Rutgers University-Webb Research Corporation

1
Rutgers University-Webb Research Corporation
Partnership for Glider Technologies
Scott Glenn, Oscar Schofield Josh Kohut,
Hugh Roarty John Kerfoot, Chip Haldeman
David Aragon, Josh Graver
Rutgers University Coastal
Ocean Observation Lab
Pat Cross
OASIS Bruce
Bricker,
Frank Bub
NAVO
Clayton Jones, Tim Scholz, Peter Collins,
Matt Palanza, Bill Wieler, Dan Webb, Doug Webb
- Webb Research Corporation Tom
Campbell, Peter Fury, Hugh Farger, Fred Jackson
Dinkum Software
2

• RU/WRC Partnership Glider Testbed Activities
• Operational testing of new hardware and software
• - First At Sea Deployments
• - Improved Steering
• - Large Vessel Deployments, Snap in wings,
  rugged tailfin
• Software Prototype Development and Testing
• - X/Y-Modem to Z-Modem Iridium for Crash
  Recovery
• - GRCS to Dockserver Glider Control
• - Mission Plan Tool Glider GUI
• - Automated System Monitoring
  Email/Cell Phone Warnings
• - Automated Decision Making Reactive
  Planning
• Data Processing
• - Web-based displays for adaptive sampling
• - Assimilation in forecast models by NRL,
  NAVO
• - Assimilation into performance
  prediction models SAIC, Metron
• Sensor Software Improvement and Testing
• - CTD Thermal lag, SAM, Hyperspectral
  spectrophotometer
• Training of operators 20 People Signed Up for
  Glider Training Jan 9-11.
• Guidance to Slocum operators for Field
  Operations

3
Glider TTI Task Execution Matrix
4
RUCOOL Global Fleet Deployments
5
Rutgers Glider Missions FY07-Present 36 Total
TTI Glider Hardening
Deployment/Recovery Pickpoint
S- Mission with snap in wing rails 16 Missions
Rugged Fin Mission 5 Missions
6
Glider Deployment/Recovery
From Coastal Boats using the cart

• From Large Ships using Pickpoint
• SHAREM 150, 151
• NSF Eddies Cruise Sargasso
• NSF LaTTE Hudson Plume
• NSF Antarctic Cruise
• NOPP Acoustics Shelf Break
• DOE Bold Cruise Mona Passage
• Liverpool Bay Cruise
• Baltic Sea Cruise

7
Snap In Wing RailsInstall or remove wings
without tools.Reduce chance breaking during
deployment and recovery.Opening Latch on
installation reduces wear.
8
Non-Rugged Fin Motor and Tail
Mission Problem
93 Fin Motor Failed
97 Tail Broke on Recovery
99 Fin Motor Failed
108 Fin Motor Failed
121 Fin Motor Failed
9
Rugged Tail Fin

• New Tail Fin Design
• - More robust
• - Breakaway rudder
• - Full range of motion to re-engage fin with
• the motor
• - Motor housed completely within
• the tail
• Side-by-Side Sea Trial Nov. 2006
• - Avg. of 1 meter significant wave height
• - No significant difference in comms
• - Manual attempt to re-engage
• new rudder failed
• New digital design in place to
• automatically re-engage the rudder.

10
Non-Rugged Fin Design Lag between Commanded and
Measured Fin Position
Rugged Fin Design NO Lag between Commanded and
Measured Fin Position
11
Problems With Leakage

• November 2006 First Rugged Fin Deployment.
• O ring seal failed from thermal expansion
• Switched from inside diameter seal to face seal
• Face seal less susceptible to expansion and
  contraction
• No leaks since design change

12
Rugged Fin
Version 3
Version 2
13
Rugged Fin
Mission Duration (days) Results
101 3 Fin performed well. Ended mission early because of offset pressure sensor.
106 12 Completed full endurance leg.
112 12 Completed full endurance leg.
122 19 Completing full endurance leg with OSU ChiPod attached.
128 28 and counting Flying Stretch Glider on Endurance Line with Additional Battery Packs up to Weight Limit
14
Proposed location but vibrations from glider
contaminated acoustic signals

• June 06 URI visits Rutgers with hydrophone
• August 06 Hydrophone attached to glider for 15
  minute segment
• Sept. 06 Hydrophone attached to glider for 14
  day mission
• April 07 URI visits Rutgers for two days of
  experiments to test beam forming capability

Final configuration with hydrophone towed behind
glider with no rigid connection between the two
15

• Sept. 06 Hydrophone attached to glider for 14
  day mission

16
Off the New Jersey shore

• Follow-up study at Tuckerton
• Glider and ITC source deployed from anchored
  vessel
• Glider travels away from source
• Glider makes 90 turn
• Glider travels perpendicular to source

Glider makes 90 turn
1 km Glide Path
2 km Glide Path
Anchored Source 1.75 kHz tone
17
External Modular Sensors
NOAA National Marine Fisheries Service (NMFS)
Collaboration Vemco Fish Finder Attached to
glider for 1 and 12 day missions
160 Acoustic pings from transmitter over 2 hour
period
18
External Modular Sensors
Oregon State University Collaboration ChiPod
attached to glider for 1 and 18 day missions
19
Chi Pod Results
Wind Stress
Temperature (sampled at 10 Hz, gradient sampled at 120 Hz)
Cox Number
Turbulence Dissipation Rate
K_T
20
Blue Before Red After
Before encounter, glider dives to 75 m and roll
is fairly steady ( -7 degrees). After
encounter, glider is able to dive to only 30 m
and roll changes direction (negative to positive)
upon inflection. This leads to conclusion of
missing wing (loss of mass loss of density
glider is unable to dive below pycnocline).
21

• At recovery
• Missing Wing
• Damaged Sensors
• Damaged Cowling/Hull Sections (no sections
  breached)
• Hardened tailfin allowed return flight even with
  multiple handicaps

A
B
C
22
Long Endurance GliderElongated payload bay to
accommodate a) batteries for increased endurance
or b) more sensors in payload bay
Stretch Glider
320 Alkaline C-cells versus 230
23
Deployment 117
Yos not symmetric
24
Deployment 128
Yos symmetric
25
Normal glider green and blue
Stretch glider red and black
We hope to move the stretch gliders up to the
normal glider
26
Real Time Glider Battery Curve Website
Voltage
Days
27
Advanced Glider Training Classes Initiated at
Rutgers
28
Glider Ballasting

• Refined System for Ballasting
• Response to operational needs
• Carefully track masses and volumes
• Extremely useful in troubleshooting
• operations and hardware during
• experiments

Glider can be statically trimmed to minimize
battery use
29
Glider Ballasting

• Refined System for Ballasting
• Response to operational needs
• Carefully track masses and volumes
• Extremely useful in troubleshooting
• operations and hardware during
• experiments

Glider can be statically trimmed to minimize
battery use
30
Predeployment Checkout Sheet

• Developed standard pre-deployment checks
• Includes hardware and software

31
Deployment Checklist
32
Control Adjustment at Launch

• At deployment, send glider on quick mission
• Use resulting glider data to adjust controller on
  deployed glider
• Improves glider navigation, speed, efficiency

33
Control Adjustments Before efficiency tuning
Desired Pitch 30 of the time
12.5 cycles
34
Control Adjustments After efficiency tuning
Desired Pitch 90 of the time
18.5 cycles
35
Online Deployment Tracking Database
36
Glider GUI Remote system monitoring and mission
planning
Glider Status
Glider Health
37
Glider Fleet Status Webpage
38
Glider Fleet Status WebpageWeb-based system for
querying Slocum sensor definitions
2
1
3
39
Diagnostics Webpage
40
Glider Fleet Status Webpage provide
mission-critical diagnostics
41
Online database for creating user-defined glider
diagnostic reports
42
Mission Planning Interface Google Earth
43
Web Display Active and Archived Deployments
Email Warnings Glider Overdue QA/QC
Checks Stay Awaypoints
http//marine.rutgers.edu/cool/auvs/
44
Data Fusion/Visualization Requirements DATA
SOURCES

• Bathymetry gridded elevation files ( 3-column
  xyz )
• Simple data format 3 column ASCII data
  lat/lon/depth
• Sources
• NAVO Data Warehouse https//idbms.navo.navy.mil/
  
• National Geophysical Data Center
  http//www.ngdc.noaa.gov/mgg/bathymetry/relief.htm
  l

• CTD Profiles KKYY formatted messages
• http//www.meds-sdmm.dfo-mpo.gc.ca/meds/Prog_Int/
  J-COMM/CODES/tesaccode_e.htm
• Platform independent ( gliders, XBT, BSP, etc. )
• Sources
• Slocum gliders ( Rutgers University )
• Seagliders ( APL University of Washington )
• Spray gliders ( SIO UC San Diego )
• XBT ( NAVO )
• BSP ( NAVO )
• Others?

• Other Operational Oceanographic Products
  HDF/NetCDF ( COARDS compliant )
• http//www.unidata.ucar.edu/software/netcdf/
• http//www.hdfgroup.org/
• Self-documenting
• Type/Sources
• Satellite products ( SST, Altimetry, Ocean
  Color, etc. )
• Rutgers University L-band and X-band ( HDF )
• NAVO altimetry
• Spatial analysis/ models
• NCOM NRL/NAVO
• NLOM NRL/NAVO
• MODAS NRL/NAVO

Pat Cross, Scott Glenn, John Kerfoot, Bruce
Bricker
45
Prototype Real-time Environment Mapping and Asset
Planning (REMAP) Matlab Interface for Valliant
Shield
46
REMAP 4D Feedback from use during Valiant Shield
2007
Requested Ability to plot 2-D glider cross-sections vs time and/or distance along track. Done - 12/11/2007
Requested Load/plot default bathymetry upon opening REMAP session. No forseen problems in implementation.
Requested Flagging bad/suspect profiles. Build new panel to handle individual CTD profile plots.
Requested Scale individual CTD profiles to 1000m depth, with the ability to change scale. Build new panel to handle individual CTD profile plots.
Requested User's manual In progress.
Requested Display product, date, time and depth of NCOM slices. Already done - make more visible.
Requested Cursor initiated panning option.  
Requested Color-code CTD profiles according to NAVOCEANO preferences. Fixed - 9/21/2007
     
Issue 3-D rotation deactivates all other panels. Done on purpose to eliminate bugs between panels while in rotation mode.
Issue Intersection of model surfaces and profiles was difficult to see. Toggle the NCOM slice transparency to view intersections.
     
Bug 3-D rotation was sufficiently jumpy that it wasn't used much. Most likely due to CPU limitation from other apps running on the machine in addition to the overhead associated with OpenGL rendering of large datasets in Matlab.
Bug Plotting polygons manually doesn't work. Works, but not explained well - will be addressed in user's manual.
47
Raw CTD
48
Velocity Independent Correction
49
Glider Flight Characteristics
50
Velocity dependent methodology
51
Global Summer/Winter SST Difference
52
Integrating Coastal Modeling and Observing
Systems using VariationalMethods and Data
AssimilationJohn WilkinGordon Zhang, Hernan
Arango, Julia Levin and Javier ZavalaCoastal
Ocean Modeling and Prediction GroupRutgers, The
State University of New Jersey
jwilkin_at_rutgers.edu
http//marine.rutgers.edu/wilkin
http//www.myroms.org
7 January 2008
53
Satellite SST CODAR gliders give an
integrated observational view of the MAB
MARCOOS Mid-Atlantic Regional Coastal Ocean
Observing System Currents and Temperature
54
Satellite SST CODAR gliders give an
integrated observational view of the MABwhich we
are melding into models using 4DVAR assimilation.
ROMS Northeast North America (NENA) shelf model
nested in HyCOM
55
Observing System Design using ROMS Adjoint and
Tangent Linear models to compute Representers of
data information content
define
Glider Profile obs along a transect that takes 3
days to complete Mooring Continuous profiling
at a single point for 3 days
Representer plots show the correlation between
the observations and the full 4-dimensional model
state. During analysis, the Adjoint Model
propagates information backward in time and
upstream into the source area, therefore
expanding the value of the observations by
introducing physics (the strong constraint)
observations tell about the ocean state at prior
times if you can infer where those waters came
from.
Analysis
56
Observing System Design using ROMS Adjoint and
Tangent Linear models to compute Representers of
data information content
define
Glider Profile obs along a transect that takes 3
days to complete Mooring Continuous profiling
at a single point for 3 days
The Tangent Linear Model computes how a
perturbations evolve going forward in time, again
using model ocean physics. By initializing the TL
model with the outcome of the backward Adjoint,
and running the TL forward, we see how the parts
of the model state that would be updated by 4DVAR
subsequently affect the model result into the
forecast window.
Analysis
Forecast
57
Observing System Design Information content
differs depending on wind direction
Mooring ensemble average
Wind direction
58
Observing System Design Information content
differs depending on wind direction
Glider ensemble average
Wind direction