7 1/2-Hour Tank Farm Fire Confined To 2 Tanks Burning Upon Arrival
In a 7 1/2-hour battle, fire fighters held a tank farm fire to the two tanks of gasoline involved upon arrival in Stockton, Calif. We used some 6000 gallons of foam concentrate, most of it AFFF, and two pumper relays to obtain sufficient water for protecting exposures. These were major factors in limiting the loss estimated at $400,000.
The Shell Oil Company tank farm, a 30-year-old facility, had eight storage tanks ranging in capacity from 155,000 to 900,000 gallons. Six of these held gasoline and the other two, diesel fuel. In addition, there were small additive tanks. The tank farm, in an area recently annexed by Stockton, had earth dikes and fire protection equipment that met minimum code standards. There was a 6-inch dead-end water main with three hydrants on the street side of the tank farm.
Although no one was on the property, the tank farm was receiving fuel by pipeline when the fire started at a 155,000-gallon gasoline tank, the oldest one, in the center of the tank farm last October 22. Fuel continued to be pumped through the pipeline during the early stages of the fire.
We believe that flammable vapors flowed from a pressurized rupture or overflow of the 155,000-gallon gasoline tank or a failure of plumbing connected to this tank. The probable cause was human failure or equipment failure or both.
Cause of fire
From the tank farm, the ground slopes toward a building used by the officer of the day at the nearby Naval Annex entrance. There was fog and no measurable wind when gasoline vapors flowed southwest about 700 feet to this building. A hot water heater in the building is believed to have been the ignition source as there was evidence of burning in the room containing the heater. In flashing back to the source of the spill, the burning vapors started a 270-foot-long grass fire.
The initial alarm was at 7:45 a.m., and Engine 1 called for a second alarm upon arrival at 7:48 a.m. The fire immediately gained such magnitude that Battalion Chief Robert Donis, who had heard the first-in report of Engine 1, requested a third alarm en route.
As I answered the fire phone at my home to be told of the incident, I heard Donis radio for a fourth alarm at 7:52 a.m. Upon hearing this, I asked the dispatcher for immediate activation of the county mutual aid agreement to get as much foam and foam equipment as we might need to handle the fire. As the fire intensified and the exposure problem increased, we extended the call for foam supplies to a radius of about 50 miles by contacting the regional command of the California Office of Emergency Services.
The extent of the fire in the southeast section of the diked area indicated a considerable spill centered around the 155,000-gallon and 600,000-gallon tanks of unleaded gasoline. The fire in the smaller tank appeared to be an open-top problem from the outset. The upper section of this vessel was heavily damaged and distorted early in the fire.
The initial attack consisted of master stream appliances on the south and east approaches while AFFF nozzles (and some protein) advanced along catwalks toward the valving on the south side of the smaller tank. At about 8:15 a.m., an all-out attack was attempted with some additional foam apparatus after we had a pretty stable water supply directed onto the exposed tanks. At this point, there were several small leaks visible at the valving attached to the center tank and a rim and vent fire on the 600,000 (full) unit.
One crew took an AFFF nozzle to the north side of the tank and succeeded in closing all the valves there. A second crew moved to the valve cluster on the south side of the tank and appeared to be effectively stopping the leaks when a vertical 1-inch pipe connected to the base of the tank broke about 5 feet above the tank bottom.
The resultant leak and flame spread forced everyone back to the nearest dike and the heavy water curtains were again relied on for cooling the involved tanks and exposures. It is presumed that the weight and stress of the distorted tank top was responsible for this pipe failure. The 1-inch line was of unknown age and it had been subjected to extreme temperature changes immediately prior to breaking.
The new fuel leak, the water buildup in the diked area, and an increase in wind velocity—added to the inaccessible location of the problem tank and the proximity of the surrounding tanks— indicated that we were in for a long period of sustained operations.
We faced a number of other problems. There was no room in other tanks for fuel transfer. The burn line on the open tank indicated the total burn time for the contents to be 2200-2300 hours— perhaps four hours less to get the fluid level below the pipe break.
Plant engineers eliminated any immediate method of subsurface foam injection. The nearest valves were 400 to 500 barrels of fuel away. To get foam into the tank by this method meant adding this amount of fuel to the severely damaged tank. A seam failure on the old tank would have drastically increased our problem, plus putting a like quantity of foam into the line to displace the trapped gasoline.
We decided to hold with a control action at this point and accumulate an effective foam supply rather than waste it in a small-scale application.
A substantial increase in wind velocity rendered the foam apparatus on the south and east ineffective, except for assisting with securing impounded water/gasoline buildup inside the dikes.
We were able to obtain only about 3000 gpm from the 6-inch dead-end main. We had to set up two teams of relay pumpers to double that amount to effectively control the exposure problems. We used five 1500-gpm pumpers for this operation. One source pumper drafted from a shipping channel 800 feet away, and another pumped from a hydrant, also 800 feet away, on a fire main in the warehouse area.
Please turn page
It was difficult to direct foam streams into the tank from apparatus stationed upwind—too far for aerial ladder operation and very hazardous on adjacent tank ladders and landings for hand line operations.
Fuel displaced by water
Our only option appeared to be to fill the tank with enough foam/water to raise the level of the gasoline above the broken pipe. We eventually did this, and when it happened—after 7 hours and 35 minutes—it seemed like a miracle. It was particularly frustrating to be unable to determine our progress in replacing the fuel in the bottom of the tank with water.
It took a lot of hard work and determination to accomplish this objective, but the silence in that tank farm was beautiful after we plugged the broken plumbing with tapered dowels. We ultimately placed two crews on the upper landings and catwalks and used large bore foam nozzles to direct streams into the open tank top. I feel that this was highly effective, but it was difficult to provide shielding and protection lines for the men exposed in positions so close to the involved tank. The dome and rim fires in the large tank presented no real problem once we extinguished the open tank fire.
There were several small fuel leaks in the valving system because of fire damage and it was necessary to maintain an all-night fire watch with charged foam lines. The foam blanket on the spill was replenished at various times and there was no reignition.
This fire reinforced our confidence in the training and fire fighting techniques that have been developed to control petroleum fires.
We are now concentrating on applying some of the knowledge gained from this fire toward preventing a similar incident. Some of these improvements consist of plumbing to enable subsurface foam application, continuous manning of fuel plants during fuel transfers, early warning systems, automatic shutdown devices at pumping stations (activated at the receiving facility), adequate local foam supplies and equipment for controlling a tank farm emergency, access roads on four sides, and acceptable methods of handling waste and spill runoff. There are more safeguards to consider, but a few of the above would go a long way toward establishing an economical and effective petroleum plant fire prevention program.
Water drafted from dike area
One last point worth mentioning is that it became necessary to set up a 1500-gpm pumper to draft from the bottom of the collecting sump to prevent overflowing the diked area. Some of the water was applied to the tanks for cooling and the remainder was piped into a hastily diked impounding area in an adjacent field.
Mutual aid response with foam equipment, manpower and foam supply was excellent. Total apparatus used was as follows:
From the Stockton Fire Department—nine pumpers, two aerial ladders, two paramedic units and five auxiliary vehicles.
Mutual aid response—three foam crash units, one each from Mather, Modesto City and Stockton Metro Airport; five pumpers, two from the Naval Command and one each from the Rancho Cordova Fire Department, Shell Oil plant at Martinez, and CID state pumper.
In addition, there were various pickup trucks, tractors, small foam apparatus and assorted vehicles of all sizes—a probable total of 16.
The Stockton Fire Department had 125 men at the fire and mutual aid agencies had 40 for a total of 175 men.
The salvage fuel consisted of approximately 100,000 gallons of unleaded gasoline from the main target tank and the surrounding additive tanks.