Search NASA⌕ Search

SEARCH · Search NASA

Results for “Robustness analysis”

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 1,045 records · Page 58
All sourcesNASA NTRSDOE OSTI
NASA NTRS2017

Ames Engineering Directorate

The Ames Engineering Directorate is the principal engineering organization supporting aerospace systems and spaceflight projects at NASA's Ames Research Center in California's Silicon Valley. The Directorate supports all phases of engineering and project management for flight and mission projects-from R&D to Close-out-by leveraging the capabilities of multiple divisions and facilities.The Mission Design Center (MDC) has full end-to-end mission design capability with sophisticated analysis and simulation tools in a collaborative concurrent design environment. Services include concept maturity level (CML) maturation, spacecraft design and trades, scientific instruments selection, feasibility assessments, and proposal support and partnerships. The Engineering Systems Division provides robust project management support as well as systems engineering, mechanical and electrical analysis and design, technical authority and project integration support to a variety of programs and projects across NASA centers. The Applied Manufacturing Division turns abstract ideas into tangible hardware for aeronautics, spaceflight and science applications, specializing in fabrication methods and management of complex fabrication projects. The Engineering Evaluation Lab (EEL) provides full satellite or payload environmental testing services including vibration, temperature, humidity, immersion, pressure/altitude, vacuum, high G centrifuge, shock impact testing and the Flight Processing Center (FPC), which includes cleanrooms, bonded stores and flight preparation resources. The Multi-Mission Operations Center (MMOC) is composed of the facilities, networks, IT equipment, software and support services needed by flight projects to effectively and efficiently perform all mission functions, including planning, scheduling, command, telemetry processing and science analysis.

Small Satellites↗
NASA NTRSDate not supplied

TPSAS-NF1676L-14336-DND

Problem Statement: Develop robust CFD-based methods for rapid design and analysis of supersonic cruise aircraft that are highly efficient, and have a low sonic boom. What is our approach? - Advanced Efficient Analysis / Design Methods - Develop computational algorithms and techniques for rapid high fidelity analysis of supersonic fluid flow phenomena, which will also contribute to multi-fidelity analysis / design tools. - Experimental Validation - Validate concepts for design optimization of both high aerodynamic efficiency and low sonic boom.

Linda Bangert↗
NASA NTRS2021

TESS-Keck Survey. V. Twin Sub-Neptunes Transiting the Nearby G Star HD 63935

We present the discovery of two nearly identically sized sub-Neptune transiting planets orbiting HD 63935, abright(V=8.6 mag), Sun-like(Teff=5560 K)star at 49 pc. TESS identified thefirst planet, HD 63935 b(TOI-509.01), in Sectors 7 and 34. We identified the second signal(HD 63935 c)in Keck High Resolution EchelleSpectrometer and Lick Automated Planet Finder radial velocity data as part of our follow-up campaign. It wassubsequently confirmed with TESS photometry in Sector 34 as TOI-509.02. Our analysis of the photometric andradial velocity data yielded a robust detection of both planets with periods of 9.0600±0.007 and 21.40±0.0019days, radii of 2.99±0.14 and 2.90±0.13R⊕, and masses of 10.8±1.8 and 11.1±2.4M⊕. We calculated densities for planets b and c consistent with a few percent of the planet mass in hydrogen/helium envelopes. Wealso describe our survey’s efforts to choose the best targets for James Webb Space Telescope atmospheric follow-up. These efforts suggest that HD 63935 b has the most clearly visible atmosphere of its class. It is the best targetfor transmission spectroscopy(ranked by the transmission spectroscopy metric, a proxy for atmosphericobservability)in the so far uncharacterized parameter space comprising sub-Neptune-sized(2.6R⊕<Rp<4R⊕),moderately irradiated(100F⊕<Fp<1000F⊕)planets around G stars. Planet c is also a viable target fortransmission spectroscopy, and given the indistinguishable masses and radii of the two planets, the system servesas a natural laboratory for examining the processes that shape the evolution of sub-Neptune planets.

Nicholas Scarsdale↗
NASA NTRSDate not supplied

Recent Advances in the U.S. in Ablative TPS for In-situ Exploration of Giant Planets

The Decadal Survey report released [1] prioritizes Uranus as the highest priority Flagship class mission to be explored with an orbiter and a probe and it also recommended Saturn Probe under New Frontiers mission class. Thermal protection system (TPS) is essential for Uranus and Saturn probe missions. The two cardinal requirements are that it must be fail-safe and yet be mass efficient. The Decadal Survey report also pointed out the readiness of heatshield for extreme entry environment technology (HEEET) at TRL 6 for probe missions at Gas Giants. HEEET relies on 3-D weaving and is shown to be a robust and mass efficient TPS through ground testing and analysis. HEEET was matured to TRL 6 in 2019. HEEET was reported at the Ice Giant Workshop in 2019 at Marseille. Since 2019, significant advances have been made primarily because of Mars Sample Return (MSR) mission. MSR baselined 3-D Woven TPS as its heatshield for the earth entry system. MSR earth entry system (EES) requirements and the resulting heatshield/TPS requirements are most stringent of all entry missions. This is a result of backward contamination protection which classifies MSR as a “restricted class 5” mission to safeguard the accidental release of potentially hazardous Mars Sample into Earth’s atmosphere. Complex requirements for the heatshield start with micro-meteor impact tolerance followed by the requirement for steep entry to minimize the size of the impact footprint which results in extreme heating. The heatshield is part of the impact attenuation system. The EES architecture does not use a parachute and is designed to tolerate impact loads. Hence heatshield TPS selection and design becomes one of the key challenges. After completion of the HEEET technology maturation in 2019, IRAD efforts focused on assessing the 3-D Woven family of TPS to MSR EES. After nearly two years of evaluation of alternate TPS such as carbon phenolic, C-C hot structure, PICA, and 3-D woven family of TPS, a single layer 3-D woven TPS derived from the dual-layer HEEET was down selected and is currently the baseline. A new loom capable of weaving the single-layer 3-D woven preform at 80” wide has been designed, assembled and is currently on the verge of weaving the MSR EES TPS. During the HEEET maturation and the follow-on development, the single-layer TPS has been tested at extreme heating conditions. As a result, single layer TPS was recommended and evaluated during the Planetary Mission Concept Studies funded by NASA in preparation for the Decadal committee. In addition, in anticipation of Saturn mission proposals, single layer was evaluated as well. Aerocapture can reduce the trip time, also allow bigger payload fraction and in addition, it allows for the possibility of probe delivery once the spacecraft is in orbit. This can lead to obtaining both in-situ data as well as data from orbit simultaneously. Since Aerocapture depletes energy/velocity, probe delivery from orbit reduces the demand on TPS. Aerocapture was mentioned in the Decadal Study report as ready for implementation, but due to perceived risk it is not adopted by the mission designers. Establishing TPS readiness for aerocapture missions will be addressed. Going one step beyond aerocapture is aerogravity assist if fast return is the goal. Recent studies [4] looked at aerogravity assist and the TPS readiness. This talk will highlight both aerocapture and aerogravity assist from a TPS perspective. The main objective of this proposed talk is to present a comprehensive picture of the SOA TPS technology including recent developments. The talk will highlight advances in manufacturing, results from the Decadal White Papers, PMCS and other studies, and aerocapture and aerogravity assist that could play a role in the near or far term in-situ exploration.

E. Venkatapathy↗
NASA NTRSDate not supplied

The Third Flight of the Marshall Grazing Incidence X-Ray Spectrometer (MaGIXS-3)

The Marshall Grazing Incidence X-ray Spectrometer (MaGIXS) is the first X-ray slitless imaging spectrograph sounding rocket instrument designed to observe spectrally dispersed soft X-ray images of the solar corona over a wide field-of-view. During the first flight of MaGIXS (MaGIXS-1), occurred on 30 July 2021, several emission lines from coronal structures including X-ray bright points were observed. Further, MaGIXS-1 analysis also demonstrated the successful inversion of overlappograms using robust unfolding algorithms. Given the demonstrated success of MaGIXS-1, the second flight of the instrument with a simplified optical design, MaGIXS-2, is scheduled for 2024 to observe high temperature diagnostic emission lines. Results from MaGIXS-1 discovered dominant missing emission lines near 15A, arising from relatively cool plasma that peaks around 2 MK. This wavelength region hosts several closely spaced satellite lines of Fe XVII, Fe XVI and Fe XV ions, which are expected to be enhanced at lower temperatures and are currently unmodeled in the CHIANTI atomic database. This wavelength region offers one of the most unique diagnostics to measure electron temperature, as well as signatures of equilibrium state of the plasma in active regions, which has not been studied so far. The goal of MaGIXS-3 mission is the to determine coronal heating parameters, such as the spatial and temporal properties of coronal heating events, by measuring discriminating observations, such as electron temperature, effective temperature, density, abundance, and departures from thermal equilibrium of the plasma, by observing the Sun in the SXR wavelength range. To meet this goal, the MaGIXS instrument will be upgraded with a new X-ray telescope mirror that will provide higher spatial and spectral resolution and throughput, allowing for spectral lines to be observed at the relevant spatial and temporal scales. In addition, MaGIXS-3 will also carry The Resolving Inversion Context X-ray Spectrometer (TRICXS), a high dispersion Bragg crystal spectrometer to spectrally resolve the lines near 14.9 to 15.9A, critical to unlock the full diagnostic potential of this wavelength range. Here we will present the preliminary concept design of MaGIXS-3 and discuss the potential observations.

X-ray Imaging↗
DOE OSTI2021

Combination of cluster number counts and two-point correlations: validation on mock Dark Energy Survey

ABSTRACT We present a method of combining cluster abundances and large-scale two-point correlations, namely galaxy clustering, galaxy–cluster cross-correlations, cluster autocorrelations, and cluster lensing. This data vector yields comparable cosmological constraints to traditional analyses that rely on small-scale cluster lensing for mass calibration. We use cosmological survey simulations designed to resemble the Dark Energy Survey Year 1 (DES-Y1) data to validate the analytical covariance matrix and the parameter inferences. The posterior distribution from the analysis of simulations is statistically consistent with the absence of systematic biases detectable at the precision of the DES-Y1 experiment. We compare the χ2 values in simulations to their expectation and find no significant difference. The robustness of our results against a variety of systematic effects is verified using a simulated likelihood analysis of DES-Y1-like data vectors. This work presents the first-ever end-to-end validation of a cluster abundance cosmological analysis on galaxy catalogue level simulations.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

ARCADE Technical Pathway and Industry Impact

The Advanced Reactor Cyber Analysis and Development Environment (ARCADE) simplifies the evaluation and assessment of robustness factor and cyber resilience that support secure-by-design for advanced reactor nuclear power plants. In this manner, ARCADE supports risk-informed performance based (RIPB) evaluations of cybersecurity through its integration of plant physics with high-fidelity emulations of control systems. This cross domain approach enables comprehensive analysis of control system sensitivities, cyber-attack scenarios, and their consequences. ARCADE has been custom developed to meet the demands identified in Tier 1 of the Tiered Cyber Analysis (TCA) as outlined in NRC Draft Regulation Guide (RG) 5.96, which provides a RIPB cybersecurity approach for new reactors.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Utilizing Semantic Big Data for realizing a National-scale Infrastructure Vulnerability Analysis System

Critical Infrastructure systems(CIs) such as energy, water, transportation and communication are highly interconnected and mutually dependent in complex ways. Robust modeling of CIs interconnections is crucial to identify vulnerabilities in the CIs. We present here a national-scale Infrastructure Vulnerability Analysis System (IVAS) vision leveraging Se- mantic Big Data (SBD) tools, Big Data, and Geographical Information Systems (GIS) tools. We survey existing ap- proaches on vulnerability analysis of critical infrastructures and discuss relevant systems and tools aligned with our vi- sion. Next, we present a generic system architecture and discuss challenges including: (1) Constructing and manag- ing a CI network-of-networks graph, (2) Performing analytic operations at scale, and (3) Interactive visualization of ana- lytic output to generate meaningful insights. We argue that this architecture acts as a baseline to realize a national-scale network based vulnerability analysis system.

Chinthavali, Supriya↗
NASA NTRS2020

Robust Conceptual Design of Transonic Airfoils

This paper describes an integrated, multi-fidelity analysis and heuristic design approach that can be used to derive initial airfoil designs for transonic flight. If successful, the final result is a geometry that can be expected to produce reasonable aerodynamic performance when used with higher order analysis methods. A key aspect of the methodology is the use of a “sonic-plateau” pressure distribution as the target distribution for inverse design. The sonic- plateau distribution is easily parameterized and has the advantage of automatically resulting in a smooth airfoil shape without any discontinuities built into the surface due to the presence of a shock in the target pressure distribution. Inverse design is performed on each airfoil using a parametrically defined pressure distribution at a reduced lift coefficient and Mach number from the design condition. The methodology is demonstrated by designing an airfoil at 38% of the wing semispan for a 737-200-like aircraft. The demonstration problem shows that the methodology is able to achieve rapid and robust convergence to the solution. The calculated designed airfoil was found to be sufficiently higher than Mach number, and the maximum thickness was close to the targeted value.

Olson, Erik D.↗
DOE OSTI2015

Comparative analysis for various redox flow batteries chemistries using a cost performance model

A robust performance-based cost model is developed for all-vanadium, iron-vanadium and iron chromium redox flow batteries. Systems aspects such as shunt current losses, pumping losses and thermal management are accounted for. The objective function, set to minimize system cost, allows determination of stack design and operating parameters such as current density, flow rate and depth of discharge (DOD). Component costs obtained from vendors are used to calculate system costs for various time frames. A 2 kW stack data was used to estimate unit energy costs and compared with model estimates for the same size electrodes. The tool has been shared with the redox flow battery community to both validate their stack data and guide future direction.

electrolyte↗
DOE OSTI2020

A robust benchmark for detection of germline large deletions and insertions

New technologies and analysis methods are enabling genomic structural variants (SVs) to be detected with ever-increasing accuracy, resolution and comprehensiveness. To help translate these methods to routine research and clinical practice, we developed a sequence-resolved benchmark set for identification of both false-negative and false-positive germline large insertions and deletions. To create this benchmark for a broadly consented son in a Personal Genome Project trio with broadly available cells and DNA, the Genome in a Bottle Consortium integrated 19 sequence-resolved variant calling methods from diverse technologies. The final benchmark set contains 12,745 isolated, sequence-resolved insertion (7,281) and deletion (5,464) calls ≥50 base pairs (bp). The Tier 1 benchmark regions, for which any extra calls are putative false positives, cover 2.51 Gbp and 5,262 insertions and 4,095 deletions supported by ≥1 diploid assembly. We demonstrate that the benchmark set reliably identifies false negatives and false positives in high-quality SV callsets from short-, linked- and long-read sequencing and optical mapping.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Robustness of topological persistence in knowledge distillation for wearable sensor data

Topological data analysis (TDA) has shown great success in various applications involving wearable sensor data. However, there are difficulties in leveraging topological features in machine learning and wearable sensors because of the large time consumption and computational resources required to extract the features. To address this problem, knowledge distillation (KD) is utilized to generate a small model and accommodate topological features with persistence image (PI) representations from the raw time series data. Deploying topological knowledge in KD enables the student to achieve better performance compared to the one trained solely on raw time series data. However, it is not yet known if there are coherent characteristics for topological features in PI, which can aid in improving the performance during KD. In this paper, we investigate the suitability and challenges of utilizing topological features in KD for wearable sensor data, thereby contributing to the advancement of the field. Our study explores the impact of transferred topological features by comparing the Teacher-to-Student framework with Multiple Teachers-to-Student where teachers utilize both time series data and persistence images obtained by TDA as inputs. Additionally, we conduct a rigorous examination of topological knowledge effects by testing under various corruptions, knowledge types, and learning strategies in the context of human activity recognition tasks. Our analysis of topological features in KD presents the optimal strategy for incorporating these features. This study includes datasets of varying scales, window lengths, and activity classes, providing a comprehensive evaluation. Our results demonstrate that leveraging topological features in KD to enhance performance across databases.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2012

Development of a Robust and Efficient Parallel Solver for Unsteady Turbomachinery Flows

The traditional design and analysis practice for advanced propulsion systems relies heavily on expensive full-scale prototype development and testing. Over the past decade, use of high-fidelity analysis and design tools such as CFD early in the product development cycle has been identified as one way to alleviate testing costs and to develop these devices better, faster and cheaper. In the design of advanced propulsion systems, CFD plays a major role in defining the required performance over the entire flight regime, as well as in testing the sensitivity of the design to the different modes of operation. Increased emphasis is being placed on developing and applying CFD models to simulate the flow field environments and performance of advanced propulsion systems. This necessitates the development of next generation computational tools which can be used effectively and reliably in a design environment. The turbomachinery simulation capability presented here is being developed in a computational tool called Loci-STREAM [1]. It integrates proven numerical methods for generalized grids and state-of-the-art physical models in a novel rule-based programming framework called Loci [2] which allows: (a) seamless integration of multidisciplinary physics in a unified manner, and (b) automatic handling of massively parallel computing. The objective is to be able to routinely simulate problems involving complex geometries requiring large unstructured grids and complex multidisciplinary physics. An immediate application of interest is simulation of unsteady flows in rocket turbopumps, particularly in cryogenic liquid rocket engines. The key components of the overall methodology presented in this paper are the following: (a) high fidelity unsteady simulation capability based on Detached Eddy Simulation (DES) in conjunction with second-order temporal discretization, (b) compliance with Geometric Conservation Law (GCL) in order to maintain conservative property on moving meshes for second-order time-stepping scheme, (c) a novel cloud-of-points interpolation method (based on a fast parallel kd-tree search algorithm) for interfaces between turbomachinery components in relative motion which is demonstrated to be highly scalable, and (d) demonstrated accuracy and parallel scalability on large grids (approx 250 million cells) in full turbomachinery geometries.

West, Jeff↗
DOE OSTI2026

Robust Automatic EXAFS First-Shell Fits

Extended X-ray absorption fine structure (EXAFS) is a widely used technique for atomic structure determination. Fourier transformation connects EXAFS in k space and R space. However, determining the appropriate k-range for the transformation can be challenging, but critical for the first-shell fit. In this study, we present an automatic method to determine the k-range using the Larch package and a Python program. The first step is to estimate spectral noise across a series of k-ranges with a fixed minimum value and identify the optimal maximum value in the k-range (k max ). The k max is determined by an empirical noise threshold that marks the point where the noise level in the Fourier transformed spectrum changes dramatically. Using the obtained k max value, the first shell is modeled to determine the minimum k value (k min ) by optimizing the background function through alignment of the spectrum with theory. The optimal k min corresponds to the point of the minimum R-factor, which quantifies the difference between the experimental and fitted spectrum. Our method was tested on various typical datasets and yielded suitable k-ranges for Fourier transformation and accurate first-shell fits. This approach helps avoid unreliable, irreproducible data analysis, especially for noisy data from diluted samples, and enables robust automatic first-shell EXAFS fitting.

EXAFS analysis↗
DOE OSTI2016

Functional Data Analysis Applied to Modeling of Severe Acute Mucositis and Dysphagia Resulting From Head and Neck Radiation Therapy

Purpose: Current normal tissue complication probability modeling using logistic regression suffers from bias and high uncertainty in the presence of highly correlated radiation therapy (RT) dose data. This hinders robust estimates of dose-response associations and, hence, optimal normal tissue–sparing strategies from being elucidated. Using functional data analysis (FDA) to reduce the dimensionality of the dose data could overcome this limitation. Methods and Materials: FDA was applied to modeling of severe acute mucositis and dysphagia resulting from head and neck RT. Functional partial least squares regression (FPLS) and functional principal component analysis were used for dimensionality reduction of the dose-volume histogram data. The reduced dose data were input into functional logistic regression models (functional partial least squares–logistic regression [FPLS-LR] and functional principal component–logistic regression [FPC-LR]) along with clinical data. This approach was compared with penalized logistic regression (PLR) in terms of predictive performance and the significance of treatment covariate–response associations, assessed using bootstrapping. Results: The area under the receiver operating characteristic curve for the PLR, FPC-LR, and FPLS-LR models was 0.65, 0.69, and 0.67, respectively, for mucositis (internal validation) and 0.81, 0.83, and 0.83, respectively, for dysphagia (external validation). The calibration slopes/intercepts for the PLR, FPC-LR, and FPLS-LR models were 1.6/−0.67, 0.45/0.47, and 0.40/0.49, respectively, for mucositis (internal validation) and 2.5/−0.96, 0.79/−0.04, and 0.79/0.00, respectively, for dysphagia (external validation). The bootstrapped odds ratios indicated significant associations between RT dose and severe toxicity in the mucositis and dysphagia FDA models. Cisplatin was significantly associated with severe dysphagia in the FDA models. None of the covariates was significantly associated with severe toxicity in the PLR models. Dose levels greater than approximately 1.0 Gy/fraction were most strongly associated with severe acute mucositis and dysphagia in the FDA models. Conclusions: FPLS and functional principal component analysis marginally improved predictive performance compared with PLR and provided robust dose-response associations. FDA is recommended for use in normal tissue complication probability modeling.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2019

A viscous vortex lattice method for analysis of cross-flow propellers and turbines

Marine hydrokinetic turbines are a promising source of renewable energy, but the costs of existing cross-flow turbine designs are too high. In order to develop new designs with reduced costs, there exists a need for a low-order computational model that is simultaneously fast, accurate, and robust. In this report we present a novel theoretical model for the analysis of cross-flow propellers and turbines based on the vortex lattice method (VLM). In order to overcome the limitation that the VLM ignores viscous effects (such as trailing-edge flow separation), we present a novel method to account for viscous-thickness-load coupling (VTLC). In order to overcome the difficulty that wake passages cause the VLM to be unstable, we present the panel-averaged induced velocity for the influence of wake vortices on the blades. Herein, we describe our model and provide verification versus RANS CFD. Results show that our VLM + VTLC model predictions are three times more accurate than prior low-order models (i.e. three times lower root mean square error to high-fidelity RANS results) yet 10,000 times faster than RANS over a wide range of operating conditions.

16 TIDAL AND WAVE POWER↗
DOE OSTI2017

Testing ΛCDM at the lowest redshifts with SN Ia and galaxy velocities

Peculiar velocities of objects in the nearby universe are correlated due to the gravitational pull of large-scale structure. By measuring these velocities, we have a unique opportunity to test the cosmological model at the lowest redshifts. We perform this test, using current data to constrain the amplitude of the "signal" covariance matrix describing the velocities and their correlations. We consider a new, well-calibrated "Supercal" set of low-redshift SNe Ia as well as a set of distances derived from the fundamental plane relation of 6dFGS galaxies. Here, analyzing the SN and galaxy data separately, both results are consistent with the peculiar velocity signal of our fiducial ΛCDM model, ruling out the noise-only model with zero peculiar velocities at greater than 7σ (SNe) and 8σ (galaxies). When the two data sets are combined appropriately, the precision of the test increases slightly, resulting in a constraint on the signal amplitude of A = 1.05 -0.21 +0.25 where A = 1 corresponds to our fiducial model. Equivalently, we report an 11% measurement of the product of the growth rate and amplitude of mass fluctuations evaluated at z eff = 0.02, f σ 8 = 0.428 -0.045 +0.048 , valid for our fiducial ΛCDM model. Here, we explore the robustness of the results to a number of conceivable variations in the analysis and find that individual variations shift the preferred signal amplitude by less than ~0.5σ. We briefly discuss our Supercal SN Ia results in comparison with our previous results using the JLA compilation.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2015

SU-E-T-131: Artificial Neural Networks Applied to Overall Survival Prediction for Patients with Periampullary Carcinoma

Purpose: Artificial neural networks (ANN) can be used to discover complex relations within datasets to help with medical decision making. This study aimed to develop an ANN method to predict two-year overall survival of patients with peri-ampullary cancer (PAC) following resection. Methods: Data were collected from 334 patients with PAC following resection treated in our institutional pancreatic tumor registry between 2006 and 2012. The dataset contains 14 variables including age, gender, T-stage, tumor differentiation, positive-lymph-node ratio, positive resection margins, chemotherapy, radiation therapy, and tumor histology.After censoring for two-year survival analysis, 309 patients were left, of which 44 patients (∼15%) were randomly selected to form testing set. The remaining 265 cases were randomly divided into training set (211 cases, ∼80% of 265) and validation set (54 cases, ∼20% of 265) for 20 times to build 20 ANN models. Each ANN has one hidden layer with 5 units. The 20 ANN models were ranked according to their concordance index (c-index) of prediction on validation sets. To further improve prediction, the top 10% of ANN models were selected, and their outputs averaged for prediction on testing set. Results: By random division, 44 cases in testing set and the remaining 265 cases have approximately equal two-year survival rates, 36.4% and 35.5% respectively. The 20 ANN models, which were trained and validated on the 265 cases, yielded mean c-indexes as 0.59 and 0.63 on validation sets and the testing set, respectively. C-index was 0.72 when the two best ANN models (top 10%) were used in prediction on testing set. The c-index of Cox regression analysis was 0.63. Conclusion: ANN improved survival prediction for patients with PAC. More patient data and further analysis of additional factors may be needed for a more robust model, which will help guide physicians in providing optimal post-operative care. This project was supported by PA CURE Grant.

60 APPLIED LIFE SCIENCES↗