Uploaded image for project: 'SAFe Program'
  1. SAFe Program
  2. SP-3744

EM Simulations in support of SKA-Low station layout choice

Change Owns to Parent OfsSet start and due date...
    XporterXMLWordPrintable

Details

    • Data Processing
    • Hide

      purposes are:

      1. to complete delivery of electromagnetic simulations of embedded element patterns and station beams for candidate SKA-Low station layouts covering the full frequency range of the instrument at 1 MHz intervals using the HARP software package and
      2. to provide software in HARP and OSKAR to support assessment of station layouts.
      Show
      purposes are: to complete delivery of electromagnetic simulations of embedded element patterns and station beams for candidate SKA-Low station layouts covering the full frequency range of the instrument at 1 MHz intervals using the HARP software package and to provide software in HARP and OSKAR to support assessment of station layouts.
    • Hide

      The outputs are a final report on layout optimisation; HDF files of simulation output covering the full SKA-Low frequency range for the 4 layouts specified above and tested harp_beam and OSKAR code.

      Show
      The outputs are a final report on layout optimisation; HDF files of simulation output covering the full SKA-Low frequency range for the 4 layouts specified above and tested harp_beam and OSKAR code.
    • 5
    • 5
    • 0
    • Team_HIPPO
    • Sprint 5
    • Hide

      The HARP simulations have been run from 50 until 252 MHz (200 frequency points including the previous PI) and not 350 MHz as required. The simulations required are computationally expensive that combined with problems listed below meant that there was not enough time to finish the simulation of all frequencies. 
      There are two main reasons for only partially completing the goals of KR1:
      1. The network switches in the Battcock Center in Cambridge were deficient from September until mid-October, which has caused FEKO licencing errors and corrupted running simulations. These have been replaced with new ones.
      2. The simulation results at several frequencies (~10-15 points) between 190 and 225 MHz were found to be inaccurate - too "ripply" - for the Vogel layout but seemed normal for the pseudo-random layout. An analysis was carried out, and we found out that the accuracy could be recovered by increasing the number of Macro Basis Functions (MBFs) and by varying a parameter in HARP, which controls how these MBFs are generated. Simulations were re-run with this new parametrisation. However, this was done by "trial-and-error". The fundamental cause could be that the underlying mesh used to represent the antenna geometry is too coarse, leading to an ill-conditioned system of equations, or that the set of generated MBFs was not complete enough.
      A script that tabulates the complex values of the electric field (both polarisations) on an azimuth/elevation grid with a separation 0.5 deg in each coordinate has been provided on the OneDrive folder shared.

      The harp_beam library was updated by the end of sprint 2 with the more optimal memory layout, and a speed-up factor of 14 was observed on CSD3 when evaluating the EEPs. An extension of this method to deal with station rotation could provide a similar factor when computing station beams.

      Show
      The HARP simulations have been run from 50 until 252 MHz (200 frequency points including the previous PI) and not 350 MHz as required. The simulations required are computationally expensive that combined with problems listed below meant that there was not enough time to finish the simulation of all frequencies.  There are two main reasons for only partially completing the goals of KR1: 1. The network switches in the Battcock Center in Cambridge were deficient from September until mid-October, which has caused FEKO licencing errors and corrupted running simulations. These have been replaced with new ones. 2. The simulation results at several frequencies (~10-15 points) between 190 and 225 MHz were found to be inaccurate - too "ripply" - for the Vogel layout but seemed normal for the pseudo-random layout. An analysis was carried out, and we found out that the accuracy could be recovered by increasing the number of Macro Basis Functions (MBFs) and by varying a parameter in HARP, which controls how these MBFs are generated. Simulations were re-run with this new parametrisation. However, this was done by "trial-and-error". The fundamental cause could be that the underlying mesh used to represent the antenna geometry is too coarse, leading to an ill-conditioned system of equations, or that the set of generated MBFs was not complete enough. A script that tabulates the complex values of the electric field (both polarisations) on an azimuth/elevation grid with a separation 0.5 deg in each coordinate has been provided on the OneDrive folder shared. The harp_beam library was updated by the end of sprint 2 with the more optimal memory layout, and a speed-up factor of 14 was observed on CSD3 when evaluating the EEPs. An extension of this method to deal with station rotation could provide a similar factor when computing station beams.

    • PI22 - UNCOVERED

    • TEAM_HIPPO

    Description

      This feature is a continuation of SP-3487. Its purposes are:

      1. to complete delivery of electromagnetic simulations of embedded element patterns and station beams for candidate SKA-Low station layouts covering the full frequency range of the instrument at 1 MHz intervals using the HARP software package and
      2. to provide software in HARP and OSKAR to support assessment of station layouts.

      The choice of antenna location within a station has important consequences for the scientific performance of SKA-Low assessed against criteria such as bandpass smoothness, far sidelobe noise and instrumental polarization. Significant problems were identified in the proposed Vogel layout, in particular a sharp dip in the zenith gain at 125 MHz, traced to a coupling effect from redundant spacings. In SP-3487, an optimised layout has been developed which essentially eliminates this problem. It is now necessary to check that this layout has good performance of the entire frequency range. The range 60-160 MHz has already been simulated and 160-200MHz is in progress. The remaining frequencies (200-350MHz with 1 MHz spacing) are now needed for 4 layouts as simulated previously: pseudo-random (LFAA CDR version), pure Vogel (for comparison with AAVS3) and 2 optimised layouts starting from Vogel but allowing maximum antenna movements of 1m and 1.5m, respectively.

      In addition, some software work is required to enable effective use of the simulations:

      1. Prepare a tool (python script or C/C++ program) to produce embedded element patterns in the same format as those produced by the Curtin and INAF groups from the HDF5 files output by HARP.  These tabulate the complex values of the electric field (both polarizations) on an azimuth/elevation grid with separation 0.5 deg in each coordinate.
      2. Improve functionality and performance of OSKAR and harp_beam. At present, a bug in code introduced to speed up processing has forced a reversion to an earlier version of OSKAR which does not support station rotation (essential for array-level imaging simulations).  A revised version should: support station rotation and creation of virtual dipoles; re-introduce and test Karel Adamek's speed-up measures and produce output for antenna EEP's, individual and multiple station beams.

      The outputs are a final report on layout optimisation; HDF files of simulation output covering the full SKA-Low frequency range for the 4 layouts specified above and tested harp_beam and OSKAR code.

      Attachments

        Issue Links

          relates to

          Feature - A Feature is a service that fulfills a stakeholder need. Each feature includes a benefit hypothesis and acceptance criteria, and is sized or split as necessary to be delivered by a single Agile Release Train (ART) in a Program Increment (PI). SP-3487 EM Harp simulations of SKA-Low layouts

          • Must have - Items labelled as "Must have" are critical to the current delivery timebox in order for it to be a success.
          • Releasing
          mentioned in

          Page Loading...

          Page Loading...

          Page Loading...

          Page Loading...

          Page Loading...

          Page Loading...

          Structure

            Activity

              People

                r.laing Laing, Robert
                f.graser Graser, Ferdl
                Votes:
                0 Vote for this issue
                Watchers:
                1 Start watching this issue

                Feature Progress

                  Story Point Burn-up: (100.00%)

                  Feature Estimate: 5.0

                  IssuesStory Points
                  To Do00.0
                  In Progress   00.0
                  Complete832.0
                  Total832.0

                  Dates

                    Created:
                    Updated:

                    Structure Helper Panel