Reinforcement learning for residential energy storage management at the neighborhood scale: A multi-benchmark evaluation

Energy infrastructure constitutes not merely technical hardware but a fundamental determinant of socioeconomic transformation and industrial evolution. While the engineering attributes of Ultra-High Voltage (UHV) transmission have been extensively documented, its role in shaping corporate behavior and catalyzing industrial upgrading remains underexplored. UHV transmission, operating at direct-current voltages of 800 kV and above or alternating-current voltages of 1000 kV and above, has been systematically deployed across China since 2007 to bridge the structural spatial mismatch between inland energy production centers and coastal consumption hubs. This study investigates the sociotechnical nexus between energy systems and firm-level innovation, employing China's large-scale UHV expansion (2007–2024) as an empirical setting with implications for global energy transition strategies. Integrating energy economics with organizational behavior theory, we adopt a staggered difference-in-differences approach to examine how enhanced energy accessibility influences automation innovation. The findings demonstrate that reliable energy infrastructure functions as a catalyst for intelligent manufacturing, operating primarily through the restructuring of business environments via cost reduction, risk mitigation, and the realization of economies of scale. Importantly, the analysis reveals substantial heterogeneity: policy effects are amplified among state-owned enterprises and regulated industries, reflecting the interplay between institutional arrangements and technology adoption. Furthermore, this study uncovers a significant resource redistribution effect whereby firms in net electricity-importing regions derive disproportionately greater benefits, suggesting that strategic infrastructure deployment can compensate for regional resource disadvantages. By situating these findings within a broader analytical framework, this research offers actionable insights for policymakers seeking to leverage energy infrastructure investments as instruments for advancing industrial modernization and fostering sustainable business practices. • Ultra-High Voltage (UHV) expansion significantly drives corporate automation. • Cost reduction and economies of scale mediate the impact. • Electricity-importing regions benefit more from the grid. • Effects are amplified in SOEs and regulated industries. • Energy infrastructure serves as a tool for industrial upgrading.

Prompt-to-policy: Leveraging large language models to guide deep reinforcement learning in public health emergencies.

The critically ill patient requires multiple extractions through catheters, implying a substantial blood loss. The discard method continues to be a commonly used method in Intensive Care Units. However, there is no homogeneity in practice. Determine the minimum blood discard volume to obtain samples through radial arterial catheter. Quasi-experimental, prospective, cross-sectional study employing the STOBE checklist. Four consecutive arterial blood gases were performed considering the flush volume (sample A with 1 dead space discarding 1 mL, sample B with 2 discard space [2 mL], sample C with 3 discard space [3 mL] and control with 4 discard space [4 mL]). The study was approved by the Ethics and Pharmacological Research Committee of the Research Institute of La paz Hospital (Law No. 2023.564). It was conducted in the Burn Intensive Care Unit and the Resuscitation and Critical Care Unit of La Paz Hospital from September 2023 to June 2024. We found values outside the Clinical Acceptance Interval (CAI) in sample A versus control in pCO2 (47.42%), HCO3 - (60.71%), Haemoglobin (67.85%), Na2+ (7.14%), K+ (71.42%) and glycaemia (21.42%). In the case of lactate, no values outside the Clinical Acceptance Interval were found. When comparing sample B versus control, we found significant differences in HCO3 - (p = 0.027) and glycaemia (p = 0.001). However, none of the parameters studied presented values outside the Clinical Acceptance Interval. When the results were related to the Reference Value of Change (RVC), the pCO2 and Haemoglobin values presented 3.57% of values outside the Clinical Acceptance Interval, respectively, although they did not present statistically significant differences (p > 0.05). When comparing sample C versus control, no significant differences were found in any parameter, nor values outside the Clinical Acceptance Interval. The volume of discard required to obtain valid values is twice the volume used to purge the arterial catheterization system (2 mL) for obtaining samples for clinical use. The calculation of the minimum amount of discard in the collection of blood samples through the arterial catheter is an intervention at no extra cost that will reduce iatrogenic anaemia, reducing the need for transfusions, the complications associated with this therapy and the reduction of costs in critically ill patients.

• Bio-methanol hits fossil parity in 2030 s, staying 3-times cheaper than e-methanol. • 10% IRR requires e- and bio-methanol selling prices of >1713 €/t and 955 €/t. • FuelEU and EU ETS enable premiums for bio-methanol; e-methanol needs support. • Reliability of European renewable methanol supply remains policy-dependent. Renewable methanol is a key option for decarbonizing European industrial sectors, particularly the chemical industry and maritime transport. Despite growing project announcements, large-scale deployment remains limited due to high production costs and investment risks. This study assesses the techno-economic feasibility of renewable methanol production in Europe by comparatively evaluating e-methanol, bio-methanol and import pathways within a harmonized modelling framework. Levelized cost of methanol, net present value and internal rate of return calculations are applied across multiple regional scenarios for Germany, Sweden and Portugal, explicitly accounting for policy boundary conditions under the European Emissions Trading System and FuelEU Maritime. Across all scenarios, bio-methanol approaches cost parity with conventional methanol by around 2030 and consistently outperforms e-methanol in terms of cost levels and economic robustness. E-methanol remains substantially more expensive under current conditions, while imports become competitive only after 2040. In the German Baseline Scenario, selling prices of 1079 €/t for bio-methanol and 2169 €/t for e-methanol are required to achieve a target internal rate of return of 10 %. Although regulatory frameworks increase willingness to pay for renewable methanol in shipping, they are insufficient to enable low-risk e-methanol investments without further cost reductions. Overall, the results indicate that renewable methanol deployment in Europe remains structurally conditioned by cost competitiveness and input cost dynamics. Bio-methanol represents the most viable near-term pathway, whereas e-methanol depends on substantial hydrogen cost reductions and sustained policy support to achieve economic viability.

Coordinating multiple UPS batteries in datacenter for load flexibility and cost reduction

Application of thermal plasma technology for melting municipal solid waste incineration fly ash: simulation and experiment.

Factors and outcomes associated with early and deferred gastrostomy tube placement in acute ischemic stroke patients in the United States.

TRIAG: Tri-reinforced infused generative agents for financial risk compliance

Optimal multistage air gap membrane distillation (AGMD) for enhanced lithium extraction from medium- and low-quality brines.

Optimizing interoperable hydrogen supply chain design: A case study in Auvergne-Rhône-Alpes

In recent years, Multi-purpose Offshore Platforms (MPOP) have emerged as a novel solution to address the increasing global food and energy demand. Beyond the conceptual and qualitative analysis, this paper develops a hybrid quantitative framework to evaluate the life-cycle performance of co-located Wave Energy Converters (WECs) and offshore aquaculture (AQ) systems. The framework integrates hydrodynamic numerical simulations and probabilistic reliability analysis into a System Dynamics (SD) model to simulate complex subsystem interactions and quantify the system productivity and economic feasibility under environmental and operational uncertainty. A case study in Southern Tasmania demonstrates that the upstream WEC farm effectively reduces the incoming wave heights by up to 23%, mitigating aquaculture mooring tensions of the downstream salmon farm by 18%. This protection effect translates into an economic benefit that the co-located configuration achieves a 30.7% reduction in Life Cycle Costs (LCC) compared to the stand-alone configuration. Furthermore, the MPOP demonstrates the robust power capacity, ensuring a continuous off-grid power supply despite long-term component degradation and fluctuating aquaculture power demand. The results validate the MPOP concept as a commercially viable solution for sustainable blue economy development and provide a comprehensive simulation and decision-making tool for exploring future offshore multi-sector cooperation. • A unified framework evaluates performance and risk of co-located renewable systems. • Environmental and operational uncertainties are explicitly incorporated. • System interactions affecting energy output and reliability are identified. • An offshore case study demonstrates practical applicability.

• A nested optimization framework for marine hybrid propulsion systems is proposed. • The framework integrates optimal power management and component sizing. • Pareto-optimal solutions minimize system investment cost and GHG emissions. • Case study shows 27% less emissions or 6% investment cost reduction compared to requirement-based design. • The framework is tested for robustness to ship power request variations. The growing demand for decarbonization of the shipping sector calls for integrated design strategies that simultaneously address energy management and propulsion system sizing. This paper presents a nested optimization framework for designing ship hybrid propulsion systems that identifies the Pareto-optimal front balancing economic and Well-to-Wake environmental performance. The framework utilizes a multi-objective genetic algorithm (NSGA-II) in the outer layer to efficiently explore the design space for battery capacity and generator sizing. For each candidate design, an inner optimization layer determines the minimum achievable greenhouse gas emissions through optimal power resource management. This nested approach ensures that each point on the resulting Pareto frontier represents a design in which both sizing and operation are simultaneously optimized. The methodological accuracy is validated by benchmarking NSGA-II against an exhaustive grid search, while its effectiveness is demonstrated in a small ferry case study by comparing results with a standard requirement-based design (RBD) approach. The results demonstrate that the optimized framework can achieve up to a 27% reduction in emissions at the same investment cost, or a 6% reduction in investment cost at the same level of emissions, compared to the RBD baseline. A sensitivity analysis is conducted to assess the method’s robustness to realistic variations in power demand and to evaluate the impact of component cost fluctuations on the resulting Pareto frontier. The proposed optimization framework serves as a decision-support tool for ship designers, enabling them to make informed, consistent choices when designing marine hybrid propulsion systems. The proposed structure is generalizable to a wide range of vessel types and operational scenarios.

• Multi-resolution deep-learning model allows training using upscaled reservoir model. • 90 % reduction in training data generation cost using upscaled reservoir model. • Proposed workflow accelerates optimization by orders of magnitude. • CCS well control at Illinois Basin Decatur Project has been optimized. • Both surface injection and zonal rate allocation optimization are conducted. The injection of CO₂ into subsurface formations entails various risks, necessitating a comprehensive evaluation of several factors such as seismic activity, seal integrity, and CO₂ leakage. Therefore, a CO₂ sequestration project involves multiple objectives which potentially exhibit trade-offs. Traditional multi-objective optimization frameworks require hundreds of forward simulations, which are computationally prohibitive for large-scale field application. In this study, we propose a deep learning-based workflow to efficiently optimize well control in CO₂ sequestration projects including surface injection and zonal rate allocation, enabling scalability to large-scale field applications. We developed a multi-resolution deep learning model based on the Fourier Neural Operator (FNO), which provides super-resolution capability. This capability allows the model to be trained on coarse-scale simulations while accurately predicting fine-scale reservoir pressure and saturation responses from permeability and injection schedules. As a result, data generation costs are substantially reduced, significantly lowering the overall cost of developing deep learning models. The original FNO architecture was modified to improve predictive accuracy across spatial resolutions, resulting in the proposed multi-resolution FNO model. This model functions as a data-driven proxy integrated with a multi-objective genetic algorithm to optimize CO₂ injection control, effectively balancing pressure management and storage efficiency. The power and efficiency of our approach are demonstrated on both synthetic and field applications, including a large-scale CO₂ injection at the Illinois Basin Decatur Project. Application to the synthetic model demonstrates the superior predictive performance of the developed multi-resolution FNO across coarse to fine-scale properties. For the field application, coarse-scale training data reduces training data generation cost by 90%, while the FNO-based proxy accurately predicts fine-scale pressure and saturation distribution, which are verified against a commercial reservoir simulator. The multi-objective optimization workflow, implemented using the FNO-based proxy model, achieves substantial improvements across multiple objectives while delivering performance orders of magnitude faster than traditional simulation-based approaches. We applied this workflow to CO₂ sequestration scenarios, including balancing pressure buildup with CO₂ injection amount, as well as optimizing surface and zonal injection rate allocations. This work introduces a novel multi-resolution FNO-based proxy model, applied to CO₂ injector control optimization. By combining FNO’s super-resolution capability with coarse-scale models, training data generation costs are greatly reduced. The proxy model accelerates forward simulations by orders of magnitude and enables efficient evaluation of multiple optimization scenarios for large-scale field applications.

• Multi energy system (MES) including a CO 2 district heating–cooling network (DHCN). • Joint design optimization of MES and CO 2 network for cost and emission reduction. • User connections to CO 2 network are chosen only when system benefits arise. • Cooling comfort flexibility enables sustainable free cooling strategies. • Full DHCN connection reduces emissions by up to 48.8% compared to isolated users. This paper presents an original procedure for the integrated optimization of the topology, design and operation of a CO 2 District Heating and Cooling Network and associated Multi-Energy System. The goal is to understand the potential of CO 2 networks to reduce the life-cycle cost and greenhouse gas emissions of the system, by determining, in one step, both the optimal size of the energy conversion and storage units, and the optimal layout and size of the network branches. The procedure (i) uses a multi-nodal approach for an adaptive selection of the users to be connected to the network, (ii) includes the CO 2 temperature optimization to minimize energy consumption of the network, and (iii) allows different users’ comfort options. The optimization model, based on the DOMES method, is tested on an urban district with 18 nodes, considering different energy conversion and storage technologies, thermal networks and demand profiles. Results show that properly adapting the users’ connection to the CO 2 network may reduce emissions by 26.6 ÷ 52.8% compared to the case without thermal network, while the impact on total costs is strongly influenced by the users’ thermal demand profiles and cooling comfort needs, ranging from a 7.0% reduction to a 23.1% increase.

This study investigated whether external effort mobilization through try-harder instructions enhances performance in complex motor tasks, as reflected by the reduction of costs associated with producing a deceptive action. Basketball passing movements with and without head fakes were examined in a reaction-time paradigm. Participants were generally instructed to initiate the movement as fast and accurately as possible, while try-harder instructions were presented in 25% of trials, prompting participants to mobilize all their cognitive resources to perform even faster. To investigate if athletic expertise modulates the potential effects of effort mobilization, basketball novices and experienced players were tested. Results demonstrated that try-harder instructions generally improved participants' performance, facilitating faster response initiation times and movement execution, as well as a specific reduction in initiation time variability. Novices benefited more than experienced players, indicating that effort results in greater improvement when complex motor actions (for example, passes with head fakes) are not yet fully stabilized and lack automated fluency. This pattern suggests that effort enhances performance in complex actions when performance is limited by the amount of cognitive capacity available, supporting the coordination of partly conflicting movement components within tight temporal constraints. The findings extend previous research on effort mobilization from simple to complex motor tasks. Try-harder instructions appear to enhance performance primarily by reducing attentional lapses rather than generally improving processing speed. Future research should investigate the effectiveness of effort mobilization in experienced athletes in situations of high concurrent cognitive load.

The California Air Resources Board (CARB) Advanced Clean Fleets (ACF) regulation mandates a complete transition to zero-emission vehicle (ZEV) sales by 2042, posing significant challenges for the heavy-duty trucking sector. Fuel cell electric vehicles (FCEVs) are promising for long-haul freight due to their extended range and fast refueling. This study builds upon the Integrating Market Penetration and Cost Technologies (IMPACT), a closed-loop simulation framework used to evaluate the levelized cost of hydrogen (LCOH) and to identify the breakeven conditions under which FCEVs can achieve total cost of ownership (TCO) parity with diesel trucks. The model integrates vehicle cost evolution, market adoption dynamics, and hydrogen infrastructure deployment across a 2025–2050 horizon, projecting fleet sizes, hydrogen demand, and refueling station utilization. Under a mature-network benchmark (UR = 1.0) representing operation at rated design throughput, LCOH declines from $11.51/kg in 2025 to $4.26/kg by 2050. The breakeven threshold for 5-year TCO parity falls within $4.90–5.85/kg, intersecting modeled cost trajectories around 2040–2045 for heavy-duty fuel-cell trucks. By 2050, statewide hydrogen demand reaches 3,200 metric tons per day under the high-adoption scenario, requiring 516 high-throughput stations to refuel 145,000 heavy-duty FCEVs. These findings indicate that hydrogen cost convergence is primarily driven by utilization scaling effects within coordinated vehicle–infrastructure dynamics, rather than being solely attributable to production cost reductions. By quantifying time-varying breakeven LCOH thresholds, this study reframes hydrogen competitiveness as a systems-level scaling challenge in zero-emission freight transitions. • Develops the Integrating Market Penetration and Cost Technologies (IMPACT) framework. • Projects LCOH declining from $11.51/kg (2025) to $4.26/kg (2050) under mature utilization conditions. • Identifies a breakeven LCOH band of $4.90–5.85/kg for heavy-duty FCEV–diesel TCO parity. • Quantifies hydrogen demand of 3,200 metric tons/day supported by 516 high-throughput stations by 2050. • Demonstrates utilization-driven scaling as a key mechanism shaping hydrogen competitiveness.

To achieve affordable, clean energy, incorporating renewable energy into existing energy systems is the key. One challenge is the fluctuating nature of renewable resources, which can be asynchronous with energy demands. Hydrogen storage, particularly metal hydride storage, is a favorable solution for balancing supply and demand. In particular, metal hydride storage, compared with pressurized or liquefied hydrogen storage, is a favorable technology choice due to its storage energy density (50-100 kg H˙2/m 3 ) and its low operating temperature and pressure. This paper presents a simulation-based framework to investigate the optimal design and operation of a coupled Electrolyzer-Fuel Cell-Metal Hydride system (SET-Unit) for minimizing operational and capital expenses in a residential application. The results show that integrating heat pumps with a metal-hydride storage system and photovoltaics can achieve 83% energy self-sufficiency and a 7.1-year payback period. Combining SET-Unit, gas boilers, and photovoltaics can result in 28% energy self-sufficiency, annual savings of over 2221 EUR, and a payback period of 7.4 years. The SET-Unit, combined with renewable energy sources such as photovoltaics, and the in-market available gas boilers or heat pumps, shows benefits in efficiency, annual energy cost reduction, and a relatively short payback period for the household. Using the low end of published values for capital expenses, economic feasibility can be achieved. • Simulation-based framework to minimize metal-hydride storage systems costs. • SET-Unit and heat pump systems increase the self-sufficiency to 83%. • SET-Unit and gas boiler systems increase the self-sufficiency to 28%. • SET-Unit and heat pump systems can save 255 EUR in energy costs annually. • SET-Unit and heat pump systems have a payback period of 7.1 years.

Adaptive optimization of combined steam and CO2 reforming for hydrogen production from variable biogas feed.

• Comprehensive feasibility studies offer an in-depth analysis of any potential barriers to technology deployment and integration. • Successful deployment of VIV-EH enhances real-time water system monitoring by providing reliable power in remote areas. • High correlation between environmental impacts and energy generation output determined by the fluid velocity. • A solution like VIV-EH offers a low-impact renewable energy generation solution suitable for harnessing energy at low velocity. With today’s focus on the transition towards a cleaner energy future, interest is growing in exploring ways to utilise 10 TWh of hidden energy potential in EU water infrastructure. Vortex-Induced Vibration Energy Harvesters (VIV-EHs) offer an innovative approach to harnessing untapped hydropower potential in urban and rural water systems, thereby supporting the transition to green energy. These devices enable decentralised energy generation, support monitoring systems in critical infrastructure, and reduce reliance on fossil fuel-powered backups. However, their economic and environmental feasibility must be carefully assessed to ensure viable deployment and integration. This study introduces a Feasibility Assessment Framework to evaluate the technical, economic, and environmental aspects of VIV-EHs, with a particular focus on do-it-yourself (DIY) design. A case study in the Mestna Gradaščica River channel in Ljubljana, Slovenia, was assessed utilizing experimental data, computational modelling, and life cycle analysis. Two configurations of the DIY VIV-EH, one with a 49 mm cylinder diameter and another with a 61 mm diameter, were evaluated for energy output, costs, and emissions. The results demonstrate that the 49 mm configuration achieved a capacity factor of 94%, generating between 131.30 kWh and 406.8 kWh over a 12-year lifespan. In comparison, the 61 mm configuration produced between 164.7 kWh and 311.0 kWh, with greater stability across a variety of velocities. The Levelized Cost of Energy (LCOE) remains high, averaging 6.5 €/kWh, indicating potential for cost reduction through optimisation. Environmental impacts were moderate, with lifecycle emissions ranging from 0.071 to 0.221 kgCO2eq/kWh, depending on velocity and configuration. These findings demonstrate that overall VIV-EHs have promise in powering remote monitoring sensors, enhancing the resilience of water and energy systems, and reducing dependence on diesel generators. Future research should focus on enhancing device efficiency and minimizing manufacturing impacts to facilitate their wider adoption as a sustainable energy solution