Laura Schaefer

Predicting and interpreting thermal performance under oscillating flow in porous structures remains a critical challenge due to the complex coupling between fluid dynamics and geometric features. This study introduces a data-driven wGAN-LBM-Nested_CV framework that integrates generative deep learning, numerical simulation based on the lattice Boltzmann method (LBM), and interpretable machine learning to predict and explain the thermal behavior in such systems. A wide range of porous structures with diverse topologies were synthesized using a Wasserstein generative adversarial network with gradient penalty (wGAN-GP), significantly expanding the design space. High-fidelity thermal data were then generated through LBM simulations across various Reynolds (Re) and Strouhal numbers (St). Among several machine learning models evaluated via nested cross-validation and Bayesian optimization, XGBoost achieved the best predictive performance for the average Nusselt number (Nu) (R^2=0.9981). Model interpretation using SHAP identified the Reynolds number, Strouhal number, porosity, specific surface area, and pore size dispersion as the most influential predictors, while also revealing synergistic interactions among them. Threshold-based insights, including Re > 75 and porosity > 0.6256, provide practical guidance for enhancing convective heat transfer. This integrated approach delivers both quantitative predictive accuracy and physical interpretability, offering actionable guidelines for designing porous media with improved thermal performance under oscillatory flow conditions.

Achieving carbon neutrality requires effective strategies to reduce CO 2 emissions, and geological sequestration of CO 2 is considered among the most promising and economically viable options. The interfacial tension (IFT) between the CO 2 and the surrounding liquid (underground salt water or brine, NaCl) is a key parameter that affects the storage capacity of CO 2 in saline aquifers; however, the experimental measurement of IFT is often time-consuming, labor-intensive, and reliant on expensive equipment, and empirical correlations demonstrate a low level of accuracy. Machine learning (ML) techniques have been suggested as an alternative approach, and the current literature related to interfacial phenomena utilizes a wide array of basic and advanced ML models for predicting IFT, though often without a comparative analysis, raising the question of which model is most appropriate for this specific application. In this work, multiple machine learning models, including linear regression (LR), support vector machine (SVM), decision tree regressor (DTR), random forest regressor (RFR), and multilayer perceptron (MLP), are used to predict the IFT of the CO 2 and aqueous solution of NaCl. Models are trained using an experimental dataset that covers a wide range of temperature, pressure, and salinity (NaCl) conditions for CO 2 -brine IFT. Hyperparameter tuning algorithms are utilized to optimize each model, and the performance is evaluated using metrics such as mean absolute error (MAE) and mean absolute percentage error (MAPE). The best-performing algorithms are found to be SVM and MLP, with a MAPE of 0.97% and 0.99% and a MAE of 0.39 mN/m and 0.40 mN/m, respectively. The linear regression model demonstrated the worst performance with a MAPE of 4.25% and an MAE of 1.7 mN/m. The feature importance analysis reveals that pressure is the main parameter affecting the IFT. Our findings indicate a notable enhancement in prediction accuracy over previous ML studies in this area. Moreover, the results from this study suggest that even the basic ML models that were investigated, when properly tuned and optimized, are sufficient for accurate IFT predictions. This demonstrates that ML models offer a cost-effective and efficient alternative to experimental methods, potentially optimizing designs for CO 2 sequestration.

Abstract Waste heat recovery is an underutilized strategy for reducing emissions and mitigating climate change. High-temperature exhaust from gas turbines (GTs) is often not utilized, leading to both energy losses and emissions. This study proposes a novel WHR system that integrates a supercritical carbon dioxide (sCO2) Brayton cycle with a concentrated solar power (CSP) component, including thermal energy storage, for GT heat recovery. The WHR system boosts overall efficiency, sustainability, and economic performance by harnessing GT exhaust and solar-derived heat. One application of GTs is providing off-grid electricity to remote mining operations. A comprehensive techno-economic simulation model is developed, and a case study for a mining site in Western Australia is presented. The results show that the proposed CSP–sCO2 WHR system can provide an additional 56,028.78 MWh/year at a levelized cost of electricity (LCOE) of $0.0597/kWh, substantially increasing the output of the GT plant while reducing emissions and costs. Compared to a standalone CSP–Rankine plant, the proposed system halves capital expenditures, cutting LCOE by $0.0246/kWh. The proposed WHR system highlights the potential for reducing emissions and fuel costs in GTs, while the comparison with CSP–Rankine demonstrates how integrating CSP with waste heat can help achieve CSP LCOE targets.

The rapid expansion of digital infrastructure has led to an increase in the number and capacity of data centers, which are now significant sources of low-grade waste heat. Traditionally, the challenge of utilizing this waste heat stems from its ultra-low temperature, limiting the effectiveness of an Organic Rankine Cycle (ORC) based recovery system. This study introduces a novel approach to waste heat recovery by integrating flat plate solar thermal collectors (FPCs) with an ORC system specifically designed for data centers. By utilizing FPCs to preheat the waste heat from data center racks, the ORC’s thermal efficiency and overall system performance are significantly improved. A techno-economic analysis, based on a case study of Ashburn, VA, the region with the highest concentration of data centers in the U.S., demonstrates that solar thermal boosting nearly doubles power output, increases ORC efficiency by over 8% during peak hours, and reduces the investment per unit of electricity by 19.09%. These findings highlight the economic and environmental potential of solar-boosted ORC systems in transforming the energy efficiency of data centers.

This study analyzes the machine learning methods for predicting CO 2 -brine interfacial tension in saline aquifers, compares their performance, and highlights implications for carbon capture and storage applications.

Concentrated solar power (CSP) technology offers a carbon-free approach to power generation, providing a relatively stable and uninterrupted electricity supply compared to other renewable sources. The effectiveness and success of CSP plants are heavily dependent upon accurate forecasting of the direct normal irradiance (DNI) from the sun. Various DNI forecasting models have been proposed to this end; however, they rely on meteorological factors that are difficult to acquire. A recent study addressed the problem of DNI forecasting without meteorological data for the first time, formulating the problem as a multi-class classification task, and proposed multiple deep neural networks to model it. Extending this research, we propose Adarmer - a novel transformer model that automatically adapts its neural architecture to the complexity of the DNI estimation task. Our findings demonstrate that Adarmer surpasses the performance of previously proposed models across the board.

The escalating energy demand and the adverse environmental impacts of fossil-fuel use necessitate a shift towards cleaner and renewable alternatives. Concentrated Solar Power (CSP) technology emerges as a promising solution, offering a carbon-free alternative for power generation. The efficiency and profitability of CSP depend on the Direct Normal Irradiance (DNI) component of solar radiation; hence, accurate DNI forecasting can help optimize CSP plants’ operations and performance. The unpredictable nature of weather phenomena, particularly cloud cover, introduces uncertainty into DNI projections. Existing DNI forecasting models use meteorological factors, which are both challenging to estimate numerically over short prediction windows and expensive to model through data at a sufficiently high spatial and temporal resolution. This research addresses the challenge by presenting a novel approach that formulates DNI prediction as a multi-class classification problem, departing from conventional regression-based methods. The primary objective of this classification framework is to identify optimal periods aligning with specific operational thresholds for CSP plants, contributing to enhanced dispatch optimization strategies. We model the DNI classification problem using four advanced deep neural networks – rectified linear unit (ReLU) networks, 1D residual networks (ResNets), bidirectional long short-term memory (BiLSTM) networks, and transformers – achieving accuracies up to 93.5% without requiring meteorological parameters. • Motivates DNI threshold forecasting to optimize CSP plant operations. • Novel approach formulating DNI prediction as a multi-class classification for CSP. • Uses advanced deep learning models to forecast DNI accurately without meteorological data. • Compares proposed solutions, highlighting distinct approaches for DNI prediction. • Enhances CSP plant operations using machine learning techniques.

In 2021, the U.S. Department of Energy's (DOE's) Solar Energy Technologies Office (SETO) funded the formation of the Heliostat Consortium (HelioCon), a five-year consortium designed to advance U.S. heliostat technologies by engaging industry, subject matter experts, and general stakeholders for direct project-level collaboration, external consulting, and mission-specific panels and workshops. HelioCon is led by the National Renewable Energy Laboratory (NREL) and Sandia National Laboratories, in partnership with the Australian Solar Thermal Research Institute. This report describes HelioCon's activities and impact in fiscal year 2024.

Abstract The escalating global energy demand has resulted in a corresponding increase in greenhouse gas emissions. Multiple energy sources of various quality levels and forms must be developed to reduce emissions. One promising approach involves utilizing low-grade heat sources such as industrial process heat. This work investigates a novel combined heating and power approach through a comprehensive techno-economic analysis of a hybrid system that combines a CO2 heat pump (HP) and an organic Rankine cycle (ORC) for utilizing low-grade waste heat. The HP is powered by photovoltaics (PV), enabling sustainable heat extraction from low-grade sources, and is further integrated with thermal energy storage (TES) for demand management. The system primarily aims to supply district heating, but it seamlessly transitions to an ORC-based power generation mode in scenarios with reduced district heating demand. The integration of PV, HP, ORC, and TES allows for the efficient utilization of diverse low-grade heat sources, with adaptability to various operational strategies and applications contingent upon demand and supply dynamics. The study’s techno-economic and parametric analysis explores component and system-level performance, with Miami, FL, serving as a case study. Lastly, multiple operational scenarios and trade-offs are presented based on heating demand, power generation requirement, and resource availability.

Abstract This work examines a combined-component, fifth-generation district heating system (DHS) with an emphasis on CO2 emission reduction and greater adaptability to diverse heat sources. There are two primary contributions resulting from this analysis. First, a mathematical framework is created to simulate a combined photovoltaic (PV)-assisted CO2 heat pump (HP) with thermal energy storage (TES) to provide domestic hot water (DHW) for a district of 13 houses. Subsequently, this paper applies a mixed-integer nonlinear optimization approach to operating the system, employing a non-dominated sorting genetic algorithm (NSGA-II). The multi-objective optimization is performed to find the optimal trade-off between maximizing the coefficient of performance (COP) of the system and maximizing system self-sufficiency from a locally installed solar-PV system. Optimization is performed over 72 hours in the Fall, using Miami as a case study. The optimal time-resolved charging profiles and HP output water temperature as decision variables are extracted from the Pareto frontier. The results of the Pareto front show that when the system’s self-sufficiency goes up from 71% to 81%, the COP decreases slightly from 4.55 to 4.36. This means a 14% increase in self-sufficiency leads to a small 4.3% decrease in COP.

Efficient methods for decreasing emissions in the energy sector include enhancing the efficacy of coal-fired power plants, augmenting the utilization nuclear energy and gas for heat and electricity generation, diversifying the application of renewable sources of energy, and allowing to consume energy in a rational manner. By adopting renewable energy sources, we not only get environmental benefits, but also strengthen the state's autonomy in fuel and energy exports. This leads to savings in foreign money and the establishment of new job possibilities. This article examines the potential, as well as limitations and challenges, of renewable sources of energy, such hydrogen, solar, wind, and biogas. This article research different factors influencing the economic viability of renewable energy potentials, including their geographical, technical, and technological characteristics. The essay also examines the benefits and drawbacks of generating biogas, wind energy, and solar thermal energy. The conclusion underlines the need of careful planning, site selection, and environmental studies to ensure the successful integration of renewable energy into the existing power system.

Abstract Due to growing concerns about the environmental impact of refrigerants, carbon dioxide (CO2) heat pumps have been increasingly evaluated as efficient alternatives for conventional heat pumps. Performance analyses of CO2 heat pump water heaters (HPWHs) have been the subject of many studies, but these are typically limited to parametric analyses of air-source HPWHs. The interrelated behavior of the supercritical and subcritical thermodynamic properties, component operation, and efficiency means that a parametric study cannot adequately capture the inherent nonlinearity. Therefore, this paper, for the first time, aims to perform a multi-objective optimization on CO2 water-sourced HPWH performance in order to minimize the total component costs, maximize gas cooler (GC) heating capacity, and maximize the coefficient of performance (COP) using two different optimization scenarios. The decision variables are defined as GC pressure (75–140 bar), evaporator temperature (−19.5–0.2 °C), and GC outlet temperature for CO2 (16–36 °C). The model performance is constrained by the practical ranges of the GC and evaporator inlet and outlet temperatures for water. A coupled simulation-optimization model through python is developed using Engineering Equation Solver (EES) software and the non-dominated sorting genetic algorithm II (NSGA-II). The result of the optimal Pareto front showed that the optimal GC heating capacity changes from 19.2 to 56.7 kW, with a lowest cost of $7771 to a highest cost of $9742, respectively. When the lower bound of the GC outlet temperature was set to 32 °C, the Pareto front showed a maximum COP of 3.23, with a corresponding GC heating capacity of 44.36 kW.

Abstract Due to growing concerns about the environmental impact of refrigerants, CO2 heat pumps have been increasingly evaluated as efficient alternatives for conventional heat pumps. Performance analyses of CO2 heat pump water heaters (HPWHs) have been the subjects of many studies, but these are typically limited to parametric analyses of air-source heat pumps. The interrelated behavior of the supercritical and subcritical thermodynamic properties, component operation, and efficiency means that a parametric study cannot adequately capture the inherent nonlinearity. Therefore, this paper, for the first time, aims to perform a multi-objective optimization on water-sourced HPWH performance in order to minimize the total component costs, maximize gas cooler (GC) heating capacity, and maximize the coefficient of performance (COP) using two different optimization scenarios. The decision variables are defined as GC pressure (75 to 140 bar), evaporator temperature (−19.5 to 0.2°C), and GC outlet temperature for CO2 (16 to 36°C). The model performance is constrained by the practical ranges of the GC and evaporator inlet and outlet temperatures for water. A coupled simulation-optimization model through Python is developed using Engineering Equation Solver (EES) software and the non-dominated sorting genetic algorithm II (NSGA-II). The result of the optimal Pareto front showed that the optimal GC heating capacity changes from 19.2 to 56.7 kW, with a lowest cost of $7,771 to a highest cost of $9,742, respectively. When the lower bound of the GC outlet temperature was set to 32°C, the Pareto front showed a maximum COP of 3.23, with a corresponding GC heating capacity of 44.36 kW.

The interest in concentrated solar power (CSP) has increased significantly over the years since it is dispatchable and requires thermal storage instead of electric storage. When compared to photovoltaics (PV), CSP has a higher Levelized Cost of Electricity (LCOE). In this paper, we present the design of a hybrid power plant using CSP and PV technologies. The hybrid system offers a lower LCOE than CSP but will still have dispatchability. The 150-MW hybrid system consists of 100-MW CSP and 50-MW PV capacity. Furthermore, the CSP system (central receiver system) has a molten salt-based thermal storage of 12 hours. The geographical focus of this study is Pakistan, which is a developing country struggling with energy crises, but which has high solar potential. The hybrid system is modeled using the System Advisor Model (SAM). The results show that by hybridizing CSP with PV, the LCOE can be reduced by 18.5%.

According to the Food and Agriculture Organization (FAO) of the United Nations, it is projected that the global population will increase by an additional 2 billion individuals by the year 2050. However, the FAO also predicts that only a mere 4% of the Earth's total surface area will be utilized for agricultural purposes. Advancements in technology and innovative solutions to existing limitations in the agricultural sector have facilitated a notable enhancement in agricultural efficiency. The extensive utilization of machine learning and Artificial Intelligence (AI) within the agricultural industry may potentially signify a significant turning point in its historical trajectory. The utilization of AI in farming presents a range of benefits for farmers, including enhanced productivity, reduced expenses, improved crop quality, and expedited go-to-market strategies. This study aims to explore the potential applications of AI in various subsectors of the agriculture industry. This study delves into the exploration of future concepts propelled by AI, while also addressing the anticipated challenges that may arise in their applications.