Journal of Research of the National Institute of Standards and Technology最新文献

Disinfection of surfaces by ultraviolet-C (UV-C) radiation is gaining importance in diverse applications. However, there is generally no accepted computational procedure to determine the minimum irradiation times and UV-C doses required for reliable and secure disinfection of surfaces. UV-C dose distributions must be comparable for devices presently on the market and future ones, as well as for the diverse surfaces of objects to be disinfected. A mathematical model is presented to estimate irradiance distributions. To this end, the relevant parameters are defined. These parameters are the optical properties of the UV-C light sources, such as wavelength and emitted optical power, as well as electrical features, like radiant efficiency and consumed power. Furthermore, the characteristics and geometry of the irradiated surfaces as well as the positions of the irradiated surfaces in relation to the UV-C light sources are considered. Because mercury (Hg) lamps are competitive with UV-C light-emitting diodes, a comparative analysis between these two light sources based on the simulation results is also discussed.

Data concerning the sensitivity of various organisms to ultraviolet (UV) radiation exposure are very important in the design of UV disinfection equipment. This review analyzes fluence data from almost 250 studies and organizes the data into a set of recommended fluence values for specific log reductions and an appendix containing all the collected data. This article was sponsored by Dianne L. Poster, Material Measurement Laboratory, and C. Cameron Miller, Physical Measurement Laboratory, National Institute of Standards and Technology (NIST). It is published in collaboration with the International Ultraviolet Association as a complement to the NIST Workshop on Ultraviolet Disinfection Technologies, 14-15 January 2020, Gaithersburg, MD. The views expressed represent those of the authors and not necessarily those of NIST.

Ultraviolet (UV) radiation in the wavelength range 200 nm ≤ λ ≤ 320 nm, which includes both the UV-C and UV-B portions of the spectrum, is known to be effective for inactivation of a wide range of microbial pathogens, including viruses. Previous research has indicated UV-C radiation to be effective for inactivation of severe acute respiratory syndrome coronavirus (SARS-CoV), the virus that caused an outbreak of SARS in 2003. Given the structural similarities of SARS-CoV and SARS-CoV-2, the cause of coronavirus disease 2019 (COVID-19), it is anticipated that UV radiation should be effective for inactivation of SARS-CoV-2 too. Recently published data support this assertion, but only for a narrow set of exposure and matrix conditions. Models based on genomic and other characteristics of viruses have been developed to provide predictions of viral inactivation responses to UV exposure at λ = 254 nm. The predictions of these models are consistent with reported measurements of viral inactivation, including for SARS-CoV-2. As such, current information indicates that UV-C irradiation should be effective for control of SARS-CoV-2, as well as for control of other coronaviruses; however, additional research is needed to quantify the effects of several important process variables, including the wavelength of radiation, the effects of relative humidity on airborne and surface-associated viruses, and the effects of the medium of exposure.

This paper describes advances in measuring the characteristic spatial distribution of surface temperature and emissivity during laser-metal interaction under conditions relevant for laser powder bed fusion (LPBF) additive manufacturing processes. Detailed descriptions of the measurement process, results, and approaches to determining uncertainties are provided. Measurement uncertainties have complex dependencies on multiple process parameters, so the methodology is demonstrated on one set of process parameters and one material. Well-established literature values for high-purity nickel solidification temperature and emissivity at the solidification temperature were used to evaluate the predicted uncertainty of the measurements. The standard temperature measurement uncertainty is found to be approximately 0.9% of the absolute temperature (16 AC), and the standard relative emissivity measurement uncertainty is found to be approximately 8% at the solidification point of high-purity nickel, both of which are satisfactory. This paper also outlines several potential sources of test uncertainties, which may require additional experimental evaluation. The largest of these are the metal vapor and ejecta that are produced as process by-products, which can potentially affect the imaging quality, reflectometry results, and thermal signature of the process, while also affecting the process of laser power delivery. Furthermore, the current paper focuses strictly on the uncertainties of the emissivity and temperature measurement approach and therefore does not detail a variety of uncertainties associated with experimental controls that must be evaluated for future generation of reference data.

To calibrate laser power and energy meters, the National Institute of Standards and Technology (NIST) uses several detector-based realizations of the scale for optical radiant flux; these realizations are appropriate for specific laser power/energy ranges and optical coupling configurations. Calibrations from 1 µW to 2 W are currently based upon calorimeters. Validation by comparisons against other primary representations of the optical watt over the last two decades suggests the instruments operate well within their typical reported uncertainty level of 0.86 % with 95 % confidence. The dominant uncertainty contribution in the instrument is attributable to light scattered by the legacy window, which was not previously recognized. The inherent electro-optical inequivalence in the calorimeter's response was reassessed by thermal modeling to be 0.03 %. The principal contributions to the overall inequivalence were corrected, yielding a shift in scale representation under 0.2 % for typical calibrations. With updates in several uncertainty contributions resulting from this reassessment, the resulting combined expanded uncertainty (k = 2) is 0.84 %, which is essentially unchanged from the previous result provided to calibration customers.

We address the problem of performing exact (tiling-error free) out-of-core semantic segmentation inference of arbitrarily large images using fully convolutional neural networks (FCN). FCN models have the property that once a model is trained, it can be applied on arbitrarily sized images, although it is still constrained by the available GPU memory. This work is motivated by overcoming the GPU memory size constraint without numerically impacting the final result. Our approach is to select a tile size that will fit into GPU memory with a halo border of half the network receptive field. Next, stride across the image by that tile size without the halo. The input tile halos will overlap, while the output tiles join exactly at the seams. Such an approach enables inference to be performed on whole slide microscopy images, such as those generated by a slide scanner. The novelty of this work is in documenting the formulas for determining tile size and stride and then validating them on U-Net and FC-DenseNet architectures. In addition, we quantify the errors due to tiling configurations which do not satisfy the constraints, and we explore the use of architecture effective receptive fields to estimate the tiling parameters.

A generative adversarial network (GAN) is an artifcial neural network with a distinctive training architecture, designed to create examples that faithfully reproduce a target distribution. GANs have recently had particular success in applications involving high-dimensional distributions in areas such as image processing. Little work has been reported for low dimensions, where properties of GANs may be better identifed and understood. We studied GAN performance in simulated low-dimensional settings, allowing us to transparently assess effects of target distribution complexity and training data sample size on GAN performance in a simple experiment. This experiment revealed two important forms of GAN error, tail underflling and bridge bias, where the latter is analogous to the tunneling observed in high-dimensional GANs.

We discuss the measurement of aggregate levels of encounters in a population, a concept we call encounter metrics. Encounter metrics are designed so that they can be deployed while preserving the privacy of individuals. To this end, encounters are labeled with a random number that cannot be linked to anything that is broadcast at the time of the encounter. Among the applications of encounter metrics is privacy-preserving exposure notifcation, a system that allows people to obtain a measure of their risk due to past encounters with people who have self-reported to be positive with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-19), the cause of coronavirus disease 2019 (COVID-19). The precise engineering of a system for exposure notifcation should be targeted to particular environments. We outline a system for use in the context of a workplace such as the National Institute of Standards and Technology (NIST).