the desorption column, and low waste heat utilization efficiency of lean solvent. With the aim of creating a 210,000 ton/year CO2 capture system for flue gas from coal-fired boilers, this study establishes a process model for CO2 capture using a blended methyldiethanolamine (MDEA)–piperazine (PZ) amine solvent via Aspen Plus software and systematically optimizes key parameters of the rich–lean heat exchanger and absorbent circulation rate. To address the bottleneck of overheating at the bottom and undercooling at the upper section of the desorption column, an interstage reheating process based on lean solvent waste heat recovery is proposed. The influences of the semi-lean solvent extraction position, extraction flow rate, and reinjection position on system energy consumption and hydrodynamic characteristics inside the column are comprehensively investigated. After optimization, the proposed interstage reheating process reduces energy consumption by 10.16% compared with the base MDEA-PZ process and achieves energy savings of 29.25% relative to the conventional monoethanolamine (MEA) process. This research provides theoretical foundations and technical data support for the engineering design and operational optimization of low-energy-consumption CO2 capture and desorption systems.

Energies doi: 10.3390/en19225416

Authors: Afonso António José Pombo Sílvio Mariano Rosário Calado

This paper presents a prospective multivariate analysis of Angola’s electricity sector over the period 2024–2050, assessing the economic, demographic, and environmental impacts of three energy transition pathways: Baseline, Green Transition, and Restricted Development. Using energy system optimization models complemented by parametric robustness tests, the study projects a rapid increase in electricity demand, driven by population growth, with Angola’s population expected to reach approximately 86 million by 2050. Under the Green Transition Scenario, electricity demand reaches 283 TWh, corresponding to a per capita consumption of 3350 kWh. Although large-scale hydropower remains the backbone of the electricity system, accounting 37–51% of total generation, solar photovoltaics emerge as the leading technology for capacity expansion, accounting for up to 33% of installed capacity and achieving the lowest levelized cost of electricity, ranging from 19.2 to 25.5 EUR/MWh. By contrast, diesel generation and new hydropower plants exhibit substantially higher generation costs. In isolated rural grids, diesel–solar hybrid systems reduce electricity generation costs by more than 50% while significantly lowering carbon intensity. Sensitivity analysis shows that overall system costs are strongly influenced by the discount rate and macroeconomic assumptions but remain comparatively robust to technological uncertainties. These findings highlight the importance of large-scale solar deployment and decentralized hybrid systems as key strategies for achieving a cost-effective, low-carbon electricity transition in Angola.

Energies doi: 10.3390/en19225415

Authors: Anshu Choudhary Partha Sarathee Bhowmik Akshay Kumar Saha

This work develops a hybrid sensorless control scheme for a two-level voltage-source-inverter (VSI)-fed interior permanent magnet synchronous motor (IPMSM) operating under constant-switching-frequency direct torque control (DTC). It combines a stator-current model-reference adaptive control (MRAC) estimator with high-frequency signal injection to secure reliable speed and position estimation over the entire operating range—including standstill and low speed, where back-EMF (electromotive force)-based MRAC breaks down. The MRAC estimator employs a parallel adjustable stator current model driven by a cross-product-error PI adaptation law for accurate tracking from medium to high speed, while at low speed, a high-frequency (HF) carrier superimposed on the reference flux exploits rotor saliency independently of the back-EMF magnitude. The principal novelty is a hysteresis-based, debounced switching logic that hands control authority between the two estimators without disturbance; it is exercised at staged low-speed set-points (20 and 30 rad/s) before transferring to the rated 52.38 rad/s reference, yielding bump-free transitions free of sustained torque or speed error. The estimator is embedded in a modified six-sector switching-table DTC that uses a carrier-intersection torque comparator and active–passive vector pairing at a fixed 2 kHz switching frequency, together with a maximum-torque-per-ampere (MTPA) flux-reference generator that curtails copper loss. The scheme is verified by processor-in-the-loop (PIL) testing on a TI F280039C C2000 LaunchPad at a 50 µs control step, with the controller compiled through Embedded Coder and run in closed loop against a Simulink plant model at a 150 V DC link, confirming stable torque regulation, phase currents contained within a ±20 A envelope, and practical deployability for encoder-free EV (electric vehicle) traction.

Energies doi: 10.3390/en19225414

Authors: Oktawia Specht

This study quantitatively evaluates the effect of the spatial resolution of a digital surface model (DSM) derived from unmanned aerial vehicle (UAV) photogrammetric data on rooftop photovoltaic potential assessment. A baseline DSM with a spatial resolution of 1.26 cm was used to generate six DSM variants with cell sizes of 2.5, 5, 10, 25, 50, and 100 cm. For each variant, roof slope and aspect were derived, annual solar irradiation was modelled, and the inclined roof area suitable for photovoltaic (PV) installation (APV), potential PV capacity (PPV), and predicted annual electricity generation (EPV) were determined. The analysis was conducted for 29 buildings by comparing the results obtained from each DSM variant with those derived from the baseline DSM. For the baseline variant, APV was 995.17 m2, PPV was 199.03 kWp, and EPV was 135,900.09 kWh/year. The smallest differences were observed for the 2.5 and 5 cm DSM variants, while the 10 cm variant also showed relatively close agreement with the baseline. For DSM variants with cell sizes of 25 cm or larger, building-level differences generally became more pronounced. Roof aspect showed particularly high sensitivity to DSM spatial resolution, with the mean absolute error (MAE) of whole-building circular mean aspect relative to the baseline DSM increasing from 6.43° for the 2.5 cm variant to 56.56° for the 100 cm variant. The median absolute relative difference increased from 0.83% to 22.86% for APV and from 0.91% to 26.81% for EPV. For the 100 cm variant, the aggregated values of APV and EPV were lower than the baseline values by 6.35% and 4.48%, respectively. The results indicate that DSM spatial resolution affects all analysed stages of rooftop photovoltaic potential assessment, while aggregated values may not fully reflect the differences observed between individual buildings. Under the baseline atmospheric assumptions, the 10 cm DSM was the largest tested cell size satisfying the adopted screening criterion, with median absolute relative differences of 5.24% for APV and 4.86% for EPV. However, the atmospheric sensitivity analysis showed that the largest qualifying cell size varied from 2.5 to 25 cm, indicating that DSM resolution selection depends on the adopted modelling assumptions.

Energies doi: 10.3390/en19225412

Authors: Dorota Makowska Marcin Sajdak Roksana Muzyka Ewa Kadela André L. Boehman

Thermal conversion offers opportunities for energy recovery from wastewater sludge, but its feasibility depends on feedstock properties. This study evaluates industrial sludges from food processing, refining and coking, alongside municipal sewage sludge, in comparison with coal, biomass and refuse-derived fuel. Proximate and ultimate analysis, ash composition and mercury determination were combined with slagging/fouling indices, preliminary ISO 21640 classification and chemometric analysis. Correlation and cluster analyses revealed marked differences among the sludge streams and showed that their comparison with reference fuels depends on data preprocessing. Elevated chlorine and sulfur contents in selected industrial sludges indicated potential corrosion and operational constraints. None of the samples met the low-tendency criteria of all eight slagging/fouling indices, although these indices require experimental validation for sludge. On a dry basis, eight out of 30 samples (27%; six refinery and two food-processing samples) fell within ISO 21640 classes 1–3 for net calorific value, chlorine and mercury; at measured moisture contents, none reached class 3 or better. The findings indicated that selected industrial sludge streams could serve as alternative fuels or fuel-blend components after appropriate drying, provided that energy requirements, contaminant contents, and process-specific operating risks are considered.

Energies doi: 10.3390/en19225411

Authors: Yi Wan Zheke Che Meiqi Zhang Donghua Yu

Electricity load forecasting is essential for the optimal operation of smart grids and effective energy scheduling. However, increasing uncertainties caused by complex power system operations, renewable energy integration, and changes in user behavior have made load series more nonstationary and nonlinear. Traditional forecasting methods and existing deep learning models still have limitations in modeling such complex nonlinear relationships. To improve the nonlinear representation capability of short-term load forecasting models, this paper proposes a CNN-KAN framework based on heterogeneous feature mapping. The proposed method first employs a convolutional neural network to extract local temporal dependency features from multivariate historical load sequences. A Kolmogorov–Arnold Network is then used to perform adaptive nonlinear function mapping. In addition, a recursive smoothing soft B-spline basis mapping mechanism is introduced to improve the continuity and adaptability of the basis representation in KAN. Experiments are conducted on the Tetouan City electricity load dataset under two forecasting scenarios, namely Quads and Combined. The model is evaluated using MSE, RMSE, MAE, R2, and MAPE. Experimental results show that CNN-KAN achieves stable forecasting performance in both scenarios. The R2 values exceed 0.99, while the MAPE values remain below 1%. The proposed model demonstrates competitive overall forecasting performance compared with the Decision Tree, LSTM, TCN, and Informer baselines across the two forecasting scenarios. These results demonstrate the effectiveness of the proposed CNN-KAN heterogeneous feature mapping framework for short-term electricity load forecasting.

Energies doi: 10.3390/en19225410

Authors: Xindi Xu Siyuan Dong Chaofan Wang Qinyun Wang

Understanding the policy drivers of green power growth is vital for advancing energy transition and achieving carbon neutrality. This study examines the impact of China’s Carbon Emissions Trading Pilots (CETP) on the development of wind and solar power using county-level data from 2005 to 2017. The results show that the CETP leads to an additional 66.7% increase in installed capacity and a 79.7% increase in power generation in the pilot counties compared with non-pilot counties. Further analyses reveal that this growth primarily results from the substitution effect between coal-fired and green power induced by the CETP. Specifically, this substitution effect is driven mainly by the increasing marginal cost under the benchmark allowance allocation based on actual power generation, rather than by the carbon price level. Hydropower attenuates the substitution effect by serving as an alternative power source, and coal consumption amplifies this substitution effect. The findings offer significant policy implications in CET design and enforcement to promote green power growth.

Energies doi: 10.3390/en19225409

Authors: Abzal Seitov Amankeldy Toleukhanov Yerzhan Belyayev Aiymzhan Baitureyeva Tangnur Amanzholov Olivier Botella Abdelhamid Kheiri Müslüm Arıcı

This study investigates the charging of a vertical latent heat thermal energy storage unit containing a single coaxial paraffin container. Measurements for smooth and internally finned containers are compared with transient conjugate simulations that account for forced and buoyancy-driven flow in water and molten paraffin. Latent heat is represented by the apparent-heat-capacity method, and motion in solidifying material is suppressed by a mushy-zone resistance term. Modeled temperatures agree with four experimental sensor records, with root mean square errors of 0.46–0.96 °C and coefficients of determination of at least 0.996. At fixed container height, increasing the PCM diameter from 40 to 80 mm increases the PCM mass fourfold and the calculated time to 99% liquid fraction from 166 to 419 min. For a 40 mm container, increasing inlet velocity from 0.005 to 0.05 m/s reduces that time from 244 to 95 min. At approximately equal PCM volume, the model predicts that annular and eight-ray star containers shorten charging by 41% and 39%, respectively, compared with a smooth cylinder. The two designs are practically equally fast (98 and 101 min), but the star container requires 1.9 times as much steel as the annulus. These results identify geometries that shorten charging within the investigated operating range; charging time is evaluated separately from full-system energy and exergy efficiency.

Energies doi: 10.3390/en19225408

Authors: Libing Tao Shaoyu Chen Jun Chen Yuanchao Guo Zihao Ye Jianan Chen Yankai Xing

Conductor galloping threatens the safe operation of overhead transmission lines, motivating non-contact methods for reconstructing its motion characteristics. However, recovering conductor displacement from electric-field measurements near the suspension point requires resolving the coupled effects of deflection and sag. This study develops an electric-field-based reconstruction method that combines local angle inversion with an arc-length constraint. A three-dimensional conductor model relates the tangential deflection angles to multipoint electric-field observations, and relative amplitude measurements provide the inputs for local Jacobian inversion. The reconstructed angles are combined with prescribed spatial modes and the inextensibility constraint to obtain horizontal and vertical displacement components while accounting for sag variation. Spectral analysis of the reconstructed motion sequences extracts the dominant frequencies. The observation model is further extended to three-phase electric-field superposition with known relative phase-voltage magnitudes and phases. Controlled single-conductor laboratory tests evaluate displacement reconstruction against scale references, while separate known-input numerical comparisons assess angle, displacement, and frequency recovery. Error analyses examine the effects of measurement perturbations, geometric uncertainty, and observation sensitivity on reconstruction. The results support recovery of the overall imposed motion patterns while identifying direction-dependent deviations and sensitivity to low-amplitude inputs. The method provides a basis for reconstructing conductor galloping characteristics from local electric-field measurements under specified geometric and measurement conditions.

Energies doi: 10.3390/en19225407

Authors: Alexis Pérez-Fargallo Mario Jiménez-Torres Carlos Rubio-Bellido Jeremy Piggot-Navarrete Jesús Alberto Pulido-Arcas

Climate zoning is widely used to establish building-envelope requirements and estimate residential heating and cooling demand. However, regulatory classifications based primarily on outdoor climatic variables may not adequately represent the diversity of indoor temperature patterns observed under actual building operation. This study assesses the correspondence between regulatory thermal zoning and measured hourly indoor air temperature profiles, using Chile as a case study. Indoor air temperature was monitored at 5 min intervals in the living rooms and bedrooms of 116 occupied dwellings distributed across 14 cities representing eight of Chile’s nine regulatory thermal zones. The measurements were aggregated into hourly profiles and analyzed separately according to room, season, and annual period using k-means clustering. Complementary linear mixed-effects models were used to assess associations between cluster membership and household energy consumption. Two characteristic clusters were identified for each room and analysis period. Their spatial distribution did not exhibit a one-to-one correspondence with the regulatory thermal zoning: geographically distant zones frequently displayed comparable indoor air temperature profiles, whereas dwellings located within some individual zones were assigned to different clusters. The clearest separation between profiles occurred during winter, when the coldest clusters reached mean early-morning air temperatures of approximately 14 °C in both living rooms and bedrooms. The energy models did not identify statistically significant consumption differences between clusters; marginal R2 values for the per-capita models ranged from 0% to 2.2%. These findings do not imply that indoor temperatures are independent of outdoor climatic conditions. Rather, they demonstrate that regulatory climate zoning alone is insufficient to represent the variability of indoor air temperatures observed under occupied residential conditions. Although the available measurements do not support formal assessments of operative temperature, thermal comfort, occupant behavior, or energy poverty, the identified data-driven profiles provide an empirical basis for refining indoor-temperature assumptions, developing more realistic baseline scenarios, and supporting the calibration and validation of residential building energy models across different climatic and regulatory contexts.

Energies doi: 10.3390/en19225406

Authors: Dariusz Siudak

This study investigates the short-term impact of tightly timed escalation shocks in the 2024 Hezbollah–Iran–Israel conflict on abnormal returns across financial and commodity markets, with particular emphasis on identifying conflict-beneficiary assets. Using a benchmark-adjusted event study analysis, we examine logarithmic returns for 80 assets across four groups–safe-haven assets, energy commodities, energy-sector equities, and the S&P 500 index–based on a 120-day estimation window and two exogenous escalation events on 18 September and 1 October 2024. The findings indicate that energy-sector equities exhibit conflict-beneficiary behavior in response to the examined geopolitical shocks: both events generate statistically significant positive abnormal returns, which accumulate over short horizons and subsequently attenuate, with the second event producing stronger and more persistent effects. In contrast, safe-haven assets, energy commodities, and the broader equity market display no systematic abnormal performance. These results remain stable across robustness checks, including non-parametric inference, shortened estimation windows, and MANOVA, which confirms significant differences in abnormal-return patterns across asset classes. The study contributes to the literature by integrating heterogeneous asset classes within an ESA framework, employing two plausibly exogenous geopolitical escalations, and formally documenting differences in market responses across asset classes. The results suggest that, under the specific geopolitical conditions examined in this study, energy-sector equities may exhibit a conditional defensive role consistent with conflict-beneficiary behavior, providing insight into how markets reprice risk during conflict-driven disruptions. These findings have practical implications for hedging and tactical allocation under closely spaced geopolitical shocks, highlighting the importance of identifying assets whose payoffs mechanically improve under conflict conditions and cautioning against unconditional reliance on traditional safe-haven assets.

Energies doi: 10.3390/en19225405

Authors: Zhao Liu Guangli He Chufu Li Wenjing Zhao Wenping Dong Xuan Li Bona Lu

Hydrogen, as a distinctive energy carrier, offers significant advantages in clean energy transition yet faces numerous technical challenges. To address limitations in ultra-high-speed rotating machinery applications in hydrogen environments, this study proposes a passive self-levitating permanent magnet support system (PPMSS). A comprehensive simulation model was established to analyze structural integrity and magnetic levitation force, integrated with Design of Experiments (DoE) methodology to develop a data prediction model for permanent magnet rotor structural strength with maximum deviation below 2.19%, significantly enhancing design efficiency. Optimization yielded optimal PPMSS parameters, achieving 15.917N support force with merely 0.917N initial comprehensive axial force. Finally, the magnetic force simulation results were validated through quasi-static experimental measurements, while the structural performance at 88,000 RPM was verified via theoretical thin-walled ring analysis. Full-assembly high-speed rotor-dynamic stability has not yet been experimentally demonstrated and is reserved for future work. The PPMSS employs magnetic forces for rotor suspension without external energy input or medium intervention, fundamentally eliminating mechanical friction and wear while featuring maintenance-free operation and extended service life, demonstrating promising applications in hydrogen energy industry development.

Energies doi: 10.3390/en19225404

Authors: Ahmed Hathat Saadi Abdelkader Abdelaziz Rabehi Mawloud Guermoui Mostefa Hamdani Mustapha Habib

Photovoltaic (PV) power forecasts are typically evaluated using aggregate error metrics computed over an entire test period, which may conceal condition-dependent accuracy differences. We evaluate whether forecasting error varies systematically with weather regime, ramp events, and regime transitions, applying an evaluation layer to existing, frozen forecast outputs from two independent PV datasets (SOLETE, Denmark; Alice Springs (DKASC), Australia) across 50 prediction cells and five forecast horizons (5–180 min). Regimes, ramps, and transitions were defined from training-period-only thresholds, with forecast-time context kept separate from retrospective conditioning; differences were assessed using episode-level Mann–Whitney U tests on contiguous condition episodes, with Holm–Bonferroni correction and bootstrap confidence intervals. Significance counts are conditional on the episode definition, and residual dependence between episodes was not modeled. Transition-conditioned differences were statistically significant in 45 of 50 cells, with higher error inside transition windows in 43 of them, and forecast-time ramp context in 18 of 50, almost all with higher error under ramps. Forecast-origin regime differences were significant in most cells (46 and 48 of 50 for variable and highly variable versus stable conditions), but nearly all of these showed lower, not higher, episode-level error under the variable regimes. At the sample level, by contrast, simpler models on Alice Springs (DKASC) showed substantially higher error under variable conditions (relative mean absolute error (MAE) increases up to about 140% across horizons, and up to 227% in individual cells), while deep-learning models showed smaller, horizon-dependent effects; this contrast reflects rank-based episode-level versus mean-dominated aggregate comparison of a right-skewed error distribution. Under cross-dataset transfer, R3-versus-R1 degradation reversed direction relative to the in-domain CNN-BiLSTM pattern (R2-versus-R1 changes were already negative in-domain and differed only in magnitude), an open question that this study does not resolve. This study contributes a systematic, statistically supported evaluation protocol for weather-regime-, ramp-, and transition-conditioned PV forecasting accuracy, showing that aggregate metrics can obscure forecasting behavior relevant to operational reliability, and that the direction of a condition-dependent effect can depend on whether it is measured at the episode or the aggregate level.

Energies doi: 10.3390/en19225403

Authors: Abniel Machín Francisco Márquez

Sodium-ion batteries are emerging as a complementary electrochemical energy storage technology to lithium-ion batteries, particularly for stationary, grid-scale, and other cost-sensitive applications. However, assessing the practical relevance of newly developed sodium-ion materials remains challenging because early-stage electrochemical evaluation is still dominated by Na-metal half cells. These configurations are indispensable for establishing reversible capacity within defined voltage windows, redox behavior, kinetic characteristics, structural evolution, initial Coulombic efficiency, and electrode-specific interfacial processes, but they do not reproduce the finite sodium inventory, electrode balancing, realistic loading, shared electrolyte environment, or coupled aging that govern practical full cells. This review develops a translational framework for evaluating sodium-ion battery materials across successive levels of experimental realism. It first defines the appropriate role of half-cell testing and then examines the transition toward realistic composite electrodes and balanced full cells, with particular emphasis on mass loading, areal capacity, electrode density, negative-to-positive capacity ratio (N/P ratio), sodium-inventory management, electrolyte amount, full-cell energy, Coulombic efficiency, impedance evolution, and reproducibility. Interfacial chemistry is considered in terms of the transition from electrode-specific solid-electrolyte interphase (SEI) and cathode–electrolyte interphase (CEI) formation to chemically coupled full-cell behavior, including interelectrode cross-talk, transition-metal migration, electrolyte-derived species, gas evolution, and their consequences for sodium loss and impedance growth. Operando and in situ characterization are discussed according to an architecture–mechanism–diagnostic framework in which X-ray methods, vibrational spectroscopies, imaging and microscopy, impedance analysis, and multimodal approaches are matched to the degradation processes they are intended to resolve under increasingly representative cell conditions. Finally, a staged validation pathway is proposed, progressing from material-level half-cell characterization to realistic electrode validation, balanced full-cell testing, operando diagnosis, and prototype-level assessment. By distinguishing what each experimental stage can establish and which questions require more representative architectures, this review provides a framework for assessing whether favorable material-level performance can be translated into practical sodium-ion battery operation.

Energies doi: 10.3390/en19225402

Authors: Mubashar Javed Laiq Khan Amin Mahmoudi Solmaz Kahourzade

The inverse-time characteristic of each relay is a degree of freedom in the coordination of directional overcurrent relays (DOCRs) that most studies leave fixed at standard inverse, optimising only the time multiplier setting (TMS) and plug setting (PS). This paper examined how the IEC 60255-151 characteristic assignment affected the sum of primary operating times on one static DG-integrated benchmark, the 33 kV section of the IEEE 30-bus network with 38 DOCRs and 62 primary/backup pairs. Six assignment rules, including a deterministic rule that selects each relay’s characteristic from its plug-setting multiplier (PSM), were compared under a common optimisation framework, in which a fractional-order, entropy-guided particle swarm–gravitational search from the authors’ earlier work explored the plug settings and a monotone relaxation solved the TMS linear programming problem. All the rules achieved zero coordination violations, and two separate findings emerged. First, characteristic-family selection had a large effect: the objective fell from 29.42 s with uniform standard-inverse settings to 9.16 s with uniform extremely inverse (EI) settings, a reduction of 68.9%. Second, the PSM-driven rule gave a smaller additional improvement, reaching 8.94 s, about 2.4% below uniform EI and the lowest value among the six rules. The relaxation agreed with a general linear programming solver on all the tested candidates, and in a preliminary quantisation check the PSM-driven rule kept a 1.5–2.0% advantage over uniform EI at the same plug settings. The findings were limited to this objective, dataset and static network configuration; implementation on protection hardware was not assessed.

Energies doi: 10.3390/en19225401

Authors: Marcin Zieliński Joanna Kazimierowicz Marcin Dębowski

Microbial carbon dioxide (CO2) utilization can supply biomass and chemicals while reducing fossil feedstock demand. This critical review compares microalgae, cyanobacteria, acetogens, hydrogen-oxidizing bacteria, and microbial electrosynthesis for integration with energy systems. Evidence is organized by measurement endpoint, reactor scale, and operating duration within a common process boundary. Photosynthetic cultivation favors sites with natural light, land, and local nutrients; annual productivity, harvesting, and contaminant transfer constrain implementation. Engineered cyanobacteria offer target metabolites when culture stability and product recovery are demonstrated. Hydrogen-oxidizing bacteria favor protein or polymer production with low-emission H2, efficient gas delivery, and explosion protection. Acetogens suit anaerobic chemical production where reductant supply and product inhibition can be managed. Recent electrosynthesis results support reactor and recovery integration, while electricity demand, selectivity, and component durability determine whole-process feasibility. Laboratory experiments, pilot demonstrations, and sustained operation provide different levels of validation. Practical selection depends on complete energy inputs, product quality, and a defined market or treatment function. The climate benefit also depends on CO2 origin, the displaced product, and subsequent carbon release; biomass formation alone does not establish durable removal.

Energies doi: 10.3390/en19225400

Authors: Jun Xiang Jiabo Gou Wei Li Weiting Xu Ao Li Chang Liu

To address low-frequency oscillations in islanded residual-pressure power generation systems caused by water-hammer dynamics and low inertia, this paper proposes a supplementary damping control strategy for the DC-voltage outer loop of a voltage-source converter (VSC), using generator-speed deviation as feedback. First, an electromechanical control model is established, and its oscillatory characteristics are examined using root-locus and frequency-domain analyses. Then, with the supplementary controller disabled, a small-step signal is applied in PSCAD. TLS-ESPRIT is used to identify an equivalent second-order model between the supplementary control signal and generator electromagnetic power, and an independent small-step record is used for validation. Damping-torque theory is subsequently employed to determine the required phase compensation. A constrained particle swarm search is used to audit controller settings against frequency- and DC-voltage-related criteria, while nonlinear verification determines whether the candidate parameters should be adopted. The implemented controller comprises a low-pass filter, a washout stage, and a single-stage lead–lag network. In a unified three-level load check, the controller maintains stable operation for the added loads of 80 and 160 kW, whereas the 240 kW case went out of the normal operating range. In a separate reduced-order hydraulic-lag sensitivity test at 160 kW, the controller maintains bounded responses when Th2 is varied from 0.36 to 0.54 s. A matched benchmark against a conventional uncompensated supplementary branch shows that the proposed phase-compensated controller reduces frequency ITAE by 18.9% and limiter residence by 22.4% relative to the selected conventional case, although its DC-voltage ITAE is 67.4% higher. The results therefore support a frequency-recovery benefit within the tested range rather than a universal performance superiority.

Energies doi: 10.3390/en19225399

Authors: Ren Gu Chen Yang Min Ding Weike Zhu Shi Li Xuelei Jiang Weiguang Chang

Long rural low-voltage feeders are prone to end-of-feeder voltage deviations. Maintaining voltage support during transitions between grid-connected and islanded operation adds a further control challenge. This paper proposes a coordinated virtual synchronous generator (VSG) control method for grid-connected operation, islanding, and reconnection. The method links feeder length, impedance, and load power to the end-of-feeder voltage. It uses line-drop compensation (LDC) with a joint capacity/current-margin constraint to balance voltage support against converter limits. A six-state coordinator manages hybrid passive–active islanding detection, control takeover after grid-connection switch (QS) position feedback, islanded voltage support, and phase-locked-loop-free (PLL-free) presynchronization. It also defines enable, interlock, fallback, and smooth-withdrawal conditions for these functions. The method is evaluated in a switching-level model and on a 1.5 kVA three-phase three-level neutral-point-clamped (NPC) prototype. Simulations show voltage regulation across the tested feeder lengths and loads and reveal the support boundary as the capacity margin decreases. On the prototype, LDC reduces the end-of-feeder line-to-line voltage deviation from 2.42% to 0.21%. The load remains continuously supplied during the tested disturbances, presynchronization, and mode transitions. These results demonstrate coordinated voltage support and grid reconnection within the tested operating range.

Energies doi: 10.3390/en19225398

Authors: Michał Bilczak Stanisław Bielski

In Poland, the nearly zero-energy building standard required by the Energy Performance of Buildings Directive became binding at the end of 2020, before an energy crisis and a steep rise in interest rates. A difference-in-differences comparison of primary and secondary market prices for 2019–2025 shows a shift of 5.75 percentage points in their growth differential, but most of it occurred in early 2021, before dwellings built to the standard could be sold; the estimate falls to 2.68 points with a linear trend and to 1.13 on central-bank transaction data and loses significance once the delayed recording of new dwellings is allowed for. The euro area shows 3.03 points. The shift is a change in relative price growth, not an identified effect of the standard. On harmonised statistics, the mean absolute annual change in new housing loans in 2020–2024 was 32.7 per cent in Poland and 14.5 per cent in the euro area. A heat pump with a seasonal coefficient of performance of 2.5 breaks even with a gas boiler at 0.581 kg CO2/kWh, close to Polish end-user intensity in 2023 and 2024; at 3.5, it emitted less throughout 2020–2024. Electrification of heating and decarbonisation of the power mix are complements.

Energies doi: 10.3390/en19225397

Authors: Sylwester Bejger

The Temporal Fusion Transformer (TFT) is presented as an interpretable-by-design architecture, whose Variable Selection Networks (VSNs) and attention weights are routinely reported in applied work as structural findings. This paper asks what those mechanisms contribute to an economist’s understanding of energy price formation and how they compare with the variable importance measures of Random Forest Regression (RFR). The testbed is Polish wholesale fuel prices, governed by an import parity pricing mechanism whose benchmark hierarchy is documented in the empirical literature, so the expected structure is independently known; TFT is treated as an off-the-shelf model and read as the applied literature reads it. Conceptually, the product of VSN and attention weights constitutes a medium-scope explanation, intermediate between global importance rankings and local SHAP attributions, describing how a variable’s influence is distributed across the lookback window. Empirically, the reference forest concentrates on the refined-product benchmark by more than five times what its feature set alone would deliver, under MDI, MDA-OOB and SHAP alike; the averaged VSN does not depart from that baseline and does not recover the benchmark hierarchy that the mechanism implies, a failure attributable in part to softmax partial identification among correlated predictors. Re-estimation under many random seeds in two independent implementations reproduces exactly at a fixed seed but not across seeds, whereas the forest’s rankings remain stable. A single TFT estimation therefore supports no reproducible structural claim under the training budget used here, and averaging across seeds removes the dispersion without altering these conclusions.