Search NASA⌕ Search

SEARCH · Search NASA

Results for “Random systems”

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 1,675 records · Page 93

Initial position optimization in molecular dynamics simulations for a Coulomb system

A new algorithm for molecular dynamics (MD) simulations is developed to optimize plasma particle distributions at given initial temperatures. By combining velocity scaling and reassignment, the method effectively eliminates the initial rise and oscillation in temperatures observed with randomly distributed positions. These rises and oscillations are undesired numerical artifacts observed in conventional plasma MD simulations, arising from unoptimized particle positions. The algorithm demonstrates temperature relaxation without initial rises or oscillations, as well as precise flow velocity relaxation, enabling accurate measurement of relaxation times. The code is accelerated using graphics processing units for parallel processing, enhancing the study of plasma dynamics. The proposed method for distributing physically valid particles in MD simulations enables accurate studies of intrinsic collision processes in plasmas, including the dynamics of strongly coupled plasmas, plasma–wave interactions, and transport phenomena in magnetized plasmas. The paper concludes with a discussion of potential applications and future enhancements to the algorithm.

Jo, Jawon (ORCID:0009000924193285)↗

Risk Assessment of Rare Events in Probabilistic Power Flow via Hybrid Multi-Surrogate Method

The risks associated with rare events threatening the security of power system operation are of paramount importance to power system planners and operators. To analyze the risks caused by high-impact, low-frequency rare events, an immensely large number of samples are typically required for the Monte-Carlo (MC) method on the high-fidelity power system model to achieve a sufficient accuracy, thereby rendering this approach computationally prohibitive. To handle this problem efficiently, it is desirable to construct a surrogate model for the power system response. However, the straightforward MC sampling of the low-fidelity surrogate can lead to biased results in the low-probability tail regions that are vital to risk assessment. Moreover, a single surrogate is unable to handle the topology uncertainties caused by random branch outages. To overcome these issues, we propose a hybrid multi-surrogate (HMS) method based on the polynomial chaos expansion (PCE) with low-probability tail events reevaluated by the high-fidelity model through a probabilistic analysis. This method improves the computational efficiency of the MC method for rare-event risk assessment by leveraging multi-fidelity models while retaining the desired accuracy. Simulations conducted in three test systems verify the excellent performances of the HMS method.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Exact Averaging of Stochastic Equations for Flow in Porous Media

It is well known that at present, exact averaging of the equations for flow and transport in random porous media have been proposed for limited special fields. Moreover, approximate averaging methods--for example, the convergence behavior and the accuracy of truncated perturbation series--are not well studied, and in addition, calculation of high-order perturbations is very complicated. These problems have for a long time stimulated attempts to find the answer to the question: Are there in existence some, exact, and sufficiently general forms of averaged equations? Here, we present an approach for finding the general exactly averaged system of basic equations for steady flow with sources in unbounded stochastically homogeneous fields. We do this by using (1) the existence and some general properties of Green's functions for the appropriate stochastic problem, and (2) some information about the random field of conductivity. This approach enables us to find the form of the averaged equations without directly solving the stochastic equations or using the usual assumption regarding any small parameters. In the common case of a stochastically homogeneous conductivity field we present the exactly averaged new basic nonlocal equation with a unique kernel-vector. We show that in the case of some type of global symmetry (isotropy, transversal isotropy, or orthotropy), we can for three-dimensional and two-dimensional flow in the same way derive the exact averaged nonlocal equations with a unique kernel-tensor. When global symmetry does not exist, the nonlocal equation with a kernel-tensor involves complications and leads to an ill-posed problem.

54 ENVIRONMENTAL SCIENCES↗

OBSERVATIONS OF THE WASP-2 SYSTEM BY THE APOSTLE PROGRAM

We present transit observations of the WASP-2 exoplanet system by the Apache Point Survey of Transit Lightcurves of Exoplanets (APOSTLE) program. Model fitting to these data allows us to improve measurements of the hot-Jupiter exoplanet WASP-2b and its orbital parameters by a factor of {approx}2 over prior studies; we do not find evidence for transit depth variations. We do find reduced {chi}{sup 2} values greater than 1.0 in the observed minus computed transit times. A sinusoidal fit to the residuals yields a timing semi-amplitude of 32 s and a period of 389 days. However, random rearrangements of the data provide similar quality fits, and we cannot with certainty ascribe the timing variations to mutual exoplanet interactions. This inconclusive result is consistent with the lack of incontrovertible transit timing variations (TTVs) observed in other hot-Jupiter systems. This outcome emphasizes that unique recognition of TTVs requires dense sampling of the libration cycle (e.g., continuous observations from space-based platforms). However, even in systems observed with the Kepler spacecraft, there is a noted lack of transiting companions and TTVs in hot-Jupiter systems. This result is more meaningful, and indicates that hot-Jupiter systems, while they are easily observable from the ground, do not appear to be currently configured in a manner favorable to the detection of TTVs. The future of ground-based TTV studies may reside in resolving secular trends, and/or implementation at extreme quality observing sites to minimize atmospheric red noise.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Development of a Computer-Aided Differential Diagnosis System to Distinguish Between Usual Interstitial Pneumonia and Non-specific Interstitial Pneumonia Using Texture- and Shape-Based Hierarchical Classifiers on HRCT Images

A computer-aided differential diagnosis (CADD) system that distinguishes between usual interstitial pneumonia (UIP) and non-specific interstitial pneumonia (NSIP) using high-resolution computed tomography (HRCT) images was developed, and its results compared against the decision of a radiologist. Six local interstitial lung disease patterns in the images were determined, and 900 typical regions of interest were marked by an experienced radiologist. A support vector machine classifier was used to train and label the regions of interest of the lung parenchyma based on the texture and shape characteristics. Based on the regional classifications of the entire lung using HRCT, the distributions and extents of the six regional patterns were characterized through their CADD features. The disease division index of every area fraction combination and the asymmetric index between the left and right lungs were also evaluated. A second SVM classifier was employed to classify the UIP and NSIP, and features were selected through sequential-forward floating feature selection. For the evaluation, 54 HRCT images of UIP (n = 26) and NSIP (n = 28) patients clinically diagnosed by a pulmonologist were included and evaluated. The classification accuracy was measured based on a fivefold cross-validation with 20 repetitions using random shuffling. For comparison, thoracic radiologists assessed each case using HRCT images without clinical information or diagnosis. The accuracies of the radiologists’ decisions were 75 and 87%. The accuracies of the CADD system using different features ranged from 70 to 81%. Finally, the accuracy of the proposed CADD system after sequential-forward feature selection was 91%.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Corrective Control to Handle Forecast Uncertainty: A Chance Constrained Optimal Power Flow

Higher shares of electricity generation from renewable energy sources and market liberalization is increasing uncertainty in power systems operation. At the same time, operation is becoming more flexible with improved control systems and new technology such as phase shifting transformers (PSTs) and high voltage direct current connections (HVDC). Previous studies have shown that the use of corrective control in response to outages contributes to a reduction in operating cost, while maintaining N-1 security. In this work, we propose a method to extend the use of corrective control of PSTs and HVDCs to react to uncertainty. We characterize the uncertainty as continuous random variables, and define the corrective control actions through affine control policies. This allows us to efficiently model control reactions to a large number of uncertainty sources. The control policies are then included in a chance constrained optimal power flow formulation, which guarantees that the system constraints are enforced with a desired probability. Lastly, by applying an analytical reformulation of the chance constraints, we obtain a second-order cone problem for which we develop an efficient solution algorithm. In a case study for the IEEE 118 bus system, we show that corrective control for uncertainty leads to a decrease in operational cost, while maintaining system security. Further, we demonstrate the scalability of the method by solving the problem for the IEEE 300 bus and the Polish system test cases.

24 POWER TRANSMISSION AND DISTRIBUTION↗

A generalized polynomial chaos based ensemble Kalman filter with high accuracy

As one of the most adopted sequential data assimilation methods in many areas, especially those involving complex nonlinear dynamics, the ensemble Kalman filter (EnKF) has been under extensive investigation regarding its properties and efficiency. Compared to other variants of the Kalman filter (KF), EnKF is straightforward to implement, as it employs random ensembles to represent solution states. This, however, introduces sampling errors that affect the accuracy of EnKF in a negative manner. Though sampling errors can be easily reduced by using a large number of samples, in practice this is undesirable as each ensemble member is a solution of the system of state equations and can be time consuming to compute for large-scale problems. In this paper we present an efficient EnKF implementation via generalized polynomial chaos (gPC) expansion. The key ingredients of the proposed approach involve (1) solving the system of stochastic state equations via the gPC methodology to gain efficiency; and (2) sampling the gPC approximation of the stochastic solution with an arbitrarily large number of samples, at virtually no additional computational cost, to drastically reduce the sampling errors. The resulting algorithm thus achieves a high accuracy at reduced computational cost, compared to the classical implementations of EnKF. Numerical examples are provided to verify the convergence property and accuracy improvement of the new algorithm. We also prove that for linear systems with Gaussian noise, the first-order gPC Kalman filter method is equivalent to the exact Kalman filter.

97 MATHEMATICS AND COMPUTING↗

Quantum Control Landscapes Beyond the Dipole Approximation: Controllability, Singular Controls, and Resources

We investigate the control landscapes of closed n-level quantum systems beyond the dipole approximation by including a polarizability term in the Hamiltonian. The latter term is quadratic in the control field. Theoretical analysis of singular controls is presented, which are candidates for producing landscape traps. The results for considering the presence of singular controls are compared to their counterparts in the dipole approximation (i.e., without polarizability). A numerical analysis of the existence of traps in control landscapes for generating unitary transformations beyond the dipole approximation is made upon including the polarizability term. An extensive exploration of these control landscapes is achieved by creating many random Hamiltonians which include terms linear and quadratic in a single control field. The discovered singular controls are all found not to be local optima. This result extends a great body of recent work on typical landscapes of quantum systems where the dipole approximation is made. We further investigate the relationship between the magnitude of the polarizability and the fluence of the control resulting from optimization. It is also shown that including a polarizability term in an otherwise uncontrollable dipole coupled system removes traps from the corresponding control landscape by restoring controllability. We numerically assess the effect of a polarizability term on a known example of a particular three-level Λ-system with a second order trap in its control landscape. It is found that the addition of the polarizability removes the trap from the landscape. The general practical control implications of these simulations are discussed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Aggregate modeling of fast-acting demand response and control under real-time pricing

This paper develops and assesses the performance of a short-term demand response (DR) model for utility load control with applications to resource planning and control design. Long term response models tend to underestimate short-term demand response when induced by prices. This has two important consequences. First, planning studies tend to undervalue DR and often overlook its benefits in utility demand management program development. Second, when DR is not overlooked, the open-loop DR control gain estimate may be too low. This can result in overuse of load resources, control instability and excessive price volatility. Our objective is therefore to develop a more accurate and better performing short-term demand response model. We construct the model from first principles about the nature of thermostatic load control and show that the resulting formulation corresponds exactly to the Random Utility Model employed in economics to study consumer choice. The model is tested against empirical data collected from field demonstration projects and is shown to perform better than alternative models commonly used to forecast demand in normal operating conditions. The results suggest that (1) existing utility tariffs appear to be inadequate to incentivize demand response, particularly in the presence of high renewables, and (2) existing load control systems run the risk of becoming unstable if utilities close the loop on real-time prices.

Aggregate Modeling↗

Generating Sequential PV Deployment Scenarios for High Renewable Distribution Grid Planning

This paper introduces a novel approach for generating solar photovoltaic (PV) plant deployment scenarios for grid integration planning. The approach guarantees consistency among scenarios of the same deployment by ensuring that higher penetration scenarios contain PV units deployed in lower penetration scenarios. It also constrains the size and spatial distribution of the PV plants and considers three placement types. A case study on a real-world distribution system proves that the precepts of scenario consistency, deployment diversity, and placement are met. The study further investigates the impact of the resulting scenarios via a stochastic hosting capacity analysis. Results indicate that the ratio between PV and load sizes, referred to as the nodal PV penetration factor (NPPF), is a key driver of the grid integration impact. By reducing the NPPF from 5 to 2, the maximum hosting capacity increased by at least 112%. The study also reveals that scenarios under random placement can lead to higher hosting capacity values.

ENERGY PLANNING, POLICY, AND ECONOMY,SOLAR ENERGY↗

COSMOLOGICAL EVOLUTION OF SUPERMASSIVE BLACK HOLES. II. EVIDENCE FOR DOWNSIZING OF SPIN EVOLUTION

The spin is an important but poorly constrained parameter for describing supermassive black holes (SMBHs). Using the continuity equation of SMBH number density, we explicitly obtain the mass-dependent cosmological evolution of the radiative efficiency for accretion, which serves as a proxy for SMBH spin. Our calculations make use of the SMBH mass function of active and inactive galaxies (derived in the first paper of this series), the bolometric luminosity function of active galactic nuclei (AGNs), corrected for the contribution from Compton-thick sources, and the observed Eddington ratio distribution. We find that the radiative efficiency generally increases with increasing black hole mass at high redshifts (z {approx}> 1), roughly as {eta}{proportional_to}M{sup 0.5}{sub .}, while the trend reverses at lower redshifts, such that the highest efficiencies are attained by the lowest mass black holes. Black holes with M{sub .} {approx}> 10{sup 8.5} M{sub Sun} maintain radiative efficiencies as high as {eta} Almost-Equal-To 0.3-0.4 at high redshifts, near the maximum for rapidly spinning systems, but their efficiencies drop dramatically (by an order of magnitude) by z Almost-Equal-To 0. The pattern for lower mass holes is somewhat more complicated but qualitatively similar. Assuming that the standard accretion disk model applies, we suggest that the accretion history of SMBHs and their accompanying spins evolves in two distinct regimes: an early phase of prolonged accretion, plausibly driven by major mergers, during which the black hole spins up, then switching to a period of random, episodic accretion, governed by minor mergers and internal secular processes, during which the hole spins down. The transition epoch depends on mass, mirroring other evidence for 'cosmic downsizing' in the AGN population; it occurs at z Almost-Equal-To 2 for high-mass black holes and somewhat later, at z Almost-Equal-To 1, for lower mass systems.

79 ASTRONOMY AND ASTROPHYSICS↗

HTR-PROTEUS Pebble Bed Experimental Program Core 4: Random Packing With a 1:1 Moderator-to-Fuel Pebble Ratio

In its deployment as a pebble bed reactor (PBR) critical facility from 1992 to 1996, the PROTEUS facility was designated as HTR-PROTEUS. This experimental program was performed as part of an International Atomic Energy Agency (IAEA) Coordinated Research Project (CRP) on the Validation of Safety Related Physics Calculations for Low Enriched HTGRs. Within this project, critical experiments were conducted for graphite moderated LEU systems to determine core reactivity, flux and power profiles, reaction-rate ratios, the worth of control rods, both in-core and reflector based, the worth of burnable poisons, kinetic parameters, and the effects of moisture ingress on these parameters. One benchmark experiment was evaluated in this report: Core 4. Core 4 represents the only configuration with random pebble packing in the HTR-PROTEUS series of experiments, and has a moderator-to-fuel pebble ratio of 1:1. Three random configurations were performed. The initial configuration, Core 4.1, was rejected because the method for pebble loading, separate delivery tubes for the moderator and fuel pebbles, may not have been completely random; this core loading was rejected by the experimenters. Cores 4.2 and 4.3 were loaded using a single delivery tube, eliminating the possibility for systematic ordering effects. The second and third cores differed slightly in the quantity of pebbles loaded (40 each of moderator and fuel pebbles), stacked height of the pebbles in the core cavity (0.02 m), withdrawn distance of the stainless steel control rods (20 mm), and withdrawn distance of the autorod (30 mm). The 34 coolant channels in the upper axial reflector and the 33 coolant channels in the lower axial reflector were open. Additionally, the axial graphite fillers used in all other HTR-PROTEUS configurations to create a 12-sided core cavity were not used in the randomly packed cores. Instead, graphite fillers were placed on the cavity floor, creating a funnel-like base, to discourage ordering effects during pebble loading. Core 4 was determined to be acceptable benchmark experiment.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

HTR-proteus pebble bed experimental program core 4: random packing with a 1:1 moderator-to-fuel pebble ratio

In its deployment as a pebble bed reactor (PBR) critical facility from 1992 to 1996, the PROTEUS facility was designated as HTR-PROTEUS. This experimental program was performed as part of an International Atomic Energy Agency (IAEA) Coordinated Research Project (CRP) on the Validation of Safety Related Physics Calculations for Low Enriched HTGRs. Within this project, critical experiments were conducted for graphite moderated LEU systems to determine core reactivity, flux and power profiles, reaction-rate ratios, the worth of control rods, both in-core and reflector based, the worth of burnable poisons, kinetic parameters, and the effects of moisture ingress on these parameters. One benchmark experiment was evaluated in this report: Core 4. Core 4 represents the only configuration with random pebble packing in the HTR-PROTEUS series of experiments, and has a moderator-to-fuel pebble ratio of 1:1. Three random configurations were performed. The initial configuration, Core 4.1, was rejected because the method for pebble loading, separate delivery tubes for the moderator and fuel pebbles, may not have been completely random; this core loading was rejected by the experimenters. Cores 4.2 and 4.3 were loaded using a single delivery tube, eliminating the possibility for systematic ordering effects. The second and third cores differed slightly in the quantity of pebbles loaded (40 each of moderator and fuel pebbles), stacked height of the pebbles in the core cavity (0.02 m), withdrawn distance of the stainless steel control rods (20 mm), and withdrawn distance of the autorod (30 mm). The 34 coolant channels in the upper axial reflector and the 33 coolant channels in the lower axial reflector were open. Additionally, the axial graphite fillers used in all other HTR-PROTEUS configurations to create a 12-sided core cavity were not used in the randomly packed cores. Instead, graphite fillers were placed on the cavity floor, creating a funnel-like base, to discourage ordering effects during pebble loading. Core 4 was determined to be acceptable benchmark experiment.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Phase Stability for the Pd-Si System. First-Principles, Experiments, and Solution-Based Modeling

Relative stabilities of the compounds in the binary Pd-Si system were assessed using first-principles calculations and experimental methods. Calculations of lattice parameters and enthalpy of formation indicate that Pd 5 Si-μ, Pd 9 Si 2 -α, Pd 3 Si-β, Pd 2 Si-γ, and PdSi-δ are the stable phases at 0 K (-273 °C). X-ray diffraction analyses (XRD) and electron probe microanalysis (EPMA) of the as-solidified and heat-treated samples support the computational findings, except that the PdSi-δ phase was not observed at low temperature. Considering both experimental data and first-principles results, the compounds Pd 5 Si-μ, Pd 9 Si 2 -α, Pd 3 Si-β, and Pd 2 Si-γ are treated as stable phases down to 0 K (-273 °C), while the PdSi-δ is treated as being stable over a limited range, exhibiting a lower bound. Using these findings, a comprehensive solution-based thermodynamic model is formulated for the Pd-Si system, permitting phase diagram calculation. Moreover, the liquid phase is described using a three-species association model and other phases are treated as solid solutions, where a random substitutional model is adopted for Pd-fcc and Si-dia, and a two-sublattice model is employed for Pd 5 Si-μ, Pd 9 Si 2 -α, Pd 3 Si-β, Pd 2 Si-γ, and PdSi-δ. Model parameters are fitted using available experimental data and first-principles data, and the resulting phase diagram is reported over the full range of compositions.

36 MATERIALS SCIENCE↗

The impact of anisotropic thermal expansion on the isothermal annealing of polycrystalline α -uranium

Although grain growth impacts microstructural evolution in a wide variety of materials systems, the effect of anisotropic thermal expansion on grain boundary mobility and texture evolution has not been widely studied. Anisotropic thermal expansion occurs in multiple non-cubic metals, and the thermomechanical processing behavior of these materials can be better understood with further study into the impact of thermal expansion on grain boundary mobility and texture evolution. In this work, we develop a mesoscale phase field model of grain growth that includes the effect of anisotropic thermal expansion, which is applied to study polycrystalline α-uranium, a highly anisotropic metal. Three-dimensional simulations on polycrystalline a-uranium with and without thermal expansion eigenstrains are performed to study the grain boundary mobility and texture evolution as a function of temperature. A strain-free temperature of 933 K is selected, and the system is studied at within the range of 873 – 933 K at intervals of ten degrees, resulting in increasing thermal eigenstrain with decreasing temperature. We also estimate a grain boundary mobility prefactor and activation energy based on existing experimental data of isothermal annealing of α-uranium. The grain boundary mobility is found to display significant deviation from Arrhenius behavior with the inclusion of thermal expansion eigenstrain as the amount of thermal eigenstrain (and thus elastic strain energy within the system) increases. This result explains an experimentally observed grain boundary mobility deviation from Arrhenius behavior. Furthermore, the texture evolution is affected, such that the grain orientations become less random with increasing thermal eigenstrain, which could explain experimentally observed texture behavior. These results indicate that the effect of thermal expansion should be considered when predicting the thermomechanical processing behavior of α-uranium and other materials with anisotropic thermal expansion.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Predicting Radiation Pneumonitis After Chemoradiation Therapy for Lung Cancer: An International Individual Patient Data Meta-analysis

Background: Radiation pneumonitis is a dose-limiting toxicity for patients undergoing concurrent chemoradiation therapy (CCRT) for non-small cell lung cancer (NSCLC). We performed an individual patient data meta-analysis to determine factors predictive of clinically significant pneumonitis. Methods and Materials: After a systematic review of the literature, data were obtained on 836 patients who underwent CCRT in Europe, North America, and Asia. Patients were randomly divided into training and validation sets (two-thirds vs one-third of patients). Factors predictive of symptomatic pneumonitis (grade {>=}2 by 1 of several scoring systems) or fatal pneumonitis were evaluated using logistic regression. Recursive partitioning analysis (RPA) was used to define risk groups. Results: The median radiation therapy dose was 60 Gy, and the median follow-up time was 2.3 years. Most patients received concurrent cisplatin/etoposide (38%) or carboplatin/paclitaxel (26%). The overall rate of symptomatic pneumonitis was 29.8% (n=249), with fatal pneumonitis in 1.9% (n=16). In the training set, factors predictive of symptomatic pneumonitis were lung volume receiving {>=}20 Gy (V{sub 20}) (odds ratio [OR] 1.03 per 1% increase, P=.008), and carboplatin/paclitaxel chemotherapy (OR 3.33, P<.001), with a trend for age (OR 1.24 per decade, P=.09); the model remained predictive in the validation set with good discrimination in both datasets (c-statistic >0.65). On RPA, the highest risk of pneumonitis (>50%) was in patients >65 years of age receiving carboplatin/paclitaxel. Predictors of fatal pneumonitis were daily dose >2 Gy, V{sub 20}, and lower-lobe tumor location. Conclusions: Several treatment-related risk factors predict the development of symptomatic pneumonitis, and elderly patients who undergo CCRT with carboplatin-paclitaxel chemotherapy are at highest risk. Fatal pneumonitis, although uncommon, is related to dosimetric factors and tumor location.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Clinical Experience and Evaluation of Patient Treatment Verification With a Transit Dosimeter

Purpose: To prospectively evaluate a protocol for transit dosimetry on a patient population undergoing intensity modulated radiation therapy (IMRT) and to assess the issues in clinical implementation of electronic portal imaging devices (EPIDs) for treatment verification. Methods and Materials: Fifty-eight patients were enrolled in the study. Amorphous silicon EPIDs were calibrated for dose and used to acquire images of delivered fields. Measured EPID dose maps were back-projected using the planning computed tomographic (CT) images to calculate dose at prespecified points within the patient and compared with treatment planning system dose offline using point dose difference and point γ analysis. The deviation of the results was used to inform future action levels. Results: Two hundred twenty-five transit images were analyzed, composed of breast, prostate, and head and neck IMRT fields. Patient measurements demonstrated the potential of the dose verification protocol to model dose well under complex conditions: 83.8% of all delivered beams achieved the initial set tolerance level of Δ{sub D} of 0 ± 5 cGy or %Δ{sub D} of 0% ± 5%. Importantly, the protocol was also sensitive to anatomic changes and spotted that 3 patients from 20 measured prostate patients had undergone anatomic change in comparison with the planning CT. Patient data suggested an EPID-reconstructed versus treatment planning system dose difference action level of 0% ± 7% for breast fields. Asymmetric action levels were more appropriate for inversed IMRT fields, using absolute dose difference (−2 ± 5 cGy) or summed field percentage dose difference (−6% ± 7%). Conclusions: The in vivo dose verification method was easy to use and simple to implement, and it could detect patient anatomic changes that impacted dose delivery. The system required no extra dose to the patient or treatment time delay and so could be used throughout the course of treatment to identify and limit systematic and random errors in dose delivery for patient groups.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Active meta-learning for predicting and selecting perovskite crystallization experiments

Autonomous experimentation systems use algorithms and data from prior experiments to select and perform new experiments in order to meet a specified objective. In most experimental chemistry situations, there is a limited set of prior historical data available, and acquiring new data may be expensive and time consuming, which places constraints on machine learning methods. Active learning methods prioritize new experiment selection by using machine learning model uncertainty and predicted outcomes. Meta-learning methods attempt to construct models that can learn quickly with a limited set of data for a new task. Here in this paper, we applied the model-agnostic meta-learning (MAML) model and the Probabilistic LATent model for Incorporating Priors and Uncertainty in few-Shot learning (PLATIPUS) approach, which extends MAML to active learning, to the problem of halide perovskite growth by inverse temperature crystallization. Using a dataset of 1870 reactions conducted using 19 different organoammonium lead iodide systems, we determined the optimal strategies for incorporating historical data into active and meta-learning models to predict reaction compositions that result in crystals. We then evaluated the best three algorithms (PLATIPUS and active-learning k-nearest neighbor and decision tree algorithms) with four new chemical systems in experimental laboratory tests. With a fixed budget of 20 experiments, PLATIPUS makes superior predictions of reaction outcomes compared to other active-learning algorithms and a random baseline.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗