Search NASA⌕ Search

Engineering topics

Ferrier, Nicola J.

Publications and source records attributed to Ferrier, Nicola J..

Optimizing cloud motion estimation on the edge with phase correlation and optical flow

Abstract. Phase correlation (PC) is a well-known method for estimating cloud motion vectors (CMVs) from infrared and visible spectrum images. Commonly, phase shift is computed in the small blocks of the images using the fast Fourier transform. In this study, we investigate the performance and the stability of the blockwise PC method by changing the block size, the frame interval, and combinations of red, green, and blue (RGB) channels from the total sky imager (TSI) at the United States Atmospheric Radiation Measurement user facility's Southern Great Plains site. We find that shorter frame intervals, followed by larger block sizes, are responsible for stable estimates of the CMV, as suggested by the higher autocorrelations. The choice of RGB channels has a limited effect on the quality of CMVs, and the red and the grayscale images are marginally more reliable than the other combinations during rapidly evolving low-level clouds. The stability of CMVs was tested at different image resolutions with an implementation of the optimized algorithm on the Sage cyberinfrastructure test bed. We find that doubling the frame rate outperforms quadrupling the image resolution in achieving CMV stability. The correlations of CMVs with the wind data are significant in the range of 0.38–0.59 with a 95 % confidence interval, despite the uncertainties and limitations of both datasets. A comparison of the PC method with constructed data and the optical flow method suggests that the post-processing of the vector field has a significant effect on the quality of the CMV. The raindrop-contaminated images can be identified by the rotation of the TSI mirror in the motion field. The results of this study are critical to optimizing algorithms for edge-computing sensor systems.

54 ENVIRONMENTAL SCIENCES↗

Prediction of Solar Irradiance and Photovoltaic Solar Energy Product Based on Cloud Coverage Estimation Using Machine Learning Methods

Cloud cover estimation from images taken by sky-facing cameras can be an important input for analyzing current weather conditions and estimating photovoltaic power generation. The constant change in position, shape, and density of clouds, however, makes the development of a robust computational method for cloud cover estimation challenging. Accurately determining the edge of clouds and hence the separation between clouds and clear sky is difficult and often impossible. Toward determining cloud cover for estimating photovoltaic output, we propose using machine learning methods for cloud segmentation. We compare several methods including a classical regression model, deep learning methods, and boosting methods that combine results from the other machine learning models. To train each of the machine learning models with various sky conditions, we supplemented the existing Singapore whole sky imaging segmentation database with hazy and overcast images collected by a camera-equipped Waggle sensor node. We found that the U-Net architecture, one of the deep neural networks we utilized, segmented cloud pixels most accurately. However, the accuracy of segmenting cloud pixels did not guarantee high accuracy of estimating solar irradiance. We confirmed that the cloud cover ratio is directly related to solar irradiance. Additionally, we confirmed that solar irradiance and solar power output are closely related; hence, by predicting solar irradiance, we can estimate solar power output. This study demonstrates that sky-facing cameras with machine learning methods can be used to estimate solar power output. This ground-based approach provides an inexpensive way to understand solar irradiance and estimate production from photovoltaic solar facilities.

14 SOLAR ENERGY↗

Derivation of Multiple Covarying Material and Process Parameters Using Physics-Based Modeling of X-ray Data

There is considerable interest in developing multimodal characterization frameworks capable of probing critical properties of complex materials by relying on distinct, complementary methods or tools. Any such framework should maximize the amount of information that is extracted from any given experiment and should be sufficiently powerful and efficient to enable on-the-fly analysis of multiple measurements in a self-consistent manner. Such a framework is demonstrated in this work in the context of self-assembling polymeric materials, where theory and simulations provide the language to seamlessly mesh experimental data from two different scattering measurements. Specifically, the samples considered here consist of diblock copolymers (BCP) that are self-assembled on chemically nanopatterned surfaces. The copolymers microphase separate into ordered lamellae with characteristic dimensions on the scale of tens of nanometers that are perfectly aligned by the substrate over macroscopic areas. These aligned lamellar samples provide ideal standards with which to develop the formalism introduced in this work and, more generally, the concept of high-information-content, multimodal experimentation. In conclusion, the outcomes of the proposed analysis are then compared to images generated by 3D scanning electron microscopy tomography, serving to validate the merit of the framework and ideas proposed here.

36 MATERIALS SCIENCE↗

Improved deconvolution of very weak confocal signals

Deconvolution is typically used to sharpen fluorescence images, but when the signal-to-noise ratio is low, the primary benefit is reduced noise and a smoother appearance of the fluorescent structures. 3D time-lapse (4D) confocal image sets can be improved by deconvolution. However, when the confocal signals are very weak, the popular Huygens deconvolution software erases fluorescent structures that are clearly visible in the raw data. We find that this problem can be avoided by prefiltering the optical sections with a Gaussian blur. Analysis of real and simulated data indicates that the Gaussian blur prefilter preserves meaningful signals while enabling removal of background noise. This approach is very simple, and it allows Huygens to be used with 4D imaging conditions that minimize photodamage.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Improved deconvolution of very weak confocal signals

Deconvolution is typically used to sharpen fluorescence images, but when the signal-to-noise ratio is low, the primary benefit is reduced noise and a smoother appearance of the fluorescent structures. 3D time-lapse (4D) confocal image sets can be improved by deconvolution. However, when the confocal signals are very weak, the popular Huygens deconvolution software erases fluorescent structures that are clearly visible in the raw data. We find that this problem can be avoided by prefiltering the optical sections with a Gaussian blur. Analysis of real and simulated data indicates that the Gaussian blur prefilter preserves meaningful signals while enabling removal of background noise. This approach is very simple, and it allows Huygens to be used with 4D imaging conditions that minimize photodamage .

Day, Kasey J.↗

Force measurement-based discontinuity detection during friction stir welding

Here, the objective of this work is to develop a method for detecting the creation of discontinuities ( i.e. , voids, volume defects) during friction stir welding. Friction stir welding is inherently cost effective, however, the need for significant weld inspection can make the process cost prohibitive. A new approach to weld inspection is required in which an in situ characterization of weld quality can be obtained, reducing the need for postprocess inspection. To this end, friction stir welds with subsurface voids and without voids were created. The subsurface voids were generated by reducing the friction stir tool rotation frequency and increasing the tool traverse speed in order to create “colder” welds. Process forces were measured during welding, and the void sizes were measured postprocess by computerized tomography ( i.e. , 3D X-ray imaging). Two parameters, based on frequency domain content and time-domain average of the force signals, were found to be correlated with void size. Criteria for subsurface void detection and size prediction were developed and shown to be in good agreement with experimental observations. Furthermore, with the proper choice of data acquisition system and frequency analyzer the occurrence of subsurface voids can be detected in real time.

47 OTHER INSTRUMENTATION↗

Quantitative three-dimensional characterization of block copolymer directed self-assembly on combined chemical and topographical prepatterned templates

Characterization of the three-dimensional (3D) structure in directed self-assembly (DSA) of block copolymers is crucial for understanding the complex relationships between the guiding template and the resulting polymer structure so DSA could be successfully implemented for advanced lithography applications. Here, we combined scanning transmission electron microscopy (STEM) tomography and coarse-grain simulations to probe the 3D structure of P2VP- b -PS- b -P2VP assembled on prepatterned templates using solvent vapor annealing. The templates consisted of nonpreferential background and raised guiding stripes that had PS-preferential top surfaces and P2VP-preferential sidewalls. The full 3D characterization allowed us to quantify the shape of the polymer domains and the interface between domains as a function of depth in the film and template geometry and offered important insights that were not accessible with 2D metrology. Sidewall guiding was advantageous in promoting the alignment and lowering the roughness of the P2VP domains over the sidewalls, but incommensurate confinement from the increased topography could cause roughness and intermittent dislocations in domains over the background region at the bottom of the film. The 3D characterization of bridge structures between domains over the background and breaks within domains on guiding lines sheds light on possible origins of common DSA defects. The positional fluctuations of the PS/P2VP interface between domains showed a depth-dependent behavior, with high levels of fluctuations near both the free surface of the film and the substrate and lower fluctuation levels in the middle of the film. As a result, this research demonstrates how 3D characterization offers a better understanding of DSA processes, leading to better design and fabrication of directing templates.

36 MATERIALS SCIENCE↗

Physics-based process model approach for detecting discontinuity during friction stir welding

The goal of this work is to develop a method for detecting the creation of discontinuities during friction stir welding. This in situ weld monitoring method could significantly reduce the need for post-process inspection. A process force model and a discontinuity force model were created based on the state-of-the-art understanding of flow around an friction stir welding (FSW) tool. These models are used to predict the FSW forces and size of discontinuities formed in the weld. Friction stir welds with discontinuities and welds without discontinuities were created, and the differences in force dynamics were observed. In this paper, discontinuities were generated by reducing the tool rotation frequency and increasing the tool traverse speed in order to create "cold" welds. Experimental force data for welds with discontinuities and welds without discontinuities compared favorably with the predicted forces. The model currently overpredicts the discontinuity size.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗