Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Detect And Avoid”

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 487 records · Page 27

An Underlay Communication Channel for 5G Cognitive Mesh Networks: Packet Design, Implementation, Analysis, and Experimental Results

This paper proposes and presents the design and implementation of an underlay communication channel (UCC) for 5G cognitive mesh networks. The UCC builds its waveform based on filter bank multicarrier spread spectrum (FB-MCSS) signaling. The use of this novel spread spectrum signaling allows the device-to-device (D2D) user equipments (UEs) to communicate at a level well below noise temperature and hence, minimize taxation on macro-cell/small-cell base stations and their UEs in 5G wireless systems. Moreover, the use of filter banks allows us to avoid those portions of the spectrum that are in use by macro-cell and small-cell users. Hence, both D2D-to-cellular and cellular-to-D2D interference will be very close to none. We propose a specific packet for UCC and develop algorithms for packet detection, timing acquisition and tracking, as well as channel estimation and equalization. We also present the detail of an implementation of the proposed transceiver on a software radio platform and compare our experimental results with those from a theoretical analysis of our packet detection algorithm.

cognitive networks↗

Characterization Techniques Investigated for Characterization of Anomalous Materials in Plutonium Oxide – 25666

The DOE has adopted a Dilute and Dispose approach for processing surplus plutonium which consists of blending plutonium oxide with adulterants and packaging it in a form which is acceptable for disposal at the Waste Isolation Pilot Plant. The feed material for the downblend process is intended to be pure plutonium oxide powder, however other objects are occasionally encountered within the oxide, particularly with legacy material. Potential technologies which could assist in the resolution of incidents where such anomalies are encountered have been investigated. Resolution of these incidents requires characterization of the anomalous object so that plutonium oxide processing can resume and so that a disposition pathway can be determined for the anomalous material. Technologies including gamma ray spectroscopy, alpha-gamma coincidence, LiDAR volumetric measurements, and surface conductivity measurements were investigated for this purpose, with a focus on systems which can be easily introduced into the processing environment when an anomaly is encountered, then removed from the environment after resolution to avoid impeding normal processing activities. This places an emphasis on small and portable measurement systems and systems that can operate in a high-background, oxide-processing environment. Gamma ray systems investigated include the GR1™a CZT detector and the MicroGe™a germanium detector, with a focus on detecting characteristic gamma rays from Pu-239 and other actinides. A gamma ray and alpha particle coincidence method was investigated with the goal of identifying actinides in the presence of a gamma ray background produced by adjacent plutonium oxide material. Leica™b BLK360 G1 and Keyence™c LJ-X8300 LiDAR systems were investigated for use in conjunction with mass measurements to gain accurate material density values, and a Foerster Sigmatest™d 2.070 was tested to determine surface conductivity. The combination of these properties would allow improved identification and characterization of a wide variety of potential anomalous material.

Munson, Justin M.↗

Theoretical UV absorption spectra of hydrodynamically escaping O{sub 2}/CO{sub 2}-rich exoplanetary atmospheres

Characterizing Earth- and Venus-like exoplanets' atmospheres to determine if they are habitable and how they are evolving (e.g., equilibrium or strong erosion) is a challenge. For that endeavor, a key element is the retrieval of the exospheric temperature, which is a marker of some of the processes occurring in the lower layers and controls a large part of the atmospheric escape. We describe a method to determine the exospheric temperature of an O{sub 2}- and/or CO{sub 2}-rich transiting exoplanet, and we simulate the respective spectra of such a planet in hydrostatic equilibrium and hydrodynamic escape. The observation of hydrodynamically escaping atmospheres in young planets may help constrain and improve our understanding of the evolution of the solar system's terrestrial planets' atmospheres. We use the dependency of the absorption spectra of the O{sub 2} and CO{sub 2} molecules on the temperature to estimate the temperature independently of the total absorption of the planet. Combining two observables (two parts of the UV spectra that have a different temperature dependency) with the model, we are able to determine the thermospheric density profile and temperature. If the slope of the density profile is inconsistent with the temperature, then we infer the hydrodynamic escape. We address the question of the possible biases in the application of the method to future observations, and we show that the flare activity should be cautiously monitored to avoid large biases.

79 ASTRONOMY AND ASTROPHYSICS↗

A BAYESIAN ESTIMATE OF THE CMB–LARGE-SCALE STRUCTURE CROSS-CORRELATION

Evidences for late-time acceleration of the universe are provided by multiple probes, such as Type Ia supernovae, the cosmic microwave background (CMB), and large-scale structure (LSS). In this work, we focus on the integrated Sachs–Wolfe (ISW) effect, i.e., secondary CMB fluctuations generated by evolving gravitational potentials due to the transition between, e.g., the matter and dark energy (DE) dominated phases. Therefore, assuming a flat universe, DE properties can be inferred from ISW detections. We present a Bayesian approach to compute the CMB–LSS cross-correlation signal. The method is based on the estimate of the likelihood for measuring a combined set consisting of a CMB temperature and galaxy contrast maps, provided that we have some information on the statistical properties of the fluctuations affecting these maps. The likelihood is estimated by a sampling algorithm, therefore avoiding the computationally demanding techniques of direct evaluation in either pixel or harmonic space. As local tracers of the matter distribution at large scales, we used the Two Micron All Sky Survey galaxy catalog and, for the CMB temperature fluctuations, the ninth-year data release of the Wilkinson Microwave Anisotropy Probe ( WMAP 9). The results show a dominance of cosmic variance over the weak recovered signal, due mainly to the shallowness of the catalog used, with systematics associated with the sampling algorithm playing a secondary role as sources of uncertainty. When combined with other complementary probes, the method presented in this paper is expected to be a useful tool to late-time acceleration studies in cosmology.

79 ASTRONOMY AND ASTROPHYSICS↗

Microstrain and Grain-Size Analysis From Diffraction Peak Width and Graphical Derivation of High-Pressure Thermomechanics

An analytical method is presented for deriving the thermomechanical properties of polycrystalline materials under high-pressure (P) and high-temperature (T) conditions. This method deals with non-uniform stress among heterogeneous crystal grains and surface strain in nanocrystalline materials by examining peak-width variation under different P-T conditions. Because the method deals directly with lattice d spacing and local deformation caused by stress, it can be applied to process any diffraction profile, independent of detection mode. In addition, a correction routine is developed using diffraction elastic ratios to deal with severe surface strain and/or strain anisotropy effects related to nano-scale grain sizes, so that significant data scatter can be reduced in a physically meaningful way. Graphical illustration of the resultant microstrain analysis can identify micro/local yields at the grain-to-grain interactions resulting from high stress concentration, and macro/bulk yield of the plastic deformation over the entire sample. This simple and straightforward approach is capable of revealing the corresponding micro and/or macro yield stresses, grain crushing or growth, work hardening or softening, and thermal relaxation under high-P-T conditions, as well as the intrinsic residual strain and/or surface strain in the polycrystalline bulk. In addition, this approach allows the instrumental contribution to be illustrated and subtracted in a straightforward manner, thus avoiding the potential complexities and errors resulting from instrument correction. Applications of the method are demonstrated by studies of [alpha]-SiC (6H, moissanite) and of micro- and nanocrystalline nickel by synchrotron X-ray and time-of-flight neutron diffraction.

43 PARTICLE ACCELERATORS↗

Light, alpha, and Fe-peak element abundances in the galactic bulge

We present radial velocities and chemical abundances of O, Na, Mg, Al, Si, Ca, Cr, Fe, Co, Ni, and Cu for a sample of 156 red giant branch stars in two Galactic bulge fields centered near (l, b) = (+5.25,–3.02) and (0,–12). The (+5.25,–3.02) field also includes observations of the bulge globular cluster NGC 6553. The results are based on high-resolution (R ∼ 20,000), high signal-to-noise ration (S/N ≳ 70) FLAMES-GIRAFFE spectra obtained through the European Southern Observatory archive. However, we only selected a subset of the original observations that included spectra with both high S/N and that did not show strong TiO absorption bands. This work extends previous analyses of this data set beyond Fe and the α-elements Mg, Si, Ca, and Ti. While we find reasonable agreement with past work, the data presented here indicate that the bulge may exhibit a different chemical composition than the local thick disk, especially at [Fe/H] ≳ –0.5. In particular, the bulge [α/Fe] ratios may remain enhanced to a slightly higher [Fe/H] than the thick disk, and the Fe-peak elements Co, Ni, and Cu appear enhanced compared to the disk. There is also some evidence that the [Na/Fe] (but not [Al/Fe]) trends between the bulge and local disk may be different at low and high metallicity. We also find that the velocity dispersion decreases as a function of increasing [Fe/H] for both fields, and do not detect any significant cold, high-velocity populations. A comparison with chemical enrichment models indicates that a significant fraction of hypernovae may be required to explain the bulge abundance trends, and that initial mass functions that are steep, top-heavy (and do not include strong outflow), or truncated to avoid including contributions from stars >40 M {sub ☉} are ruled out, in particular because of disagreement with the Fe-peak abundance data. For most elements, the NGC 6553 stars exhibit abundance trends nearly identical to comparable metallicity bulge field stars. However, the star-to-star scatter and mean [Na/Fe] ratios appear higher in the cluster, perhaps indicating additional self-enrichment.

79 ASTRONOMY AND ASTROPHYSICS↗

Seamless Wireless Communication Platform for Internet of Things Applications

The rapid growth of the Internet of Things (IoT) devices resulted in the proliferation of wireless technologies to cater to their increasing data rate requirements and support multiple applications. However, such ever-increasing wireless technologies present numerous challenges such as incompatible wireless standards, increased energy consumption, and insecure communication. The traditional gateways proposed in the literature suffers from limitations such as computational complexity, resource requirements, increased cost, and device size. We vision an era of seamless wireless communication to alleviate the aforementioned challenges in IoT applications. through three inter-dependent functionalities namely detection and identification of wireless technologies, energy-efficient transmit power control, and secure end-to-end communication. To prove the concept, a new gateway is proposed to achieve these three functionalities with only physical layer measurements so that the different communication protocols in the higher layers can be avoided. Novel schemes are conceptualized for resource-limited seamless IoT applications. Moreover, the conceptual seamless IoT platform is validated through software-based computer simulation and software-defined radio-based testbed implementation. Finally, the preliminary analysis demonstrates that the proposed platform has great potential in advancing seamless IoT applications.

97 MATHEMATICS AND COMPUTING↗

Production regimes for Self-Interacting Dark Matter

In the context of Self-Interacting Dark Matter as a solution for the small-scale structure problems, we consider the possibility that Dark Matter could have been produced without being in thermal equilibrium with the Standard Model bath. We discuss one by one the following various dark matter production regimes of this kind: freeze-in, reannihilation and dark freeze-out. We exemplify how these mechanisms work in the context of the particularly simple Hidden Vector Dark Matter model. In contrast to scenarios where there is thermal equilibrium with the Standard Model bath, we find two regimes which can easily satisfy all the laboratory and cosmological constraints. These are dark freeze-out with 3-to-2 annihilations and freeze-in via a light mediator. In the first regime, different temperatures in the visible and the Dark Matter sectors allow us to avoid the constraints coming from cosmic structure formation as well as the use of non-perturbative couplings to reproduce the observed relic density. For the second regime, different couplings are responsible for Dark Matter relic density and self-interactions, permitting to surpass BBN, X-ray, CMB and direct detection constraints.

79 ASTRONOMY AND ASTROPHYSICS↗

The Use of Long-wave Infrared Cameras for Hazardous Waste Remediation - 20457

For over fifty years, infrared cameras have been used in military applications, nondestructive testing, condition monitoring, and predictive maintenance. As infrared cameras continue to become more sophisticated and less expensive, they are providing value in an ever-increasing variety of unique applications, including hazardous waste remediation. This paper describes several instances of how infrared cameras have been used within the US Department of Defense and the US Department of Energy in support of waste remediation projects, including the author's recent use of an infrared camera in support of the Calcine Retrieval Project at the Idaho National Laboratory. Infrared cameras provide images of infrared radiation, or heat energy, which is otherwise invisible to the unaided eye. Infrared radiation is part of the electromagnetic spectrum, which includes visible light. But unlike visible light, infrared has wavelengths longer than the human eye can detect. Infrared is emitted by everything with a temperature above absolute zero (-273 deg. C, or -459 deg. F); the higher the temperature, the greater the infrared thermal radiation, or heat, that is emitted. Even objects that feel cold to us, like ice, emit thermal radiation and can be imaged by infrared cameras. These cameras are typically used to look for abnormally hot or cold spots on a component or target area under normal operating conditions. The method provides a rapid, wide-area, noncontact technique for identifying problems associated with a temperature differential. All infrared cameras can provide qualitative thermal information by displaying relative differences in temperatures within a two-dimensional image. More expensive infrared cameras can also provide quantitative information where an absolute temperature value is assigned to each pixel associated with the displayed two-dimensional image. Proper camera calibration and a solid understanding of heat transfer and thermography techniques are required when using an infrared camera to obtain quantitative information. Case studies outlined in this paper include the rapid, non-intrusive detection of hazardous decontamination solution within one-ton shipping containers at Pine Bluff Arsenal, the non-intrusive identification of residual elemental sodium within the cooling loops of the Experimental Breeder Reactor II (EBR-II) reactor at the Idaho National Laboratory, process monitoring of heat exchanger melt-and-drain efforts during EBR-II decommissioning, and the remote detection of internal steel supports within calcine storage bins prior to bin penetration. For each use of thermography, the author describes a summary of the waste remediation effort, the infrared camera used, the thermal imaging technique employed, and the results obtained. The paper concludes with a discussion on common mistakes to avoid for similar applications of thermography. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Measuring the Density of Different Materials by Using the Collimated Fast Neutron Beam

It was demonstrated in the previous work that various threat materials could be detected inside the sea going cargo container by measuring the three variables, carbon and oxygen concentration and density of investigated material. Density was determined by measuring transmitted neutrons, which is not always practical in terms of setting up the instrument geometry. In order to enable more geometry flexibility, we have investigated the possibility of using the scattered neutrons in cargo material identification. For that purpose, the densities of different materials were measured depending on the position of neutron detectors and neutron generator with respect to the target position. One neutron detector was put above the target, one behind and one in front of the target, above the neutron generator. It was shown that all three positions of neutron detectors can be successfully used to measure the target density, but only if the detected neutrons are successfully discriminated from the gamma rays. Although the associated alpha particle technique/associate particle imaging (API) was used to discriminate the neutrons from the gamma rays, it is believed that the same results would be obtained by using the pulse shape discrimination method. In that way API technique can be avoided and the neutron generator which produces much higher beam intensity than 10{sup 8} n/s can be used. (authors)

36 MATERIALS SCIENCE↗

Scalable Tensor Methods for Nonuniform Hypergraphs

While multilinear algebra appears natural for studying the multiway interactions modeled by hypergraphs, tensor methods for general hypergraphs have been stymied by theoretical and practical barriers. A recently proposed adjacency tensor is applicable to nonuniform hypergraphs, but is prohibitively costly to form and analyze in practice. We develop tensor times same vector (TTSV) algorithms for this tensor which improve complexity from $O(n^r)$ to a low-degree polynomial in $r$, where $n$ is the number of vertices and $r$ is the maximum hyperedge size. Our algorithms are implicit, avoiding formation of the order $r$ adjacency tensor. Here, we demonstrate the flexibility and utility of our approach in practice by developing tensor-based hypergraph centrality and clustering algorithms. We also show these tensor measures offer complementary information to analogous graph-reduction approaches on data, and are also able to detect higher-order structure that many existing matrix-based approaches provably cannot.

97 MATHEMATICS AND COMPUTING↗

Safeguards Modeling for Advanced Nuclear Facility Design.

Future nuclear fuel cycle facilities will see a significant benefit from considering materials accountancy requirements early in the design process. The Material Protection, Accounting, and Control Technologies (MPACT) working group is demonstrating Safeguards and Security by Design (SSBD) for a notional electrochemical reprocessing facility as part of a 2020 Milestone. The idea behind SSBD is to consider regulatory requirements early in the design process to provide more optimized systems and avoid costly retrofits later in the design process. Safeguards modeling, using single analyst tools, allows the designer to efficiently consider materials accountancy approaches that meet regulatory requirements. However, safeguards modeling also allows the facility designer to go beyond current regulations and work toward accountancy designs with rapid response and lower thresholds for detection of anomalies. This type of modeling enables new safeguards approaches and may inform future regulatory changes. The Separation and Safeguards Performance Model (SSPM) has been used for materials accountancy system design and analysis. This paper steps through the process of designing a Material Control and Accountancy (MC&A) system, presents the baseline system design for an electrochemical reprocessing facility, and provides performance metrics from the modeling analysis. The most critical measurements in the electrochemical facility are the spent fuel input, electrorefiner salt, and U/TRU product output measurements. Finally, material loss scenario analysis found that measurement uncertainties (relative standard deviations) for Pu would need to be at 1% (random and systematic error components) or better in order to meet domestic detection goals or as high as 3% in order to meet international detection goals, based on a 100 metric ton per year plant size.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

Physics-Based Machine Learning Methods for U-235 Forensics Signatures

Signatures of low-intensity U-235 sources have been recently studied by utilizing a variety of machine learning (ML) classifiers using features derived from gamma spectral measurements collectedunder structured campaigns. Several ML classifiers, such as ensemble of tress and classification trees, revealed misleadingly-optimistic training error due to over-fitting, and furthermore,their performance is not directly relatable to the physical properties due to their data-driven, opaque designs. We present a regression-based ML method that first estimates the inverse distanceto the source and then utilizes a threshold to infer its presence, by representing the background as a source located at an infinite distance. For the inverse distance estimation, we study the ensembleof trees and Gaussian process regression methods, and a hyper parameter auto-tuning and selection method that employs five regression estimators. These methods avoid the over-fittingobserved in several ML classifiers, while providing the classification error nearly comparable to them based on independent test data. Their error is directly related to estimates of the inversephysical distance to source, and the precision of error determines the seperability property that determines the false alarm and missed detection rates. The property of monotonic decrease of thesource strength with increasing detector distance combined with Poisson distribution of measurements is utilized to analytically validate these methods by deriving the generalization equations ofunderlying regression methods.

Rao, Nageswara↗

WE-AB-BRB-04: A Novel Monolithic Silicon 2D Detector Array for Use in Stereotactic Applications

Purpose: To assess the capability of a novel 2D monolithic silicon detector array in measuring stereotactic photon fields. Methods: The silicon array detector used in this work, named Magic Plate-512 (MP512), is a thin monolithic silicon wafer (52 × 52 × 0.47 mm{sup 3}) with 512 ion-implanted diodes (0.5 × 0.5 mm{sup 2}). Adjacent pixels are spaced evenly with 2 mm pitch, covering a maximum detection area 46 mm wide. Its fast, FPGA based read-out system is synchronised with the linac to allow readout of all pixels pulse-by-pulse. A clinical SABR lung plan (consisting of 9 single segment beams, 6MV) was measured with the array at 1.5 cm depth in a solid water phantom (100 cm SSD). The typical field size was in the range of 3 × 3 cm{sup 2} to 4 × 4 cm{sup 2}. Each beam was delivered at perpendicular incidence to the detector plane so as to avoid the need for angular dependence corrections. The fields were measured under the same conditions using Gafchromic EBT3 film for comparison. The film was scanned at 72 dpi resolution, with the red channel data used for analysis. Results: Average gamma passing rates of (92.3 ± 1.8) % for 2%/2mm criteria, and (86.6 ± 2.3) % for 1%/2mm criteria were achieved for MP512, using EBT3 film as the reference distribution. The detector array was able to accurately measure the full-width-at-half-maximum (FWHM), to within (0.77 ± 0.01) mm accuracy when compared to film. The penumbral widths (80%-20%) were measured to within (0.30 ± 0.01) mm accuracy to film. Conclusion: The MP512 is a feasible option for measurement of stereotactic photon fields, with its high density of detection points making it useful for small field applications. The prototype array has demonstrated merit; in the future the development of a larger array detection area would be beneficial for clinical applications.

60 APPLIED LIFE SCIENCES↗

Scannerless laser range imaging using loss modulation

A scannerless 3-D imaging apparatus is disclosed which utilizes an amplitude modulated cw light source to illuminate a field of view containing a target of interest. Backscattered light from the target is passed through one or more loss modulators which are modulated at the same frequency as the light source, but with a phase delay .delta. which can be fixed or variable. The backscattered light is demodulated by the loss modulator and detected with a CCD, CMOS or focal plane array (FPA) detector to construct a 3-D image of the target. The scannerless 3-D imaging apparatus, which can operate in the eye-safe wavelength region 1.4-1.7 .mu.m and which can be constructed as a flash LADAR, has applications for vehicle collision avoidance, autonomous rendezvous and docking, robotic vision, industrial inspection and measurement, 3-D cameras, and facial recognition.

47 OTHER INSTRUMENTATION↗

Contrasting Time-Frequency Representations for Unknown Waveform Detection

Identifying unseen electromagnetic waveforms is critical for many applications, like interference management, electronic warfare and spectrum management. Traditionally this is done using statistical methods for anomaly detection, which has evolved to deep learning models for identifying the unseen data, formally termed as open set recognition. Some prior methods use a generative model to emulate open set data, which face challenges in generating synthetic samples for open set while simultaneously selecting an optimal discriminator for accurate classification. To alleviate this issue, we propose a discriminative model that effectively combines time and frequency domain features of communication signals for accurate predictions. We further introduce a cosine similarity loss that makes the domain specific features unique to enhance the prediction rate. Additionally, our model avoids generic feature vectors by extracting class-specific features during training, resulting in improved class representation. The experiment results show that this combined feature approach with cosine loss outperforms single-domain models and improves accuracy by 10% over models without cosine loss.

99 - GENERAL AND MISCELLANEOUS↗

In-situ Detection of Glovebox Glove Degradation Prior to Glove Failure; Effects on Safety, Waste, and Operational Costs - 20402

Glovebox glove pressure decay leak testing and analysis of the data can greatly improve glovebox safety, minimize Transuranic (TRU) glove waste, and minimize operational cost. It is well known, the weakest point of containment on a glovebox is the glove. Mitigating unplanned openings in gloves is critical to minimizing operational and safety costs in glovebox operations. Mitigating unplanned glove openings due to glove failures will be discussed in this paper. The lack of an engineered solution to determine a gloves operational life in nuclear operations, has lead to use of theoretical and manual subjective methods to determine the life and safety of a glovebox glove. Whether the glove is inspected visually prior to use, or gloves are changed regularly to theoretically avoid failure, both the operators PPE and the room are being exposed to contamination, or un-necessary amounts of TRU glove waste are being generated. Millions of dollars yearly in operational costs are spent on glove failures due to loss of production, incident analysis, regulatory audits, clean up, and paperwork. In a nuclear application, a glove failure could contaminate an operator and laboratory, shut down the operation for weeks, potentially create a regulatory audit, and potentially create a media frenzy; the damages to the organization can be astronomical. Elimination of glove failures is possible by tracking glove material degradation and setting limits to allow glove change when it becomes necessary, prior to glove failure. Analytical pressure decay leak testing allows the detection of material degradation, and in turn the ability to limit operational glove failures. A German based company called MK Versuchsanlagen, e.K., has developed an advanced Glove Integrity Testing System. The system, with its use of proprietary RFID and special software technology, is unique in its capabilities. The system is capable to perform highly accurate regular glove testing on any size glovebox line in minimal time. The system records and can analyze a tremendous amount of system data that in turn can determine, for example, a safe glove change time prior to a glove material failure. This paper will show how an advanced Glove Integrity Testing System can detect glove material degradation and help determine a safe change point prior to glove failure. Pressure decay curves of new gloves and recorded effects of accelerated aging on the glove over time will be shown. The use of this technology in nuclear facilities can greatly improve glovebox operational safety, minimize glovebox glove TRU waste, and save organizations tremendous costs in downtime, clean up, documentation, and potential regulatory review. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Multistage mixed precision iterative refinement

Abstract Low precision arithmetic, in particular half precision (16‐bit) floating point arithmetic, is now available in commercial hardware. Using lower precision can offer significant savings in computation and communication costs with proportional savings in energy. Motivated by this, there has been a renewed interest in mixed precision iterative refinement schemes for solving linear systems , and new variants of GMRES‐based iterative refinement have been developed. Each particular variant with a given combination of precisions leads to different condition number‐based constraints for convergence of the backward and forward errors, and each has different performance costs. The constraints for convergence given in the literature are, as an artifact of the analyses, often overly strict in practice, and thus could lead a user to select a more expensive variant when a less expensive one would have sufficed. In this work, we develop a multistage mixed precision iterative refinement solver which aims to combine existing mixed precision approaches to balance performance and accuracy and improve usability. For a user‐specified initial combination of precisions, the algorithm begins with the least expensive approach and convergence is monitored via inexpensive computations with quantities produced during the iteration. If slow convergence or divergence is detected using particular stopping criteria, the algorithm switches to use a more expensive, but more reliable variant. A novel aspect of our approach is that, unlike existing implementations, our algorithm first attempts to use “stronger” GMRES‐based solvers for the solution update before resorting to increasing the precision(s). In some scenarios, this can avoid the need to refactorize the matrix in higher precision. We perform extensive numerical experiments on a variety of random dense problems and problems from real applications which confirm the benefits of the multistage approach.

Oktay, Eda↗