Search NASA⌕ Search

SEARCH · Search NASA

Results for “Model Predictive Control”

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 1,441 records · Page 80

DSPS: Differentiable stellar population synthesis

ABSTRACT Models of stellar population synthesis (SPS) are the fundamental tool that relates the physical properties of a galaxy to its spectral energy distribution (SED). In this paper, we present DSPS: a python package for SPS. All of the functionality in DSPS is implemented natively in the JAX library for automatic differentiation, and so our predictions for galaxy photometry are fully differentiable, and directly inherit the performance benefits of JAX, including portability onto GPUs. DSPS also implements several novel features, such as i) a flexible empirical model for stellar metallicity that incorporates correlations with stellar age, ii) support for the Diffstar model that provides a physically-motivated connection between the star formation history of a galaxy (SFH) and the mass assembly of its underlying dark matter halo. We detail a set of theoretical techniques for using autodiff to calculate gradients of predictions for galaxy SEDs with respect to SPS parameters that control a range of physical effects, including SFH, stellar metallicity, nebular emission, and dust attenuation. When forward modelling the colours of a synthetic galaxy population, we find that DSPS can provide a factor of 5 speed-up over standard SPS codes on a CPU, and a factor of 300-400 on a modern GPU. When coupled with gradient-based techniques for optimization and inference, DSPS makes it practical to conduct expansive likelihood analyses of simulation-based models of the galaxy–halo connection that fully forward model galaxy spectra and photometry.

79 ASTRONOMY AND ASTROPHYSICS↗

On-Line Prediction of Resistant Force During Soil–Tool Interaction

For off-road vehicles such as excavators and wheel loaders, a large portion of energy is consumed to overcome the soil resistant force in the digging process. For optimal control of the digging tool, a high-fidelity model of the soil–tool interaction force is important to reduce energy consumption. Here, in this paper, an on-line soil resistant force prediction method is proposed. In this method, a hybrid model, which combines a physical model and a data-driven model, is used for the force prediction. In addition, the parameters of the hybrid model can be updated on-line based on real-time data. Comparisons with experimental data demonstrate that the proposed prediction method has an average error of around 12.7%.

42 ENGINEERING↗

A localized ensemble of approximate Gaussian processes for fast sequential emulation

More attention has been given to the computational cost associated with the fitting of an emulator. Substantially less attention is given to the computational cost of using that emulator for prediction. This is primarily because the cost of fitting an emulator is usually far greater than that of obtaining a single prediction, and predictions can often be obtained in parallel. In many settings, especially those requiring Markov Chain Monte Carlo, predictions may arrive sequentially and parallelization is not possible. In this case, using an emulator procedure which can produce accurate predictions efficiently can lead to substantial time savings in practice. In this paper, we propose a global model approximate Gaussian process framework via extension of a popular local approximate Gaussian process (laGP) framework. Our proposed emulator can be viewed as a treed Gaussian process where the leaf nodes are laGP models, and the tree structure is learned greedily as a function of the prediction stream. The suggested method (called leapGP) has interpretable tuning parameters which control the time‐memory trade‐off. One reasonable choice of settings leads to an emulator with a training cost and makes predictions rapidly with an asymptotic amortized cost of .

97 MATHEMATICS AND COMPUTING↗

Network-cognizant voltage droop control

The present disclosure provides techniques for network-cognizant droop control in power systems, such as a power distribution system. An example device includes a processor configured to determine, based on (i) a model representing a structure of a power system that includes a plurality of energy resources and (ii) an indication of predicted uncontrollable power injections in the power system, for each controllable energy resource in the plurality of energy resources, a respective value of a first droop coefficient and a respective value of a second droop coefficient. The processor may be further configured to cause at least one controllable energy resource in the plurality of energy resources to modify an output power of the at least one energy resource based on the respective value of the first droop coefficient and the respective value of the second droop coefficient.

24 POWER TRANSMISSION AND DISTRIBUTION↗

The predictability of advection-dominated flux-transport solar dynamo models

Space weather is a matter of practical importance in our modern society. Predictions of forecoming solar cycles mean amplitude and duration are currently being made based on flux-transport numerical models of the solar dynamo. Interested in the forecast horizon of such studies, we quantify the predictability window of a representative, advection-dominated, flux-transport dynamo model by investigating its sensitivity to initial conditions and control parameters through a perturbation analysis. We measure the rate associated with the exponential growth of an initial perturbation of the model trajectory, which yields a characteristic timescale known as the e-folding time τ {sub e}. The e-folding time is shown to decrease with the strength of the α-effect, and to increase with the magnitude of the imposed meridional circulation. Comparing the e-folding time with the solar cycle periodicity, we obtain an average estimate for τ {sub e} equal to 2.76 solar cycle durations. From a practical point of view, the perturbations analyzed in this work can be interpreted as uncertainties affecting either the observations or the physical model itself. After reviewing these, we discuss their implications for solar cycle prediction.

79 ASTRONOMY AND ASTROPHYSICS↗

Twinac: initiation of a community-driven accelerator digital twin framework

We present the initiation of a community-driven framework for the integration of accelerator digital twins into control systems: Twinac. Few facilities have fully integrated accelerator digital twins like at Cornell’s CHESS. Many facilities have active research to employ surrogate models to aid in operational decisions like at Argonne’s ALS, MSU’s FRIB, SLAC’s LCLS-II, and Fermilab’s FAST/IOTA, PIP-II, and main complex. To lower the barrier to entry for all accelerator facilities to build and benefit from a digital twin of their own accelerators, we propose the following software framework. Twinac will provide the capability to compose one’s own digital twin using reusable components engineered at other facilities. With this model in place, Twinac will also support tools for (1) predictive maintenance systems; (2) discovery of correlated but uncontrolled environmental factors, like seasonal temperature variations causing performance changes on power supplies, magnets, etc.; and (3) prototyping and updating sophisticated optimization and controls algorithms. The Twinac framework will enable sharing and simplified deployment of modeled components and control algorithms at all facilities. With an inter-facility team to build and support the Twinac framework, it will be easy to publish and try out the latest advancements at one’s own facility.

Miceli, Tia [Fermilab]↗

Delamination-informed lifecycle decisions: A dielectric and machine learning framework for composite sorting and recycling

Composite materials are widely used in aerospace, marine, and automotive sectors due to their high strength-to-weight ratio and durability. However, their long-term reliability can be compromised by damage accumulation. Specifically, delamination initiation serves as a precursor to structural failure, which is often difficult to detect during damage inspection. Identifying and sorting delamination initiation in samples not only increases operational safety while providing critical information for end-of-life decisions, which influences both the service life extension value and the efficiency of fiber extraction during recycling. This research addresses two challenges: (1) developing a nondestructive, ex-situ framework to sort composite materials based on damage severity, particularly delamination, and (2) understanding how damage in composites influences resin removal during pyrolysis. Both experimental work and finite element analysis were performed to predict critical stress levels that are associated with delamination onset. Based on these results, three loading levels 50 %, 75 %, and 90 % of maximum stress, were selected for controlled experiments, generating composite samples with varying extents of damage for machine learning model training. Microscopic imaging of these samples confirmed the damage progression from matrix cracking to delamination, validating the computational predictions. We explored supervised machine learning using dielectric measurements to classify damage states. Preliminary results show an artificial neural network can identify early delamination which is a potential precursor to failure, with 94.44 % accuracy on our dataset. A parallel investigation into the effect of damage severity on pyrolysis recycling showed that heavily delaminated samples required significantly less energy for comparable matrix removal than undamaged samples.

dielectric variables↗

Stochastic Predictions of Cell Kill During Stereotactic Ablative Radiation Therapy: Do Hypoxia and Reoxygenation Really Matter?

Purpose: To simulate stereotactic ablative radiation therapy on hypoxic and well-oxygenated in silico tumors, incorporating probabilistic parameter distributions and linear-quadratic versus linear-quadratic-cubic methodology and the evaluation of optimal fractionation schemes using biological effective dose (BED{sub α/β=10} {sub or} {sub 3}) comparisons. Methods and Materials: A temporal tumor growth and radiation therapy algorithm simulated high-dose external beam radiation therapy using stochastic methods. Realistic biological proliferative cellular hierarchy and pO{sub 2} histograms were incorporated into the 10{sup 8}-cell tumor model, with randomized radiation therapy applied during continual cell proliferation and volume-based gradual tumor reoxygenation. Dose fractions ranged from 6-35 Gy, with predictive outcomes presented in terms of the total doses (converted to BED) required to eliminate all cells that could potentially regenerate the tumor. Results: Well-oxygenated tumor control BED{sub 10} outcomes were not significantly different for high-dose versus conventional radiation therapy (BED{sub 10}: 79-84 Gy; Equivalent Dose in 2 Gy fractions with α/β of 10: 66-70 Gy); however, total treatment times decreased from 7 down to 1-3 weeks. For hypoxic tumors, an additional 28 Gy (51 Gy BED{sub 10}) was required, with BED{sub 10} increasing with dose per fraction due to wasted dose in the final fraction. Fractions of 9 Gy compromised well for total treatment time and BED, with BED{sub 10}:BED{sub 3} of 84:176 Gy for oxic and 132:278 Gy for non-reoxygenating hypoxic tumors. Initial doses of 12 Gy followed by 6 Gy further increased the therapeutic ratio. When delivering ≥9 Gy per fraction, applying reoxygenation and/or linear-quadratic-cubic cell survival both affected tumor control doses by a significant 1-2 fractions. Conclusions: The complex temporal dynamics of tumor oxygenation combined with probabilistic cell kinetics in the modeling of radiation therapy requires sophisticated stochastic modeling to predict tumor cell kill. For stereotactic ablative radiation therapy, high doses in the first week followed by doses that are more moderate may be beneficial because a high percentage of hypoxic cells could be eradicated early while keeping the required BED{sub 10} relatively low and BED{sub 3} toxicity to tolerable levels.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Impacts of elevated dissolved CO 2 on a shallow groundwater system: Reactive transport modeling of a controlled-release field test

We report one of the risks that CO 2 geological sequestration imposes on the environment is the impact of potential CO 2 /brine leakage on shallow groundwater. The reliability of reactive transport models predicting the response of groundwater to CO 2 leakage depends on a thorough understanding of the relevant chemical processes and key parameters affecting dissolved CO 2 transport and reaction. Such understanding can be provided by targeted field tests integrated with reactive transport modeling. A controlled-release field experiment was conducted in Mississippi to study the CO 2 -induced geochemical changes in a shallow sandy aquifer at about 50 m depth. The field test involved a dipole system in which the groundwater was pumped from one well, saturated with CO 2 at the pressure corresponding to the hydraulic pressure of the aquifer, and then re-injected into the same aquifer using a second well. Groundwater samples were collected for chemical analyses from four monitoring wells before, during and after the dissolved CO 2 was injected. In this paper, we present reactive transport models used to interpret the observed changes in metal concentrations in these groundwater samples. A reasonable agreement between simulated and measured concentrations indicates that the chemical response in the aquifer can be interpreted using a conceptual model that encompasses two main features: (a) a fast-reacting but limited pool of reactive minerals that responds quickly to changes in pH and causes a pulse-like concentration change, and (b) a slow-reacting but essentially unlimited mineral pool that yields rising metal concentrations upon decreased groundwater velocities after pumping and injection stopped. During the injection, calcite dissolution and Ca-driven cation exchange reactions contribute to a sharp pulse in concentrations of Ca, Ba, Mg, Mn, K, Li, Na and Sr, whereas desorption reactions control a similar increase in Fe concentrations. Finally, after the injection and pumping stops and the groundwater flow rate decreases, the dissolution of relatively slow reacting minerals such as plagioclase drives the rising concentrations of alkali and alkaline earth metals observed at later stages of the test, whereas the dissolution of amorphous iron sulfide causes slowly increasing Fe concentrations.

58 GEOSCIENCES↗

Divertor Plasma Detachment Control Neural Network

DivControlNN is a state-of-the-art software tool that leverages advanced machine learning techniques to predict and control divertor plasma behavior in fusion reactors. Plasma, a highly energetic and electrically charged gas, requires meticulous management to protect reactor components and maintain optimal energy production. Conventional simulation methods, although extremely detailed, typically demand extensive computational time-making them unsuitable for real-time control scenarios. DivControlNN addresses this challenge by learning from tens of thousands of high-fidelity simulations, thereby creating a rapid surrogate model that can deliver near-instantaneous predictions. At the core of its functionality is a sophisticated technique known as latent space mapping, which condenses complex, high-dimensional plasma data into a compact, lower-dimensional representation. This streamlined representation enables the system to quickly forecast essential plasma properties and determine the precise conditions required for effective detachment. Detachment is a crucial process in which the plasma is cooled before reaching the divertor plates, thereby reducing heat loads and mitigating material erosion. In recent experiments conducted on the KSTAR tokamak in South Korea, DivControlNN successfully guided the detachment process without any fine-tuning-even when applied to a new tungsten divertor configuration. By achieving a computational speed-up of over one hundred million times compared to traditional simulation methods while maintaining low prediction errors, DivControlNN stands to significantly enhance real-time control and diagnostic capabilities in future fusion reactors. This breakthrough paves the way for safer, more reliable reactor operation and represents a major advancement toward realizing fusion energy as a practical, sustainable, and clean power source.

Xu, Xueqiao [Lawrence Livermore National Laborator↗

Multi-decadal variation of ENSO forecast skill since the late 1800s

Diagnosing El Niño-Southern Oscillation (ENSO) predictability within operational forecast models is hindered by computational expense and the need for initialization with three-dimensional fields generated by global data assimilation. We instead examine multi-year ENSO predictability since the late 1800s using the model-analog technique, which has neither limitation. We first draw global coupled model states from pre-industrial control simulations, from the Coupled Model Intercomparison Project Phase 6, that are chosen to initially match observed monthly sea surface temperature and height anomalies in the Tropics. Their subsequent 36-month model evolution are the hindcasts, whose 20th century ENSO skill is comparable to twice-yearly hindcasts generated by a state-of-the-art European operational forecasting system. Despite the so-called spring predictability barrier, present throughout the record, there is substantial second-year ENSO skill, especially after 1960. Overall, ENSO exhibited notably high values of both amplitude and skill towards the end of the 19th century, and again in recent decades.

54 ENVIRONMENTAL SCIENCES↗

The kinetics of static recovery by dislocation climb

Abstract The initial microstructure of a wide range of structural materials is conditioned by thermo-mechanical treatments such as hot-working, tempering, or solution annealing. At the elevated temperatures associated with these treatments the dislocation microstructure evolves, usually decreasing in density through a process known as static recovery. Despite its technological relevance, static recovery is not fully characterized from a theoretical standpoint, with even the controlling mechanisms subject to debate. In this study, a climb-enabled discrete dislocation dynamics (DDD) capability is leveraged to explore the kinetics of static recovery in pure Fe when controlled by dislocation climb. Quantitative data from these simulations is used to develop a revised static recovery law, and provides the parameters appropriate for predictive microstructure models in Fe. This law differs from previous analytical derivations invoking climb of dislocations, following the logarithmic trends typical of experimental observations where prior work did not. Direct comparison between the recovery law derived from DDD to experimental recovery data in alpha Fe shows strong agreement across a range of temperatures, and suggests that climb is the controlling mechanism for static recovery in pure metals.

36 MATERIALS SCIENCE↗

Online Learning of Effective Turbine Wind Speed in Wind Farms

To develop better wind farm controllers that can meet more complex objectives, methods of modeling the wind turbine wakes at low computational expense are needed. Gaussian process (GP) regression offers a computationally inexpensive framework for learning complex functions from noisy measurements with very few datapoints. In this work, an online learning approach is presented to learn the rotor-averaged wind velocity at downstream wind turbines with GPs, using the available datastream of wind field measurements and wind turbine control set-points. This framework can readily be integrated into model-based controls methods because the model a) is updated online at low computational expense, b) assumes a mathematically favorable Gaussian form, and c) explicitly quantifies the stochastic nature of the wake field so that the trade-off between exploration and exploitation, and the uncertainty in the prediction, can be utilized. We show that a GP-learned model can match true values with errors within 0.5% on average, with as few as 5 training data points.

Gaussian process↗

Evaluating the limitations of Bayesian metabolic control analysis

AbstractBayesian Metabolic Control Analysis (BMCA) has emerged as a promising framework for inferring metabolic control coefficients in data-limited scenarios by integrating Bayesian inference with linlog rate laws. However, its predictive accuracy and limitations remain underexplored. This study systematically evaluates BMCA’s ability to infer elasticity values, flux control coefficients (FCCs), and concentration control coefficients (CCCs) under varying data availability conditions using three synthetic metabolic network models. Our findings highlight the strengths and weaknesses of BMCA, guiding its application in metabolic engineering and emphasizing the need for methodological refinements.Author summaryUnderstanding how enzymes control metabolic pathways is crucial for optimizing biomanufacturing and synthetic biology applications. Bayesian Metabolic Control Analysis (BMCA) is a promising computational method that integrates Bayesian inference with metabolic control analysis to estimate key control parameters, even in cases with limited experimental data. However, the accuracy and limitations of BMCA remain unclear. In this study, we systematically evaluate BMCA using three synthetic metabolic networks to determine how different types of physiological data impact its predictive performance. We find that BMCA requires flux and enzyme concentration data for accurate predictions, while external metabolite concentrations contribute little. Additionally, BMCA fails to predict elasticity values beyond a magnitude of 1.5 and reliably infer allosteric regulation, even when strong regulatory interactions exist. In addition, BMCA does not accurately rank metabolic control points, which may limit its utility in identifying key enzymes in engineered pathways. Our work provides practical insights into when and how BMCA can be applied, guiding future research in metabolic modeling and control analysis.

Shin, Janis (ORCID:0000000216572455)↗

Predicting and Controlling Corrosion (Abbreviated Final Report)

Corrosion accumulates significant replacement costs in the transportation, utility, manufacturing, and infrastructure sectors. It also underpins several of LLNL’s core missions in stockpile stewardship, defense, and energy security. Whereas most corrosion and aging models are empirically parameterized to describe well-defined conditions late in the reaction, the factors that determine the early stages of corrosion—during which mitigation could be most impactful— are poorly understood. This project addressed the critical need for new approaches to predict the kinetics of corrosion initiation based on firm physical and chemical understanding. The activities encompassed degradation of relevant metals in both hydrogen-rich and environmental corrosion scenarios. The team integrated state-of-the-art multiscale simulation, in situ characterization, and data science within three technical thrusts: hydriding of Ti alloys; aqueous corrosion of Al and Ni-Cr alloys; and degradation of additively manufactured 316L stainless steel. In each case, novel capabilities were developed to identify and track the impacts of key atomistic, compositional, and microstructural features on the metal systems. For hydriding, protocols were developed to tightly integrate multiscale models, advanced multimodal characterization, and machine learning to determine how hydrogen interacts with native passivating surface oxides and nucleates new undesired phases, shedding new light on the critical role of grain boundaries, interfaces, and atomically disordered regions. For aqueous corrosion, the project demonstrated methods to predict dissolution rates of metal surfaces in corrosive solutions, to measure and understand microstructural and grain orientation effects on corrosion susceptibility, and to investigate competing growth and dissolution kinetics of surface oxides. For additively manufactured metals, analysis using state-of-the-art microscopy techniques revealed the role of specific heterogeneities invoked during laser processing, including cellular structure, dislocations, and precipitates, on corrosion susceptibility. In addition to new capabilities and understanding, the project provided an avenue for workforce development, as well as key partnerships with stakeholders in corrosion science.

08 HYDROGEN↗

A quantitative model for a nanoscale switch accurately predicts thermal actuation behavior

Manipulation of temperature can be used to actuate DNA origami nano-hinges containing gold nanoparticles. We develop a physical model of this system that uses partition function analysis of the interaction between the nano-hinge and nanoparticle to predict the probability that the nano-hinge is open at a given temperature. The model agrees well with experimental data and predicts experimental conditions that allow the actuation temperature of the nano-hinge to be tuned over a range of temperatures from 30 °C to 45 °C. Additionally, the model identifies microscopic interactions that are important to the macroscopic behavior of the system, revealing surprising features of the system. This combination of physical insight and predictive potential is likely to inform future designs that integrate nanoparticles into dynamic DNA origami structures or use strand binding interactions to control dynamic DNA origami behavior. Furthermore, our modeling approach could be expanded to consider the incorporation, stability, and actuation of other types of functional elements or actuation mechanisms integrated into nucleic acid devices.

36 MATERIALS SCIENCE↗

Triplet-pair spin signatures from macroscopically aligned heteroacenes in an oriented single crystal

The photo-driven process of singlet fission generates coupled triplet pairs (TT) with fundamentally intriguing and potentially useful properties. The quintet 5 TT 0 sublevel is particularly interesting for quantum information because it is highly entangled, is addressable with microwave pulses, and could be detected using optical techniques. Previous theoretical work on a model Hamiltonian and nonadiabatic transition theory, called the JDE model, has determined that this sublevel can be selectively populated if certain conditions are met. Among the most challenging, the molecules within the dimer undergoing singlet fission must have their principal magnetic axes parallel to one another and to an applied Zeeman field. Here, we present time-resolved electron paramagnetic resonance (TR-EPR) spectroscopy of a single crystal sample of a tetracenethiophene compound featuring arrays of dimers aligned in this manner, which were mounted so that the orientation of the field relative to the molecular axes could be controlled. The observed spin sublevel populations in the paired TT and unpaired (T+T) triplets are consistent with predictions from the JDE model, including preferential 5 TT 0 formation at z ‖ B 0 , with one caveat—two 5 TT spin sublevels have little to no population. This may be due to crossings between the 5 TT and 3 TT manifolds in the field range investigated by TR-EPR, consistent with the intertriplet exchange energy determined by monitoring photoluminescence at varying magnetic fields.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Monitoring Noble Gases (Xe and Kr) and Aerosols (Cs and Rb) in a Molten Salt Reactor Surrogate Off-Gas Stream Using Laser-Induced Breakdown Spectroscopy (LIBS)

In this study with surrogate materials we show that laser-induced breakdown spectroscopy (LIBS) is a robust tool with promising capability toward monitoring gaseous (Xe and Kr) and aerosol (Cs and Rb) species in an off-gas stream from a molten salt reactor (MSR). MSRs will continually evolve fission products into the cover gas flowing across the reactor headspace. The cover gas entrains Xe and Kr gases, along with aerosol particles, before passing into an off-gas treatment system. Univariate models of Xe and Kr peaks showed a strong correlation to concentration indicated by their coefficients of determination of 0.983 and 0.997, respectively. Multivariate models were built for all four analytes using partial least squares regression coupled with preprocessing steps including normalization, trimming, and/or genetic algorithm derived filters. The models were evaluated by predicting the concentrations of the analytes in four validation samples, in which all calibration models were successfully validated at a confidence interval of 99.9%. Finally, pressure controllers were used to regulate the mass flow rate of Kr flowing into the measurement cell in sinusoidal and stepwise waveforms to test the real-time monitoring capabilities of the regression models. Both univariate and partial least squares Kr models were able to successfully quantify the gas concentration in the real-time evaluation. The root mean squared error of prediction (RMSEP) values for these real-time tests were calculated to be 0.051, 0.060, and 0.121 mol% demonstrating the measurement systems’ capability to perform online monitoring with acceptable accuracy.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗