Search

Conventional neural networks are universal function approximators, but because they are unaware of underlying symmetries or physical laws, they may need impractically many training data to approximate nonlinear dynamics. Recently introduced Hamiltonian neural networks can efficiently learn and forecast dynamical systems that conserve energy, but they require special inputs called canonical coordinates, which may be hard to infer from data. Here we significantly expand the scope of such networks by demonstrating a simple way to train them with any set of generalised coordinates, including easily observable ones., Comment: 7 pages, 4 figures

We detail how incorporating physics into neural network design can significantly improve the learning and forecasting of dynamical systems, even nonlinear systems of many dimensions. A map building perspective elucidates the superiority of Hamiltonian neural networks over conventional neural networks. The results clarify the critical relation between data, dimension, and neural network learning performance., Comment: 7 pages, 9 figures

Conventional artificial neural networks are powerful tools in science and industry, but they can fail when applied to nonlinear systems where order and chaos coexist. We use neural networks that incorporate the structures and symmetries of Hamiltonian dynamics to predict phase space trajectories even as nonlinear systems transition from order to chaos. We demonstrate Hamiltonian neural networks on the canonical Henon-Heiles system, which models diverse dynamics from astrophysics to chemistry. The power of the technique and the ubiquity of chaos suggest widespread utility., Comment: 5 pages and 5 figures

A finite element formulation is developed for a poroelastic medium consisting of an incompressible hyperelastic skeleton saturated by an incompressible fluid. The governing equations stem from mixture theory and the application is motivated by the study of interstitial fluid flow in brain tissue. The formulation is based on the adoption of an ALE perspective. We focus on a flow regime in which inertia forces are negligible. The stability and convergence of the formulation is discussed, and numerical results demonstrate agreement with the theory.

In the spirit of historic astronomical endeavors, we invited school groups across the globe to collaborate in a solar distance measurement using the rare June 5/6th transit of Venus. In total, we recruited 19 school groups spread over 6 continents and 10 countries to participate in our Hetu'u Global Network. Applying the methods of French astronomer Joseph-Nicolas Delisle, we used individual second and third Venus-Sun contact times to calculate the distance to the Sun. Ten of the sites in our network had amiable weather; 8 of which measured second contact and 5 of which measured third contact leading to consistent solar distance measurements of 152+/-30 million km and 163+/-30 million km respectively. The distance to the Sun at the time of the transit was 152.25 million km; therefore, our measurements are also consistent within 1sigma of the known value. The goal of our international school group network was to inspire the next generation of scientists using the excitement and accessibility of a rare astronomical event. In the process, we connected hundreds of participating students representing a diverse, multi-cultural group with differing political, economic, and racial backgrounds., Comment: 19 pages; 7 Figures; 1 Table; Accepted for publication in Astronomy Education Review (AER) For more information see http://www.das.uchile.cl/~drodrigu/easter/index_en.html

The extraplanar diffuse ionized gas (eDIG) in 17 nearby, edge-on disk galaxies is studied using deep Taurus Tunable Filter Halpha and [N II] lambda 6583 images and conventional interference filter Halpha + [N II] lambda 6548, 6583 images which reach flux levels generally below about 1 x 10^{-17} erg s^{-1} cm^{-2} arcsec^{-2}. [N II] lambda 6583/Halpha excitation maps are available for 10 of these objects. All but one galaxy in the sample exhibit eDIG. The contribution of the eDIG to the total Halpha luminosity is relatively constant, of order 12 +/- 4%. The Halpha scale height of the eDIG derived from a two-exponential fit to the vertical emission profile ranges from 0.4 kpc to 17.9 kpc, with an average of 4.3 kpc. This average value is noticeably larger than the eDIG scale height measured in our Galaxy and other galaxies. This difference in scale height is probably due in part to the lower flux limits of our observations. A quantitative analysis of the topology of the eDIG confirms that several galaxies in the sample have a highly structured eDIG morphology. The distribution of the eDIG emission is often correlated with the locations of the H II regions in the disk. A strong correlation is found between the star formation rate (IR luminosity) per unit disk area and the extraplanar ionized mass, providing further support for a strong connection between the disk and eDIG components in these galaxies. The excitation maps confirm that the [N II]/Halpha ratios are systematically higher in the eDIG than in the disk. More detail is given in Paper II. (abridged), Comment: 34 pages + 17 jpg figures. Accepted for publication in the Astrophysical Journal Supplements, Vol. 148, October 2003 issue

The results from deep long-slit spectroscopy of nine edge-on spiral galaxies with known extraplanar line emission are reported. Emission from Halpha, [N II] lambda 6548, 6583, and [S II] lambda 6716, 6731 is detected out to heights of a few kpc in all of these galaxies. Several other fainter diagnostic lines such as [O I] lambda 6300, [O III] lambda 4959, 5007, and He I lambda 5876 are also detected over a smaller scale. The relative strengths, centroids and widths of the various emission lines provide constraints on the electron density, temperature, reddening, source(s) of ionization, and kinematics of the extraplanar gas. In all but one galaxy, photoionization by massive OB stars alone has difficulties explaining all of the line ratios in the extraplanar gas. Hybrid models that combine photoionization by OB stars and another source of ionization such as photoionization by turbulent mixing layers or shocks provide a better fit to the data. The (upper limits on the) velocity gradients measured in these galaxies are consistent with the predictions of the galactic fountain model to within the accuracy of the measurements., Comment: 25 pages + several jpg figures. Accepted for publication in the Astrophysical Journal, Vol. 592, July 20 2003

The NASA Engineering Safety Center (NESC) Systems Engineering (SE) Technical Discipline Team (TDT) initiated the Model Based Systems Engineering (MBSE) Pathfinder effort in FY16. The goals and objectives of the MBSE Pathfinder include developing and advancing MBSE capability across NASA, applying MBSE to real NASA issues, and capturing issues and opportunities surrounding MBSE. The Pathfinder effort consisted of four teams, with each team addressing a particular focus area. This paper focuses on Pathfinder team 1 with the focus area of architectures and mission campaigns. These efforts covered the timeframe of February 2016 through September 2016. The team was comprised of eight team members from seven NASA Centers (Glenn Research Center, Langley Research Center, Ames Research Center, Goddard Space Flight Center IV&V Facility, Johnson Space Center, Marshall Space Flight Center, and Stennis Space Center). Collectively, the team had varying levels of knowledge, skills and expertise in systems engineering and MBSE. The team applied their existing and newly acquired system modeling knowledge and expertise to develop modeling products for a campaign (Program) of crew and cargo missions (Projects) to establish a human presence on Mars utilizing In-Situ Resource Utilization (ISRU). Pathfinder team 1 developed a subset of modeling products that are required for a Program System Requirement Review (SRR)/System Design Review (SDR) and Project Mission Concept Review (MCR)/SRR as defined in NASA Procedural Requirements. Additionally, Team 1 was able to perform and demonstrate some trades and constraint analyses. At the end of these efforts, over twenty lessons learned and recommended next steps have been identified.