Search NASA⌕ Search

SEARCH · Search NASA

Results for “Latent”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 271 records · Page 15
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

A simple introduction to the SiMPL method for density-based topology optimization

We introduce a novel method for solving density-based topology optimization problems: Sigmoidal Mirror descent with a Projected Latent variable (SiMPL). The SiMPL method (pronounced as “the simple method”) optimizes a design using only first-order derivative information of the objective function. The bound constraints on the density field are enforced with the help of the (negative) Fermi–Dirac entropy, which is also used to define a non-symmetric distance function called a Bregman divergence on the set of admissible designs. This Bregman divergence leads to a simple update rule that is further simplified with the help of a so-called latent variable. Because the SiMPL method involves discretizing the latent variable, it produces a sequence of pointwise-feasible iterates, even when high-order finite elements are used in the discretization. Numerical experiments demonstrate that the method outperforms other popular first-order optimization algorithms. In conclusion, to outline the general applicability of the technique, we include examples with (self-load) compliance minimization and compliant mechanism optimization problems.

Calculus of Variations and Optimization↗
DOE OSTI2022

Estimation of the Surface Fluxes for Heat and Momentum in Unstable Conditions with Machine Learning and Similarity Approaches for the LAFE Data Set

Abstract Measurements of three flux towers operated during the land atmosphere feedback experiment (LAFE) are used to investigate relationships between surface fluxes and variables of the land–atmosphere system. We study these relations by means of two machine learning (ML) techniques: multilayer perceptrons (MLP) and extreme gradient boosting (XGB). We compare their flux derivation performance with Monin–Obukhov similarity theory (MOST) and a similarity relationship using the bulk Richardson number (BRN). The ML approaches outperform MOST and BRN. Best agreement with the observations is achieved for the friction velocity. For the sensible heat flux and even more so for the latent heat flux, MOST and BRN deviate from the observations while MLP and XGB yield more accurate predictions. Using MOST and BRN for latent heat flux, the root mean square errors (RMSE) are 107 Wm $$^{-2}$$ - 2 and 121 Wm $$^{-2}$$ - 2 , respectively, as well as the intercepts of the regression lines are $$\approx 110$$ ≈ 110 Wm $$^{-2}$$ - 2 . For the ML methods, the RMSEs reduce to 31 Wm $$^{-2}$$ - 2 for MLP and 33 Wm $$^{-2}$$ - 2 for XGB as well as the intercepts to just 4 Wm $$^{-2}$$ - 2 for MLP and $$-1$$ - 1 Wm $$^{-2}$$ - 2 for XGB with slopes of the regression lines close to 1, respectively. These results indicate significant deficiencies of MOST and BRN, particularly for the derivation of the latent heat flux. In fact, in contrast to the established theories, feature importance weighting demonstrates that the ML methods base their improved derivations on net radiation, the incoming and outgoing shortwave radiations, the air temperature gradient, and the available water contents, but not on the water vapor gradient. The results imply that further studies of surface fluxes and other turbulent variables with ML techniques provide great promise for deriving advanced flux parameterizations and their implementation in land–atmosphere system models.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Fused filament fabrication of novel phase-change material functional composites

Additively manufactured polymer heat exchangers are of recent interest in the thermal sciences due to their lightweight and intricate heat-transfer-enhancing geometrical features. The goal of the present research is to directly encapsulate phase-change material (PCM) into polymer filament for the purpose of 3D printing polymer heat exchangers capable of latent heat thermal energy storage and management. PCMs have the ability to absorb large amounts of latent heat while undergoing a solid-to-liquid phase change at a nearly constant temperature. Encapsulation or containment of PCMs is necessary to prevent leaking since the material continuously cycles between a liquid and a solid state. An organic-based PCM and high-density polyethylene (HDPE) were combined and extruded into a functional composite filament. For the first time, with the aid of a heated enclosure and HDPE build plates, it was demonstrated that these functional composite materials can be printed using fused filament fabrication (FFF). Printing HDPE is known to be difficult; however, based on what was observed in the current study, the PCM and HDPE composite was easier to print than pure HDPE. Thermal properties, such as latent heat of fusion, phase-change temperature, and thermal conductivity, of the composite filament and 3D printed samples were investigated and compared with compression-molded bulk material. Finally, the microstructures of the composite filament and 3D printed samples were visualized, and the basic structures of the HDPE and PCM within both the filament and 3D printed parts were clearly delineated.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Experimental investigation on phase change material–based finned tube heat exchanger for thermal energy storage and building envelope thermal management

Phase change materials (PCMs) are attractive solutions for thermal energy storage (TES) applications by absorbing and releasing large amounts of latent heat during solid–liquid phase transitions. However, their relatively low thermal conductivity requires novel heat exchanger–based solutions to improve the power density and overall energy storage efficiency of the TES system. This work presents the design and experimental results of a finned tube heat exchanger to store collected natural thermal energy from a building envelope in a latent-based TES and to release it later for building heating/cooling applications. We experimentally evaluate the finned tube heat exchanger and evaluate the performance of TES in reducing building heating and cooling loads over 3–4 h of desired time of operation (e.g., peak load). The optimized design allows for maximum energy density by minimizing the heat exchanger volume, and the system is evaluated experimentally using commercially available heat exchanger materials and an organic PCM. Here, the experimental results reveal that the TES system is able to charge and discharge stored latent energy within 3–4 h, matching peak building electricity demand duration under an average fluid flow rate of 0.136 kg/s and temperature difference of 5.55 °C. Importantly, such optimized designs illuminate a path toward TES designs that are low-cost, scalable, and optimized for thermal energy and power availability under the desired time of operation.

25 ENERGY STORAGE↗
DOE OSTI2024

STSR-INR: Spatiotemporal super-resolution for multivariate time-varying volumetric data via implicit neural representation

Implicit neural representation (INR) has surfaced as a promising direction for solving different scientific visualization tasks due to its continuous representation and flexible input and output settings. We present STSR-INR, an INR solution for generating simultaneous spatiotemporal super-resolution for multivariate time-varying volumetric data. Inheriting the benefits of the INR-based approach, STSR-INR supports unsupervised learning and permits data upscaling with arbitrary spatial and temporal scale factors. Unlike existing GAN- or INR-based super-resolution methods, STSR-INR focuses on tackling variables or ensembles and enabling joint training across datasets of various spatiotemporal resolutions. Here we achieve this capability via a variable embedding scheme that learns latent vectors for different variables. In conjunction with a modulated structure in the network design, we employ a variational auto-decoder to optimize the learnable latent vectors to enable latent-space interpolation. To combat the slow training of INR, we leverage a multi-head strategy to improve training and inference speed with significant speedup. We demonstrate the effectiveness of STSR-INR with multiple scalar field datasets and compare it with conventional tricubic+linear interpolation and state-of-the-art deep-learning-based solutions (STNet and CoordNet).

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Bayesian reduced-order deep learning surrogate model for dynamic systems described by partial differential equations

We propose a reduced-order deep-learning surrogate model for dynamic systems described by time-dependent partial differential equations. This method employs space–time Karhunen–Loève expansions (KLEs) of the state variables and space-dependent KLEs of space-varying parameters to identify the reduced (latent) dimensions. Subsequently, a deep neural network (DNN) is used to map the parameter latent space to the state variable latent space. An approximate Bayesian method is developed for uncertainty quantification (UQ) in the proposed KL-DNN surrogate model. The KL-DNN method is tested for the linear advection–diffusion and nonlinear diffusion equations, and the Bayesian approach for UQ is compared with the deep ensembling (DE) approach, commonly used for quantifying uncertainty in DNN models. It was found that the approximate Bayesian method provides a more informative distribution of the PDE solutions in terms of the coverage of the reference PDE solutions (the percentage of nodes where the reference solution is within the confidence interval predicted by the UQ methods) and log predictive probability. The DE method is found to underestimate uncertainty and introduce bias. For the nonlinear diffusion equation, we compare the KL-DNN method with the Fourier Neural Operator (FNO) method and find that KL-DNN is 10% more accurate and needs less training time than the FNO method.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Thermal energy storage composites with preformed expanded graphite matrix and paraffin wax for long-term cycling stability and tailored thermal properties

Harvesting solar energy, preventing hot spots in electronics, transport of temperature-sensitive materials, and capture and repurposing of thermal energy require a latent heat thermal energy storage (TES) system to store/discharge heat repeatedly. For the practical application of phase change material (PCM) composites within TES systems, reliable thermal performance throughout its operational lifetime is essential. Nevertheless, the reliability of thermal conductivity in multi-phase composites over relevant numbers (>10 3 ) of melt/freeze cycles has barely been studied, particularly for composites containing fillers for thermal conductivity enhancement. Here, we introduce a preform-type expanded graphite (EG)/paraffin wax composite possessing highly robust heat transfer and storage properties even after 10,000 melt/freeze cycles. To achieve such excellent reliability, comparative studies on the combined influence of fabrication process, particle size, EG vol%, binder amount, and compaction on both magnitude and robustness of thermal conductivity were undertaken. Our parametric study has yielded a trade-off between thermal conductivity and latent heat. Based on our modeling, 20 vol% EG approaches the case where all EG particles are well-connected thermally while 10 vol% EG is close to loosely connected fillers in the matrix. Thermal conductivity of our paraffin composites containing 20 vol% EG (25.1 W·m -1 ·K -1 ) is highest among other EG/paraffin composites without aligned EG in the literature. After 10,000 thermal cycling, negligible conductivity fading was observed for the 10 vol% EG composite, while reduction in latent heat remained within 10% for all 10, 14, 17 and 20 vol% EG samples. Here we anticipate this work provides insight on suitable recipe for desirable magnitude and robustness of thermal conductivity of EG/paraffin composites.

25 ENERGY STORAGE↗
DOE OSTI2026

Self-supervised and multi-fidelity learning for extended predictive soil spectroscopy

Infrared spectroscopy is a cost-effective, non-destructive, and environmentally benign technology that is increasingly recognized as an important solution for meeting the global demand for soil data. While both near-infrared (NIR) and mid-infrared (MIR) diffuse reflectance spectroscopy enable rapid estimation of soil properties, they present a significant trade-off: NIR offers superior scalability and lower operational costs, whereas MIR provides higher analytical fidelity by capturing fundamental molecular vibrations. In this study, we propose a self-supervised, multi-fidelity learning framework designed to bridge this gap. Our approach leverages large-scale MIR spectral libraries to learn a compact, transferable latent representation, into which NIR spectra are subsequently aligned for downstream prediction. The workflow consists of pretraining a latent model on a large MIR library, adapting the representation using a smaller paired NIR–MIR dataset, and evaluating generalization on an independent external test set. Across a range of chemical and physical soil properties, we found that MIR-derived embeddings improved prediction accuracy relative to baseline models that used raw MIR inputs. Predictions derived from the spectrum conversion (NIR to MIR) task did not match the performance of the original MIR spectra but were similar or superior to predictive performance of NIR-only models, suggesting the unified spectral latent space can effectively leverage the larger and more diverse MIR dataset for prediction of soil properties not well represented in current NIR libraries.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

CO 2 storage site characterization using ensemble-based approaches with deep generative models

Estimating spatially distributed properties such as permeability from available sparse measurements is a great challenge in efficient subsurface CO 2 storage operations. In this paper, a deep generative model that can accurately capture complex subsurface structure is tested with an ensemble-based inversion method for accurate and accelerated characterization of CO 2 storage sites. We chose Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) for its realistic reservoir property representation and Ensemble Smoother with Multiple Data Assimilation (ES-MDA) for its robust data fitting and uncertainty quantification capability. WGAN-GP are trained to generate high-dimensional permeability fields from a low-dimensional latent space and ES-MDA then updates the latent variables by assimilating available measurements. Several subsurface site characterization examples including Gaussian, channelized, and fractured reservoirs are used to evaluate the accuracy and computational efficiency of the proposed method and the main features of the unknown permeability fields are characterized accurately with reliable uncertainty quantification. Furthermore, the estimation performance is compared with a widely-used variational, i.e., optimization-based, inversion approach, and the proposed approach outperforms the variational inversion method in several benchmark cases. We explain such superior performance by visualizing the objective function in the latent space: because of nonlinear and aggressive dimension reduction via generative modeling, the objective function surface becomes extremely complex while the ensemble approximation can smooth out the multi-modal surface during the minimization. This suggests that the ensemble-based approach works well over the variational approach when combined with deep generative models at the cost of forward model runs unless convergence-ensuring modifications are implemented in the variational inversion.

42 ENGINEERING↗
DOE OSTI2021

Experimental study of sprays with isooctane, hexane, ethanol and their binary mixtures under different flash boiling intensities

Here, this work investigated the behaviors of single-component and multi-component sprays under different flash boiling intensities. Isooctane, hexane, ethanol, isooctane-hexane, and isooctane-ethanol mixtures were used as test fuels. The pressure ratio of the ambient pressure over saturation pressure (R p ) was used as the indicator of the superheated degree. Macroscopic characteristics of the sprays under 10 different R p (1.1 to 0.14) were investigated using Diffused Backlight Imaging (DBI). Droplet size, which revealed the spray microscopic characteristics, was also investigated using Phase Doppler Anemometry (PDA). The results showed that ambient pressure reduction led to higher flash boiling intensity, significant spray morphology changes, and smaller droplet size, regardless of the fuel type and blend ratios. Spray collapse was observed for all sprays once R p dropped below 0.3. Spray collapse dramatically increased the liquid penetration length, expanded the near-field spray angle, and reduced the far-field angle for sprays with all the fuels. Fuel properties, such as the latent heat of vaporization, had significant impacts on droplet size distributions, thus altering the liquid penetration length and spray angles. Under similar R p , sprays with high ethanol content generally had larger droplet sizes compared with other fuel sprays due to the high latent heat of vaporization. The larger droplets caused the high ethanol content sprays to be more resistant to the plume trajectory change, resulting in a smaller far-field spray angle variations with changing R p . At R p of 0.28, sprays with the similar latent heat of vaporization have similar droplet sizes, and droplet sizes at different radial locations were also similar to each other.

42 ENGINEERING↗
DOE OSTI2024

Influence of alloy solidification path on melt pool behavior in additive manufacturing

Numerical models used to study transport phenomena in laser-powder bed fusion processes often rely on assumptions and simplifications to reduce their computational expense. One common simplification is in the description of latent heat evolution during the solid-liquid phase change (i.e., the solidification pathway), justified by the fact that the mushy zone thickness is similar to the numerical grid spacing used for continuum transport models. The lack of resolution of transport phenomena in the mushy zone motivates the use of computationally convenient solidification paths such as linear or sigmoidal relationships over pathways derived from fundamental solidification theory such as equilibrium or Scheil models. In the present work, an uncertainty quantification (UQ) framework is used to analyze the influence of solidification pathway selection on the solidification dynamics and melt pool geometries in laser based additive manufacturing (AM) of IN625. Results show the solidification pathway has a quantifiable influence on the cooling rate at the liquidus isotherm, mushy zone thickness, and solidification time. Due to similarities in the latent heat evolution at the beginning of solidification, the equilibrium and Scheil models predict similar cooling rates near the liquidus isotherm, however the wider freezing range of Scheil leads to a wider mushy zone compared to equilibrium. The non-physical latent heat release profiles of sigmoidal and linear paths lead to significant overpredictions of cooling rates at the liquidus isotherm compared to equilibrium and Scheil. Finally, these results indicate that careful consideration should be given to the choice of solidification pathway to ensure reliable model predictions.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Hydrogen-Steam Separation Using Mechanical Vapor Recompression Cycle

Solar thermochemical hydrogen and fuel production is a promising pathways for producing sustainable fuels and chemicals. One of the main challenges in the development of these systems is their low steam conversion extent, dictated by its restrictive thermodynamics requiring extremely high temperatures and low oxygen partial pressure to obtain conversions over 10 \%. While condensing the unreacted steam is technically simple, the latent heat is thus lost, which can be larger than the producd hydrogen higher heating value. We propose to use a modified mechanical vapor recompression cycle, allowing to recover the latent heat by compressing the steam-hydrogen mixture prior to the condensation process, thus creating a temperature different between the hot exhaust and cold inlet streams. We show that this method can recover over 90 % of the latent heat, thus increasing the viability of solar thermochemical hydrogen production cycles even under limited conversion conditions.

hydrogen technology↗
DOE OSTI2024

A deep learning-based workflow for fast prediction of 3D state variables in geological carbon storage: A dimension reduction approach

Deep learning (DL) models are extensively used as surrogate models for high-fidelity simulations of multiphase fluid flow in porous media at large scales, enabling fast forecasts of the spatial–temporal evolution of three-dimensional (3D) state variables in geological carbon storage (GCS). However, training these models in high-dimensional space remains computationally demanding and prone to overfitting because of limited training data. This paper presents a novel workflow to address these challenges by integrating dimension reduction (DR) methods. Here, the proposed workflow employed pre-trained DR models to extract the latent variables of geological models and state variables and utilized the multi-layer perceptron (MLP) for constructing mapping functions between the input and output variables in latent spaces. Subsequently, the pre-trained reconstruction models converted the MLP-predicted latent state variables to their original high-dimensional form. Furthermore, we proposed a novel strategy for the DR and reconstruction of 3D saturation fields to account for the unique data characteristics of sparsity, nonuniformity, and discontinuity. The proposed strategy applied PCA and inverse PCA for 2D average saturation fields and developed a DL-based 3D reconstruction model, leveraging three 2D average saturation fields as input to produce a 3D saturation field as output. The pre-training of DR and reconstruction models and training of MLP models were conducted on 84 Gulf of Mexico (GoM) simulations and evaluated on 12 testing simulations. Each simulation contained 720 monthly time steps, with the first 360 months as the injection period and the rest as the post-injection period. The proposed workflow, incorporating DR and DL models, accurately predicts the normalized 3D pressure fields, achieving mean square error (MSE) of 2.92 × 10 -7 compared to the ground truth obtained from a full-physics simulator. Furthermore, the proposed strategy outperformed PCA and convolutional autoencoder (CAE) models on 3D saturation fields, resulting in minor workflow prediction errors with an MSE of 2.93 × 10 -5 . The results suggest the proposed workflow provides sufficient predictive fidelity across temporal and spatial scales, and enables a speedup of 160 times compared to the full-physics simulator, facilitating improved decision-making and risk assessment for large-scale GCS management in real-time scenarios.

3D reconstruction model↗
DOE OSTI2023

An artificial intelligence’s interpretation of complex high-resolution in situ transmission electron microscopy data

Complicated nano- and atomic-scale processes with sub-angstrom spatial resolution and millisecond time resolution visualized by in situ transmission electron microscopy (TEM) are often highly dynamical and time consuming to analyze and interpret. Here, we report how variational autoencoders (VAEs) can provide an artificial intelligence’s interpretation of high-resolution in situ TEM data by condensing and deconvoluting complicated atomic-scale dynamics into a latent space with reduced dimensionality. We designed a VAE model with high latent dimensions capable of deconvoluting information from complex high-resolution TEM data. We demonstrate how this model, with high latent dimensions trained on atomically resolved TEM images of lead sulfide (PbS) nanocrystals, is able to capture movements and perturbations of periodic lattices in both simulated and real in situ TEM data. Importantly, the VAE model shows the capability of detecting and deconvoluting dynamical nanoscale physical processes, such as the rotation of crystal lattices and intra-particle ripening during the annealing of semiconductor nanocrystals.

47 OTHER INSTRUMENTATION↗
DOE OSTI2022

Energy and exergy analysis of multi-stage vacuum membrane distillation integrated with mechanical vapor compression

Membrane distillation (MD) is a promising candidate for desalinating hypersaline brine, but its poor energy efficiency has remained a major barrier for widespread application. One possible solution to this issue is to recover the latent heat in the process. In this work, a multi-stage vacuum MD (MSVMD) was integrated with a mechanical vapor compressor (MVC) to enhance the latent heat recovery, and the energetic and exergetic performance of this integrated process was examined. A comprehensive energy and exergy analysis is provided to compare MSVMD and MSVMD-MVC processes for desalination of hypersaline brine. This analysis was conducted by examining the effect of the compression ratio on the energetic and exergetic performance, and the findings are reported in terms of specific thermal energy consumption (STEC), specific electricity consumption (SEC), and exergetic efficiency. The energy analysis shows that thermal energy consumption can be reduced as the compression ratio increases, due to the enhancement of latent heat recovery. The MSVMD-MVC process can be operated in a steady-state condition, without the need for thermal heat input; with STEC and SEC of 0 and 49 kWh/m 3 at the feed temperature of 50 °C and MVC compression ratio of 2.14. Moreover, exergy analysis demonstrates the efficacy of the eNRTL model in exergy calculation. Exergy destruction can be greatly reduced by increasing the compression ratio to an optimal value. For high salinity brine (124 g/L), MSVMD-MVC achieved a higher exergetic efficiency of 6.85%, compared to 2.42% in MSVMD. Furthermore, the result suggests that the application of MVC can intensify the energy efficiency and exergetic efficiency of the MSVMD system, although this process cannot outperform the current desalination technologies from the standard primary energy point of view.

42 ENGINEERING↗
DOE OSTI2022

Multitask Machine Learning of Collective Variables for Enhanced Sampling of Rare Events

Computing accurate reaction rates is a central challenge in computational chemistry and biology because of the high cost of free energy estimation with unbiased molecular dynamics. In this work, a data-driven machine learning algorithm is devised to learn collective variables with a multitask neural network, where a common upstream part reduces the high dimensionality of atomic configurations to a low dimensional latent space and separate downstream parts map the latent space to predictions of basin class labels and potential energies. Here, the resulting latent space is shown to be an effective low-dimensional representation, capturing the reaction progress and guiding effective umbrella sampling to obtain accurate free energy landscapes. This approach is successfully applied to model systems including a 5D Müller Brown model, a 5D three-well model, the alanine dipeptide in vacuum, and an Au(110) surface reconstruction unit reaction. It enables automated dimensionality reduction for energy controlled reactions in complex systems, offers a unified and data-efficient framework that can be trained with limited data, and outperforms single-task learning approaches, including autoencoders.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

A Deep Learning-Driven Sampling Technique to Explore the Phase Space of an RNA Stem-Loop

The folding and unfolding of RNA stem-loops are critical biological processes; however, their computational studies are often hampered by the ruggedness of their folding landscape, necessitating long simulation times at the atomistic scale. Here, we adapted DeepDriveMD (DDMD), an advanced deep learning-driven sampling technique originally developed for protein folding, to address the challenges of RNA stem-loop folding. Although tempering- and order parameter-based techniques are commonly used for similar rare-event problems, the computational costs or the need for a priori knowledge about the system often present a challenge in their effective use. DDMD overcomes these challenges by adaptively learning from an ensemble of running MD simulations using generic contact maps as the raw input. DeepDriveMD enables on-the-fly learning of a low-dimensional latent representation and guides the simulation toward the undersampled regions while optimizing the resources to explore the relevant parts of the phase space. We showed that DDMD estimates the free energy landscape of the RNA stem-loop reasonably well at room temperature. Our simulation framework runs at a constant temperature without external biasing potential, hence preserving the information on transition rates, with a computational cost much lower than that of the simulations performed with external biasing potentials. Here, we also introduced a reweighting strategy for obtaining unbiased free energy surfaces and presented a qualitative analysis of the latent space. This analysis showed that the latent space captures the relevant slow degrees of freedom for the RNA folding problem of interest. Finally, throughout the manuscript, we outlined how different parameters are selected and optimized to adapt DDMD for this system. We believe this compendium of decision-making processes will help new users adapt this technique for the rare-event sampling problems of their interest.

Gupta, Ayush↗
DOE OSTI2022

Non‐Linear Dimensionality Reduction With a Variational Encoder Decoder to Understand Convective Processes in Climate Models

Abstract Deep learning can accurately represent sub‐grid‐scale convective processes in climate models, learning from high resolution simulations. However, deep learning methods usually lack interpretability due to large internal dimensionality, resulting in reduced trustworthiness in these methods. Here, we use Variational Encoder Decoder structures (VED), a non‐linear dimensionality reduction technique, to learn and understand convective processes in an aquaplanet superparameterized climate model simulation, where deep convective processes are simulated explicitly. We show that similar to previous deep learning studies based on feed‐forward neural nets, the VED is capable of learning and accurately reproducing convective processes. In contrast to past work, we show this can be achieved by compressing the original information into only five latent nodes. As a result, the VED can be used to understand convective processes and delineate modes of convection through the exploration of its latent dimensions. A close investigation of the latent space enables the identification of different convective regimes: (a) stable conditions are clearly distinguished from deep convection with low outgoing longwave radiation and strong precipitation; (b) high optically thin cirrus‐like clouds are separated from low optically thick cumulus clouds; and (c) shallow convective processes are associated with large‐scale moisture content and surface diabatic heating. Our results demonstrate that VEDs can accurately represent convective processes in climate models, while enabling interpretability and better understanding of sub‐grid‐scale physical processes, paving the way to increasingly interpretable machine learning parameterizations with promising generative properties.

54 ENVIRONMENTAL SCIENCES↗