Understanding current energy consumption behavior in communities is critical for informing future energy use decisions and enabling efficient energy management. Urban energy models, which are used to simulate these energy use patterns, require large datasets with detailed building characteristics for accurate outcomes. However, such detailed characteristics at the individual building level are often unknown and costly to acquire, or unavailable. Through this work, we propose using a generative modeling approach to generate realistic building attributes to fill in the data gaps and finally provide complete characteristics as inputs to energy models. Our model learns complex, building-level patterns from training on a large-scale residential building stock model containing 2.2 million buildings. We employ a tabular diffusion-based framework that is designed to handle heterogeneous (discrete and continuous) features in tabular building data, such as occupancy, floor area, heating, cooling, and other equipment details. We develop a capability for conditional diffusion, enabling the imputation of missing building characteristics conditioned on known attributes. We conduct a comprehensive validation of our conditional diffusion model, firstly by comparing the generated conditional distributions against the underlying data distribution, and secondly, by performing a case study for a Baltimore residential region, showing the practical utility of our approach. Our work is one of the first to demonstrate the potential of generative modeling to accelerate building energy modeling workflows.

Contrastive Language-Image Pretraining (CLIP) models excel at understanding image-text relationships but struggle with adapting to new data without forgetting prior knowledge. To address this, models are typically fine-tuned using both new task data and a memory buffer of past tasks. However, CLIP's contrastive loss suffers when the memory buffer is small, leading to performance degradation on previous tasks. We propose a memory-efficient, distributionally robust method that dynamically reweights losses per class during training. Our approach, tested on class incremental settings (CIFAR-100, ImageNet1K) and a domain incremental setting (DomainNet) adapts CLIP models quickly while minimizing catastrophic forgetting, even with minimal memory usage.

Increased electricity consumption combined with new forms of generation is testing the reliability of our grid infrastructure. This work describes a method to improve the reliability of the grid through large-scale advanced building control. This paper develops a bi-level distributed control framework to shift the load of 153 buildings to achieve a system-level objective of tracking a power reference signal. This bi-level control is based on the previously-developed ANPV-MPC, a predictive controller that uses a Bayesian neural network to generate an accurate control model and adapt to changing conditions over time. By shifting the building electricity demand to better match the available power, the grid system supplying the buildings is more reliable as evidenced by the analysis of node voltages across an IEEE 13-bus distribution system. The proposed bi-level control framework tracks the system-level power reference with enough accuracy to regulate node voltages across the IEEE 13-bus distribution system within ANSI limits of ±5%. Additionally, the adaptive nature of ANPV-MPC allows each building across the system to adapt to changing conditions, further amplifying the system-level reliability.

Atmospheric CO2 separation using solid-sorbent adsorption systems is strongly influenced by ambient temperature and humidity, which vary over hourly and seasonal timescales. However, detailed cyclic adsorption models that resolve adsorption-desorption dynamics are computationally intensive, limiting their application to long-term, ambient condition-resolved analysis and system-scale optimization. In this work, we develop a physics-informed dynamic reduced-order model (ROM) for a solid-sorbent temperature-vacuum swing adsorption system. The model retains the dominant physical mechanisms governing cyclic operation, including CO2-H2O co-adsorption, adsorption–desorption kinetics, incomplete regeneration, and cycle-to-cycle sorbent-state memory. It is formulated as a computationally efficient, control-oriented state-transition model rather than an equipment-level representation of a specific Direct air capture plant. Using this framework, approximately 120,000 cycle-resolved simulations are generated across environmental conditions, initial sorbent states, system designs, and operating strategies. The resulting dataset reveals a structured performance landscape and supports the development of a neural-network surrogate for rapid cyclic prediction and optimization. Results demonstrate that system performance is highly sensitive to environmental variability and that cycle-resolved optimization under time-varying conditions can simultaneously reduce energy demand and improve CO2 productivity compared with fixed operating strategies. Overall, this work presents a scalable physics-to-surrogate workflow that links adsorption physics, reduced-order dynamic modeling, and neural-network-assisted optimization for atmospheric CO2 separation systems operating under realistic environmental variability.

Adaptive Computing is an application-agnostic outer loop framework to strategically deploy simulations and experiments to guide decision making for scale-up analysis. Resources are allocated over successive batches, which makes the allocation adaptive to some objective such as optimization or model training. The framework enables the characterization and management of uncertainties associated with predictive models of complex systems when scale-up questions lead to significant model extrapolation. A key advancement of this framework is its integration of multi-fidelity surrogate modeling, uncertainty management, and automated orchestration of various computing and experimentation resources into a single integrated software package. This enables efficient multi-fidelity modeling across multiple computing resources by incorporating real-world constraints such as relative queue times and throughput on individual machines into the multi-fidelity sampling decision. We discuss applications of this framework to problems in the renewable energy space, including biofuels production, material synthesis, perovskite crystal growth, and building electrical loads.

Management of the electrical grid is becoming more complex due to the increased penetration of alternative energy generation technologies and a broadening diversity of electric loads. This complexity creates challenges in balancing demand and generation that can increase the potential for grid instabilities. One effective way to address this issue is to leverage previously unexploited demand flexibility through advanced control strategies. In this work, we propose an advanced control method, called adaptive neural parameter-varying model predictive control (ANPV-MPC), to control the temperature and energy consumption of a building via its Heating, Ventilation, and Air Conditioning system. ANPV-MPC combines key ideas in parameter-varying control, adaptive control, and online learning strategies to bridge the gap between computationally efficient linear model predictive control and more accurate nonlinear model predictive control. The novelty in ANPV-MPC is the use of a physics-inspired Bayesian neural network to estimate the coefficients of the parameter-varying linear control model. The Bayesian neural network additionally provides uncertainty estimates, triggering online training to capture evolving building system conditions. We show that ANPV-MPC can approximate the building system dynamics with a 28.39% higher accuracy than traditional linear model predictive control, resulting in 36.23% better control performance without increasing complexity of the optimal control problem. ANPV-MPC also adapts in real time to previously unseen conditions using online learning, further improving its performance.

Accurate modeling of tropical cyclone wind fields is essential for the design, risk assessment, and operational planning of offshore energy infrastructure. While mesoscale simulations are widely used thanks to their computational efficiency, they lack the necessary resolution to capture key features such as wind shear and veer profiles as well as the distribution turbulent kinetic energy (TKE). High-fidelity large-eddy simulation (LES) models on the other hand, can resolve turbulent structures and provide a more accurate representation of the complex wind field, albeit at a higher computational cost. To address this modeling gap, we introduce a two-part generative framework to enhance the resolution and physics-capturing ability of mesoscale simulations. First, a reduced-order model based on Karhunen–Loève (KL) decomposition is used to extract dominant spatial modes from one-dimensional mean wind profiles. A multilayer perceptron (MLP) is trained to map mesoscale mode weights to their LES counterparts, enabling accurate reconstruction of vertical velocity profiles. Second, a conditional Diffusion Denoising Probabilistic Model (DDPM) is developed to super-resolve coarse and low-fidelity mesoscale velocity fields, recovering fine-scale turbulence structures and stress distributions. The framework is evaluated across different tropical cyclone intensity categories defined by the Saffir–Simpson scale and demonstrates strong performance in both interpolation and extrapolation tasks. The generated fields accurately reproduce spatial coherence, stress distributions, and spectral energy characteristics observed in LES data. By bridging the fidelity gap between mesoscale and LES outputs, this approach offers a scalable, data-driven solution for enhancing the representation of tropical cyclone wind fields, enabling more robust offshore energy infrastructure systems design in tropical-cyclone-prone areas.

In recent years, with the advent of mature machine learning products like ChatGPT, Stable Diffusion, and Sora, the world has witnessed tremendous changes driven by the rapid development of generative artificial intelligence (GAI). Beyond applications in text, speech, image, and video creation, deep generative models (DGMs) underpinning these cutting-edge technologies have also been employed by domain researchers to address scientific and engineering challenges. This paper aims to fill a gap in the research community by providing a comprehensive review of how DGMs have been utilized in energy system applications. Based on five of the most popular DGMs, we review and categorize 228 research articles into five focus areas: data generation, forecasting, situational awareness, modeling, and optimal decision-making. Through this classification, we uncover trends in how DGMs are employed for each type of problem, highlighting GAI techniques that contribute to breakthroughs over traditional methods. We discuss limitations in existing literature, engineering challenges, and propose future directions, all tailored to the unique nature of problems in energy system engineering. Our goal is to offer insights for energy system domain researchers, providing a comprehensive view of existing studies and potential future opportunities.

With a potentially increasing share of the electricity grid relying on wind to provide generating capacity and energy, there is an expanding global need for historically accurate, spatiotemporally continuous, high-resolution wind data. Conventional downscaling methods for generating these data based on numerical weather prediction have a high computational burden and require extensive tuning for historical accuracy. In this work, we present a novel deep learning-based spatiotemporal downscaling method using generative adversarial networks (GANs) for generating historically accurate high-resolution wind resource data from the European Centre for Medium-Range Weather Forecasting Reanalysis version 5 data (ERA5). In contrast to previous approaches, which used coarsened high-resolution data as low-resolution training data, we use true low-resolution simulation outputs. We show that by training a GAN model with ERA5 as the low-resolution input and Wind Integration National Dataset Toolkit (WTK) data as the high-resolution target, we achieved results comparable in historical accuracy and spatiotemporal variability to conventional dynamical downscaling. This GAN-based downscaling method additionally reduces computational costs over dynamical downscaling by two orders of magnitude. We applied this approach to downscale 30 km, hourly ERA5 data to 2 km, 5 min wind data for January 2000 through December 2023 at multiple hub heights over Ukraine, Moldova, and part of Romania. With WTK coverage limited to North America from 2007–2013, this is a significant spatiotemporal generalization. The geographic extent centered on Ukraine was motivated by stakeholders and energy-planning needs to rebuild the Ukrainian power grid in a decentralized manner. This 24-year data record is the first member of the super-resolution for renewable energy resource data with wind from the reanalysis data dataset (Sup3rWind).

If clean energy pathways are to harness massive increases in wind power, innovations with broad geographic viability will be needed to support buildout in diverse locations. However, geodiversity in impact potential is seldom captured in technology assessment. Here we propose a scalable approach to plant-level optimization using artificial intelligence to evaluate land sparing and economic benefits of wake steering at more than 6,800 plausible onshore wind locations in the USA. This emerging controls strategy optimizes plant energy production by directing turbine wakes. On the basis of estimates from our artificial intelligence model trained on engineering wind flow simulations, co-optimizing plant layouts with wake steering can reduce land requirements by an average of 18% per plant (site-specific benefits range from 2% to 34%), subject to errors and uncertainties in the flow model, wind resource estimates, buildout scenario and geographic factors. According to model estimates, wake steering is predicted to increase power production during high-value (relatively low wind) periods, boosting the annual revenue of individual plants by up to US$3.7 million (equivalent to US$13,000 MW-1 yr-1) but producing negligible gains in some settings. Consideration of wake steering's geographic potential reveals divergent nationwide prospects for improved economics and siting flexibility. Wind farms would benefit from optimization of their design and operation. Harrison-Atlas et al. report an artificial intelligence-aided optimization approach that shows the potential of wake steering strategies to minimize land requirements and costs.

To plan and optimize energy storage demands that account for Li-ion battery aging dynamics, techniques need to be developed to diagnose battery internal states accurately and rapidly. This study seeks to reduce the computational resources needed to determine a battery's internal states by replacing physics-based Li- ion battery models - such as the single-particle model (SPM) and the pseudo-2D (P2D) model - with a physics-informed neural network (PINN) surrogate. The surrogate model makes high-throughput techniques, such as Bayesian calibration, tractable to determine battery internal parameters from voltage responses. This manuscript is the first of a two-part series that introduces PINN surrogates of Li-ion battery models for parameter inference (i.e., state-of-health diagnostics). In this first part, a method is presented for constructing a PINN surrogate of the SPM. A multi-fidelity hierarchical training, where several neural nets are trained with multiple physics-loss fidelities is shown to significantly improve the surrogate accuracy when only training on the governing equation residuals. The implementation is made available in a companion repository (https://github.com/NREL/PINNSTRIPES). The techniques used to develop a PINN surrogate of the SPM are extended in Part II (Hassanaly et al., 2024) for the PINN surrogate for the P2D battery model, and explore the Bayesian calibration capabilities of both surrogates.

In recent years, machine learning has made sig-nificant progress in clinical outcome prediction, demonstrating increasingly accurate results. However, the substantial resources required for hospitals to train these models, such as data collection, labeling, and computational power, limit the feasibility for smaller hospitals to develop their own models. An alternative approach involves transferring a machine learning model trained by a large hospital to smaller hospitals, allowing them to fine-tune the model on their specific patient data. However, these models are often trained and validated on data from a single hospital, raising concerns about their generalizability to new data. Our research shows that there are notable differences in measurement distributions and frequencies across various regions in the United States. To address this, we propose a benchmark that tests a machine learning model's ability to transfer from a source domain to different regions across the country. This benchmark assesses a model's capacity to learn meaningful information about each new domain while retaining key features from the original domain. Using this benchmark, we frame the transfer of a machine learning model from one region to another as a domain in-cremental learning problem. While the task of patient outcome prediction remains the same, the input data distribution varies, necessitating a model that can effectively manage these shifts. We evaluate two popular domain incremental learning methods: data replay, which stores examples from previous data sources for fine-tuning on the current source, and Elastic Weight Consolidation (EWC), a model parameter regularization method that maintains features important for both data sources. Finally, we propose a new domain incremental learning method that combines EWC and data replay with the ability to adjust the number of updates utilizing data from previous sources. Our results show that this proposed method outperforms EWC and data replay alone. We also highlight specific shortcomings related to model transferability in the clinical setting, underscoring the need for further research and development in this area.

As renewable energy generation increases, the impacts of weather and climate on energy generation and demand become critical to the reliability of the energy system. However, these impacts are often overlooked. Global climate models (GCMs) can be used to understand possible changes to our climate, but their coarse resolution makes them difficult to use in energy system modelling. Here we present open-source generative machine learning methods that produce meteorological data at a nominal spatial resolution of 4 km at an hourly frequency based on inputs from 100 km daily-average GCM data. These methods run 40 times faster than traditional downscaling methods and produce data that have high-resolution spatial and temporal attributes similar to historical datasets. We demonstrate that these methods can be used to downscale projected changes in wind, solar and temperature variables across multiple GCMs including projections for more frequent low-wind and high-temperature events in the Eastern United States. Global climate models are challenging to integrate in energy system models because their output data resolution is too coarse. Buster et al. generate high-resolution meteorological data with climate change impacts from global climate model datasets using generative machine learning.

As urbanization and climate change progress, urban heat becomes a priority for climate adaptation efforts. High temperatures concentrated in urban heat can drive increased risk of heat-related death and illness as well as increased energy demand for cooling. However, estimating the effects of urban heat is an ongoing field of research typically burdened by an imprecise description of the built environment, significant computational cost, and a lack of high-resolution estimates of the impacts of climate change. Here, we present open-source, computationally efficient machine learning methods that can improve the accuracy of urban temperature estimates when compared to historical reanalysis data. These models are applied to residential buildings in Los Angeles, and we compare the energy benefits of heat mitigation strategies to the impacts of climate change. We find that cooling demand is likely to increase substantially through midcentury, but engineered high-albedo surfaces could lessen this increase by more than 50

Machine learning has revolutionized the modeling of clinical timeseries data. Using machine learning, a Deep Neural Network (DNN) can be automatically trained to learn a complex mapping of its input features for a desired task. This is particularly valuable in Electronic Health Record (EHR) databases, where patients often spend extended periods in intensive care units (ICUs). Machine learning serves as an efficient method for extract meaningful information. However, many state-of-the-art (SOTA) methods for training DNNs demand substantial volumes of labeled data, posing significant challenges for clinics in terms of cost and time. Self-supervised learning offers an alternative by allowing practitioners to extract valuable insights from data without the need for costly labels. Yet, current SOTA methods often necessitate large data batches to achieve optimal performance, increasing computational demands. This presents a challenge when working with long clinical timeseries data. To address this, we propose an efficient method of contrastive pretraining tailored for long clinical timeseries data. Our approach utilizes an estimator for negative pair comparison, enabling effective feature extraction. We assess the efficacy of our pretraining using standard self-supervised tasks such as linear evaluation and semi-supervised learning. Additionally, our model demonstrates the ability to impute missing measurements, providing clinicians with deeper insights into patient conditions. We demonstrate that our pretraining is capable of achieving better performance as both the size of the model and the size of the measurement vocabulary scale. Finally, we externally validate our model, trained on the MIMIC-III dataset, using the eICU dataset. We demonstrate that our model is capable of learning robust clinical information that is transferable to other clinics.