SEARCH · Search NASA
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.
Surrogate-based 0-th Order Global Optimization for black-box problems. This software is comprised of active learning algorithms based on surrogate models to solve black-box optimization problems and other scientific applications. See also: https://pypi.org/project/soogo/
Deep learning models enable the prediction of clinical endpoints from whole-slide images (WSIs), but many such models function as “black boxes”, lacking transparency about whether and which histomorphological patterns drive their predictions, hindering interpretability and clinical adoption. Here we propose a human-in-the-loop explanation framework, MorphoXAI, which provides both local and global interpretability for deep learning models by incorporating human-expert interpretations. At the global level, it reveals the histomorphological patterns on which the model consistently relies to distinguish between classes of WSIs, as well as the patterns associated with confusion between classes. At the local level, it indicates which of these patterns are used in the prediction of an individual WSI and which regions within the slide correspond to such patterns. We validated our method across multiple deep learning–based WSI analysis tasks spanning different tissue types. The results show that our framework generates explanations that accurately reflect the histomorphology underlying the model’s predictions at both global and local levels. For interpretability and clinical utility in diagnostic contexts, human evaluation results showed that our explanations were easy to interpret, rich in diagnostic features, and directly helpful for diagnostic decision-making, thereby enhancing pathologist-AI collaboration. Our work highlights that unifying global and local explanations and grounding them in expert-interpreted morphology enhances the interpretability and verifiability of deep learning models, thereby facilitating the transparent deployment of such models in clinical practice.
Photovoltaic (PV) inverter manufacturers use custom, proprietary control approaches and topologies in their inverter design. The proprietary nature of these approaches makes it challenging to share electromagnetic transients (EMT) domain models for system studies. This research work presents an approach to develop EMT models from experimental data. We use novel approach in experimental design, high fidelity data collection, use of learning-based modeling, and co-simulation to reduce the time taken to develop an EMT model for an inverter under test (IUT). We used a 20 kW off-the-shelf grid following PV inverter and subjected the inverter to controlled tests. The tests include voltage and frequency step changes, as well as solar irradiance variations. The recorded high frequency data were used to train a neural network model representing the dynamic behavior of the IUT. The model was subsequently imported into an EMT tool using co-simulation techniques, and thus completing the modeling effort.
The meteoric rise of generative artificial intelligence (AI) large language models (LLMs) has created an opportunity to utilize them to increase efficiencies in a multitude of industries. While LLMs carry great potential to revolutionize the manner in which work is performed, numerous known deficiencies limit their utility, including the black box nature of the models, the stochastic nature of the response (i.e., presenting the same prompt multiple times results in different responses), and the potential for hallucination. Widespread adoption of LLMs in safety-critical industries such as nuclear will require some form of explainability to assure end users that the LLM’s response to a given query is valid. Model uncertainty is inherently linked to the concepts of trust and explainability, and can be used to identify situations in which the model is insufficiently certain about its answer. Although uncertainty is not enough in and of itself to determine the suitability of an answer—a model can be very certain of an inaccurate answer—it still provides valuable supporting information. Practical methodologies for gauging or quantifying the uncertainty in LLM outputs are presented herein, along with examples based on nuclear-specific prompts.
Photovoltaic inverter (PV) inverter manufacturers use custom, proprietary control approaches and topologies in their inverter design. Due to this proprietary nature, it is not possible to share EMT domain models for system studies. This research work presents a novel approach in experimental design, high fidelity data collection, use of learning-based modeling, and co-simulation to enhance the PV inverter modeling. We used a 20 kW off-the-shelf grid following PV inverter and subjected the inverter to controlled tests including voltage and frequency step changes, as well as solar irradiance variations. The recorded high frequency data was used in learning-based model training. This learning-based model was imported into an Electromagnetic Transient (EMT) simulation tool using co-simulation techniques to complete the modeling effort and integrate the model into an EMT simulation tool. The three key components in this research work are the design of experimental setup, use of learning-based approach for model development and use of co-simulation to complete the approach. The proposed approach will allow users to develop a model in a really short period of time and achieve reasonable inverter models.
Real-time hybrid simulation (RTHS) divides a structural system into an analytical and experimental substructure. The former is based on a well-established analytical model while the latter consists of a physical model in the laboratory, for which there is not a well-established analytical model. This paper extends real-time hybrid simulation to monopile-type Offshore Wind Turbines (OWTs) to enable the investigation of their behavior considering the response of pile foundations under operational and more severe conditions. The embedded foundation and surrounding soil of the OWT are modeled physically in a soil box in the laboratory while the remaining parts of the system and loading are modeled analytically. The program OpenFAST, developed by the National Renewable Energy Laboratory (NREL), is linked to the RTHS coordinator to determine the hydrodynamic and aerodynamic loads acting on the OWT, along with modeling the dynamics of the electric power generation equipment and associated controller for the OWT. The RTHS framework along with its initial implementation and validation are described in this paper. RTHSs of a 5 MW OWT subjected to operational and more severe conditions are performed to experimentally validate the framework. The framework offers a realistic approach to investigate the behavior of OWT structures supported on monopiles. Furthermore, this approach accounts for the coupled response of the OWT structure with its foundation, while experimentally capturing the nonlinearities of the soil-foundation interaction in real-time.
Anomalous behavior is ubiquitous in subsurface solute transport due to the presence of high degrees of heterogeneity at different scales in the media. Although fractional models have been extensively used to describe the anomalous transport in various subsurface applications, their application is hindered by computational challenges. Simpler nonlocal models characterized by integrable kernels and finite interaction length represent a computationally feasible alternative to fractional models; yet, the informed choice of their kernel functions still remains an open problem. We propose a general data-driven framework for the discovery of optimal kernels on the basis of very small and sparse data sets in the context of anomalous subsurface transport. Using spatially sparse breakthrough curves recovered from fine-scale particle-density simulations, we learn the best coarse-scale nonlocal model using a nonlocal operator regression technique. Predictions of the breakthrough curves obtained using the optimal nonlocal model show good agreement with fine-scale simulation results even at locations and time intervals different from the ones used to train the kernel, confirming the excellent generalization properties of the proposed algorithm. A comparison with trained classical models and with black-box deep neural networks confirms the superiority of the predictive capability of the proposed model.
With recent advances in the fields of satellite imagery and machine learning we now have the ability to develop explainable deep learning models that enhance critical infrastructure analysis. Funded through Idaho National Laboratory’s (INL) Laboratory Directed Research and Development (LDRD) office we are in the process of developing a deep learning model capable of identifying critical infrastructure facilities and embedded features within those facilities. Utilizing current limit of practice techniques in the machine learning areas of explainability and transfer learning our model, once complete, will have the capacity to be used on multiple different imagery data sets and produce results that not only classify critical infrastructure facilities, but also explain why a critical infrastructure facility was classified as a certain type of facility. These advancements eliminate the ‘black box’ approach deep learning models have had in the past, where a user will have to trust the conclusion of a model without understanding what reasoning when into the model’s classification process. They also expand a model’s usefulness, traditionally a deep learning model will have to use the same data set it was originally trained on. Given the long training times of deep learning models this is impractical in a number of scenarios. By utilizing transfer learning advancements, we are eliminating the need to train our model on the same data set it is then run on to classify critical infrastructure facilities. We are also enabling the analysis and classification of data sets that are potentially too small to be divided into a training and testing data set. Once completed this model can provide a foundation to enhanced critical infrastructure analysis, dependency analysis, and potential disaster relief efforts.
Managing and referencing design knowledge is a critical activity in the design process. However, reliably retrieving useful knowledge can be a frustrating experience for users of knowledge management systems due to inherent limitations of standard keyword-based searches. In this research, we consider the task of retrieving relevant lessons learned from the NASA Lessons Learned Information System (LLIS). To this end, we apply a state-of-the-art natural language processing (NLP) technique for information retrieval: semantic search with sentence-BERT, which is a modification of a Bidirectional Encoder Representations from Transformers (BERT) model that uses siamese and triplet network architectures to obtain semantically meaningful sentence embeddings. While the pretrained sBERT model shows excellent out-of-the-box performance, we further fine-tune the model on data from the LLIS so that it learns on design engineering-relevant vocabulary. We quantify the improvement in query results using both standard sBERT and fine-tuned sBERT over the LLIS’s built-in keyword search. Additionally, we demonstrate a use case for the query system by searching for lessons learned relevant to specific requirements from a NASA project as part of a broader knowledge management and retrieval system. Results indicate that applying state-of-the-art natural language processing techniques, especially when fine-tuned using engineering data, to design information retrieval tasks shows significant promise in modernizing design knowledge management systems.
One way to reduce the time of conducting optimization studies is to evaluate designs in parallel rather than just one-at-a-time. For expensive-to-evaluate black-boxes, batch versions of Bayesian optimization have been proposed. They work by building a surrogate model of the black-box to simultaneously select multiple designs via an infill criterion. Still, despite the increased availability of computing resources that enable large-scale parallelism, the strategies that work for selecting a few tens of parallel designs for evaluations become limiting due to the complexity of selecting more designs. It is even more crucial when the black-box is noisy, necessitating more evaluations as well as repeating experiments. Here we propose a scalable strategy that can keep up with massive batching natively, focused on the exploration/exploitation trade-off and a portfolio allocation. We compare the approach with related methods on noisy functions, for mono and multi-objective optimization tasks. These experiments show orders of magnitude speed improvements over existing methods with similar or better performance.
Abstract Bayesian optimization (BO) is a low-cost global optimization tool for expensive black-box objective functions, where we learn from prior evaluated designs, update a posterior surrogate Gaussian process model, and select new designs for future evaluation using an acquisition function. This research focuses upon developing a BO model with multiple black-box objective functions. In the standard multi-objective (MO) optimization problem, the weighted Tchebycheff method is efficiently used to find both convex and non-convex Pareto frontiers. This approach requires knowledge of utopia values before we start optimization. However, in the BO framework, since the functions are expensive to evaluate, it is very expensive to obtain the utopia values as a prior knowledge. Therefore, in this paper, we develop a MO-BO framework where we calibrate with multiple linear regression (MLR) models to estimate the utopia value for each objective as a function of design input variables; the models are updated iteratively with sampled training data from the proposed MO-BO. These iteratively estimated mean utopia values are used to formulate the weighted Tchebycheff MO acquisition function. The proposed approach is implemented in two numerical test examples and one engineering design problem of optimizing thin tube geometries under constant loading of temperature and pressure, with minimizing the risk of creep-fatigue failure and design cost, along with risk-based and manufacturing constraints. Finally, the model accuracy with frequentist, Bayesian and without MLR-based calibration are compared to true Pareto solutions.
The work performed on global dynamic simulation and noise correlation of gear transmission systems at the University of Akron is outlined. The objective is to develop a comprehensive procedure to simulate the dynamics of the gear transmission system coupled with the effects of gear box vibrations. The developed numerical model is benchmarked with results from experimental tests at NASA Lewis Research Center. The modal synthesis approach is used to develop the global transient vibration analysis procedure used in the model. Modal dynamic characteristics of the rotor-gear-bearing system are calculated by the matrix transfer method while those of the gear box are evaluated by the finite element method (NASTRAN). A three-dimensional, axial-lateral coupled bearing model is used to couple the rotor vibrations with the gear box motion. The vibrations between the individual rotor systems are coupled through the nonlinear gear mesh interactions. The global equations of motion are solved in modal coordinates and the transient vibration of the system is evaluated by a variable time-stepping integration scheme. The relationship between housing vibration and resulting noise of the gear transmission system is generated by linear transfer functions using experimental data. A nonlinear relationship of the noise components to the fundamental mesh frequency is developed using the hypercoherence function. The numerically simulated vibrations and predicted noise of the gear transmission system are compared with the experimental results from the gear noise test rig at NASA Lewis Research Center. Results of the comparison indicate that the global dynamic model developed can accurately simulate the dynamics of a gear transmission system.
Here, this study quantifies the impact of automated feature engineering methods (feature extraction and selection) on the quality and accuracy of machine learning models that predict building energy consumption. The case study compares model performance for three main scenarios: baseline (no feature extraction and selection), feature extraction only, and feature extraction combined with feature selection (filter and/or wrapper methods) for fully trained machine learning models for 200 metered/sub-metered energy measurements across 118 real buildings. For consistency, the same machine learning model architecture (a black box deep learning neural network with probabilistic forecast output) was used for all scenarios. Based on results, all feature engineering methods provided noticeable prediction accuracy improvements (e.g., 29%-68% median prediction improvement) compared to baseline scenarios. However, in this application, feature selection methods provide little practical value due to their limited performance gains and high computational cost. Smarter algorithm development supported by better computational environments will be needed before feature selection methods can reliably and efficiently improve predictive model performance.
Graph deep learning models, which incorporate a natural inductive bias for atomic structures, are of immense interest in materials science and chemistry. Here, we introduce the Materials Graph Library (MatGL), an open-source graph deep learning library for materials science and chemistry. Built on top of the popular Deep Graph Library (DGL) and Python Materials Genomics (Pymatgen) packages, MatGL is designed to be an extensible “batteries-included” library for developing advanced model architectures for materials property predictions and interatomic potentials. At present, MatGL has efficient implementations for both invariant and equivariant graph deep learning models, including the Materials 3-body Graph Network (M3GNet), MatErials Graph Network (MEGNet), Crystal Hamiltonian Graph Network (CHGNet), TensorNet and SO3Net architectures. MatGL also provides several pre-trained foundation potentials (FPs) with coverage of the entire periodic table, and property prediction models for out-of-box usage, benchmarking and fine-tuning. Finally, MatGL integrates with PyTorch Lightning to enable efficient model training.
This report contains results of structural analyses performed in support of the NASA structural testing of an all-composite stitched/RFI (resin film infusion) wing stub box. McDonnell Douglas Aerospace Company designed and fabricated the wing stub box. The analyses used a global/local approach. The global model contains the entire test article. It includes the all-composite stub box, a metallic load-transition box and a metallic wing-tip extension box. The two metallic boxes are connected to the inboard and outboard ends of the composite wing stub box, respectively. The load-transition box was attached to a steel and concrete vertical reaction structure and a load was applied at the tip of the extension box to bend the wing stub box upward. The local model contains an upper cover region surrounding three stringer runouts. In that region, a large nonlinear deformation was identified by the global analyses. A more detailed mesh was used for the local model to obtain more accurate analysis results near stringer runouts. Numerous analysis results such as deformed shapes, displacements at selected locations, and strains at critical locations are included in this report.
Deep reinforcement learning (DRL) algorithms have successfully solved many challenging problems in various power system control scenarios. However, their decision-making process is usually regarded as black-boxes. Furthermore, how DRL models interact with human intelligence remains an open problem. Thus, this paper proposes a policy extraction framework to extract a complex DRL model into an explainable policy. This framework includes three parts: 1) DRL training and data generation. We train an agent for a specific control task and generate data, which contains the control policy of the agent. 2) Policy extraction. We propose an information gain rate based weighted oblique decision tree (IGR-WODT) for DRL policy extraction. 3) Policy evaluation. We define three metrics to evaluate the performance of the proposed approach. A case study for the under-voltage load shedding problem shows that the IGR-WODT presents a performance enhancement compared with DRL, weighted oblique decision tree, and univariate decision tree. The proposed policy extraction method could provide an intuitive explanation of the neural network decision-making process to the dispatchers when making final decisions on power grid operation. Also, the resulted rule-based controller could replace the deep neural network-based controller in many field edge devices with limited computing resources, providing comparable performance.
Arctic Amplification (AA), the amplified surface warming in the Arctic relative to the global mean, is a robust and impactful feature of climate change. While the basic physical picture of AA has been depicted, a clear understanding of how the degree of AA is determined has not been established. Here, by deciphering the intricate role of atmospheric heat transport (AHT), we build a two-box energy-balance model of AA and derive that the degree of AA is a simple nonlinear function of the Arctic and global feedbacks, the meridional heterogeneity in radiative forcing, and the partial sensitivities of AHT to global mean warming and meridional warming gradient. The formula captures the varying degree of AA in individual climate models and attributes the variation to specific physical factors. It further conveys a concise picture of how essential physics mutually determine the degree of AA and limit the range within 1.5~3.5. Our results articulate AHT as both forcing and feedback to AA, highlighting its partial sensitivities instead of total change as the key parameters for understanding AA. We also find that the effect of lapse rate feedback, a widely-recognized major contributor to AA, is fully offset by the effect of water vapor feedback.
The objective of this paper is to derive a method of constructing semi-analytic solutions to the Noh problem when the equation of state is a black box. Such solutions can be used for verification tests of hydrodynamics codes. We present the underlying theory, the method for finding solutions, and several examples of derived semi-analytic solutions. We end by performing a classic verification convergence test comparing numerical results from a hydrodynamics code against a non-trivial semi-analytic solution.