Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “constrained learning”

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
NASA NTRSDate not supplied

Enhancing NASA Earth Science Data Discovery from Scientific Publications

Earth observations from space borne instruments have evolved explosively in the past decades. Following closely are reanalysis systems assimilating model and observational data, yielding even longer records and larger number of variables. Thanks to advances in internet technology, it is now easier than ever to visualize and analyze these data using web interfaces. On the other hand, it also becomes an increasingly daunting task to build upon the existing knowledge published in various peer reviewed sources, and navigate toward the most relevant data, analysis, and visualization. We present an analysis of a subset of publications that utilized a popular visualization web interface at the NASA Goddard Earth Science Data and Information Services Center. Known as "Giovanni", it allows researchers from wide backgrounds to work with hundreds of variables from space observations and assimilation systems. Since coming online more than a decade ago, Giovanni has been credited in more than 100 papers per year, and the total count now is estimated to be nearly 1,500. Many of these papers contain valuable information about when, where and how Giovanni has been used, and hence forge an opportunity to learn and share the knowledge of which variables were used for what research projects. The purpose of our work is to retrieve the information from the papers and organize it as a knowledge repository which links together datasets, variables, places, dates and phenomena all of which reflect the essence of the published research. Since the publications are unstructured texts, we use natural language processing along with machine learning methods in the retrieval process. One of the challenges is deciphering the dataset names, because in many cases researchers refer to variables, rather than the datasets containing them. To constrain the number of terms, we deploy Earth Science ontologies as dictionaries for the term extraction. We demonstrate that storing these terms and underlying ontologies, along with datasets, variables and papers in the knowledge graph database, enables various linkages between all these entities facilitating the data discovery. Thus, we are setting a qualitatively new stage in improvements of web data interfaces, where machine learning techniques are used to establish and optimize usage-based discovery of data.

Irina V Gerasimov↗
DOE OSTI2022

Machine Learning with Gradient-Based Optimization of Nuclear Waste Vitrification with Uncertainties and Constraints

Gekko is an optimization suite in Python that solves optimization problems involving mixed-integer, nonlinear, and differential equations. The purpose of this study is to integrate common Machine Learning (ML) algorithms such as Gaussian Process Regression (GPR), support vector regression (SVR), and artificial neural network (ANN) models into Gekko to solve data based optimization problems. Uncertainty quantification (UQ) is used alongside ML for better decision making. These methods include ensemble methods, model-specific methods, conformal predictions, and the delta method. An optimization problem involving nuclear waste vitrification is presented to demonstrate the benefit of ML in this field. ML models are compared against the current partial quadratic mixture (PQM) model in an optimization problem in Gekko. GPR with conformal uncertainty was chosen as the best substitute model as it had a lower mean squared error of 0.0025 compared to 0.018 and more confidently predicted a higher waste loading of 37.5 wt% compared to 34 wt%. The example problem shows that these tools can be used in similar industry settings where easier use and better performance is needed over classical approaches. Future works with these tools include expanding them with other regression models and UQ methods, and exploration into other optimization problems or dynamic control.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗
DOE OSTI2023

Building thermal dynamics modeling with deep transfer learning using a large residential smart thermostat dataset

Understanding thermal dynamics and obtaining the computational model of residential buildings enable its scaled application in energy retrofits, control optimization and decarbonization. In this paper, we present a deep learning approach to model building thermal dynamics with smart thermostat data collected from residential buildings, with the goal to investigate model generalizability. In the first stage, we developed and compared different Deep Learning architectures including Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) models and CNN-LSTM to predict indoor air temperature in a multi-step time horizon. In the second stage, we implemented a Transfer Learning (TL) process, which aims to improve the prediction performance on a new set of buildings (targets), exploiting the knowledge of related or similar buildings (sources). Different TL strategies and source model identification methods were investigated. The study showed that the CNN-LSTM performed the best among the architectures compared, with an average Mean Absolute Error (MAE) of 0.26 °C for one-hour-ahead (twelve 5-min future steps) predictions. Furthermore, the results showed that freezing the LSTM layer and fine-tuning the other layers of the CNN-LSTM achieved the best performance among four TL strategies, which further improved the performance with respect to a machine learning approach by 10%, and proving the effectiveness and generalizability of the proposed approach. A comparison of three different source model identification methods showed that randomly selecting source models constrained by similar building characteristics can provide good TL performance while retaining simplicity comparing with other quantitative source identification methods.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Reinforcement Learning-based Output Structured Feedback for Distributed Multi-Area Power System Frequency Control

Load frequency control (LFC) is a key factor to maintain the stable frequency in multi-area power systems. As the modern power systems evolve from centralized to decentralized paradigm, LFC needs to consider the decentralized scheme that considers limited information from the information-exchange graph for the generator control of each interconnected area. This paper aims to solve a data-driven constrained LQR problem with mean-variance risk constraints and output structured feedback, and applies this framework to solve the LFC problem in multi-area power systems. By reformulating the constrained optimization problem into a minimax problem, the stochastic gradient descent max-oracle (SGDmax) algorithm with zero-order policy gradient (ZOPG) is adopted to find the optimal feedback gain from the learning, while guaranteeing the convergence. In addition, to improve the adaptation of the proposed learning method to new or varying models, we construct an emulator grid that approximates the dynamics of a physical grid and performs training based on this model. Once the feedback gain is obtained from the emulator grid, it is applied to the physical grid with a robustness test to check whether the controller from the approximated emulator applies to the actual system. Numerical tests show that the obtained feedback controller can successfully control the frequency of each area, while mitigating the uncertainty from the loads, with reliable robustness that ensures the adaptability of the obtained feedback gain to the actual physical grid.

Kwon, Kyung-bin↗
DOE OSTI2022

Dynamic machine learning-based optimization algorithm to improve boiler efficiency

With decreasing computational costs, improvement in algorithms, and the aggregation of large industrial and commercial datasets, machine learning is becoming a ubiquitous tool for process and business innovations. Machine learning is still lacking applications in the field of dynamic optimization for real-time control. This work presents a novel framework for performing constrained dynamic optimization using a recurrent neural network model combined with a metaheuristic optimizer. The framework is designed to augment an existing control system and is purely data-driven, like most industrial Model Predictive Control applications. Several recurrent neural network models are compared as well as several metaheuristic optimizers. Hyperparameters and optimizer parameters are tuned with parameter sweeps, and the resulting values are reported. Further, the best parameters for each optimizer and model combination are demonstrated in closed-loop control of a dynamic simulation, and several recommendations are made for generalizing this framework to other systems. Up to 0.953% improvement is realized over the non-optimized case for a simulated coal-fired boiler. While this is not a large improvement in percentage, the total economic impact is $991,000 per year, and this study builds a foundation for future machine learning with dynamic optimization.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
NASA NTRS2019

Machine Learning, Markov Chain Monte Carlo, and Optimal Algorithms to Characterize the AdvACT Kilopixel Transition-Edge Sensor Arrays

Next-generation focal planes comprising dozens of kilopixel transition-edge sensor (TES) arrays require new methods to rapidly screen candidate arrays, evaluate array non-idealities in the field, identify outlier devices for removal, and optimize the array performance in the field. We demonstrate robust methods to estimate TES parameters (critical temperatures and thermal conductivity parameters) and their uncertainties using a custom Markov Chain Monte Carlo (MCMC) algorithm. We also constrain systematic effects in estimating the TES parameters from non-isothermal current-voltage curves (IVs) at approximately a ~3% level. Additionally, for the first time, we have applied Machine Learning (ML) algorithms to tune detector arrays and optimize their performance.

Maria Salatino↗
DOE OSTI2021

Constrained Block Nonlinear Neural Dynamical Models

Neural network modules conditioned by known priors can be effectively trained and combined to represent systems with nonlinear dynamics. This work explores a novel formulation for data-efficient learning of deep control-oriented nonlinear dynamical models by embedding local model structure and constraints. The proposed method consists of neural network blocks that represent input, state, and output dynamics with constraints placed on the network weights and system variables. For handling partially observable dynamical systems, we utilize a state observer neural network to estimate the states of the system's latent dynamics. We evaluate the performance of the proposed architecture and training methods on system identification tasks for three nonlinear systems: a continuous stirred tank reactor, a two tank interacting system, and an aerodynamics body. Models optimized with a few thousand system state observations accurately represent system dynamics in open loop simulation over thousands of time steps from a single set of initial conditions. Experimental results demonstrate an order of magnitude reduction in open-loop simulation mean squared error for our constrained, block-structured neural models when compared to traditional unstructured and unconstrained neural network models.

Skomski, Elliott↗
DOE OSTI2025

Automated Construction of Artificial Lattice Structures with Designer Electronic States

Manipulating matter with a scanning tunneling microscope (STM) enables the creation of atomically defined artificial structures that host designer quantum states. However, the time-consuming nature of the manipulation process, coupled with the sensitivity of the STM tip, constrains the exploration of diverse configurations and limits the size of the designed features. In this study, we present a reinforcement learning (RL)-based framework for creating artificial structures by spatially manipulating carbon monoxide (CO) molecules on a copper substrate by using the STM tip. The automated workflow combines molecule detection and manipulation, employing deep-learning-based object detection to locate CO molecules and linear assignment algorithms to allocate these molecules to designated target sites. We initially perform molecule maneuvering based on randomized parameter sampling for sample bias, tunneling current set point, and manipulation speed. This data set is then structured into an action trajectory used to train an RL agent. The model is subsequently deployed on the STM for real-time fine-tuning of the manipulation parameters during structure construction. Our approach incorporates path-planning protocols coupled with active drift compensation to enable atomically precise fabrication of structures with significantly reduced human input while realizing larger-scale artificial lattices with the desired electronic properties. Furthermore, using our approach, we demonstrate the automated construction of an extended artificial graphene lattice and confirm the existence of a characteristic Dirac point in its electronic structure. Further challenges regarding the RL-based structural assembly scalability are discussed.

Algorithms↗
DOE OSTI2024

Exact enforcement of temporal continuity in sequential physics-informed neural networks

The use of deep learning methods in scientific computing represents a potential paradigm shift in engineering problem solving. One of the most prominent developments is Physics-Informed Neural Networks (PINNs), in which neural networks are trained to satisfy partial differential equations (PDEs). While this method shows promise, the standard version has been shown to struggle in accurately predicting the dynamic behavior of time-dependent problems. To address this challenge, methods have been proposed that decompose the time domain into multiple segments, employing a distinct neural network in each segment and directly incorporating continuity between them in the loss function of the minimization problem. In this work we introduce a method to exactly enforce continuity between successive time segments via a solution ansatz. This hard constrained sequential PINN (HCS-PINN) method is simple to implement and eliminates the need for any loss terms associated with temporal continuity. The method is tested for a number of benchmark problems involving both linear and non-linear PDEs. Examples include various first order time dependent problems in which traditional PINNs struggle, namely advection, Allen–Cahn, and Korteweg–de Vries equations. Furthermore, second and third order time-dependent problems are demonstrated via wave and Jerky dynamics examples, respectively. Notably, the Jerky dynamics problem is chaotic, making the problem especially sensitive to temporal accuracy. Finally, the numerical experiments conducted with the proposed method demonstrated superior convergence and accuracy over both traditional PINNs and the soft-constrained counterparts.

42 ENGINEERING↗
DOE OSTI2020

A Survey of Constrained Gaussian Process: Approaches and Implementation Challenges

Gaussian process regression is a popular Bayesian framework for surrogate modeling of expensive data sources. As part of a larger effort in scientific machine learning, many recent works have incorporated physical constraints or other a priori information within Gaussian process regression to supplement limited data and regularize the behavior of the model. We provide an overview and survey of several classes of Gaussian process constraints, including positivity or bound constraints, monotonicity and convexity constraints, differential equation constraints provided by linear PDEs, and boundary condition constraints. We compare the strategies behind each approach as well as the differences in implementation, concluding with a discussion of the computational challenges introduced by constraints.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Neuromorphic overparameterisation and few-shot learning in multilayer physical neural networks

Abstract Physical neuromorphic computing, exploiting the complex dynamics of physical systems, has seen rapid advancements in sophistication and performance. Physical reservoir computing, a subset of neuromorphic computing, faces limitations due to its reliance on single systems. This constrains output dimensionality and dynamic range, limiting performance to a narrow range of tasks. Here, we engineer a suite of nanomagnetic array physical reservoirs and interconnect them in parallel and series to create a multilayer neural network architecture. The output of one reservoir is recorded, scaled and virtually fed as input to the next reservoir. This networked approach increases output dimensionality, internal dynamics and computational performance. We demonstrate that a physical neuromorphic system can achieve an overparameterised state, facilitating meta-learning on small training sets and yielding strong performance across a wide range of tasks. Our approach’s efficacy is further demonstrated through few-shot learning, where the system rapidly adapts to new tasks.

Science & Technology - Other Topics↗
DOE OSTI2025

Diagnostic-free onboard battery health assessment

Diverse usage patterns induce complex and variable aging behaviors in lithiumion batteries, complicating accurate health diagnosis and prognosis. Separate diagnostic cycles are often used to untangle the battery’s current state of health from prior complex aging patterns. However, these same diagnostic cycles alter the battery’s degradation trajectory, are time-intensive, and cannot be practically performed in onboard applications. Here, in this work, we leverage portions of operational measurements in combination with an interpretable machine learning model to enable rapid, onboard battery health diagnostics and prognostics without offline diagnostic testing and the requirement of historical data. We integrate mechanistic constraints within an encoder-decoder architecture to extract electrode states in a physically interpretable latent space and enable improved reconstruction of the degradation path. The health diagnosis model framework can be flexibly applied across diverse application interests with slight fine-tuning.

battery aging reconstruction↗
NASA NTRSDate not supplied

Towards Quantifying and Correlating RTV Intumescence to Entry-Relevant Conditions

Room Temperature Vulcanizing silicone (RTV) is a high-temperature adhesive used as a gap-filler between Thermal Protection System (TPS) tiles for heatshields on numerous missions. It is also adopted to bond instrumentation plugs of temperature and pressure sensors into heatshield’s tiles. While RTV has been traditionally assumed to be a non-porous and non-ablating material, numerous experiments have shown that RTV pyrolyzes and becomes highly porous as it is heated. Experimental data has also shown heating rate-dependent swelling (or intumescence) and shrinking of RTV, these volume changes may cause roughness-induced boundary layer transition. Outgassing and surface oxide formation of RTV upon decomposition also occur, which can be sources of contamination of heat shield sensors. This combination of detrimental effects motivates a critical need to develop a high-fidelity model for RTV ablation. As a first critical step in RTV modeling, a comprehensive material properties database for RTV was collected that account for the lack of data in properties such as pyrolysis char yield, shape change, virgin and char porosity. The next step is to quantify the volume change of RTV as a function of temperature. While this was done in previous campaigns, limited data was collected which showed a possible heating rate dependent intumescent behavior. To further understand the intumescence phenomenon of RTV, numerous dedicated experiments were performed at Beamline 8.3.2 of the Advanced Light Source, where RTV samples were heated while X-ray scans were taken in situ. Micro-Computed Tomography scans were taken in situ for low heating rates (≤ 60 °C/min), whereas continuous radiography scans were taken for higher heating rates (≤ 1500 ºC/min). Unconstrained samples and samples constrained in different holders (quartz, graphite, FiberForm and PICA) were tested to determine the change in intumescence due to confinement. Using deep learning techniques, the tomographies and radiographies are segmented to obtain volume change, porosity, pore size distributions and connectivity. From the results obtained, modifications to the previous RTV intumescence model are proposed, along with preliminary comparisons to testing RTV at high-enthalpy facilities.

RTV↗
DOE OSTI2024

Active learning for the design of polycrystalline textures using conditional normalizing flows

Generative modeling has opened new avenues for solving previously intractable materials design problems. However, these new opportunities are accompanied by a drastic increase in the required amount of training data. This is in stark juxtaposition to the high expense and difficulty in curating such large materials datasets. In this work, we propose a novel framework for integrating generative models within an active learning loop. Further, this enables the training of generative models with datasets significantly smaller than what has previously been demonstrated, providing a direct route for their application in data constrained environments. The functionality of this framework is then demonstrated by addressing the challenge of designing polycrystalline textures associated with target anisotropic mechanical properties. The developed protocol exhibited a cost reduction between 14 to 18 times over a randomly sampled experimental design.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Online energy consumption forecast for battery electric buses using a learning-free algebraic method

Accurately predicting the energy consumption plays a vital role in battery electric buses (BEBs) route planning and deployment. Based on the algebraic derivative estimation, we present a novel method to forecast the energy consumption in real time. In contrast to the mainstream machine-learning-based methods, the proposed method does not require access to the historical energy consumption data. It eliminates the time-consuming and computationally expensive offline training. Consequently, its prediction performance is not constrained by the quantity and quality of the training data. Moreover, the method can swiftly adapt to new situations not included in the previous driving cycles, which makes it especially suitable for emerging transport modes, e.g., on-demand transit services. In addition, its online execution only involves algebraic calculations, yielding superior calculation efficiency. Using real-world data, we comprehensively compare the performance of the proposed learning-free algebraic method with multiple representative machine-learning-based methods. Finally, the advantages and limitations of the proposed method are discussed in detail.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2023

Neutron Leakage Spectra of the EUCLID Experiment [Abstract]

Integral experiments with sub-critical and critical configurations of special nuclear material are performed in support of nuclear data validation and adjustment. Different nuclear data evaluations may have different values for individual cross sections due to uncertainties in (or lack of) differential experiments, but compensating errors in these data sets can lead to the same k eff results for one application while vastly different for another application. One example of this is 239 Pu, where both ENDF/B-VIII.0 and JEFF-3.3 correctly compute k eff of the Jezebel critical assembly, but the individual contributions from each reaction are vastly different. To help resolve this specific case, the Experiments Underpinned by Computational Learning for Improvements in Nuclear Data (EUCLID) project used machine learning to design a set of experiments to help resolve the compensating errors in 239 Pu. A total of six responses were measured during the experimental campaign, which constrain the data in ways that k eff alone cannot and will be used for adjustment of the nuclear data. One of these responses is the neutron leakage spectrum, which recent work has shown to be useful for constraining the prompt fission neutron spectrum and inelastic scattering. The neutron leakage spectra were measured utilizing a 3 in. right cylinder EJ-301D detector. The measured signal in the detector was deconvoluted using spectrum unfolding techniques, which are presented and compared to simulations.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2025

Improving the Parameterization of Cloud and Rain Microphysics in E3SM using Novel Observationally-Constrained Bayesian Approach (Final Technical Report)

In this project, we sought to develop new cloud and rain microphysics frameworks within the Energy Exascale Earth System Model (E3SM). This work encompassed two primary avenues of research: 1) Further development of a Bayesian-based scheme called BOSS (Bayesian Observationally-constrained Statistical-physical Scheme) to represent cloud and rain microphysics, testing it in realistic high-resolution cloud models, and implementing it in E3SM; 2) Development of a methodology utilizing machine learning to enable computationally tractable use of tractable use of Markov chain Monte Carlo sampling for Bayesian parameter estimation in Earth system and cloud models. In this project, we adapted the BOSS microphysics scheme, originally formulated for rain-only, to include all liquid-phase microphysical processes for cloud and rain, in particular the processes that mediate between these two categories, for example the conversion from cloud to rain through collision and coalescence of drops. We constrained the scheme via comparison and testing against a detailed model that explicitly represents the evolution of cloud and rain particles, called a bin microphysics scheme.

54 ENVIRONMENTAL SCIENCES↗