For Want of a Nail.....

There is an old parable: For want of a nail the shoe was lost;
For want of a shoe the horse was lost;
For want of a horse the battle was lost;
For the failure of battle the kingdom was lost—All for the want of a horse-shoe nail.

It seems to quite accurately describe the version of events the Space Ex has officially released to explain the loss of its Falcon 9 in mid flight. Space Ex believes that a bolt failed on a strut assembly that holds down the Ghe COPV's inside of the second stage LOX tank. The failure of one of these bolts set off a chain event that within one second resulted in the overpressurization and explosion of the second stage LOX tank and destroyed the Falcon 9.

After Space Ex was able to triangulate high speed accelerometer data and locate the original "sound" that corresponded in time with the loss of Ghe pressure in the COPV's inside the second stage LOX tank they quickly began conducting tensile strength tests on strut assemblies they had in stock. Although some of these strut assemblies failed at around 6000 pounds of force instead of the rated 10,000 pounds of the strut it still didn't explain how one could have failed at 2000 pounds of force which is the calculated load on the struts and maximum G force during ascent.

Eventually, Space Ex was able to find a bolt for one of these assemblies that failed at 2000 pounds of force. Therefore, they concluded that it was probably a bolt that holds the strut assembly to the tank that actually failed. During subsequent press conferences Elon Musk went into a lot of detail explaining how the strut assembly was not manufactured by Space Ex but was instead specified and bought based upon this manufacturers strength specifications. When pressed by reporters to name the manufacturer Musk declined and explained that it would not help the situation. Musk went on to explain that further metallurgical testing on the bolt itself showed improper grain forging which could have also led to a similar failure on the Falcon 9 that exploded. He seemed to have found the smoking gun for the failure sequence that led to the loss of his rocket.

When asked about the materials that the assembly was made of he also declined to go into details but put forward the information that it was a type of steel and was rated for a much higher failure pressure than it would have ever seen on flight. This is very curious, especially when in the same interview he went on to explain that they were considering going to an inconel material which is much harder to come by and astronomically more expensive to buy and manufacture. In the short term he suggested that they didn't actually test the strut assemblies previous to the loss of this vehicle but would certainly do so in the future. He was careful to explain that they had used the manufacturer's strength rating in lieu of actual testing but that at least one improperly forged bolt had been found in their inventory.

A "type of steel" is a red flag for anyone familiar with cryogenic atmospheres. Carbon steel materials lose all of their tensile strength when exposed to cryogenic temperatures. The immediate embrittlement that such temperatures cause in the granular structure of carbon steel is instantaneous and catostrophic. Just as an example, we experienced a catastrophic failure on a high pressure pipeline because of very short term exposure to LN2 (-320 degrees F). We had a heat exchanger controller failure that led to a small amount of LN2 trickling into the 3" 6000 PSIG pipeline previous to starting pumps to pressurize the system. The pipeline soon violently exploded at less than 200 PSIG in a location where it was in mechanical tension.

Even stainless steels with their lower proportion of carbon lose proportional amounts of tensile strength when exposed to cryogenic temperatures which is why careful control of metallurgy, forging process, and purity is required in all such systems. Different grades of stainless (depending largely on the amount of carbon to nickel composite) have different reactions to such temperatures. What has since become public knowledge is that Space Ex was buying off the shelf strut assemblies from a manufacturer that was not aware of the environment they were to be used in. Space Ex took some standard temperature de-rating tables based upon the assumed metallurgy of the struts they were buying and effectively load rated their strut system by analysis. There is nothing wrong with load rating by analysis as long as strict compliance of material pedigrees are observed but that does not appear to be what happened in this case. Since the manufacturer was not aware that these strut assemblies would be exposed to cryogen temperatures they do not track forging processes and metallurgy necessary to do such de-rating by analysis. This also explains why ultra weight conscious rocket ship designers used 10000 psi rated struts in a 2000 psi application.

As I have already mentioned in this thread, Space Ex has long struggled with configuration control of the hardware they are flying. It is a byproduct of being in a tremendous hurry to launch payloads. To be fair Space Ex is driven to this frenzy by the government that is punishing them monetarily for delays on manifests. Add to that the open competition they find themselves in to gain more launch manifests and you begin to see the extreme pressure they are under to launch vehicles. Under such pressure there is little wonder that they took the shortcut of buying off the shelf hardware and downrating it by analysis. It doesn't make it a good practice but it is understandable.

It doesn't make Musk's claim that a bolt "snuck" through the system accurate. It may have, but the real fault was in using assemblies in environments they were not made to be used in without at least explaining to the supplier that they needed to control the metallurgy and purity to make Space Ex's analysis hold up.

Unfortunately for Space Ex there is no proof that any of this is what caused the loss of the Falcon 9 mid-flight. While such a sequence of events seems to fit most of the data that Space Ex has, there are also parts of this story that do not fit the data. This is why neither NASA's independent investigation signed off on this theory, nor did any of the members of Space Ex's investigation that do not directly work for Space Ex. The broken strut assembly scenario is one of several fault tree sequences that could explain what happened. However, there is at least on major piece of data that most definitely is not explainable by this scenario.

The first actual warning that something was amiss on the Falcon launch was a minor drop in pressure on the Ghe pressure of the COPV's mounted in the second stage LOX tank. Musk's first public statements about the investigation brought out the fact that the data seemed counterintuitive. The data, which was taken at a relatively low rate of speed, showed a drop in pressure followed by a return to "normal" pressure. The transducer that monitored this pressure was mounted on a manifold assembly tied to several of the COPV's in the LOX tank. Through the years of doing this type of work I have been exposed to several instances where the data doesn't seem to match any logical sequence of events. It is frustrating to say the least. At some point it is not unusual in such a case to believe that you may have just found THE exception to the universal laws of physics. Reason always prevails but that thought can occur.

The drop in pressure that Space Ex saw in their data was very short, less than a second, before the overpressurization and rupture of the LOX tank occurred. The best scenario that Space Ex was able to come up with was that the initial "sound" that high speed accelerometer data was used to triangulate to the area of the COPV's seemed to occur at the same time as the drop in pressure. If this was indeed a strut or bolt breaking then the COPV would begin to rapidly rise in the LOX tank, ripping the connecting tubing that attached it to the other COPV's as it moved. Space Ex engineers theorized that the tubing could have kinked and shut off the leaking Ghe for a few milliseconds which would explain the return towards "normal" pressure on the Ghe manifold.

I think we have reached the point described above concerning an exception to the universal laws of physics with this explanation. Or... as a friend I worked with for many years used to explain in a crude way; bull$%^t. He would usually do this quite openly in a feigned sneeze at high volume; bull$%^t as he covered his mouth. In order to understand my incredulous disbelief let me explain a little further.

A water hose can indeed kink in such a manner that you can shut off the flow of water at low pressures. I have even seen brake lines of small diameter kink such that hydraulic fluid can be restricted enough to defeat the balance on a brake system. What I have never seen and will defy anyone to produce is a 5500 PSIG Ghe line that is kinked enough to shut off the leak of Helium. In the configuration that Space Ex uses of their COPV's inside LOX tanks there are several different tanks tied to one tubing manifold. If one tank becomes detached and begins to rapidly rise the line would have to kink in two directions at once for the pressure to return to "normal." It would have to shut off leakage from the rising tank and the tank that it was tubed to at the same time. If one such kink is impossible I don't know how to describe the statistical impossibility that two simultaneous kinks would represent.

There is more....

In the constant effort to save weight Space Ex used titanium tubing to make these manifolds. Titanium is both lighter and stronger than the stainless tubing usually used in such applications so they were able to use extremely thin walled tubing for this assembly. In other words, this tubing has great strength for retaining internal pressure but almost no shear strength to resist tearing apart. If a strut assembly were to break and the COPV to start a rapid ascent in the tank it would immediately tear the tubing apart instantaneously releasing a large volume of helium into a tank with a very small ullage. In other words, the LOX tank would rupture AND there would be no rise in Ghe pressure after the initial drop in pressure.

Add to all of this the fact that all preliminary data suggests that the same type of COPV in the same second stage LOX tank just experienced a "massive breach" that destroyed a second Falcon 9 as it was filled with propellants while sitting on the launch pad and the story gets even harder to believe. Preliminary data also suggests that there was a large drop in pressure followed by a similar rise in pressure in this same Ghe system right before this vehicle exploded.

Is there a scenario that matches this sequence of events? It turns out there is and it isn't counterintuitive at all once you understand the COPV failure mechanism. More on that tomorrow......