 |
SOCRATES PlusSatellite Orbital Conjunction Reports Assessing Threatening Encounters in Space |
|
Top 10 Conjunctions by Maximum Probability
Top 10 Conjunctions by Minimum Range
- Introduction
- Methodology
- Changes with SOCRATES Plus
Back in 2004, there was no source of information on which satellites were active, so we decided to screen all payloads against the full catalog. Since then, CelesTrak has worked hard, with a growing number of collaborators, to provide more accurate operational status and we are preparing to perform a comprehensive review that will make that even more reliable. As such, SOCRATES will now screen only active satellites against the full catalog.
As of today (2023 Mar 2), 7,292 of the 10,154 payloads in Earth orbit that we receive GP data for are considered active. CelesTrak maintains operational status codes for each object in the SATCAT, which is updated on a daily basis.
Another change since 2004 has been the rapid (and then unforseen) growth in the orbital population. With the growth of large constellations, such as Starlink or OneWeb, we are seeing the overwhelming majority of conjunctions being reported just between their own satellites. It makes little sense to screen fully operational satellites within the same constellation (fleet) against each other using GP data when that satellite operator has a much more accurate and timely view of what is going on. SOCRATES Plus will no longer screen for intra-fleet conjunctions among fully operational satellites (we will, however, screen them against partially operational or dead satellites from that fleet). Doing that for Starlink and OneWeb reduces the total number of conjunctions (on 2023 Mar 2) from 143,997 to 56,343.
This change not only makes practical sense, it gives CelesTrak some breathing room to improve the efficiency of our computational process. Back on 2005 Oct 22, we screened 2,694 payloads against a total catalog of 8,626 objects and that took 1h 31m 43s (on a desktop computer of the day). Today (2023 Mar 2), SOCRATES screened 10,154 payloads against a catalog of 24,139 objects and that took 6h 46m 27s, including screening for intra-fleet conjunctions. Not only are the results not timely, soon we will be unable to complete a run every 8 hours as we have in the past. The corresponding SOCRATES Plus run took 4h 33m 36s, a result that we are continuing to work to improve.
Each SOCRATES Plus run now dynamically determines which fully operational satellites belong to each constellation. It also creates 'constellations' for objects which are physically docked to each other, such as for the ISS, CSS, and the two MEV satellites (since these objects can have different GP data). All of this is based on the latest ops status and other information in CelesTrak's SATCAT.
SOCRATES Plus now also adds screening for analyst satellites in the 8xxxx and 27xxxx ranges used by 18 SDS. The original SOCRATES was tied to objects in the main SATCAT and now that restriction has been removed. In addition, we are using the CSV format of the GP data behind the scenes—which STK Version 12 supports—to allow screening of objects beyond the range of 5-digit catalog numbers. This change means SOCRATES Plus is ready to handle the continuing rapid growth of the SATCAT.
The data format for the raw data has also changed somewhat. First, we have made the comma-delimited text file an RFC 4180-compliant CSV file, primarily by adding a header row. That will make it easier for users, who may want to load the data into a spreadsheet for additional analysis, to see what each column represents. The new format and associated queries are described in the SOCRATES Format Documentation.
We have also modified the content of the file somewhat. First, we have removed the Time In and Time Out data, which seems to be of limited utility and only added additional computational overhead. More importantly, we added the ops status (in brackets) to the name field in the raw data. This information has been displayed in the table of SOCRATES search results, but in the past it had been dynamically generated using the latest status. But that meant if you went back to the historical data, the ops status at the time of the report was not captured. Now it will be.
In line with the changes to the data, the table structure of the search results has been modified (primarily to remove the Time In/Time Out fields) and the HTML contents have been streamlined to remove hidden fields that are no longer of use. That change makes the results load much faster for large queries.
The query structure to produce search results has also been simplified (you can see the details in the SOCRATES Format Documentation). These queries are automatically produced from the Search page and can be bookmarked for future reference. But the new query format should make it more intuitive to customize these or use them in dynamic web pages, if desired.
Searches have been improved, too. It is still possible to get results sorted by minimum range or maximum probability, and the time in search option has been replaced by time of closest approach (TCA). And now there is an option to sort by relative speed (a measure of the specific kinetic energy), which can be an indicator of the potential consequences (and hence risk) of a collision. And there is an option to sort by NORAD Catalog Number.
Once you have performed a search, a link for each of the other sorts can be found at the top of the search results table, so that you don't have to go back to the search page to perform a different sort (e.g., to switch from a sort by minimum range to one by relative speed for the same name(s) or catalog number(s)).
- Enhancements Under Development/Consideration
[Added 2023 Apr 5] Today, SOCRATES Plus allows searches by one or two names or one or two catalog numbers. Soon, we hope to add searching by International Designator, as well. That would allow easier screening for all objects from a particular launch, which may not have common names. It would also facilitate searches related to key debris events.
CelesTrak is considering replacing 18 SDS GP data we currently use with SupGP data (when available) in the standard screening. That would allow using data that more accurately reflects recent and upcoming maneuvers. But we want to ensure we don't do that without applying some quality checks first, like ensuring the SupGP data is continuing to update.
Another option under consideration is supporting on-demand screening for limited cases. For example, if we perform screening using SupGP data, the user may want to see the results using GP data, or vice versa. The user may also wish to see the results with the latest data instead of the data used at the time of the run, which could have been started some hours earlier. To do this will require using our validated SGP4 propagator in JavaScript, which will then run the calculations on the user's machine.
CelesTrak would like to include risk-based results in SOCRATES, as well, as was discussed in the paper Consideration of Collision "Consequence" in Satellite Conjunction Assessment and Risk Analysis. To do this properly requires much better data on satellite size and mass, along with the public availability of orbital data with covariance. CelesTrak continues to work with the community to find ways to improve that data.
Finally, we are considering how we might produce some minimum-required data set that would allow a user to reproduce our results and share them with other users. Notionally, that would include the designator (catalog number and name), ops status, hard body radii, and raw GP data for the two objects along with the calculated TCA for the event. Everything else can be computed from that data (e.g., range, relative velocity, maximum probability, apogee, perigee). In fact, it could be used to dynamically generate a CCSDS Conjunction Data Message (CDM) quite easily, if needed for standards compliance. It would be far easier to use this minimum-required set than to transmit more verbose data sets containing mostly derived information.
Overall, the goal of these improvements is to continue to show how conjunction analysis can be more effective at improving safety of flight and protecting the space environment. We do that by helping satellite operators focus on what are the most significant threats, not by producing as much data as possible.
- Additional Resources
As a service to the satellite operator community, CelesTrak offers SOCRATES—Satellite Orbital Conjunction Reports Assessing Threatening Encounters in Space. This service provides regular information on pending conjunctions on orbit over the coming week. Because of the potentially catastrophic consequences of such conjunctions going unnoticed, it is hoped that this service will help satellite operators avoid undesired close approaches through advanced mission planning.
Three times each day, CelesTrak runs a list of all active satellite payloads on orbit against a list of all objects on orbit using the full catalog of all unclassified GP (general perturbations) element sets (often referred to as TLEs) releasable to the public to look for satellite conjunctions over the next seven days. The reason for limiting the search to conjunctions with active payloads is to give satellite operators an opportunity to plan—or obtain planning services—to consider moving their satellites out of harm's way (if their satellite is maneuverable), or to take other appropriate measures.
The runs are made using STK/CAT—STK's Conjunction Analysis Tools—and the SGP4 propagator in STK Version 12.2. STK/CAT is set to look for all conjunctions which are within 5 km at time of closest approach (TCA) and reports both minimum distance and maximum probability for the conjunction. Because the minimum distance method ignores position covariance information and can lead to an exaggerated assessment of the true risk, CelesTrak believes the maximum probability method provides a more reasonable (although still conservative) assessment of the true probability. For more details on the maximum probability method, see Dr. Sal Alfano's paper titled "Relating Position Uncertainty to Maximum Conjunction Probability" (2,493,912 bytes).
For more information on the original SOCRATES—which has been running since May 2004—see the papers, briefings, and animations on "Satellite Orbital Conjunction Reports Assessing Threatening Encounters in Space (SOCRATES)" presented at the 15th AAS/AIAA Space Flight Mechanics Conference, the Fourth European Conference on Space Debris, and the Sixth US-Russian Space Surveillance Workshop.
SOCRATES has been running as originally designed since May 2004. While running quite reliably over the past almost 20 years, a number of things have changed since then that have motivated us to update the original approach. For now, we are calling that SOCRATES Plus and we plan to run it in parallel with the original SOCRATES (before we switch over in another month or so) to allow users to explore the new features while we continue to improve performance on the back end. Here are some of the differences we've made in SOCRATES Plus: