SEARCH · Search NASA

Results for “Data Intensive Computing”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 865 records · Page 48

All sources NASA NTRS DOE OSTI

DOE OSTI2023

Systemwide Planning with a Branch-and-Price Algorithm for Pavement-Marking Assessment Data Collection via the Mobile Retroreflectivity Unit Routing Model

The visibility of pavement markings is one of the most critical factors for traffic safety, and a periodical assessment plan is crucial for maintaining this function. Traditional assessment methods, such as visual windshield surveys or manual testing using handheld devices, are unsafe, time-consuming, and labor-intensive. In recent years, transportation agencies have begun to adopt the use of mobile retroreflectivity units (MRUs) for condition assessment of pavement markings. MRUs, different from other manual methods, can be utilized to collect large-scale retroreflectivity data in an efficient manner. However, no relevant research has yet proposed a mathematical optimization model for arranging the evaluation schedule and paths of MRUs. This study aims to propose a MRU routing model, and an efficient solution methodology. A branch-and-price algorithm, including column generation and branch-and-bound, was implemented. Computational experiments have been conducted based on actual tasks from the Florida MRU program for validation. In conclusion, results show that the proposed solution methodology with a set partitioning model in this study not only finds the optimal solution for problems with tasks less than 60, but also effectively narrows the solution gap to be within 1.0% for problems with tasks less than 131.

42 ENGINEERING↗

DOE OSTI2022

An Eulerian crystal plasticity framework for modeling large anisotropic deformations in energetic materials under shocks

Here, this paper demonstrates a novel Eulerian computational framework for modeling anisotropic elastoplastic deformations of organic crystalline energetic materials (EM) under shocks. While Eulerian formulations are advantageous for handling large deformations, constitutive laws in such formulations have been limited to isotropic elastoplastic models, which may not fully capture the shock response of crystalline EM. The present Eulerian framework for high-strain rates, large deformation material dynamics of EM incorporates anisotropic isochoric elasticity via a hypo-elastic constitutive law and visco-plastic single-crystal models. The calculations are validated against atomistic calculations and experimental data and benchmarked against Lagrangian (finite element) crystal plasticity computations for shock-propagation in a monoclinic organic crystal, octahydro-1,3,5,7-tetranitro-1,3,5,7 tetrazocine (β-HMX). The Cauchy stress components and the resolved shear stresses calculated using the present Eulerian approach are shown to be in good agreement with the Lagrangian computations for different crystal orientations. The Eulerian framework is then used for computations of shock-induced inert void collapse in β-HMX to study the effects of crystal orientations on hotspot formation under different loading intensities. The computations show that the hotspot temperature distributions and the collapse profiles are sensitive to the crystal orientations at lower impact velocities (viz., 500 m/s); when the impact velocity is increased to 1000 m/s, the collapse is predominantly hydrodynamic and the role of anisotropy is modest. The present methodology will be useful to simulate energy localization in shocked porous energetic material microstructures and other situations where large deformations of single and polycrystals govern the thermomechanical response.

42 ENGINEERING↗

DOE OSTI2023

Count Every Trip: Finding the Uncertainty in Energy Estimates Made from Inferred Travel Modes

To properly inform transport policy and infrastructure changes, transportation related metrics need both measured values and uncertainties of those values. Travel monitoring smartphone apps can record people's travel behavior, but trip data quality is limited by sensor errors, user labeling rates and the accuracy of inference algorithms used for travel diary creation. We discuss the use of phone app recorded travel diary data to estimate energy consumption, and propose the use of propagation of variance to find error bars for such estimates. We define energy consumption for one trip as trip length times the energy intensity per distance unit of the travel mode used. We characterize trip length errors with relative error and inferred trip mode errors with confusion matrix columns. The resulting variances of each measurement are then propagated to the final calculated energy consumption. We tested our uncertainty methods on a dataset that used phone app data combined with prompted recall, consisting of 92,234 labeled trips for over 500,000 miles. Accounting for uncertainty using expected energy intensities and variance propagation gives a dataset-wide aggregate energy consumption percent error of about 9%, within one standard deviation from the truth. Future work could involve applying similar methods to other travel diary based metrics.

ADVANCED PROPULSION SYSTEMS,MATHEMATICS AND COMPUT↗

DOE OSTI2012

CATHARE calculations of Phenix ultimate natural convection test

The Phenix Sodium cooled Fast Reactor (SFR) started operation in 1973 and it was stopped in 2009. Before the reactor was definitively shutdown, in order to collect experimental data for code assessments in the frame of Generation IV an intensive program of ultimate tests was set up. Among these ultimate experiments, two thermal hydraulic tests were performed: an asymmetrical test consisting in a trip on one secondary pump and a natural convection test in the primary circuit. The natural convection test has been used for an international benchmark on system codes in the frame of the IAEA. The CATHARE code - initially developed for water cooled reactors and now extended for safety analyses for other kinds of reactors, including Sodium Fast Reactor - was used by CEA for this benchmark. The paper reminds briefly the Phenix reactor with the main physical parameters and the instrumentation used during the natural convection test. Main test results are also briefly reminded including the evolution of the core and the heat exchangers inlet and outlet temperatures, and some local temperature measurements. The main developments to perform CATHARE SFR computations and the strategy of system code assessment are presented. Then the CATHARE modelling of Phenix reactor is depicted and the various assumptions are pointed out. CATHARE encountered no problem to predict the initial nominal state. Afterwards, the whole transient scenario is calculated and CATHARE calculations are compared to the Phenix measurements. The global trend is rather well predicted by the CATHARE code. Nevertheless, due to complex flow phenomena occurring in large plena and components, the system code encountered physical limitations, leading to remaining discrepancies between code prediction and plant data. Various sensitivity calculations are presented and they bring partial answers. Additional analyses are in progress to understand more deeply the complex 3D phenomena involved during the different phases of the natural convection test. Additional work for coupling CATHARE system code and TRIO-U CFD code is in progress and will bring useful information to better understand the physical phenomena involved during the natural convection test and to improve system modeling for future SFR safety analysis. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

DOE OSTI2018

PDFgetN3: Atomic Pair Distribution Functions From Neutron Powder Diffraction Data Using ad hoc Corrections

We describe PDFgetN3, a new software tool for the extraction of PDF from neutron powder diffraction intensity data. We demonstrate its use with constant wavelength neutron data measured at the new powder diffractometer PEARL at the Delft University of Technology (TU), Delft. PDFgetN3 uses an ad hoc data collection protocol similar to PDFgetX3. Here, the quality of the resulting PDFs is assessed by structure refinement and by comparison with established results from synchrotron x-ray scattering.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2015

WE-AB-BRB-05: Toward a 2D Water-Equivalent Dosimetry Panel Using KCl:Eu2+

Purpose: KCl:Eu2+ storage phosphor shows promise for radiation therapy dosimetry. The purpose of this work is to investigate several important aspects of this material for potential commercial use. Methods: KCl:Eu2+ chips were fabricated and a conformal coating using Parylene was applied. Material’s dose response in a 6 MV beam was investigated using Monte-Carlo simulations. We attempted to micronize the materials using a spiral jet mill. As we did not have a water-free glovebox, we used commercially available non-hygroscopic BaFBr0.85I0.15:Eu2+ computed radiography material to test if a homogeneous panel can be made using micron-sized phosphors. Results: Dosimeters remained intact and showed no change in PSL intensity after eight hrs of submersion in water. We then optically bleached the samples for reuse, irradiated and immersed for another 24 hrs. We observed marginal worsening of the PSL signal for both the soaked and un-soaked chips. By contrast, we were unable to measure PSL intensity of the un-coated pellets due to these pellets dissolving within minutes of being immersed in water. MC data indicate that the micron-sized KCl:Eu2+ is predicted to have a nearly water-equivalent response. KCl:Eu2+ particles with a median size of 3 microns can be produced using a jet mill, which could be reduced further if necessary. While the particles tend to agglomerate over time when stored in a desiccator, they still possess favorable d50’s and d99’s even after 100 minutes, providing an adequate time window for making a panel via tape casting. A panel cast using optimized methods exhibits nearly perfect particle arrangement. Conclusions: Data shown here support ongoing efforts in fabricating a reusable, high resolution dosimetry panel in a water-free glovebox using micron-sized KCl:Eu2+ particles separated by water-equivalent polymers. The conformal coating thereafter will provide good humidity resistance. HL is the founder of DoseImaging, LLC that is exclusively dedicated to commercializing this technology.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2025

PyCMG-based Simulation of Volumetric Concrete Microstructure

Concrete is a complex, heterogeneous material with a microstructure composed of aggregates, cement paste, and pores spanning multiple length scales. Understanding this microstructure is critical for advancing the performance, durability, and modeling of concrete-based systems. While experimental imaging such as X-ray computed tomography (XCT) provides valuable insights, generating large datasets with detailed ground truth annotations is both costly and labor-intensive due to challenges in segmenting similar phases, such as aggregates and cement paste, that often share similar attenuation properties. To address this, we developed a pipeline to simulate realistic 3D concrete microstructures using the open-source Python package PyCMG. This simulation effort focuses on generating high-fidelity, annotated microstructures that can serve as training or benchmarking datasets for image analysis, segmentation algorithms, and machine learning models, particularly in scenarios where experimental data is scarce.

Ziabari, Amir [Oak Ridge National Laboratory; ORNL↗

DOE OSTI2024

Optimization Studies of Radiation Shielding for PIP-II Project at Fermilab

The Proton Improvement Plan-II (PIP-II) at Fermilab represents a significant advancement in the quest to answer some of the most profound questions about our universe using the world's most intense high-energy neutrino beam. The project requires the construction of a new addition to the Fermilab accelerator complex – an 800-MeV high-intensity superconducting linear accelerator. Ensuring the safety and regulatory compliance of this ambitious project is paramount, necessitating thorough dose rate assessments under both normal operational and accidental scenarios to align with the Fermilab Radiological Control Manual (FRCM) standards. Our approach included a shielding optimization used for the simulations with the Monte Carlo code MARS [1,2,3] to incorporate new magnet and collimator designs, essential for reflecting the current state of PIP-II infrastructure. The implementation of high-resolution detector planes, despite their computational demands, enabled us to gather detailed radiation field data crucial for optimizing shielding configurations. To overcome the significant computational demands, we developed a branching code that drastically reduced simulation runtimes while maintaining statistical integrity. This was achieved through geometry splitting and the application of Russian Roulette techniques, tailored to prioritize regions of interest based on predefined importances and weight limits.

43 PARTICLE ACCELERATORS↗

DOE OSTI2024

Optimization Studies of Radiation Shielding for PIP-II Project at Fermilab

Makovec, Alajos↗

DOE OSTI2020

Enabling Highly Efficient Capsule Networks Processing Through A PIM-Based Architecture Design

In recent years, the CNNs have achieved great successes in the image processing tasks, e.g., image recognition and object detection. Unfortunately, traditional CNN's classication is found to be easily misled by increasingly complex image features due to the usage of pooling operations, hence unable to preserve accurate position and pose information of the objects. To address this challenge, a novel neural network structure called Capsule Network has been proposed, which introduces equivariance through capsules to signicantly enhance the learning ability for image segmentation and object detection. Due to its requirement of performing a high volume of matrix operations, CapsNets have been generally accelerated on modern GPU platforms that provide highly optimized software library for common deep learning tasks. However, based on our performance characterization on modern GPUs, CapsNets exhibit low effciency due to the special program and execution features of their routing procedure, including massive unshareable intermediate variables and intensive syn- chronizations, which are very dicult to optimize at software level. To address these challenges, we propose a hybrid computing architecture design named PIM-CapsNet. It preserves GPU's on-chip computing capability for accelerating CNN types of layers in CapsNet, while pipelining with an off-chip in-memory acceleration solution that effectively tackles routing procedure's ineffciency by leveraging the processing-in-memory capability of today's 3D stacked memory. Using routing procedure's inherent parallellization feature, our design enables hierarchical improvements on CapsNet inference effciency through minimizing data movement and maximizing parallel processing in memory. Evaluation results demonstrate that our proposed design can achieve substantial improvement on both performance and energy savings for CapsNet inference, with almost zero accuracy loss. The results also suggest good performance scalability in optimizing the routing procedure with increasing network size.

Zhang, Xingyao↗

DOE OSTI2011

A METHOD FOR THE STUDY OF ACCRETION DISK EMISSION IN CATACLYSMIC VARIABLES. I. THE MODEL

We have developed a spectrum synthesis method for modeling the ultraviolet (UV) emission from the accretion disk from cataclysmic variables (CVs). The disk is separated into concentric rings, with an internal structure from the Wade and Hubeny disk-atmosphere models. For each ring, a wind atmosphere is calculated in the comoving frame with a vertical velocity structure obtained from a solution of the Euler equation. Using simple assumptions, regarding rotation and the wind streamlines, these one-dimensional models are combined into a single 2.5-dimensional model for which we compute synthetic spectra. We find that the resulting line and continuum behavior as a function of the orbital inclination is consistent with the observations, and verify that the accretion rate affects the wind temperature, leading to corresponding trends in the intensity of UV lines. In general, we also find that the primary mass has a strong effect on the P Cygni absorption profiles, the synthetic emission line profiles are strongly sensitive to the wind temperature structure, and an increase in the mass-loss rate enhances the resonance line intensities. Synthetic spectra were compared with UV data for two high orbital inclination nova-like CVs-RW Tri and V347 Pup. We needed to include disk regions with arbitrary enhanced mass loss to reproduce reasonably well widths and line profiles. This fact and a lack of flux in some high ionization lines may be the signature of the presence of density-enhanced regions in the wind, or alternatively, may result from inadequacies in some of our simplifying assumptions.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2015

TRIO: Burst Buffer Based I/O Orchestration

The growing computing power on leadership HPC systems is often accompanied by ever-escalating failure rates. Checkpointing is a common defensive mechanism used by scientific applications for failure recovery. However, directly writing the large and bursty checkpointing dataset to parallel filesystem can incur significant I/O contention on storage servers. Such contention in turn degrades the raw bandwidth utilization of storage servers and prolongs the average job I/O time of concurrent applications. Recently burst buffer has been proposed as an intermediate layer to absorb the bursty I/O traffic from compute nodes to storage backend. But an I/O orchestration mechanism is still desired to efficiently move checkpointing data from bursty buffers to storage backend. In this paper, we propose a burst buffer based I/O orchestration framework, named TRIO, to intercept and reshape the bursty writes for better sequential write traffic to storage severs. Meanwhile, TRIO coordinates the flushing orders among concurrent burst buffers to alleviate the contention on storage server bandwidth. Our experimental results reveal that TRIO can deliver 30.5% higher bandwidth and reduce the average job I/O time by 37% on average for data-intensive applications in various checkpointing scenarios.

Wang, Teng↗

DOE OSTI2019

PERFORMANCE CHARACTERISTICS OF A VECTOR SENSOR ARRAY IN AN ENERGETIC TIDAL CHANNEL

A compact three-dimensional vector sensor array (termed the NoiseSpotter) was recently developed with the goal of characterizing and localizing sources of sound in energetic environments such as those where marine renewable energy devices are likely to be located. The three vector sensors on the NoiseSpotter are mounted on a bottom platform with each sensor suspended inside an optional flow noise removal shield. The multiple pressure and particle velocity channels on the NoiseSpotter are simultaneously sampled and stored on-board to enable coherent array processing of vector sensor array data. A series of field trials were conducted where the NoiseSpotter was deployed on a bottom platform in an 8 m deep tidal channel in Sequim Bay, Washington. Pure tones were transmitted from a drifting vessel whose distance from the bottom platform varied between 50 m and 500 m. The effect of flow noise on vector sensor array performance is quantified along with the detectability of low intensity sound in an energetic environment. Flow noise removal efficiency for each channel of every vector sensor is demonstrated using multiple metrics derived from the active and reactive intensities computed using the pressure-velocity cross-spectral density matrix. A flow noise removal shield is shown to improve the detectability of low intensity sound such as those that are expected to be emitted by marine renewable energy devices.

Raghukumar, Kaustubha↗

DOE OSTI2022

An Online Dynamic Amplitude-Correcting Gradient Estimation Technique to Align X-ray Focusing Optics

High-brightness X-ray pulses, as generated at synchrotrons and X-ray free electron lasers (XFELs), are used in a variety of scientific experiments. At these facilities, measurements often require optical equipment, e.g Compound Refractive Lenses (CRLs) to be precisely aligned and focused. The lateral alignment of CRLs to a beamline requires precise positioning along four axes: two translational, and the two rotational. At a synchrotron, alignment is often accomplished manually. However, XFEL beamlines present a beam brightness that fluctuates stochastically, making manual alignment a time-consuming endeavor. Automation using simplex or classic stochastic descent often fails, given the errant gradient estimates. Herein we present a dynamic-amplitude correction to the usual gradient based on the combination of a generalized finite difference stencil and a time-dependent sampling pattern. Intensity is recorded periodically, then used to normalize numerical derivatives against fluctuations. Error expectation is analyzed, and efficacy is demonstrated on classic benchmarks. We provide a proof of concept by laterally aligning optics on a simulated XFEL beamline using data recorded at both synchrotron and XFEL facilities.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2013

Assessment of Proton Deflectometry for Exploding Wire Experiments

This project provides the first demonstration of the application of proton deflectometry for the diagnosis of electromagnetic field topology and current-carrying regions in Z-pinch plasma experiments. Over the course of this project several milestones were achieved. High-energy proton beam generation was demonstrated on the short-pulse high-intensity Leopard laser, (10 Joules in ~350 femtoseconds, and the proton beam generation was shown to be reproducible. Next, protons were used to probe the electromagnetic field structure of short circuit loads in order to benchmark the two numerical codes, the resistive-magnetohydrodynamics (MHD) code, Gorgon, and the hybrid particle-in-cell code, LSP for the interpretation of results. Lastly, the proton deflectometry technique was used to map the magnetic field structure of pulsed-power-driven plasma loads including wires and supersonic jets formed with metallic foils. Good agreement between the modeling and experiments has been obtained. The demonstrated technique holds great promise to significantly improve the understanding of current flow and electromagnetic field topology in pulsed power driven high energy density plasmas. Proton probing with a high intensity laser was for the first time implemented in the presence of the harsh debris and x-ray producing z-pinch environment driven by a mega-ampere-scale pulsed-power machine. The intellectual merit of the program was that it investigated strongly driven MHD systems and the influence of magnetic field topology on plasma evolution in pulsed power driven plasmas. The experimental program involved intense field-matter interaction in the generation of the proton probe, as well as the generation of plasma subjected to 1 MegaGauss scale magnetic fields. The computational aspect included two well-documented codes, in combination for the first time to provide accurate interpretation of the experimental results. The broader impact included the support of 2 graduate students, one at UCSD and one at NTF, who were exposed to both the experimental physics work, the MHD and PIC modeling of the system. A first generation college undergraduate student was employed to assist in experiments and data analysis throughout the project. Data resulting from the research program were broadly disseminated by publication in scientific journals, and presentation at international and national conferences and workshops.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2021

Performance Analysis of PIConGPU: Particle-in-Cell on GPUs using NVIDIA’s NSight Systems and NSight Compute

PIConGPU, Particle In Cell on GPUs, is an open source simulations framework for plasma and laser-plasma physics used to develop advanced particle accelerators for radiation therapy of cancer, high energy physics and photon science. While PIConGPU has been optimized for at least 5 years to run well on NVIDIA GPU-based clusters, there has been limited exploration by the development team of potential scalability bottlenecks using recently updated and new tools including NVIDIA’s NVProf tool and the brand-new NVIDIA NSight Suite (Systems and Compute) tools. PIConGPU is a highly optimized application that runs production jobs at scale on a system Oak Ridge Leadership Facility’s (OLCF) Summit supercomputer (using the full machine at 4600 nodes; at 98% of GPU utilization on all ~28000 NVIDIA Volta GPUs). PIConGPU has been selected as one of the the eight applications for OLCF’s coveted Center for Accelerated Application Readiness (CAAR) program aimed at the facility’s Frontier supercomputer (OLCF’s first exascale system to launch in 2021), to partner with our vendors (primary vendors: AMD and Cray/HPE) ensuring that Frontier will be able to perform large-scale science when it opens to users in 2022. To this effect, performance engineers on the PIConGPU team wanted to dive deep into the application to understand at the finest granularity, which portions of the code could be further optimized to exploit the hardware on Summit at it’s maximum potential and also to elucidate which key kernels should be tracked and optimized for the CAAR effort to port this code to Frontier. Any bottlenecks that are observed via performance profiling on Summit are likely to also impact scalability on the Frontier-dev system and the Frontier Early Access (EA) system. Additionally, the engineers wanted to take a closer look at the newest NVIDIA profiling tools which allows us to identify the most useful features on these tools and will provide an opportunity to compare it to new AMD and Cray’s performance analysis tool releases and provide feedback to our vendor partners on what features are most important and mission critical for CAAR efforts. The primary goal of this report is to focus on the evaluation of PIConGPU’s most time-intensive kernels using NVProf and NSight Suite. Three kernels, Current Deposition (also known as Compute Current), Particle Push (Move and Mark), and Shift Particles are known to be some of the most time-consuming kernels in PIConGPU. The Current xi Deposition kernel and Particle Push kernel both set up the particle attributes for running any physics simulation with PIConGPU, so it is crucial to improve the performance of these two kernels. In this report, we measure single GPU metrics for the three kernels, offer high level takeaways from the conducted analysis, and compare the profiling data from NSight Compute to that of NVProf. This analysis was performed using a grid size of 240 x 272 x 224, and 10 time steps with the Mid-November Figure of Merit (FOM) run setup. The Traveling Wave Electron Acceleration (TWEAC) science case used in this run is a representative science case for PIConGPU. This execution can also be used for baseline analysis on AMD MI50/ MI60 systems. As of the time of writing, the PIConGPU application has limited use for features of NSight Systems, so this report will mainly focus on insights garnered from NSight Compute. For this analysis, we run the “full” metric set available in NSight Compute version 2020.1.2 and use NSight Systems version 2020.3.1 to generate the application timeline.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2026

Measuring the Stress Factors for Photovoltaic (PV) Backsheet Degradation

Back sheet failure has resulted in power loss and large-scale recall of photovoltaic modules, resulting in billions of dollars in lost revenue. The light exposure on the backside of a photovoltaic module comes primarily from reflected light which alters the distribution of natural sunlight. Because of this, modelling the backside exposure and duplicating the exposure is much more difficult than modeling the frontside exposure. This project aims to study how various back sheets and junction box materials degrade under different conditions and to develop Python code to help model and predict degradation. The stress factors for back-sheet degradation must be quantified to extrapolate accelerated stress tests to the field. Test samples were placed in the A3, A4, and A5 conditions, as defined in IEC 62788-7-2, to assess the temperature and humidity dependence of ultraviolet (UV) induced degradation. We are utilizing a custom chamber with exposure from 0.5 UV-suns to 5 UV-suns to understand the dependence of degradation on light intensity. A group of samples put in the A3 condition had glass filters with 50% UV cut-offs of 320 nm, 335 nm, and 360 nm to assess the wavelength dependence of UV degradation. All this data is necessary to assess the impact of non-standard UV light exposure. The material evaluation tests include gloss measurements, attenuated total internal reflectance Fourier transform infrared spectroscopy (ATR-FTIR), UV-visible reflectance/transmittance utilizing a Cary Ci7000 spectrophotometer, and a nano-indenter for surface hardness and modulus measurements. Alongside the experimental work, there is a computational effort using raytracing and Python open-source tools in PVDeg , PVLib, and Bifacial_Radiance. This code will create specific exposure scenarios and enable the evaluation of chamber degradation relative to field degradation. Equation 1 is a strawman equation used to model degradation on the backside of a PV module. We will create simplified code, based on the results of ray-tracing calculations, which uses a view factor approach to provide fast calculations for the most common exposure scenarios.

14 SOLAR ENERGY↗

DOE OSTI2024

Stress intensity factor models using mechanics-guided decomposition and symbolic regression

The finite element method can be used to compute accurate stress intensity factors (SIFs) for cracks with complex geometries and boundary conditions. In contrast, handbook solutions act as surrogate SIF models that provide significantly faster evaluation times. However, the development of conventional surrogate SIF models relies on manual development based on low-order parameterizations. This limits surrogate model accuracy and generalizability. Here, in this paper, we develop a framework for the automated development of mechanics-guided handbook SIF solutions by using interpretable machine learning via genetic programming for symbolic regression (GPSR). Formalizing the mechanics-based approach of Raju and Newman, SIF training data is decomposed into multiple subsets. This decomposition enables parallel GPSR model development of subfunctions, each of which accounts for specific geometrical corrections with respect to a known analytical model. Using this mechanics-based approach with GPSR allows for equations to be learned with improved accuracy and reduced complexity relative to the Raju Newman equations while maintaining the inherent interpretability of mathematical expressions. In this paper, we present equations that match the complexity of the Raju Newman equations while having reduced error, as well as equations with similar errors and reduced complexity.

42 ENGINEERING↗