Search NASA⌕ Search

SEARCH · Search NASA

Results for “statistical methods for engineering applications”

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 163 records · Page 9

High-order asynchrony-tolerant finite difference schemes for partial differential equations

Synchronizations of processing elements (PEs) in massively parallel simulations, which arise due to communication or load imbalances between PEs, significantly affect the scalability of scientific applications. We have recently proposed a method based on finite-difference schemes to solve partial differential equations in an asynchronous fashion – synchronization between PEs is relaxed at a mathematical level. While standard schemes can maintain their stability in the presence of asynchrony, their accuracy is drastically affected. In this work, we present a general methodology to derive asynchrony-tolerant (AT) finite difference schemes of arbitrary order of accuracy, which can maintain their accuracy when synchronizations are relaxed. We show that there are several choices available in selecting a stencil to derive these schemes and discuss their effect on numerical and computational performance. We provide a simple classification of schemes based on the stencil and derive schemes that are representative of different classes. Their numerical error is rigorously analyzed within a statistical framework to obtain the overall accuracy of the solution. Results from numerical experiments are used to validate the performance of the schemes.

42 ENGINEERING↗

Fully consistent Eulerian Monte Carlo fields method for solving probability density function transport equations in turbulence modeling

The probability density function (PDF) transport equation method is a sophisticated model for the closure of turbulent mixing and turbulent reactive flows. An efficient solution approach for solving the PDF transport equation has been vital for the method to be widely used in applications. The Eulerian Monte Carlo fields (EMCF) method has been developed to solve the PDF transport equation efficiently for decades. A recent work by Wang et al. revealed a serious issue of the EMCF method for not being fully consistent with the PDF transport equation for which the method is designed to solve. This work advances the state of the art by introducing fully consistent EMCF methods for solving the PDF transport equation. The fully consistent EMCF formulations are derived for two different PDF equation forms. The consistency of the EMCF formulations is mathematically confirmed by examining the derived moment transport equations from the EMCF formulations and from the PDF transport equation. The method of manufactured solutions is employed to further verify the consistency and convergence of the different EMCF formulations numerically. The newly introduced EMCF formulations bring the EMCF method to full consistency with the PDF transport equations for the first time.

42 ENGINEERING↗

An Applied Strategy for Using Empirical and Hybrid Models in Online Monitoring

The monitoring of plant equipment for failure prediction is one of the key contributors to operation and maintenance (O&M) costs for a nuclear power plant (NPP) because O&M monitoring depends on labor-intensive activities that are required to meet high equipment reliability standards. These activities rely primarily on humans for information gathering, condition diagnosis, and predictive analysis. Online monitoring aims to automate these activities by relying on sensors to replace human information gathering and machine learning to replace human analysis and decision making. To facilitate automated monitoring, a systematic strategy for anomaly detection is needed to optimally use the available sensor data, empirical models, and physics-supported models. This strategy is essential to provide credible reasoning on why and when an empirical (i.e., purely data-driven) versus hybrid (i.e., physics-supported) approach should be used and to determine the ideal mix of these two approaches for a defined anomaly detection scope. The extant methods usually adopt an ad hoc trial-and-error approach that, in addition to being time-consuming and costly, is also highly subjective; it is impacted by the background and the skill set of the personnel making the decisions. Thus, such an approach cannot guarantee an optimum outcome. This represents the motivation of the current research effort, which is focused on devising a scientifically supported strategy for the optimum selection of anomaly detection methods. This report presents a detailed assessment of the main anomaly detection techniques within the empirical or hybrid method streams. Empirical methods include pattern, statistical, and causal inference. Hybrid methods include the use of physics models to train and test data methods, reduce data dimensionality, reduce data-model complexity, augment data, and reduce empirical uncertainty; hybrid methods also include the use of data to tune physics models. The listed techniques within these two streams represent the vast majority of techniques performed for anomaly detection. Using the techniques as outcomes, a strategy was developed to enable a systematic decision-making process to lead to one of these techniques. The strategy is driven by key decision points related to data relevance, simple modeling feasibility, data inference, physics-modeling value, data dimensionality, physics knowledge, method of validation, performance, data availability and suitability for training and testing, cause-effect, entropy inference, and model fitting. Each of these decision points in the strategy is explained in detail in this report with examples, along with the scientific basis behind the decisions and outcomes in common and simplified terminology. The strategy is developed for use by any NPP staff with basic engineering or science knowledge. A user-friendly graphical state flow diagram was also developed as a visual presentation of the strategy. The strategy was tested and demonstrated through two pilot projects for the application of anomaly detection at an NPP. Each pilot had two use cases: an initial case in which certain decisions were made that resulted in one or more empirical techniques and a revised use case where one or more key decisions were modified resulting in using a set of hybrid methods.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

A general probabilistic approach for the quantitative assessment of LES combustion models

The Wasserstein metric is introduced as a probabilistic method to enable quantitative evaluations of LES combustion models. The Wasserstein metric can directly be evaluated from scatter data or statistical results through probabilistic reconstruction. The method is derived and generalized for turbulent reacting flows, and applied to validation tests involving a piloted turbulent jet flame with inhomogeneous inlets. It is shown that the Wasserstein metric is an effective validation tool that extends to multiple scalar quantities, providing an objective and quantitative evaluation of model deficiencies and boundary conditions on the simulation accuracy. Several test cases are considered, beginning with a comparison of mixture-fraction results, and the subsequent extension to reactive scalars, including temperature and species mass fractions of CO and CO2. To demonstrate the versatility of the proposed method in application to multiple datasets, the Wasserstein metric is applied to a series of different simulations that were contributed to the TNF-workshop. Analysis of the results allows to identify competing contributions to model deviations, arising from uncertainties in the boundary conditions and model deficiencies. These applications demonstrate that the Wasserstein metric constitutes an easily applicable mathematical tool that reduce multiscalar combustion data and large datasets into a scalar-valued quantitative measure.

42 ENGINEERING↗

Stochastic flux freezing and magnetic dynamo

Magnetic flux conservation in turbulent plasmas at high magnetic Reynolds numbers is argued neither to hold in the conventional sense nor to be entirely broken, but instead to be valid in a statistical sense associated to the ''spontaneous stochasticity'' of Lagrangian particle trajectories. The latter phenomenon is due to the explosive separation of particles undergoing turbulent Richardson diffusion, which leads to a breakdown of Laplacian determinism for classical dynamics. Empirical evidence is presented for spontaneous stochasticity, including numerical results. A Lagrangian path-integral approach is then exploited to establish stochastic flux freezing for resistive hydromagnetic equations and to argue, based on the properties of Richardson diffusion, that flux conservation must remain stochastic at infinite magnetic Reynolds number. An important application of these results is the kinematic, fluctuation dynamo in nonhelical, incompressible turbulence at magnetic Prandtl number (Pr{sub m}) equal to unity. Numerical results on the Lagrangian dynamo mechanisms by a stochastic particle method demonstrate a strong similarity between the Pr{sub m}=1 and 0 dynamos. Stochasticity of field-line motion is an essential ingredient of both. Finally, some consequences for nonlinear magnetohydrodynamic turbulence, dynamo, and reconnection are briefly considered.

42 ENGINEERING↗

A Data-Driven Framework for Power System Event Type Identification via Safe Semi-Supervised Techniques

Herein this paper investigates the use of phasor measurement unit (PMU) data with deep learning techniques to construct real-time event identification models for transmission networks. Increasing penetration of distributed energy resources represents a great opportunity to achieve decarbonization, as well as challenges in systematic situational awareness. When high-resolution PMU data and sufficient manually recorded event labels are available, the power event identification problem is defined as a statistical classification problem that can be solved by numerous cutting-edge classifiers. However, in real grids, collecting tremendous high-quality event labels is quite expensive. Utilities frequently have a large number of event records without in-depth details (i.e., unlabeled events). To bridge this gap, we propose a novel semi-supervised learning-based method to improve the performance of event classifiers trained with a limited number of labeled events by exploiting the information from massive unlabeled events. In other words, compared to existing data-driven methods, our method requires only a small portion of labeled data to achieve a similar level of accuracy. Meanwhile, this work discusses and addresses the performance degradation caused by class distribution mismatch between the training set and the real applications. Based on the proposed safe learning mechanism, our model does not directly use all unlabeled events during model training, but selectively uses them through a comprehensive evaluation procedure. Numerical studies on a sizable PMU dataset have been used to validate the performance of the proposed method.

42 ENGINEERING↗

A SIMPLE METHOD FOR MODELING COLLISION PROCESSES IN PLASMAS WITH A KAPPA ENERGY DISTRIBUTION

We demonstrate that a nonthermal distribution of particles described by a kappa distribution can be accurately approximated by a weighted sum of Maxwell–Boltzmann distributions. We apply this method to modeling collision processes in kappa-distribution plasmas, with a particular focus on atomic processes important for solar physics. The relevant collision process rate coefficients are generated by summing appropriately weighted Maxwellian rate coefficients. This method reproduces the rate coefficients for a kappa distribution to an estimated accuracy of better than 3%. This is equal to or better than the accuracy of rate coefficients generated using “reverse-engineering” methods, which attempt to extract the needed cross sections from the published Maxwellian rate coefficient data and then reconvolve the extracted cross sections with the desired kappa distribution. Our approach of summing Maxwellian rate coefficients is easy to implement using existing spectral analysis software. Moreover, the weights in the sum of the Maxwell–Boltzmann distribution rate coefficients can be found for any value of the parameter κ, thereby enabling one to model plasmas with a time-varying κ. Tabulated Maxwellian fitting parameters are given for specific values of κ from 1.7 to 100. We also provide polynomial fits to these parameters over this entire range. Several applications of our technique are presented, including the plasma equilibrium charge state distribution (CSD), predicting line ratios, modeling the influence of electron impact multiple ionization on the equilibrium CSD of kappa-distribution plasmas, and calculating the time-varying CSD of plasmas during a solar flare.

79 ASTRONOMY AND ASTROPHYSICS↗

Transforming microseismic clouds into near real-time visualization of the growing hydraulic fracture

SUMMARY Microseismic observations during unconventional reservoir stimulation are typically seen as a proxy for clusters of hydraulic fractures and the extent of the stimulated reservoir. Such straightforward interpretation is often misleading and fails to provide a physically reasonable image of the fracturing process. This paper demonstrates the application of a physics-based machine learning algorithm which enables a rapid and accurate fracture mapping from the microseismic data. Our training and validation data set relies on a history-matched geomechanical modelling workflow implemented in GEOS software for the Hydraulic Fracturing Test Site 1 (HFTS-1) project. For this study we augmented the simulated fracture growth through geostatistical modelling of induced seismicity, so that the synthetic microseismic catalogue matches the main statistical properties of the field observations. We formulated the problem of mapping the actual fracture in the clutter of events to parallel common video segmentation workflows: several past video frames (microseismic density snapshots) are passed through a deep convolutional network to classify whether a given voxel is associated with a fracture or intact rock. We found that for accurate fracture mapping, the network’s input and architecture must be augmented to incorporate the fluid injection parameters (pressure, rate, concentration of proppant, and location of the perforation within the cluster). The error rate for the network reached as little as 10 per cent of the fracture area, while a conventional microseismic interpretation approach yielded ∼300 per cent. Our approach also yields must faster predictions than conventional methods (minutes instead of weeks), and could enable engineers to make rapid decisions regarding engineering parameters (pumping rate, viscosity) in real time during stimulation.

58 GEOSCIENCES↗

MISPR : an open-source package for high-throughput multiscale molecular simulations

Computational tools provide a unique opportunity to study and design optimal materials by enhancing our ability to comprehend the connections between their atomistic structure and functional properties. However, designing materials with tailored functionalities is complicated due to the necessity to integrate various computational-chemistry software (not necessarily compatible with one another), the heterogeneous nature of the generated data, and the need to explore vast chemical and parameter spaces. The latter is especially important to avoid bias in scattered data points-based models and derive statistical trends only accessible by systematic datasets. Here, we introduce a robust high-throughput multi-scale computational infrastructure coined MISPR (Materials Informatics for Structure–Property Relationships) that seamlessly integrates classical molecular dynamics (MD) simulations with density functional theory (DFT). By enabling high-performance data analytics and coupling between different methods and scales, MISPR addresses critical challenges arising from the needs of automated workflow management and data provenance recording. The major features of MISPR include automated DFT and MD simulations, error handling, derivation of molecular and ensemble properties, and creation of output databases that organize results from individual calculations to enable reproducibility and transparency. In this work, we describe fully automated DFT workflows implemented in MISPR to compute various properties such as nuclear magnetic resonance chemical shift, binding energy, bond dissociation energy, and redox potential with support for multiple methods such as electron transfer and proton-coupled electron transfer reactions. The infrastructure also enables the characterization of large-scale ensemble properties by providing MD workflows that calculate a wide range of structural and dynamical properties in liquid solutions. MISPR employs the methodologies of materials informatics to facilitate understanding and prediction of phenomenological structure–property relationships, which are crucial to designing novel optimal materials for numerous scientific applications and engineering technologies.

36 MATERIALS SCIENCE↗

High-Throughput Computation and the Applicability of Monte Carlo Integration in Fatigue Load Estimation of Floating Offshore Wind Turbines

Long-term fatigue loads for floating offshore wind turbines are hard to estimate because they require the evaluation of the integral of a highly nonlinear function over a wide variety of wind and wave conditions. Current design standards involve scanning over a uniform rectangular grid of metocean inputs (e.g., wind speed and direction and wave height and period), which becomes intractable in high dimensions as the number of required evaluations grows exponentially with dimension. Monte Carlo integration offers a potentially efficient alternative because it has theoretical convergence proportional to the inverse of the square root of the number of samples, which is independent of dimension. In this paper, we first report on the integration of the aeroelastic code FAST into NREL's systems engineering tool, WISDEM, and the development of a high-throughput pipeline capable of sampling from arbitrary distributions, running FAST on a large scale, and postprocessing the results into estimates of fatigue loads. Second, we use this tool to run a variety of studies aimed at comparing grid-based and Monte Carlo-based approaches with calculating long-term fatigue loads. We observe that for more than a few dimensions, the Monte Carlo approach can represent a large improvement in computational efficiency, but that as nonlinearity increases, the effectiveness of Monte Carlo is correspondingly reduced. The present work sets the stage for future research focusing on using advanced statistical methods for analysis of wind turbine fatigue as well as extreme loads.

17 WIND ENERGY↗

Application of unsupervised deep learning to image segmentation and in-situ contact angle measurements in a CO 2 -water-rock system

Rock surface wettability is a critical property that regulates multiphase flows in porous media, which can be quantified using the surface contact angle (CA). X-ray micro-computed tomography (μCT) provides an effective approach to in-situ measurements of surface CAs. However, the CA measurement accuracy depends significantly on the quality of CT image segmentation, which is the clustering of CT pixels into separate phases. Inspired by this, we developed a deep learning (DL)-based CA measurement workflow. Motivated by the recent tremendous progress in unsupervised learning techniques and aiming to avoid expensive manual data annotations, an unsupervised DL pipeline for CT image segmentation was proposed and implemented, which includes unsupervised model training and post-processing. The unsupervised model training was driven by a novel loss function constrained with feature similarity and spatial continuity and implemented by iterative forward and backward paths; the former clustered the pixel-wise feature vectors extracted by convolution neural networks, whereas the latter updated the parameters using gradient descent. An over-segmentation strategy was adopted for model training. The post-processing steps based on agglomerative hierarchical clustering (AHC) were implemented to further merge the over-segmented model output to the desired cluster number, which is intended to improve the efficiency of image segmentation. The developed unsupervised DL pipeline was compared with other commonly-used image segmentation methods using pixel-wise and physics-based evaluation metrics on a synthetic raw-image dataset, which had a known ground truth. The unsupervised DL pipeline showed the best performance. Next, the segmented images were input to an automatic CA measurement tool, and the results were validated by comparisons with manual measurements. The CA values from the manual and automatic measurements showed similar distributions and statistical properties. The automatic measurement demonstrated a wider spectrum because of the much larger number of measurement data points. The primary novelty of the unsupervised DL pipeline developed in this study lies in the novel loss function and the over-segmentation strategy associated with AHC post-processing. Finally, the workflow has been proven an efficient tool for pore-scale wettability characterization, which has a wide range of applications in fundamental studies of multiphase flows in natural porous media, which have critical implications to geological carbon sequestration, hydrocarbon energy recovery, and contaminant transport in groundwater.

42 ENGINEERING↗

Day-ahead photovoltaic power production forecasting methodology based on machine learning and statistical post-processing

A main challenge towards ensuring large-scale and seamless integration of photovoltaic systems is to improve the accuracy of energy yield forecasts, especially in grid areas of high photovoltaic shares. The scope of this paper is to address this issue by presenting a unified methodology for hourly-averaged day-ahead photovoltaic power forecasts with improved accuracy, based on data-driven machine learning techniques and statistical post-processing. More specifically, the proposed forecasting methodology framework comprised of a data quality stage, data-driven power output machine learning model development (artificial neural networks), weather clustering assessment (K-means clustering), post-processing output optimisation (linear regressive correction method) and the final performance accuracy evaluation. The results showed that the application of linear regression coefficients to the forecasted outputs of the developed day-ahead photovoltaic power production neural network improved the performance accuracy by further correcting solar irradiance forecasting biases. The resulting optimised model provided a mean absolute percentage error of 4.7% when applied to historical system datasets. Finally, the model was validated both, at a hot as well as a cold semi-arid climatic location, and the obtained results demonstrated close agreement by yielding forecasting accuracies of mean absolute percentage error of 4.7% and 6.3%, respectively. Finally, the validation analysis provides evidence that the proposed model exhibits high performance in both forecasting accuracy and stability.

42 ENGINEERING↗

Accelerated statistical failure analysis of multifidelity TRISO fuel models

Statistical nuclear fuel failure analysis is critical for the design and development of advanced reactor technologies. Although Monte Carlo Sampling (MCS) is a standard method of statistical failure analysis for fuels, the low failure probabilities of some advanced fuel forms and the correspondingly large number of required model evaluations limit its application to low-fidelity (e.g., 1-D) fuel models. In this paper, we present four other statistical methods for fuel failure analysis in Bison, considering tri-structural isotropic (TRISO)-coated particle fuel as a case study. The statistical methods considered are Latin hypercube sampling (LHS), adaptive importance sampling (AIS), subset simulation (SS), and the Weibull theory. Using these methods, we analyzed both 1-D and 2-D representations of TRISO models to compute failure probabilities and the distributions of fuel properties that result in failures. The results of these methods compare well across all TRISO models considered. Overall, SS and the Weibull theory were deemed the most efficient, and can be applied to both 1-D and 2-D TRISO models to compute failure probabilities. Moreover, since SS also characterizes the distribution of parameters that cause TRISO failures, and can consider failure modes not described by the Weibull criterion, it may be preferred over the other methods. Finally, a discussion on the efficacy of different statistical methods of assessing nuclear fuel safety is provided.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Using flow cytometry and multistage machine learning to discover label-free signatures of algal lipid accumulation

Most applications of flow cytometry or cell sorting rely on the conjugation of fluorescent dyes to specific biomarkers. However, labeled biomarkers are not always available, they can be costly, and they may disrupt natural cell behavior. Label-free quantification based upon machine learning approaches could help correct these issues, but label replacement strategies can be very difficult to discover when applied labels or other modifications in measurements inadvertently modify intrinsic cell properties. Here we demonstrate a new, but simple approach based upon feature selection and linear regression analyses to integrate statistical information collected from both labeled and unlabeled cell populations and to identify models for accurate label-free single-cell quantification. We verify the method’s accuracy to predict lipid content in algal cells (Picochlorum soloecismus) during a nitrogen starvation and lipid accumulation time course. Our general approach is expected to improve label-free single-cell analysis for other organisms or pathways, where biomarkers are inconvenient, expensive, or disruptive to downstream cellular processes.

42 ENGINEERING↗

Connectivity-informed drainage network generation using deep convolution generative adversarial networks

Abstract Stochastic network modeling is often limited by high computational costs to generate a large number of networks enough for meaningful statistical evaluation. In this study, Deep Convolutional Generative Adversarial Networks (DCGANs) were applied to quickly reproduce drainage networks from the already generated network samples without repetitive long modeling of the stochastic network model, Gibb’s model. In particular, we developed a novel connectivity-informed method that converts the drainage network images to the directional information of flow on each node of the drainage network, and then transforms it into multiple binary layers where the connectivity constraints between nodes in the drainage network are stored. DCGANs trained with three different types of training samples were compared; (1) original drainage network images, (2) their corresponding directional information only, and (3) the connectivity-informed directional information. A comparison of generated images demonstrated that the novel connectivity-informed method outperformed the other two methods by training DCGANs more effectively and better reproducing accurate drainage networks due to its compact representation of the network complexity and connectivity. This work highlights that DCGANs can be applicable for high contrast images common in earth and material sciences where the network, fractures, and other high contrast features are important.

42 ENGINEERING↗

Statistical Performance Evaluation Of Soft Seat Pressure Relief Valves

Risk-based inspection methods enable estimation of the probability of failure on demand for spring-operated pressure relief valves at the United States Department of Energy's Savannah River Site in Aiken, South Carolina. This paper presents a statistical performance evaluation of soft seat spring operated pressure relief valves. These pressure relief valves are typically smaller and of lower cost than hard seat (metal to metal) pressure relief valves and can provide substantial cost savings in fluid service applications (air, gas, liquid, and steam) providing that probability of failure on demand (the probability that the pressure relief valve fails to perform its intended safety function during a potentially dangerous over pressurization) is at least as good as that for hard seat valves. The research in this paper shows that the proportion of soft seat spring operated pressure relief valves failing is the same or less than that of hard seat valves, and that for failed valves, soft seat valves typically have failure ratios of proof test pressure to set pressure less than that of hard seat valves.

42 ENGINEERING↗

Application and Certification of Comparative Vacuum Monitoring Sensors for Structural Health Monitoring of 737 Wing Box Fittings

Multi-site fatigue damage, hidden cracks in hard-to-reach locations, disbonded joints, erosion, impact, and corrosion are among the major flaws encountered in today's extensive fleet of aging aircraft and space vehicles. The use of in-situ sensors for real-time health monitoring of aircraft structures are a viable option to overcome inspection impediments stemming from accessibility limitations, complex geometries, and the location and depth of hidden damage. Reliable, structural health monitoring systems can automatically process data, assess structural condition, and signal the need for human intervention. Prevention of unexpected flaw growth and structural failure can be improved if on-board health monitoring systems are used to continuously assess structural integrity. Such systems are able to detect incipient damage before catastrophic failures occurs. Condition-based maintenance practices could be substituted for the current time-based maintenance approach. Other advantages of on-board distributed sensor systems are that they can eliminate costly, and potentially damaging, disassembly, improve sensitivity by producing optimum placement of sensors and decrease maintenance costs by eliminating more time- consuming manual inspections. This report presents a Sandia Labs-aviation industry effort to move SHM into routine use for aircraft maintenance. This program addressed formal SHM technology validation and certification issues so that the full spectrum of concerns, including design, deployment, performance and certification were appropriately considered. The Airworthiness Assurance NDI Validation Center (AANC) at Sandia Labs, in conjunction with Boeing, Delta Air Lines, Structural Monitoring Systems Ltd., Anodyne Electronics Manufacturing Corp. and the Federal Aviation Administration (FAA) carried out a certification program to formally introduce Comparative Vacuum Monitoring (CVM) as a structural health monitoring solution to a specific aircraft wing box application. Validation tasks were designed to address the SHM equipment, the health monitoring task, the resolution required, the sensor interrogation procedures, the conditions under which the monitoring will occur, the potential inspector population, adoption of CVM into an airline maintenance program and the document revisions necessary to allow for routine use of CVM as an alternate means of performing periodic structural inspects. To carry out the validation process, knowledge of aircraft maintenance practices was coupled with an unbiased, independent evaluation. Sandia Labs designed, implemented, and analyzed the results from a focused and statistically-relevant experimental effort to quantify the reliability of the CVM system applied to the Boeing 737 Wing Box fitting application. All factors that affect SHM sensitivity were included in this program: flaw size, shape, orientation and location relative to the sensors, as well as operational and environmental variables. Statistical methods were applied to performance data to derive Probability of Detection (POD) values for CVM sensors in a manner that agrees with current nondestructive inspection (NDI) validation requirements and also is acceptable to both the aviation industry and regulatory bodies. This report presents the use of several different statistical methods, some of them adapted from NDI performance assessments and some proposed to address the unique nature of damage detection via SHM systems, and discusses how they can converge to produce a confident quantification of SHM performance An important element in developing SHM validation processes is a clear understanding of the regulatory measures needed to adopt SHM solutions along with the knowledge of the structural and maintenance characteristics that may impact the operational performance of an SHM system. This report describes the major elements of an SHM validation approach and differentiates the SHM elements from those found in NDI validation. The activities conducted in this program demonstrated the feasibility of routine SHM usage in general and CVM in particular for the application selected. They also helped establish an optimum OEM-airline-regulator process and determined how to safely adopt SHM solutions. This formal SHM validation will allow aircraft manufacturers and airlines to confidently make informed decisions about the proper utilization of CVM technology. It will also streamline the regulatory actions and formal certification measures needed to assure the safe application of SHM solutions.

42 ENGINEERING↗

Agentic AI vs ML-Based Autotuning: A Comparative Study for Loop Reordering Optimization

High Performance Computing (HPC) applications rely heavily on code optimizations to achieve good performance on modern CPU and GPU architectures. Traditional Machine Learning auto-tuning approaches have demonstrated success in exploring high-dimensional spaces, but they often require expensive compile-run evaluations and lack adaptability for large HPC applications. The recent advances in Large Language Models (LLMs) and Agentic AI systems raise intriguing questions about the potential of these approaches to address specific optimization methodologies. This work aims to answer an essential question for the HPC community: “How Agentic AI Systems Compare to Traditional ML Autotuning Techniques?” To address this question, we present a comparative analysis between a traditional ML-based optimization approach and an Agentic AI system, evaluating their respective capabilities and limitations for loop-level optimization. In addition, we introduced a new Agentic AI system named LoopGen-AI using three different Large Language Models: GPT-4.1, Claude 4.0, and Gemini 2.5. A key finding is that LoopGen-AI achieves competitive per-formance with only a few program runs, the reasoning logs from the agents revealed that their decisions rely heavily on the combination of semantic understanding of the target kernel with dynamic feedback from the environment, highlighting a promising new dimension in performance tuning. In contrast, ML-based autotuners focus on statistical exploration, and require orders of magnitude more runs to reach peak performance. Additionally, our analysis shows that prompt engineering, particularly using Persona + Context Manager patterns, significantly impacts the effectiveness of Agentic AI. Our results indicate that while Agentic AI systems are not yet a complete replacement for ML-based autotuners, it can effectively complement traditional methods.

Rosas, Miguel Romero↗