Wednesday, February 10, 2010

Clerical Error – AN Safe

The Royal Canadian Mounted Police (RCMP) announced Wednesday afternoon that the two missing one-tonne (metric ton) bags of ammonium nitrate reported to be missing last December are not really missing. According to an article on, an RCMP spokesman said that their investigation concluded that the “most reasonable explanation for the discrepancy is administrative error”. Kinder Morgan, the distributor handling the material, reported the bags as missing on December 31st. Their own internal investigation pointed to a clerical error and they reported that to the RCMP on January 6th. But, with the Winter Olympics opening later this week, and ammonium nitrate being a well known component for improvised explosive devices, the RCMP was taking no chances. According to the article:
“The investigating team, which included the Integrated National Security Enforcement Team (E-INSET), members of the Vancouver 2010 Joint Intelligence Group, Public Works Canada forensic accountants and Natural Resources Canada, conducted more than 200 interviews of drivers and others involved in transporting the material.”
Loss of Product In their advanced notice of proposed rule making (ANPRM) for the ammonium nitrate security program, DHS indicated that they intended to include in the final regulation a requirement to report the loss, theft or diversion of ammonium nitrate. While theft and diversion are relatively easy to define, the problem of loss is much more complicated as this Canadian incident shows. When inventorying discrete items, most people would expect it to be a relatively simple matter of counting the items. Of course, in the real world things are always more complicated. With a large number of items stored in multiple locations it is relatively easy to have counting errors. This is usually dealt with by having multiple counting teams conducting independent counts. Discrepancies are resolved by recounts until both counting teams come up with the same number. Things get more complicated when there are shipping and receiving operations going on during the count. This is why every organization establishes acceptable levels of inventory discrepancies; inventory inconsistencies below the acceptable level are not investigated. Things are much more difficult with frangible inventory items. Chemicals used in manufacturing, for example, are typically removed from inventory by weight not by container count. Process designers will frequently attempt to size production runs to consume full containers (or multiples of full containers). If, however, there are multiple chemicals going into a product (usually the case) it is practically impossible to size the process to consume full containers of all the chemicals involved. This means that every time the process is run, there will be containers of partially consumed chemicals left in inventory; typically called ‘piece containers’. Organizations with strong inventory control procedures will require that each piece container is re-marked with the current weight of the contents of the container. This is typically determined by taking the initial net weight and subtracting the amount consumed. Unfortunately, all industrial weighments will be close approximations rather than exact amounts. Scale accuracy and operational procedures will affect just how accurately those weighments will be made. An acceptable variation of 1% is not unusual in most of the chemical manufacturing industry. Now a 1% acceptable variation does not seem to be much to worry about. With a drum containing 450 lbs of a chemical that is only 4.5 lbs. Of course, 4.5 lbs of an explosive can be quite devastating; remember the Underwear Bomber only had a few ounces of PETN explosives and intended on bringing down an aircraft with it. For an 80,000 pound rail car load that could be 800 lbs of allowable variation; about half of the amount of ammonium nitrate used in the attack on the Murrah Building in Oklahoma City. Now ammonium nitrate is an industrial chemical, but most of it sold in the United States is sold as fertilizer. Most of it is sold in bulk as a dry powder. At manufacturer’s facilities it is loaded using conveyor belts onto trucks, railcars and barges. I’m not aware of what the fertilizer industry standards are for that type loading, but I would suspect that it would be more than 1%. At all but the largest distributors, the material is transferred into trucks by frontend loaders; 1% weighment variations in those types of operations would be all but impossibly tight. Reportable Loss So, how do you determine what level of ‘loss’ needs to be reported? In the Canadian incident they reported an apparent loss of 0.033%; most companies would consider that an allowable and unremarkable inventory error. And, in fact, that is what the RCMP concluded. Will DHS have the manpower, time and money to conduct a similar investigation for every inventory discrepancy of even 1%? Don’t answer too quick to answer; the 0.033% discrepancy of 6,000 metric tons of ammonium nitrate, if it had been real, it would have made significantly more than 4,800 lbs of improvised explosive device. That would be a very big, a very loud, and a very deadly explosion.

No comments:

/* Use this with templates/template-twocol.html */