That was a good place to start — the second stage and the PAF (payload attach fitting). Those parts would have the most contact with the ILIAD payload and the most likely location to contain a link to Bronze Knot.
Parkowski dug into the engineering database in Mosu.
The second stage and PAF were both common between the Shrike Heavy and the Shrike 9 single-core rocket. They had been flying with small modifications for hundreds of flights over the last fifteen to twenty years. There were gigabytes and gigabytes of data for those two assembly-level parts of the launch vehicle.
She filtered out everything except the SH-21 data. Everything looked nominal.
The post-launch data analysis looked great. The second stage had nailed the orbit insertion for the transplanetary path that would take the ILIAD probe to Venus. The three first stages had all been recovered. Everything looked as good as it could get, at least to Parkowski’s level of understanding.
Prior to launch, OuterTek had delivered data packages on every part on the bill of materials (BOM) for the launch vehicle for review by Space Systems Command and the federally-funded Aerospace Corporation. The packages were standard, sterile, like they had put the information together and submitted them dozens of times before — which they had. This was normal.
Parkowski got out another water and took a sip. There was something here, there had to be; some clue as to what Bronze Knot was.
Or, maybe that was just wishful thinking on her part.
She went to the BOM itself inside of Sangam. It was the list of parts on the rocket itself, complete with part numbers and descriptions. It was DePresti’s most-accessed file according to the OuterTek interface. Parkowski checked it carefully, nothing was out of the ordinary.
Just for fun, she compared it to the pre-launch acceptance test packages that OuterTek had given to the government for review. All of the part numbers matched and lined up, nothing was out of the ordinary.
On a hunch, she checked the post-flight reports.
Most of the reports for each of the second stage’s main components were extremely detailed, down to the part number and qualification history. Some of the part numbers were different from what was in DePresti’s BOM.
In fact, roughly a third of the pieces on the second stage and PAF — but not any of the three first-stage cores — were different from what was supposed to be on the launch vehicle.
And, they were big parts too: the fuel tank, the nozzle, all of the avionics parts. All of the smaller parts were correct, but the large ones were all different.
What did that mean?
OuterTek was notorious for replacing parts, even on the launch pad, and not informing their customer or even their upper management. DePresti had complained about it during the entire launch campaign. Was this just another case of that?
Parkowski was no launch expert. She would have to ask her boyfriend.
She switched to the Mosu tool and started poking around for those part numbers. Maybe she could find a work order or a ticket that explained why those parts had been replaced with different ones, or where the parts that were supposed to have been on the rocket had gone.
Her first few attempts got her nowhere. The system wasn’t smart enough to filter out just those parts. The search feature was much less helpful than she had anticipated.
Gritting her teeth, she tried a different tactic — search by date. She checked the computer’s calendar and found all of the work order tickets that had been created or closed out during the SH-21 launch campaign.
There were thousands and thousands. Too many to go through one by one.
She focused on those closest to the launch date. If OuterTek made a parts switch, it would have to be near the end, when the government oversight was the most overtaxed and exhausted from the long launch campaign. Likely, it would have been backdated and documented after the fact, so that no one from the government would catch it.
There were a number of tickets, numbered sequentially, with a “BK” header on them, from three days before launch.
BK.
Bronze Knot.
She had found the jackpot.
Parkowski dug in.
It appeared to her that several parts on the second stage were changed — all due to BK requirements. Bronze Knot requirements, but without the trigraph that would catch someone’s suspicion.
The notes were straightforward, basic one-for-one swaps of small parts, not the large ones that were incorrect in the BOM. The instructions in the tickets told the next person on the workflow to update it in their internal part tracking system that fed the BOM.
The parts were small, sensors and thermistors and wiring harnesses and whatnot. But, all of them had tickets at the top stating that they were “BK,” which to Parkowski had to mean Bronze Knot.
She was finding more questions than answers. The only thing she knew for sure was that OuterTek was definitely involved.
Maybe there was an updated BOM with all of the changes made? Or they had changed everything and just included some of the part swaps in Mosu.
Parkowski went back to the post-flight data. Sure enough, there was a BOM provided, but it had been updated with the changes made in the Mosu tickets. It still didn’t match the pre-flight one though — the part numbers that she found in the post-flight reports were there in the post-flight BOM.
How could a rocket’s parts change from processing to the post-flight data review?
She shook her head. None of it made sense.
Parkowski checked the clock — it was 4:30 PM — and went back to the post-flight reports. Maybe she could find something she had missed.
And, in a “propulsion performance” folder, she found it.
CHAPTER THIRTY-FIVE
Parkowski was curious how the rocket had actually done performance-wise on its mission. She had seen the executive-level summaries, how they had “nailed the bullseye” and whatnot, but as an engineer she wanted to see the raw data and analyses that had led everyone to conclude that the mission was a success. There were processed data in the post-flight report, but this was the raw, unfiltered data that had been used to produce the summaries
There was only one engine on the second stage, as opposed to nine on each first-stage booster, so there was significantly less data for it. But, in the folder with all of the engine data and analysis, there were two copies of everything. One set had raw data, spreadsheets, and charts had “DELIVERY” in all-caps before the real filename.
The other set did not.
Curious, she pulled up a chart deck with both the DELIVERY tag and the one without and compared the two.
The one that said DELIVERY was exactly like she had expected. The second stage had separated from the first and immediately did its first burn, getting on its transfer orbit for the Hohmann maneuver that would end with it on its interplanetary orbit to Venus. Then, once it reached the second planet, the second stage would burn again to reach an orbit around Venus before separating the payload and putting itself in a disposal orbit that would decay until it burned up in a few years in Venus’ hellish atmosphere.
That one made sense.
What didn’t make any sense was the other chart package.
It showed a completely different burn sequence, with much longer coasts and a longer mission duration than Parkowski remembered from when she followed the launch closely.