Remembering Columbia

10 years ago today, the Space Shuttle Columbia exploded during re-entry, ending the lives of seven astronauts and instigating an investigation into the cause. AGI’s Bob Hall was part of a team of AGI engineers who sprang into action following the tragedy to support both the media’s explanations to a devastated public, as well as the Columbia Accident Investigation Board (CAIB). I had no personal connection with the Colombia crew or with anyone directly working the shuttle program. So, in that sense, my sense of loss is akin to what the rest of America felt at the loss of seven astronauts and a significant blow to our national space program. What follows relates the experiences I had following the tragedy. Sat 1 Feb I remember being awoken by a phone call Saturday morning around 9:15 a.m. Although it was unusual for me to be asleep at that time of day, it was a luxury I had in my single, pre-fatherhood days. And in some ways it was an indicator of how numb we as a country and even some of us in the space industry had become over shuttle landings. Who was going to wake up to see it land? This was post-Challenger, sure, but I guess I was thinking that liftoff was where the risk was (how true). On the other end of the line was Paul Graziani, AGI’s CEO. Due to my space operations background, Paul often reaches out to me with timely space-related news and events. Still, as I tried to collect my thoughts I couldn’t wrap my head around what Paul was trying to tell me, that the shuttle was lost. A few of us quickly assembled at AGI and got to work trying to put together a scenario showing the landing portion of the mission. Earlier in the shuttle program, we had had access to near real-time shuttle position and velocity data from NASA, but that was no longer the case in 2003. So we set out to build the most accurate re-entry and landing scenario we could. Our goal was two-fold: to lay the foundation for potential detailed anomaly investigation work and to put together a simulation that could be used early on to communicate to the public what had taken place. We had a very good relationship with Miles O’Brien of CNN, and in early conversations that day we began collaborating with him, pulling together information and discussing ways to depict events for CNN. I worked with a team of people to pull together a valid shuttle ephemeris for a similar re-entry. Most-re-entries at the time seemed to be the ascending node flight path, which took the shuttle over the Gulf of Mexico, approaching Florida from the southwest. For Columbia, the re-entry was a descending node flight path, which meant it crossed over Hawaii and came over the California coast, flying over the continental US before approaching Florida from the west. So it took some time for us to dig up a similar historical shuttle re-entry ephemeris file. As we did this we were digesting the piecemeal information coming out throughout the day – the different ground-based video feeds of the break-up and debris re-entry, the Louisiana weather radar that picked up the debris cloud, the reports of what may have gone wrong. It was Miles who first told us, that day, of the foam strike during liftoff 16 days prior. As Saturday morning turned into Saturday afternoon, we began talking seriously with Miles about me flying down to CNN headquarters in Atlanta to support his work and potentially present some analyses on the air. Paul had actually mentioned this as a possibility to me pretty early that day but I had discounted it. In any case, somewhere around 4 o’clock I gathered up everything we had amassed by that point and headed home to pack and to the airport. All the while, we were spooling up a team of AGI engineers behind the scenes to track down different threads and pieces of the puzzle. Imagine my surprise when I landed in Atlanta after dinner and tried to sync up with Miles, only to learn from his producer that while I was flying to Atlanta, he was flying to Houston, to work out of the Johnson Space Center. My initial felling was that I should follow him to Houston, since I felt I could be most responsive and productive by collaborating in person. For some reason, we did not pursue that path and I stayed in Atlanta. Sunday 2 Feb I worked all day Sunday at CNN headquarters, in constant touch with a growing team at AGI in Pennsylvania, building a few different STK scenarios. Since STK is an engineering tool, we couldn’t just ‘fake it’ with some animation. We needed real ephemeris and attitude data to represent what the shuttle was doing. We put together a landing scenario and iteratively used a breakup model with our propagators to depict the debris re-entry over Texas and Louisiana. We were overlaying the time-based weather radar imagery onto our globe simultaneously. The other scenario we were working on was for liftoff. Here we had plenty of historical shuttle ascent data and we worked to simulate a piece of the external tank foam breaking off and hitting the orbiter. This was all created and updated in a fairly frenetic fashion as Miles was able to garner new information and feed it to us. As Sunday started to wind down, it became clearer that we were not going to broadcast anything that day. Monday 3 Feb Monday was my first on-air segment with Miles. I was in the Atlanta studio while he was remote in Houston, but we did it conversationally while I ran STK live to show a hypothetical launch debris strike. What we showed that day seems rudimentary today, but at the time no one outside of NASA had access to the great launch footage showing the debris strike up close and in slow motion. We chose to show the foam debris strike in the area of the left main landing gear door, since Miles had ascertained that was where some of the early anomalous indications were. This scenario allowed for engineering analysis of the line-of-sight available from the various camera locations to see what should have been seen. Tuesday 4 Feb By Tuesday we had matured our debris re-entry scenario enough so that I could show the “S” turns the shuttle performed over the southwestern U.S. to help decelerate along with the point of breakup west of Dallas and the resultant debris continuing to the ground throughout Texas and Louisiana. I continued to work at CNN headquarters that day, and was working with the producers on another segment for Wednesday. Unfortunately, some time late Tuesday we came to a joint decision that wouldn’t happen, and on Wednesday I flew home instead. All the time I was wishing I had something more technically sound or definitive to show that I could really use to leverage STK and explain events. Unfortunately in the first few days, solid engineering information was not generally available outside of NASA, despite Miles working hard to dig up things for us. Subsequent weeks Once home, I continued to work on this project. We continued to expand our knowledge of events and evolve our scenarios in support of Miles. From a NASA standpoint, our approach at AGI is to do whatever it takes to help a customer in need, and clearly there was a need. As time went on, the Columbia Accident Investigation Board (CAIB) was formed and it became clear that we needed our engineering point of contact, system engineer Josh Lane, co-located with them in Houston. As Josh really got immersed in the CAIB activities, we were able to get the latest information the board was assembling and fuse it into an STK scenario. Some of the data types we combined into our single analysis included communications access, telemetered data status, vehicle attitude, vehicle position, debris locations and radar data. The scenarios we built and maintained were used for comprehensive engineering understanding of events as well as for intra-team and high-level CAIB briefings to explain what took place. Probably the most compelling depiction we provided included a multi-shot screen which included synchronized ground track and up-close vehicle views along with a customized view showing a cutaway of the left wing and a series of over 30 on-board sensors. The sensors were depicted as color-coded balls and we programmed them to change color over time as the telemetered values went outside of nominal range, high or even off-scale (indicating a break in the sensor telemetry path as the wing suffered more damage). This comprehensive view provided tremendously valuable situational awareness which was useful not only in post-pass or anomaly. In the end, I was proud to be able to help both in communicating technical subjects to the public and in providing detailed analyses and analytical tools to the team charged with investigating what happened. By enhancing the overall understanding of what took place, we assisted the investigation which helped prevent a future recurrence – ultimately honoring the crew the best way we could.

Systems Tool Kit (STK)

Modeling and simulation software for digital mission engineering and systems analysis.