Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Neural Network”

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 883 records · Page 49

Sparse Convolutional Neural Networks for particle classification in ProtoDUNE-SP events

Deep Learning (DL) methods and Computer Vision are becoming important tools for event reconstruction in particle physics detectors. In this work, we report on the use of submanifold sparse convolutional neural networks (SparseNets) for the classification of track and shower hits from a DUNE prototype liquid-argon detector at CERN (ProtoDUNE-SP). By taking advantage of the three-dimensional nature of the problem we use a set of nine input features to classify sparse and locally dense hits associated to track or shower particles. The SparseNet has been trained on a test sample and shows promising results: efficiencies and purities greater than 90%. This has also been achieved with a considerable speedup and substantially less resource utilization with respect to other DL networks such as graph neural networks. This method offers great scalability advantages for future large neutrino detectors such as the planned DUNE experiment.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Using convolutional neural networks to detect edge localized modes in DIII-D from Doppler backscattering measurements

In H-mode tokamak plasmas, the plasma is sometimes ejected beyond the edge transport barrier. These events are known as edge localized modes (ELMs). ELMs cause a loss of energy and damage the vessel walls. Understanding the physics of ELMs, and by extension, how to detect and mitigate them, is an important challenge. In this paper, we focus on two diagnostic methods—deuterium-alpha (D α ) spectroscopy and Doppler backscattering (DBS). The former detects ELMs by measuring Balmer alpha emission, while the latter uses microwave radiation to probe the plasma. DBS has the advantages of having a higher temporal resolution and robustness to damage. These advantages of DBS diagnostic may be beneficial for future operational tokamaks, and thus, data processing techniques for DBS should be developed in preparation. In sight of this, we explore the training of neural networks to detect ELMs from DBS data, using D α data as the ground truth. With shots found in the DIII-D database, the model is trained to classify each time step based on the occurrence of an ELM event. The results are promising. When tested on shots similar to those used for training, the model is capable of consistently achieving a high f1-score of 0.93. Furthermore, this score is a performance metric for imbalanced datasets that ranges between 0 and 1. We evaluate the performance of our neural network on a variety of ELMs in different high confinement regimes (grassy ELM, RMP mitigated, and wide-pedestal), finding broad applicability. Beyond ELMs, our work demonstrates the wider feasibility of applying neural networks to data from DBS diagnostic.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Neural network denoising of x-ray images from high-energy-density experiments

Noise is a consistent problem for x-ray transmission images of High-Energy-Density (HED) experiments because it can significantly affect the accuracy of inferring quantitative physical properties from these images. We consider experiments that use x-ray area backlighting to image a thin layer of opaque material within a physics package to observe its hydrodynamic evolution. The spatial variance of the x-ray transmission across the system due to changing opacity serves as an analog for measuring density in this evolving layer. The noise in these images adds nonphysical variations in measured intensity, which can significantly reduce the accuracy of our inferred densities, particularly at small spatial scales. Denoising these images is thus necessary to improve our quantitative analysis, but any denoising method also affects the underlying information in the image. In this paper, we present a method for denoising HED x-ray images via a deep convolutional neural network model with a modified DenseNet architecture. In our denoising framework, we estimate the noise present in the real (data) images of interest and apply the inferred noise distribution to a set of natural images. These synthetic noisy images are then used to train a neural network model to recognize and remove noise of that character. We show that our trained denoiser network significantly reduces the noise in our experimental images while retaining important physical features.

47 OTHER INSTRUMENTATION↗

Cascading a systolic array and a feedforward neural network for navigation and obstacle avoidance using potential fields

A technique is developed for vehicle navigation and control in the presence of obstacles. A potential function was devised that peaks at the surface of obstacles and has its minimum at the proper vehicle destination. This function is computed using a systolic array and is guaranteed not to have local minima. A feedfoward neural network is then used to control the steering of the vehicle using local potential field information. In this case, the vehicle is a trailer truck backing up. Previous work has demonstrated the capability of a neural network to control steering of such a trailer truck backing to a loading platform, but without obstacles. Now, the neural network was able to learn to navigate a trailer truck around obstacles while backing toward its destination. The network is trained in an obstacle free space to follow the negative gradient of the field, after which the network is able to control and navigate the truck to its target destination in a space of obstacles which may be stationary or movable.

Plumer, Edward S.↗

Neural Network Approach to Locating Cryptography in Object Code

Finding and identifying cryptography is a growing concern in the malware analysis community. In this paper, artificial neural networks are used to classify functional blocks from a disassembled program as being either cryptography related or not. The resulting system, referred to as NNLC (Neural Net for Locating Cryptography) is presented and results of applying this system to various libraries are described.

97 MATHEMATICS AND COMPUTING↗

A simple method to derive bounds on the size and to train multilayer neural networks

A new derivation is presented for the bounds on the size of a multilayer neural network to exactly implement an arbitrary training set; namely, the training set can be implemented with zero error with two layers and with the number of the hidden-layer neurons equal to no.1 is greater than p - 1. The derivation does not require the separation of the input space by particular hyperplanes, as in previous derivations. The weights for the hidden layer can be chosen almost arbitrarily, and the weights for the output layer can be found by solving no.1 + 1 linear equations. The method presented exactly solves (M), the multilayer neural network training problem, for any arbitrary training set.

Sartori, Michael A.↗

Parallel physics-informed neural networks via domain decomposition

Here we develop a distributed framework for the physics-informed neural networks (PINNs) based on two recent extensions, namely conservative PINNs (cPINNs) and extended PINNs (XPINNs), which employ domain decomposition in space and in time-space, respectively. This domain decomposition endows cPINNs and XPINNs with several advantages over the vanilla PINNs, such as parallelization capacity, large representation capacity, efficient hyperparameter tuning, and is particularly effective for multi-scale and multi-physics problems. Here, we present a parallel algorithm for cPINNs and XPINNs constructed with a hybrid programming model described by MPI + X, where X ∈ {CPUs, GPUs}. The main advantage of cPINN and XPINN over the more classical data and model parallel approaches is the flexibility of optimizing all hyperparameters of each neural network separately in each subdomain. We compare the performance of distributed cPINNs and XPINNs for various forward problems, using both weak and strong scalings. Our results indicate that for space domain decomposition, cPINNs are more efficient in terms of communication cost but XPINNs provide greater flexibility as they can also handle time-domain decomposition for any differential equations, and can deal with any arbitrarily shaped complex subdomains. To this end, we also present an application of the parallel XPINN method for solving an inverse diffusion problem with variable conductivity on the United States map, using ten regions as subdomains.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Parameter estimation in space systems using recurrent neural networks

The identification of time-varying parameters encountered in space systems is addressed, using artificial neural systems. A hybrid feedforward/feedback neural network, namely a recurrent multilayer perception, is used as the model structure in the nonlinear system identification. The feedforward portion of the network architecture provides its well-known interpolation property, while through recurrency and cross-talk, the local information feedback enables representation of temporal variations in the system nonlinearities. The standard back-propagation-learning algorithm is modified and it is used for both the off-line and on-line supervised training of the proposed hybrid network. The performance of recurrent multilayer perceptron networks in identifying parameters of nonlinear dynamic systems is investigated by estimating the mass properties of a representative large spacecraft. The changes in the spacecraft inertia are predicted using a trained neural network, during two configurations corresponding to the early and late stages of the spacecraft on-orbit assembly sequence. The proposed on-line mass properties estimation capability offers encouraging results, though, further research is warranted for training and testing the predictive capabilities of these networks beyond nominal spacecraft operations.

Parlos, Alexander G.↗

Using neural networks to assist in OPAD data analysis

The space shuttle main engine (SSME) became the subject of plume emission spectroscopy in 1986 when researchers from NASA-Marshall Space Flight Center (MSFC), Arnold Engineering Development Center (AEDC), and Rocketdyne went to the SSME test stands at the NASA-Stennis Space Center and at Rocketdyne's Santa Susan Field Laboratory to optically observe the plume. Since then, plume spectral acquisitions have recorded many nominal tests and the qualitative spectral features of the SSME plume are now well established. Significant discoveries made with both wide-band and narrow-band plume emission spectroscopy systems led MSFC to promote the Optical Plume Anomaly Detection (OPAD) program with a goal of instrumenting all SSME test stands with customized spectrometer systems. A prototype OPAD system is now installed on the SSME Technology Test Bed (TTB) at MSFC. The OPAD system instrumentation consists of a broad-band, optical multiple-channel analyzer (OMA) and a narrow-band device called a polychrometer. The OMA is a high-resolution (1.5-2.0 Angstroms) 'super-spectrometer' covering the near-ultraviolet to near-infrared waveband (2800-7400 Angstroms), providing two scans per second. The polychrometer consists of sixteen narrow-band radiometers: fourteen monitoring discrete wavelengths of health and condition monitoring elements and two dedicated to monitoring background emissions. All sixteen channels are capable of providing 500 samples per second. To date, the prototype OPAD system has been used during 43 SSME firings on the TTB, collecting well over 250 megabytes of plume spectral data. One goal of OPAD data analysis is to determine interatively with the help of a computer code, SPECTRA4, developed at AEDC. Experience has shown that iteration with SPECTRA4 is an incredibly labor intensive task and not one to be performed by band. What is really needed is the 'inverse' of SPECTRA4 but the mathematical model for this inverse mapping is tenuous at best. However, the robustness of PSECTRA4 run in the 'forward' direction means that accurate input/output mappings can be obtained. If the mappings were inverted (i.e., input becomes output and output becomes input) then an 'inverse' of SPECTRA4 would be at hand but the 'model' would be specific to the data utilized and would in no way be general. Building a generalized model based upon known input/output mappings while ignoring the details of the governing physical model is possible through the use of a neural network. The research investigation described involves the development of a neural network to provide a generalized 'inverse' of SPECTRA4. The objectives of the research were to design an appropriate neural network architecture, train the network, and then evaluate its performance.

Whitaker, Kevin W.↗

Embedded training of neural-network subgrid-scale turbulence models

We report that the weights of a deep neural-network model are optimized in conjunction with the governing flow equations to provide a model for subgrid-scale stresses in a temporally developing plane turbulent jet at Reynolds number Re 0 = 6000 . The objective function for training is first based on the instantaneous filtered velocity fields from a corresponding direct numerical simulation, and the training is by a stochastic gradient descent method, which uses the adjoint Navier-Stokes equations to provide the end-to-end sensitivities of the model weights to the velocity fields. In-sample and out-of-sample testing on multiple dual-jet configurations show that its required mesh density in each coordinate direction for prediction of mean flow, Reynolds stresses, and spectra is half that needed by the dynamic Smagorinsky model for comparable accuracy. The same neural-network model trained directly to match filtered subgrid-scale stresses, without the constraint of being embedded within the flow equations during the training, fails to provide a qualitatively correct prediction. The coupled formulation is generalized to train based only on mean-flow and Reynolds stresses, which are more readily available in experiments. The mean-flow training provides a robust model, which is important, though a somewhat less accurate prediction for the same coarse meshes, as might be anticipated due to the reduced information available for training in this case. The anticipated advantage of the formulation is that the inclusion of resolved physics in the training increases its capacity to extrapolate. This is assessed for the case of passive scalar transport, for which it outperforms established models due to improved mixing predictions.

42 ENGINEERING↗

Inverse Design of Photonic Surfaces via High throughput Femtosecond Laser Processing and Tandem Neural Networks

Abstract This work demonstrates a method to design photonic surfaces by combining femtosecond laser processing with the inverse design capabilities of tandem neural networks that directly link laser fabrication parameters to their resulting textured substrate optical properties. High throughput fabrication and characterization platforms are developed that generate a dataset comprising 35280 unique microtextured surfaces on stainless steel with corresponding measured spectral emissivities. The trained model utilizes the nonlinear one‐to‐many mapping between spectral emissivity and laser parameters. Consequently, it generates predominantly novel designs, which reproduce the full range of spectral emissivities (average root‐mean‐squared‐error < 2.5%) using only a compact region of laser parameter space 25 times smaller than what is represented in the training data. Finally, the inverse design model is experimentally validated on a thermophotovoltaic emitter design application. By synergizing laser‐matter interactions with neural network capabilities, the approach offers insights into accelerating the discovery of photonic surfaces, advancing energy harvesting technologies.

36 MATERIALS SCIENCE↗

Probabilistic neural networks for fluid flow surrogate modeling and data recovery

We consider the use of probabilistic neural networks for fluid flow surrogate modeling and data recovery. This framework is constructed by assuming that the target variables are sampled from a Gaussian distribution conditioned on the inputs. Consequently, the overall formulation sets up a procedure to predict the hyperparameters of this distribution which are then used to compute an objective function given training data. We demonstrate that this framework has the ability to provide for prediction confidence intervals based on the assumption of a probabilistic posterior, given an appropriate model architecture and adequate training data. The applicability of the present framework to cases with noisy measurements and limited observations is also assessed. To demonstrate the capabilities of this framework, we consider canonical regression problems of fluid dynamics from the viewpoint of reduced-order modeling and spatial data recovery for four canonical data sets. The examples considered in this study arise from (i) the shallow-water equations, (ii) a two-dimensional cylinder flow, (iii) the wake of a NACA0012 airfoil with a Gurney flap, and (iv) the NOAA sea surface temperature data set. Furthermore, the present results indicate that the probabilistic neural network not only produces a machine-learning-based fluid flow surrogate model but also systematically quantifies the uncertainty therein to assist with model interpretability.

42 ENGINEERING↗

A neural network for the identification of measured helicopter noise

The results of a preliminary study of the components of a novel acoustic helicopter identification system are described. The identification system uses the relationship between the amplitudes of the first eight harmonics in the main rotor noise spectrum to distinguish between helicopter types. Two classification algorithms are tested; a statistically optimal Bayes classifier, and a neural network adaptive classifier. The performance of these classifiers is tested using measured noise of three helicopters. The statistical classifier can correctly identify the helicopter an average of 67 percent of the time, while the neural network is correct an average of 65 percent of the time. These results indicate the need for additional study of the envelope of harmonic amplitudes as a component of a helicopter identification system. Issues concerning the implementation of the neural network classifier, such as training time and structure of the network, are discussed.

Cabell, R. H.↗

Classifying Radio Galaxies with the Convolutional Neural Network

We present the application of a deep machine learning technique to classify radio images of extended sources on a morphological basis using convolutional neural networks (CNN). In this study, we have taken the case of the Fanaroff–Riley (FR) class of radio galaxies as well as radio galaxies with bent-tailed morphology. We have used archival data from the Very Large Array (VLA)—Faint Images of the Radio Sky at Twenty Centimeters survey and existing visually classified samples available in the literature to train a neural network for morphological classification of these categories of radio sources. Our training sample size for each of these categories is ∼200 sources, which has been augmented by rotated versions of the same. Our study shows that CNNs can classify images of the FRI and FRII and bent-tailed radio galaxies with high accuracy (maximum precision at 95%) using well-defined samples and a “fusion classifier,” which combines the results of binary classifications, while allowing for a mechanism to find sources with unusual morphologies. The individual precision is highest for bent-tailed radio galaxies at 95% and is 91% and 75% for the FRI and FRII classes, respectively, whereas the recall is highest for FRI and FRIIs at 91% each, while the bent-tailed class has a recall of 79%. These results show that our results are comparable to that of manual classification, while being much faster. Finally, we discuss the computational and data-related challenges associated with the morphological classification of radio galaxies with CNNs.

79 ASTRONOMY AND ASTROPHYSICS↗

SSME Condition Monitoring Using Neural Networks and Plume Spectral Signatures

For a variety of reasons, condition monitoring of the Space Shuttle Main Engine (SSME) has become an important concern for both ground tests and in-flight operation. The complexities of the SSME suggest that active, real-time condition monitoring should be performed to avoid large-scale or catastrophic failure of the engine. In 1986, the SSME became the subject of a plume emission spectroscopy project at NASA's Marshall Space Flight Center (MSFC). Since then, plume emission spectroscopy has recorded many nominal tests and the qualitative spectral features of the SSME plume are now well established. Significant discoveries made with both wide-band and narrow-band plume emission spectroscopy systems led MSFC to develop the Optical Plume Anomaly Detection (OPAD) system. The OPAD system is designed to provide condition monitoring of the SSME during ground-level testing. The operational health of the engine is achieved through the acquisition of spectrally resolved plume emissions and the subsequent identification of abnormal emission levels in the plume indicative of engine erosion or component failure. Eventually, OPAD, or a derivative of the technology, could find its way on to an actual space vehicle and provide in-flight engine condition monitoring. This technology step, however, will require miniaturized hardware capable of processing plume spectral data in real-time. An objective of OPAD condition monitoring is to determine how much of an element is present in the SSME plume. The basic premise is that by knowing the element and its concentration, this could be related back to the health of components within the engine. For example, an abnormal amount of silver in the plume might signify increased wear or deterioration of a particular bearing in the engine. Once an anomaly is identified, the engine could be shut down before catastrophic failure occurs. Currently, element concentrations in the plume are determined iteratively with the help of a non-linear computer code called SPECTRA, developed at the USAF Arnold Engineering Development Center. Ostensibly, the code produces intensity versus wavelength plots (i.e., spectra) when inputs such as element concentrations, reaction temperature, and reaction pressure are provided. However, in order to provide a higher-level analysis, element concentration is not specified explicitly as an input. Instead, two quantum variables, number density and broadening parameter, are used. Past experience with OPAD data analysis has revealed that the region of primary interest in any SSME plume spectrum lies in the wavelength band of 3300 A to 4330 A. Experience has also revealed that some elements, such as iron, cobalt and nickel, cause multiple peaks over the chosen wavelength range whereas other elements (magnesium, for example) have a few, relatively isolated peaks in the chosen wavelength range. Iteration with SPECTRA as a part of OPAD data analysis is an incredibly labor intensive task and not one to be performed by hand. What is really needed is the "inverse" of the computer code but the mathematical model for the inverse mapping is tenuous at best. However, building generalized models based upon known input/output mappings while ignoring details of the governing physical model is possible using neural networks. Thus the objective of the research project described herein was to quickly and accurately predict combustion temperature and element concentrations (i.e., number density and broadening parameter) from a given spectrum using a neural network. In other words, a neural network had to be developed that would provide a generalized "inverse" of the computer code SPECTRA.

Hopkins, Randall↗

Adaptive enhanced sampling by force-biasing using neural networks

A machine learning assisted method is presented for molecular simulation of systems with rugged free energy landscapes. The method is general and can be combined with other advanced sampling techniques. In the particular implementation proposed here, it is illustrated in the context of an adaptive biasing force approach where, rather than relying on discrete force estimates, one can resort to a self-regularizing artificial neural network to generate continuous, estimated generalized forces. By doing so, the proposed approach addresses several shortcomings common to adaptive biasing force and other algorithms. Specifically, the neural network enables smooth estimates of generalized forces in sparsely sampled regions, force estimates in previously unexplored regions, and continuous force estimates with which to bias the simulation, as opposed to biases generated at specific points of a discrete grid. The usefulness of the method is illustrated with three different examples, chosen to highlight the wide range of applicability of the underlying concepts. In all three cases, the new method is found to enhance considerably the underlying traditional adaptive biasing force approach. Furthermore, the method is also found to provide improvements over previous implementations of neural network assisted algorithms.

97 MATHEMATICS AND COMPUTING↗

Estimating Lossy Compressibility of Scientific Data Using Deep Neural Networks

Simulation based scientific applications generate increasingly large amounts of data on high-performance computing (HPC) systems. To allow data to be stored and analyzed efficiently, data compression is often utilized to reduce the volume and velocity of data. However, a question often raised by domain scientists is the level of compression that can be expected so that they can make more informed decisions, balancing between accuracy and performance. In this letter, we propose a deep neural network based approach for estimating the compressibility of scientific data. To train the neural network, we build both general features as well as compressor-specific features so that the characteristics of both data and lossy compressors are captured in training. Our approach is demonstrated to outperform a prior analytical model as well as a sampling based approach in the case of a biased estimation, i.e., for SZ. However, for the unbiased estimation (i.e., ZFP), the sampling based approach yields the best accuracy, despite the high overhead involved in sampling the target dataset.

97 MATHEMATICS AND COMPUTING↗

A neural network based 3D/3D image registration quality evaluator for the head-and-neck patient setup in the absence of a ground truth

Purpose: To develop a neural network based registration quality evaluator (RQE) that can identify unsuccessful 3D/3D image registrations for the head-and-neck patient setup in radiotherapy. Methods: A two-layer feed-forward neural network was used as a RQE to classify 3D/3D rigid registration solutions as successful or unsuccessful based on the features of the similarity surface near the point-of-solution. The supervised training and test data sets were generated by rigidly registering daily cone-beam CTs to the treatment planning fan-beam CTs of six patients with head-and-neck tumors. Two different similarity metrics (mutual information and mean-squared intensity difference) and two different types of image content (entire image versus bony landmarks) were used. The best solution for each registration pair was selected from 50 optimizing attempts that differed only by the initial transformation parameters. The distance from each individual solution to the best solution in the normalized parametrical space was compared to a user-defined error threshold to determine whether that solution was successful or not. The supervised training was then used to train the RQE. The performance of the RQE was evaluated using the test data set that consisted of registration results that were not used in training. Results: The RQE constructed using the mutual information had very good performance when tested using the test data sets, yielding the sensitivity, the specificity, the positive predictive value, and the negative predictive value in the ranges of 0.960-1.000, 0.993-1.000, 0.983-1.000, and 0.909-1.000, respectively. Adding a RQE into a conventional 3D/3D image registration system incurs only about 10%-20% increase of the overall processing time. Conclusions: The authors' patient study has demonstrated very good performance of the proposed RQE when used with the mutual information in identifying unsuccessful 3D/3D registrations for daily patient setup. The classifier had very good generality and required only to be trained once for each implementation. When the RQE is incorporated with an automated 3D/3D image registration system, it can improve the robustness of the system.

60 APPLIED LIFE SCIENCES↗