1) Definition of Major Leak

Trying to deal successfully with a major leak is formidable task. The two most discussed possibilities are a direct hit by an aircraft and a planned act of sabotage. The latter is usually dismissed on the basis that proper security measures can provide a necessary deterrence. The air crash scenario is usually dismissed as an improbability. However, an air crash accompanied by exploding fuel would result in the following: The aircraft impact or subsequent explosion would probably rupture the chlorine container(s). The ensuring fire would instantly vaporize the liquid chlorine, and the chlorine vapor would rise quickly with the heat of the fire.

This sequence of events would serve to greatly diminish or eliminate chlorine exposure in the surrounding area. This was clearly demonstrated when a freight train derailment severely damaged a 90-ton chlorine tank car. The car was ruptured by the couplers from an adjoining butane tanker. All 90 tons of chlorine were released. The butane tank car exploded, and all of its contents were consumed by fire. The heat from this fire vaporized the liquid chlorine which disappeared into the upper atmosphere because of the rising hot air from the butane fire. A subsequent investigation revealed that no one in the surrounding area or at the scene of the accident was found who had experienced any exposure to chlorine. The consensus definition of the most probable major leak is a guillotine break in the liquid chlorine header between the chlorine supply system and the chlorine evaporators.

2) Important Aspects of a Major Leak

If a leak is to be considered a major one, there has to be a liquid spill. Any major gas leak can be dealt with quickly by the use of container kits and the proper use of chlorinator injectors to evacuate the vapor that is leaking. When a liquid spill is involved, the designer must make provisions for collecting the liquid in a confined sump and be able to hustle it off to either a scrubber or an absorption tank.

A major leak will never create a high atmospheric pressure condition in the room where the leak has occurred. Because of the enormous cooling effect in the leak area due to the liquid chlorine attempting to vaporize, the room pressure will be negative. This situation assures the flow of fresh outside air into the leak area. Therefore, a containment room for chlorine storage should always be designed for proper continuous outside ventilation. The fresh air from the outside should enter the storage room at ground level and exit at rooftop level. However, during a major leak event, the fresh air should enter at ceiling level and discharge to the scrubber at floor level. The scrubber system should be a one-pass system; never use a recirculating system, as it would cause untold corrosion damage in the room where the leak occurred.

3) Liquid Chlorine Collection System

This part of the design focuses on the storage room floor configuration. The floor should have a dramatic slope(2-1/2 in./10 ft) to a common point terminating in the liquid collection sump. The collecting slots should be narrow (2 inches max.) and deep (5-6 inches) to shield the liquid from room temperature. This will significantly diminish the liquid evaporation rate. When liquid chlorine spills on a flat surface, about 20 percent will "flash off" as vapor. This causes a thin sheet of chlorine hydrate ice to form on the remaining liquid, which prevents further evaporation until the ambient temperature melts the icy film. During the freezing and thawing cycle the vaporization rate of the remaining liquid is typically 8 lb of chlorine per square foot per hour.

The collection slots in the floor should terminate in the lowest part of the sloping floor (see fig. 3-1). At this point the floor should be constructed to accept either a pump or an eductor. (Pumps are available from both the Duriron Co. and Powell. These pumps routinely handle liquid chlorine.) Dealing with the liquid chlorine spill in the design of a neutralizing system is a number one priority because it reduces by a factor of ten the time required for a scrubber system to complete its objective.

4) Fundamentals of Estimating Leak Rates

One of the major errors usually committed when calculating leak rates from one-ton container, railcars, and/or bulk storage tanks is trying to estimate the relevance of the physical dimensions of a given leak. When a guillotine break in a one-inch chlorine header is mentioned, this dimension is used as the area of the chlorine leak. In reality the chlorine liquid has to pass through a long series of restrictions to the flow. These are: the 22 inches of 1/2-inch tubing inside the ton container, the ton container shutoff valve, the auxiliary container valve, 4 ft of 9/32-inch inside diameter flexible copper tubing (maybe an auxiliary header valve), and a header valve.

There is no method of quantifying the restrictions in terms of making it possible to calculate the chlorine leak rate from a broken or ruptured one inch-diameter pipe. The only possible way that this problem can be solved is by simulating such a leak. Such a simulation was performed at the EBMUD, Oakland, California WWTP in the early 1950s - but for an entirely different reason. Operators needed to know the max. liquid chlorine withdrawal rate from a single one-ton container in order to verify the necessity to go to bulk storage.

The amount of liquid chlorine that the one-ton container could deliver to three 6000 lb/day chlorinator was only 10,200 lb/day with a 45-lb pressure drop between the container and the chlorinator. Converting this flow rate to the pressure drop due to a header rupture, assuming a worst case of 120 psi pressure drop, was only 11.4 lb/min. or 16,416 lb/day as described previously. The same approach has to be made for the case of non insulated bulk storage tanks and insulated railcars. These restrictions limit the ability to calculate major leak patterns. The only way to arrive at reasonable leak flow rates is to simulate a leak on site and use scales for actual weight loss.