Conjunction Event Predictions for Operational ESA Satellites

1
Conjunction Event Predictions for Operational ESA
Satellites

• SpaceOps 2004, 17-21 May, Montreal, Canada

2
Index

• 1. Introduction
• 2. The Collision Risk Assessment Tool (CRASS)
• 3. CRASS Data
• 4. CRASS Products
• 5. CRASS Operations
• 6. CRASS Off-line
• 7. Collision Avoidance Procedure
• 8. Conclusions

3
Introduction

• ESA has been monitoring conjunction events of
  ERS-1, ERS-2, and ENVISAT with members of the
  trackable, unclassified Catalogue population.

• Sun-synchronous orbit at an altitude near 800 km
  is an area of high object densities

4
The Collision Risk Assessment tool (CRASS)

• Conjunction detection
• Smart Sieve technique fast and robust filter
  algorithms
• Extraction of events within a user-defined
  consider-ellipsoid
• Covariance propagation
• Orbit
• Object properties
• Geopotential
• Drag
• Solar radiation pressure
• Collision probability estimation
• Conjunction geometry
• Collision cross-section
• Propagated covariance

5
CRASS Data (1 of 4)

• From ESOC-FD (Flight Dynamics Division)
• Binary orbit files
• Manoeuvre files
• OD Covariance files
• From ESOC-MA (Mission Analysis Office)
• Space object data from DISCOS (periodic updates)
• Covariance look-up tables
• Satellite property files
• Solar Flux Data (NOAA ESOC)
• From USNO (via anonymous ftp)
• Earth Orientation data and Leap Seconds
• From OIG (NASA Goddard)
• Unclassified catalog data (TLEs)

6
CRASS Data (2 of 4)
7
CRASS Data (3 of 4)

• All input files have standard ASCII formats
  (portability)
• Pre-processing tools in CRASS suite
• Create a catalog snapshot with unique TLEs
• Transform binary orbit files to TEME frame
• Merge orbit and manoeuvre files into a
  standardised ASCII file
• Transform Covariance files to (U,V,W) frame
• Merge Earth Orientation data and Leap Second info
• Pre-processing tools run independently

8
CRASS Data (4 of 4)

• Manoeuvres
• Nominal manoeuvre introduced in the orbit
  prediction 3 days before execution
• Avoid manoeuvring the satellite into a
  conjunction event
• Effect of manoeuvre on orbit dispersion modelled
  by the software
• Covariance look-up tables
• Modelled during development phase
• Sorted by eccentricity, perigee height, and
  inclination
• Modulated with radar cross-section
• Radar data from FGAN or Monge
• Optional for high risk events
• Confirm/reject high risk

9
CRASS Products (1 of 4)

• CRASS generates a conjunction file for each
  target in a project
• Conjunction files contain a user-defined number
  of top ranking events, according to
• Collision risk
• Fly-by distance
• Automatic issue of mails
• CRASS Newsletters routine forecasts
• CRASS Warning
• Warning thresholds are exceeded
• Sufficient reaction time
• Several lists of mail recipients per project
  (adapted workload)

10
CRASS Products (2 of 4)

• Conjunction files organised in tables
• Table 0 General Job Execution Parameters
• Table 1 Conjunction Geometry, Miss Distance,
  and Collision Risk
• Table 2 Chaser Approach Geometry and Target
  Geodetic Position at Conjunctions
• Table 3 Target and Chaser Osculating Kepler
  Elements at Conjunction Epochs
• Table 4 Target and Chaser Position
  Uncertainties at Conjunction Epochs
• Table 5 Projection of the combined covariance
  onto the B-plane
• Table 6 Description of Target and Chaser
  Objects
• Table 7 Target and Chaser Object Properties
• Table 8 Conjunction Histories of Events Listed
  in Table 1 (up to 7 results per event for daily
  predictions over a one-week prediction time span)

11
CRASS Products (3 of 4)

• Graphical output
• 3-D target and chaser trajectories
• Contour plots of collision probability in the
  B-plane
• Error ellipsoids at the time of conjunction
• Position/velocity errors /- 0.5 orbits about
  conjunction epoch
• Position/velocity errors up to conjunction

12
CRASS Products (4 of 4)

• Archiving
• Cyclic storage of input data
• Archiving of historic output
• Mail archive
• Software execution log
• All errors trapped and reported on standard
  output
• In case of corrupted or insufficient orbit data
• Transition to TLE
• Consider effect on orbit dispersion
• Message to ESOC-FD
• In case of outdated catalog data
• Message to ESOC-MA

13
CRASS Operations (1 of 2)

• Forecasts generated on a daily basis
• Prediction intervals of one week
• Histories of 7 forecasts for each conjunction
  event
• The tool runs automatically (crontab) over the
  night-time, during the computer idle time

14
CRASS Operations (2 of 2)

• Update the TLE Catalog file
• Retrieve Earth rotation parameter file and leap
  second file from the US Naval Observatory
• Merge Earth rotation parameter file and leap
  second file into a single file in CRASS Earth
  rotation parameter file format.
• Update dates in the master input file for the
  current run
• Merge orbit files and manoeuvre files and convert
  them to CRASS orbit file format
• Execute the collision risk assessment
• Create files with obsolete TLEs, decayed objects
  according to DISCOS and objects decaying in the
  forecast time span
• Dispatch CRASS e-mails and warnings to the
  service users. Send out warning messages to
  ESOC-FD and ESOC-MA in case of corrupted or
  outdated input data

15
Collision Avoidance Procedure (1 of 4)

• Adopted thresholds
• Event probability gt 1/10000
• Closest miss lt 300 m
• Consider provision of dedicated radar data
• FGAN (34-m S-band antenna, Germany)
• Monge (10-m C-band antenna, France)
• Only event probability can trigger an avoidance
  manoeuvre
• Manoeuvre options
• Discard manoeuvre
• Wait for more data (mass penalty)
• Advance or delay nominal manoeuvre
• Dedicated avoidance manoeuvre
• The Project Office takes the final decision

16
Collision Avoidance Procedure (2 of 4)
ELSE
END
ELSE
END
17
Collision Avoidance Procedure (3 of 4)
ELSE
THEN
END
18
Collision Avoidance Procedure (4 of 4)
ELSE
19
CRASS Off-line

• Each program in the CRASS suite may be
  independently executed
• Off-line set-up shares common input structure
  with operational set-up
• Alternative output directory structure
• Faster reproduction of specific conjunction event
  results
• Check effectiveness of proposed avoidance
  manoeuvres
• Check that avoidance manoeuvre does not result in
  additional high-risk conjunctions
• Monte-Carlo validation of CRASS results by
  stochastic methods is also available

20
Conclusions

• CRASS is a useful operational system
• CRASS data acquisition and program execution are
  automated
• CRASS is fast, robust and accurate
• CRASS performs checks on all inputs and reports
  anomalies
• CRASS archives all inputs and outputs
• CRASS automatically issues newsletters or
  warnings depending on risk criteria
• A specific procedure is triggered by collision
  warnings
• The procedure may end up with the implementation
  of an avoidance manoeuvre