Aging Bulk Fuel Terminals

Feb. 13, 2009
Experience at Newark Liberty heightens challenges facing many storage locations

Aging transportation infrastructure is looming as a potential near-term problem at many transportation hubs. Concrete forms a major component of many transportation facilities, a large portion of which may be more than 20 years old. This article focuses on what may be a widespread but yet largely undiscovered problem at bulk fuel terminals utilizing concrete containment structures for aboveground petroleum storage tanks. It focuses on an upgrade project currently underway at the bulk fuel facility at Newark Liberty International Airport, which was originally constructed in the 1970s.

At Newark, jet fuel is received by pipeline and stored in aboveground tanks. Delivery to aircraft is via tankers loaded at truck racks in the bulk fuel farm or by a series of underground hydrant pipelines to fueling positions adjacent to aircraft gates where hydrant trucks meter fuel into aircraft.

Each of the 24 large vertical storage tanks at this facility has a secondary containment system which includes a concrete annular dike floor approximately four feet below grade, and a circular dike ringwall approximately 16 feet tall. The lower four feet of the ringwall is constructed from reinforced concrete, which is below grade (outside surface only) and the remainder of the upper dike wall is fabricated from rolled steel plates welded together. (See Photo 1 below depicting the inside of the dike wall, with the lower portion being concrete like the dike floor, and the upper portion being painted steel.)

The lower concrete ringwall supports and stabilizes the upper steel plate ringwall, which is connected to the concrete ringwall via a cast in place steel channel. The concrete/steel dikes were designed to provide liquid tight secondary containment for the primary tanks.

Deterioration
In 2005 hairline cracks started to appear on the exposed inside surface of the concrete ringwalls and a study was performed to determine the extent of the concrete deterioration. The resulting report indicated that many of the concrete dike ringwalls exhibited internal failure of the concrete.

Alternative repair options ranged in cost from $20,000 per dike (repair voids in the concrete by local patching with epoxy mortar) to $450,000 per dike (replacement of all concrete and rebar in the dike ringwall). The study recommendations were not implemented at that time.

Due to more extensive cracks and spalling, a new dike assessment study was carried out in 2007 to determine the extent of further concrete deterioration, and to develop a repair methodology, budget and schedule. A consultant specializing in the preservation of concrete structures was retained. Their report indicated that the extent of concrete ringwall deterioration was much more extensive than indicated in the 2005 report. This was due either to additional deterioration between 2005 and 2007, or differing inspection methodologies used in the 2005 and 2007 reports.

The data collected during the 2007 dike assessment survey indicated that various factors, outlined below, contributed to the presence of severe internal deterioration of the concrete dike ringwalls.

1) The root cause of the concrete deterioration was unequal thermal expansion stresses caused by solar heating of the steel and concrete ringwalls as the sun passes from east to west during daytime. This was exacerbated by several factors:

  • The concrete dike’s temperature changes are moderated by the damp backfill in contact with its exterior surface, whereas the steel ringwall is exposed to solar heating and air temperature on both surfaces.
  • The upper steel ringwall is painted dark brown and will absorb more solar radiation that the lighter concrete ringwall.
  • The coefficient of expansion of steel and concrete are significantly different.
  • Concrete is weak in tension, compared to steel.

Therefore, daytime solar preferential heating of the dark steel ringwall, and nighttime radiation cooling produce a diurnal expansion/contraction cycle that applied high-tension stress to the top eight inches of the concrete ringwall, where the steel ringwall channel is cast into the concrete. This led to the stress crack failures in the concrete, approximately eight inches below the top of the wall, adjacent to the lower flange of the embedded steel channel (see Photo 2).

2) Inadequate maintenance of caulking of the steel-to-concrete ringwall joint allowed the penetration of stormwater into the top of the ringwall concrete. Under winter conditions, this led to freeze/thaw cycles which caused vertical cracks in the concrete, parallel to the inner face of the ringwall. Over time these cracks have propagated approximately three feet down from the top of the wall.

3) The original concrete was of low quality and remarkably non-homogeneous, even within the same dike. According to the petrographic analysis, the original concrete was defective in that:

  • Compressive strength ranged from 2,350 to 4,550 psi, and much of it was below the required 4,000 psi;
  • Entrained air was low;
  • Entrapped air was high; and
  • Water cement ratio was high.

The aggregates contained 4 percent chert which caused an alkali–silica reaction. The interior of the 10-inch thick ringwall was essentially rotted out and only an outer one-inch thick layer of sound concrete on the inner and outer surfaces was holding the wall together.

4) The lack of waterproofing on the exterior of the dike ringwall allowed water from the surrounding damp clay soil to saturate the concrete and accelerate the internal disintegration processes via the freeze thaw mechanism and acceleration of the alkali–aggregate reaction.

The report characterized the extent of the deterioration and indicated that all the concrete dikes required immediate attention to remain liquid-tight and stay in service plus a major repair or replacement within one to two years.

Making repairs
Operational constraints would not allow taking any tanks off line for the dike repairs; thus, the project was separated into two phases:

  • Immediate concrete repairs to restore the dikes to liquid tight condition in 2007;
  • Planned major upgrades of the 24 tanks between 2007 and 2009 to extend the service life of the dikes for another 20 years.

The immediate repairs were determined to be:

  • Re-caulk dike floor to ringwall joints;
  • Re-caulk interior ringwall top to steel shell joint (minimize water intrusion for freeze/thaw);
  • Route and caulk new cracks to stabilize the ringwall;
  • Chip and trowel mortar to restore adequate strength; and
  • Install flashing at exterior ringwall top to steel shell joint to limit water intrusion.

The immediate repairs were implemented and a design study was performed to determine the most cost-effective, long-term upgrade for the containment dike ringwalls.

The objective was to determine the most cost-effective way to extend the service life of the existing 24 containment dikes by 20 years, without taking any of the dikes or tanks off line during the upgrade project.

The conceptual solutions suitable for upgrading the existing dikes focused on the joint between the lower concrete ringwall and the upper steel wall. If the entire ringwall were to be replaced, this problematic joint could be eliminated by designing an all-steel ringwall or an all-concrete ringwall.

Another option was to redesign this steel-to-concrete joint so that differential expansion of the concrete and steel walls could not lead to failure of the concrete portion of the ringwall due to differential expansion stresses.

Single Material Dike Wall
A single material containment dike wall would eliminate the root cause of the failure of the steel/concrete dike ringwall. There are two ways that this could be accomplished:

1) Replacement of the existing 17-foot high concrete/steel material containment dike with an all-concrete dike wall or an all-steel dike wall would be very expensive as it would require demolition of the existing steel shell and the concrete ringwall, and would put the contained tank off line for several weeks, due to the lack of containment. Because the bulk fuel facility was operating at rated capacity, dikes and/or storage tanks could not be taken off-line for a period of several weeks. Therefore this solution was eliminated based upon high capital cost and unacceptable operational impact.

2) A hybrid solution was developed to result in a single material (steel) dike wall without compromising the function of the dike during the upgrade project. This was done by retrofitting a structural steel wall inside the four-foot interior above-grade portion of the concrete ringwall using a steel shell “downward extension” connected to the bottom of the existing upper steel ringwall and supported by the existing concrete dike floor.

As indicated in Photo 3, dual steel angles were shop-rolled to the inner ringwall radius in 30-foot sections, and welded to the bottom of the existing steel ringwall. A matching rolled angle was bolted to the dike floor, using a formed-in-place gasket and mounting studs epoxied into the concrete dike floor. (See Photo 4, which depicts the steel floor angle in place.)

The next step was to weld rolled steel plates approximately four feet high to fill the vertical gap between the upper and lower rolled angles that were already attached to the existing dike structure.

Then one quarter-inch thick plates, 30-feet long (precut and rolled to the ringwall radius) were hoisted over the dike wall and supported vertically between the new top angle and bottom angle and welded into position (as indicated in Photo 5). This process was repeated until the entire dike interior was lined with steel plates (See Photo 7).

The plates were pre-primed prior to installation and painted in place with a primer and top coat.

This containment dike upgrade was designed to be retrofitted without taking the tanks or dikes off line as the existing concrete ringwall did not have to be demolished or repaired as part of this solution. This steel dike shell extension was designed with one inch of clearance from the existing concrete wall interior and is structurally independent of the concrete ringwall. Therefore, continued deterioration of the concrete ringwall will not impact the structural integrity of the steel dike shell extension. The backside of the new steel dike wall is provided with cathodic protection as the existing upper steel dike shells are connected into the facility’s cathodic protection grid.

The upgrade was designed to meet ASCE 7-02 Minimum Design Loads for Buildings and Other Structures, which is the basis of current IBC building code.

Conclusions
Aging concrete containment structures should be inspected by concrete structural inspectors on a regular basis, as interior degradation of concrete containment structures can go undetected for years until the surface layers start to crack. At that point there may be very little structural integrity remaining in the concrete which makes repair very difficult and expensive, as internal rebar is frequently severely corroded and it must be replaced along with the concrete.

Local patching even with modified epoxy mortars typically may not have a sufficiently long service life since the originate concrete may continue to degrade around the new patch.

Two prototype containment dike upgrades were completed at Newark in June 2007 and a total of 12 tank dikes were upgraded by the end of 2008. The remaining 12 dikes will be completed by mid 2009. The project is on schedule and budget and no tanks have been taken off-line during the upgrade project. The projected service life of the upgraded containment dikes has been estimated in excess of 20 years.

During the 2007 investigation it became apparent that the alkali-silica reaction played a major role in the internal deterioration of the concrete. In many concretes, the aggregates are chemically inert. However, some aggregates react with the alkali hydroxides in cement, causing expansion and cracking over a period of many years. The alkali-aggregate reaction has two forms: alkali-silica reaction and alkali-carbonate reaction.

The alkali-silica reaction is of more concern as aggregates containing silica reactive materials are more common. The aggregates containing certain types of silica will react with alkali hydroxide in concrete to form a gel that swells as it absorbs water from the surrounding cement paste or the environment. These gels can swell and cause sufficient expansive pressure to damage concrete.

Concrete that is exposed to damp conditions (such as dike floors) will deteriorate more quickly. Typical indicators of this type of degradation are random cracking and spalling. Petrographic examination of the concrete can determine if this reaction is present.

In new structures, the alkali-silica reaction can be controlled by use of certain concrete additives in the mix design. Since this reaction was not widely known several decades ago when much of the existing concrete tank farm infrastructure was installed, its occurrence in older concrete structures is largely determined by the quality of aggregate that was used in the concrete mix.

This reaction may start on the damp buried side of an older concrete containment structure and progress through the center to the exposed face. Once the telltale random cracking is evident at the exposed face, there is a good chance that the internal degradation is so complete that the structure cannot be patched and complete replacement of defective concrete is required. The alkaline nature of this degradation also corrodes internal rebar which was not typically epoxy coated on older concrete structures (see Photo 2).

About the Author
Dennis Eryou is a professional engineer in private practice with 25 years experience in aviation fueling systems design and environmental issues. He is a registered PE in 21 states, and can be reached at [email protected].