Thursday, April 18, 2013

Fertilizer Explosion in West, Texas

Post-Scriptum 1-2-15:

I am utterly amazed at the continuing readership of this article,  obsolete as it is.  

The verdict on this incident is "in".  A building fire that should not have happened consumed most but not all of the AN in the building and an adjacent railroad boxcar.  Some of the AN in the building was in flammable wooden bins (which it should not have been stored in).  The explosion happened when the burning building collapsed upon,  and confined,  the remaining burning AN.  

It could have been a lot worse:  most of the 100+ tons of AN had already been consumed in the fires.  If instead it had detonated early,  the entire town would have been completely destroyed.  

There were plenty of mistakes:  (1) a very large flammable wooden-framed building unprotected by a fire sprinkler system,  (2) storing AN in the same building as feeds and grains,  which are flammable fuel,  (3) storing AN in flammable wooden bins,  (4) issuing building permits to construct habitations,  schools,  and businesses immediately adjacent to an agricultural business where fertilizers (including AN) will inevitably be stored,  and (5) the state fire marshal not having the authority to do anything about these other mistakes.  

Not all of these mistakes were the fault of the business,  as you can see.  There is plenty of blame to go around.  The most egregious one in West was not planning to handle a fire in the building where the AN was located.  Once the fire goes out of control,  some explosion at some level becomes inevitable.  It is only an unpredictable matter of time.  People will die.  

The only mistake NOT made at West (but already made elsewhere in some other incidents) was in using dynamite to break free AN caked onto transport auger blades.  Caking happens when the fertilizer encounters moisture,  you've all seen it in your homeowner's lawn fertilizer spreaders.  (The AN-based fertilizer blends sold for your lawn spreader are not shock-sensitive explosives,  though.)  

The problem with AN caking on transport auger blades is that straight AN is a shock-sensitive explosive. Dynamite is what you use to set it off,  for crying out loud.  

Ignorance does indeed kill.


Final Update 5-11-13:

The earliest postings and updates (4-19 and older),  have been made obsolete by information that came out publicly since.  From 4-20-13 forward in time,  what I have posted has turned out to be very accurate.  I have seen nothing to date to contradict those conclusions of a fire-induced AN detonation. 

I don't even believe all of the maximum-possible 270 tons were involved,  as the damage to West much have been far higher if it had (I think the majority of it was consumed in fire-induced decomposition for several minutes before the actual detonation occurred).  

This was based upon before/after aerial photo comparisons,  and upon frame-by-frame analysis of two separate cell phone videos of the actual blast.  I am not specifically an expert in AN (or other high explosive) detonation forensics,  but I do have years of experience with very similar forensics after rocket motor explosion accidents.  The basic comparison techniques,  and the background in shock waves versus subsonic blast waves,  is the same.  

 I also have experiences putting post-explosion rocket motor fragments back together for a forensics evaluation (very similar to air crash investigations).  That doesn't come into play with what I have done regarding the West incident,  because I have had no access on the ground to the site.  The various agencies doing the investigation are doing precisely those very up-close-and-personal forensics in the debris.  

The legalities are beginning to surface big-time,  what with suits already filed,  and a flood of advertising by various law firms looking for business windfalls.  I suggest that most such activity is  fruitless (at least in the sense of "real" moral justice),  until the investigators release their final report.  Then (and only then) will we know how that AN got immersed in a building fire.  

That's the key:  if that building had not caught on fire,  there would have been no explosion.  And it wasn't supposed to be able to catch fire,  based on what we learned from Texas City back in 1947.

Remember:  the cause of this thing will classify in one of three ways:  (1) it is something new,  in which case we had better learn from it,  in order to prevent it from happening again,  (2)  it is something already known,  in which case we need to do a better job adhering to the rules we already know,  or (3) this was some kind of deliberate act,  perhaps (or perhaps not) demanding enhanced security against future similar acts.  The investigators will tell us which it was.  

Update 5-4-13:

Text of my column published in the Waco "Trib"  Tuesday 4-30-13 on the safety of AN fertilizer operations is appended here.   A lot of folks out there are looking for "what was done wrong" to report,  or to use in court.  I suggest we let the investigators find out what really happened in their own good time.  

Meanwhile,  there are plants just like this in every rural community in the US.  Based on the history until recently,  there has been nothing to fear.  And,  ammonia thefts and other small-time crime and vandalism present no real danger,  although certainly they should be prevented.  

"Ammonium nitrate, which appears to be at the root of the devastating explosion in West, is a solid granulated compound made from ammonia and nitric acid. It is an explosive — much more so when mixed with the proper amount of fuel oil.
Unlike most of the high explosives, it is also insensitive in these ways: it is not friction-sensitive. It cannot be detonated by simple application of flame. That said, it has a large minimum diameter for effective detonation, some six to nine times larger than with dynamite.

It is shock-sensitive, which means one needs a detonating blasting cap and a booster charge to set off the explosion. And we have learned there is a "back door" by which exposure to flame does set off the detonation.

Ammonium nitrate exposed to fire does two things: it melts and it decomposes, which produces oxidizing gases that feed the fire. Decomposition proceeds through the material like a wave, with a speed measured in miles per hour. This is true of both the granulated solid and melted form.

If there is "confinement," that decomposition wave suddenly transitions into a detonation wave traveling at a speed measured in miles per second. That’s the "back door" for ammonium nitrate detonation in a building fire. The entire pile "burns" instantaneously, which is what happens when ammonium nitrate or any other high explosive detonates.

Confinement happens in two ways:

(1) The decomposing dry material is stacked too high.

(2) The decomposing melt flows down a drain (or other hole) to fill a pipe or a cavity.

If stacked too high, the very weight of the overlying material provides the confinement.  For melt inside a pipe or other closed cavity, the confinement is the container itself.

It is difficult to prevent either of these in a building fire engulfing stored ammonium nitrate. Even properly stacked, ammonium nitrate can fall into too deep a pile as floors and storage racks collapse in a blaze. It is difficult to predict, much less control, where molten ammonium nitrate will flow and collect as a building collapses in flames. The real key is to prevent the fire in the first place.

Investigation into the explosion’s cause in West continues. It really gets down to how the building caught fire and why it continued to burn. Once fire ensued, explosion was inevitable.

Investigation of the famous 1947 Texas City ammonium nitrate explosion disaster that killed nearly 600 people identified this very confinement issue. Bagged ammonium nitrate was stacked too deeply in the holds of two ships. Once the ships caught fire, the explosions — there were two, a day apart — became inevitable. Since then, safely handling ammonium nitrate fertilizer has focused on limiting stack heights; controlling where the melt can puddle; and, most importantly, preventing fire in the first place.

Before Texas City, fire-induced ammonium nitrate fertilizer explosions were common. Since Texas City there have been none until West. That is a testament to the common-sense handling and safety practices developed after Texas City. It wasn’t till the 1970s any regulatory agencies assumed a watchdog role involving this material.

In short, risks were successfully handled at ag supply businesses in every small town in the nation with nothing more than those three common-sense points. These are far more important than dotting i’s and crossing t’s on some permit or report, as the unregulated decades prove.

Decisions in the inevitable court cases resulting from West explosion should reflect this, but odds are they won’t. Given proper observance of the three common-sense precautions, ammonium nitrate really is safe despite what just happened in West. Same goes for toxic but not-very-explosive ammonia.

There have been recent disasters unrelated to fire-induced explosions. These mainly seem to be people using explosives to try to break up the big solid clumps of fertilizer that result when granulated ammonium nitrate is exposed to water. Considering that ammonium nitrate has been known to be a shock-sensitive explosive for over a century, that seems to be more than a little stupid. Ignorance really does kill.

Yet now something new must be considered: terrorists breaking into an agricultural supply business at night with blasting cap and dynamite, deliberately using that shock sensitivity to start explosions. No fire is required. This is a lot more sinister and dangerous than even druggies breaking in to steal ammonia for making meth."

Update 4-22-13:

I changed the title,  and added some safety and handling data below for ammonium nitrate (AN).  

Update Sat 4-20-13:

This thing is a work-in-progress.  I have seen a lot of very-recently published photo and video information.  Now I pretty much know what happened,  with good evidence to back those conclusions up. 
It was an AN detonation blast that demolished the east side of town.  The viral videos,  analyzed frame-by-frame,  show 3 distinct events separated by milliseconds,  which is why it looks like one explosion to the human eye (we just don't see that fast).  

The ammonia tanks are the 4 horizontal white tanks on the south side of the "fertilizer factory" (really just a receiving warehouse).  All 4 are still there,  and at least 3 are quite intact.  So also are all (or most) of the farm-trailer "nurse" tanks,  scattered all over the site,  quite intact. 

The "fertilizer factory" building itself was burning quite energetically,  which is just the AN decomposing and adding oxygen to the fire.  As long as the conditions for detonation are not satisfied,  you can literally burn AN in a fire quite safely!  I think without proof that a fair portion of the max possible permitted 270 tons of AN on the site was consumed non-explosively in the fire. 

The first explosive event took place not anywhere in the AN-containing building,  but to the north,  out of camera view of any of the viral videos posted on the internet.  A couple of these show some sort of fiery glow propagating southward toward the AN building about 1 (or at most 2) video frames before the building blew up.  I think,  without much proof,  that that was the explosion of some AN in a railroad boxcar not in camera view.  There seems to be another rail car overturned in that same vicinity. 

It is not clear whether that first event had anything to do with the "fertilizer factory" exploding.  You can see the enormous amount of surface material propelled into the sky in that second event,  unlike the other two.  That second event is the powerful one,  the one that damaged the town.  It correlates precisely to the location of the north end of that "fertilizer factory". 

The third event is 3-5 frames later:  a bright fireball but no dirt propelled skyward.  That could be some AN in the south end of the building,  chemicals in the tank farm on its south porch,  or a BLEVE event with one of the ammonia tanks.  It's not possible to know,  with the data I have.  Doesn't matter,  it's a weak event,  not the one that damaged the town

Pre- and post-explosion aerial photography confirms this interpretation.  You can see the blast point at the north end of the receiving warehouse,  complete with radial scrape scars.  That part of its foundation is gone,  while the rest is still there. 

The south porch chemical tank farm is scattered about,  but the 4 ammonia tanks are all still there.  Only the nearest one looks to be off its mountings.  They may have vented,  but it appears they did not explode. 

There is another partially-obscured and "weaker" blast scar,  centered on the rail siding north of the receiving warehouse.  It is adjacent to what appears to be an overturned rail car.  The nearby grain bins were blown off their foundations,  but did not explode (no blackening on their foundations). 

The furthest grain bin was not even shredded by the blast shock,  which gives you a rough radius to limit the true supersonic shock effects.  Most of the town received not a supersonic shock wave,  just a subsonic overpressure wave.  Otherwise,  damage and casualties would have been much higher.  It takes very little overpressure to collapse all kinds of residential structures:  1 or 2 psi is enough. 

The damage to the town looks just about like a 20,000-30,000 lb TNT "daisy cutter" bomb from Vietnam: just about a quarter mile diameter of visually-obvious massive destruction.  Allowing for a lack of fuel oil (for the full power of the AN detonation),  I'd guess that something like about 30-45 tons (factor 2 error on tons conversion here,  was "90" originally) of AN actually detonated in the second event.  The rest of the AN in the fire in the north end of that building most likely just fed oxygen to the fire without exploding

I could be wrong,  but that's the way it looks right now.

The proximate cause here is a fire in the north end of that building where the AN was,  and also perhaps around a presumed rail car (with some AN aboard),  just north of that building.  The real question is what started that fireIf that hadn't gotten started,  none of this would have happened. 

There are basically 3 possibilities the authorities must investigate:  (1) this was something we have never seen before,  in which case we must understand it,  in order to prevent future similar incidents,  (2) this was neglect of something we already knew about,  in which case we just do a better job observing all the known precautions,  and (3) this was some sort of deliberate act,  which is the most sinister item.  It may be a year before we know. 


Update 4-22-13:

Ammonium nitrate (AN) is a very insensitive explosive,  but it is an explosive nonetheless.  It is also a very effective source of fixed nitrogen as fertilizer.  Given proper handling,  it is quite safe to store and use AN as fertilizer,  or as the base for fertilizer blends.  

AN is shock-sensitive,  but not friction-sensitive,  nor is an AN explosion ignitable by a flame alone.  The "shock" in shock-sensitive refers to a very strong event,  such as a blasting cap plus a piece of a dynamite stick.  An extreme hammer blow against a solid anvil has the same effect.

What makes AN explode in a fire situation is "confinement".  When heated in a fire,  the material melts,  and it decomposes.  The decomposition releases oxidizing materials into the fire,  making it even hotter and brighter.  Without confinement of some kind,  the AN can be completely consumed by decomposition,  and quite safely.  

With confinement,  the decomposition "wave" proceeding through the material at speeds measured in miles/hour,  will suddenly transition into a very fast "burn" called a shock detonation,  with an effective burn speed measured in miles/second.  To human senses,  the entire pile vanishes in one incredibly-powerful explosion.

With AN,  confinement happens in either of two ways: (1) it is stacked too high,  and (2) the melt runs down a drain into a pipe or other cavity.  In Europe,  the safety code says stack AN no more than 2 meters (6.5 feet) deep (which really applies to bagged AN).  Based on recounted experiences with AN in the making of rocket propellant,  I personally would limit stacks to under 4 feet.

AN is shipped in steel drums or in paper bags.  The drums are their own confinement limiter:  just the depth inside the drum.  You need to keep it dry,  or the material will clump together,  becoming unusable for a granulated fertilizer.

Storage of AN requires a building and environment that is non-flammable.  You never want a fire to occur where AN is stored.  Proper housekeeping is important,  and attention must be paid to keeping the roof in good repair,  because you cannot have floor drains.  

In planning for the fire you hope will never occur,  you must observe the bag stack height limit,  and you must remember that steel drums on racks can fall into a pile as the racks fail in the heat,  and then the drums themselves will fail in the heat,  releasing the most-likely-molten-and-decomposing AN into a pile that is too deep.  If you have floor drains,  the melt will flow into them.  

In both outcomes,  you now have confined AN that is ready to detonate.  Once that occurs there is no stopping the disaster,  it is only a matter of time.  

The trick is to prevent the fire in the first place.  If you do that,  AN is quite safe to store and handle.            

Foam is probably a better fire-fighting agent than plain water,  but either can be used.  After all,  foam is water-based.  AN does not explode on contact with water,  that risk is with powdered magnesium (and some other light metals like lithium,  sodium,  and calcium).

Update Fri 4-19-13:

In some of the post-incident aerial footage KWTX-TV has been showing I thought I saw a blast point where the 3 grain storage bins were located north of the warehouse building.  That would be out of camera view to the left (downwind,  north) in the famous clip of the explosion.  It’s very hard to say for sure,  though. 

What I see in the famous clip of the blast,  but not until viewed frame by frame,  is the light/fire-glow from what might be an explosion the camera was not pointed at.  It came into view from the left,  more or less from where those grain bins were,  if the camera was pointing at a fire in the receiving warehouse building,  which is where the AN would be.  That’s a big “if”.  That unexplained glow appeared in only one single frame in that video clip. 

I'm going over to KWTX-TV tomorrow and look closely at their images and video,  and maybe see what really happened.  My contacts at the Waco "Trib" newspaper tell me that this ag place was permitted for two 10,000 gal ammonia tanks (I do see four ammonia tanks in pre-incident photography,  but two of those might just be temporary blending tanks,  not real storage tanks),  and up to 270 tons of AN-based dry fertilizer. 

I have definitely seen at least one ammonia tank still there (and apparently intact) in the post-incident aerial footage.

Especially if all the ammonia tanks are still there,  then an ammonia BLEVE is not a feasible explanation for what happened.  I'm not sure how an AN detonation could have been triggered by nothing but a fire (seeing as how there was at most a couple of hundred tons' worth of confinement),  although some tell me the fire-fighters were spraying water on the burning building where it was. 

An AN decomposition explosion seems more feasible,  but quite a bit less energetic than a full detonation.  However,  even a decomposition is more powerful than an ammonia BLEVE.  The damage to the town is more than I first understood.  More than just a BLEVE happened here,  that’s for sure. 

The third possibility is a grain dust explosion in one of those bins,  triggered by the fire.  Not a detonation,  but still quite powerful.  Those were sheet metal bins.  Not concrete. And that unexplained glow did seem to come from that direction. 
Update Thurs 4-18-13:  Here's an excerpt from an email I sent my sister when she asked if we were OK:

We're fine.  Didn't really hear it or feel it in McGregor.

It was an ammonia tank-in-a-fire explosion,  what the fire and safety guys call a "BLEVE" event.  Not a mass detonation.  Crudely the same sort of thing as a superheated steam boiler explosion.  Propane tanks can do exactly the same thing. 

None of the local news reporters seem to know the difference between ammonia and ammonium nitrate.  If this had been AN,  the whole town of West would be a crater.  AN is a mass-detonable explosive,  ammonia is a flammable,  corrosive,  poisonous gas.

It was bad enough as it is.  Not sure how many or how big the tanks were,  but not all of them exploded (there were ammonia vapors drifting all over West,  forcing evacuation).  The source was a farm ag place:  small farm ammonia trailers,  a couple of on-site storage tanks,  and maybe a railroad tank car (or maybe not).  Casualties not known with certainty:  5-15 dead so far,  over 160 sent to hospitals,  some critical.

The hard thing to imagine is how a fire actually got to the ammonia tanks.  Ammonia is flammable in air,  and could have been released by a safety vent valve,  but usually does not self-sustain without additional ignition heat coming from somewhere else.  The tanks are usually isolated from everything,  and that's to prevent what just happened. 


Here's the dry technical stuff:

Ammonia (NH3) is a somewhat-flammable,  very poisonous substance that is normally a gas at ordinary conditions.  It can be stored under pressure as a liquid,  like propane,  and presents a similar explosion and fire hazard to the threat presented by propane.  However,  unlike propane,  ammonia is also a very serious poison,  and is very corrosive.  

Anhydrous ammonia is a flammable fuel,  but it is not by itself an actual “high explosive”.  The liquid can be used as an agricultural fertilizer,  sprayed directly onto the field.  Diluted with sufficient water,  ammonia is non-flammable,  and not so very toxic,  and in fact is a good household cleaner. 

The related compound ammonium nitrate (NH4NO3) is the real “high explosive”.  This is a white granular solid material,  also known as 100-0-0 fertilizer.  It is a monopropellant explosive,  and is capable of mass detonation,  when triggered by the shock wave of a blasting cap.  The explosive yield is about 3 times higher (about the same as an equivalent mass of dynamite) when mixed with the correct amount of fuel oil. That is the explosive in the “fertilizer bomb” used in Oklahoma City,  not ammonia.  

Ammonia storage tanks can undergo something called a “boiling-liquid expanding vapor explosion” (BLEVE), particularly in a situation where the tank is exposed to an external heat source,  such as a facility fire.  The safety valve that releases tank contents cannot be too small,  or too large.  Both can lead to a sudden catastrophic bursting of the pressure storage tank.  This is an explosion,  but not the more-devastating “full detonation” of a real “high explosive”.  Because the suddenly-released material is also flammable,  it rapidly burns as it contacts the air,  adding to the explosion strength,  very similar to an exploding gasoline tank.  

The North American Emergency Response Guidebook lists ammonia in solutions from 50% strength all the way to anhydrous ammonia as ID number 1005 (on the hazardous material shipping placard),  and recommends its general emergency response guide number 125 for “corrosive gases”.  For the danger from a facility fire,  the guide says in part “containers may explode when heated”,  which is exactly what happened in West.  

It also recommends an initial evacuation distance of “1600 meters or 1 mile” in the event of a tank fire situation.  The toxicity of released ammonia gas may actually require a larger evacuation distance to stay under allowable dose limits,  but (unlike propane) ammonia vapors are lighter than air,  and tend to rise as they blow downwind.  

The explosion danger posed by a quantity like 50,000 pounds of anhydrous ammonia is very considerable (as was just seen in West,  Texas,  Wednesday,  4-17-13),  but is nothing like the danger posed by 50,000 pounds of the real explosive,  ammonium nitrate (or even worse,  ammonium nitrate-fuel oil as at Oklahoma City).  That would have been much more than 10 times worse in all conceivable ways.