Environmental regulations continue to be a driving force for change in the way storage tanks are managed by the petroleum industry. Though recent regulations don't affect the main characteristics of floating roofs for storage tanks, they do impact details such as rim seals.

The US Environmental Protection Agency (EPA) in the January 14, 2000, edition of the Federal Register announced a regulation that will require industry action, says Terry Gallagher, product design manager for Chicago Bridge & Iron Company, Plainfield, Illinois.

The announcement reaffirms that slotted guide poles for petroleum and volatile organic liquid storage tanks are subject to the “no visible gap” clause of EPA Regulations on Standards of Performance for New Stationary Sources Title 40, Part 60, sub-parts K, Ka, and Kb. These sub-parts comprise the regulations for storage tanks.

The opening in the storage tank roof through which the guide pole passes, as well as the slots in the guide pole, constitutes “visible gaps: that must be maintained in a closed position at all times except when the device is in actual use.

“We expect that there will be a need to modify many of the external floating roof tanks currently in operation in the US refinery and terminal facilities over the next few years,” Gallagher says. “New equipment such as low-loss slotted guide poles is available. They have been available for several years and have proven performance records. In many cases, these components can be implemented economically and without removing the tank from product service.”

Floating-roof tanks use a significant amount of mechanical equipment to ensure safe and efficient operation, he says. Floating-roof primary and secondary seals, adjustable roof supports, mixers, swing-lines, special fitting details, and rolling ladders all must be designed to operate with the floating roof without compromising the integrity of the storage tank. Rolling ladders, roof drains, and fire foam systems must be designed and installed so that the floating roof remains in a balanced operating condition.

Tank Maintenance

Floating-roof tanks are not maintenance free, Gallagher adds, but with the proper roof selection and use of well-designed rim seals, fittings, and related equipment, the new floating roof can provide many years of trouble-free operation. Many tanks built more than 45 years ago are still in operation.

Periodic visual inspections of the floating roof will help eliminate surprises, Gallagher says. Operators should look for indications of product on the deck, unusual water ponding, deck ballooning, manhole covers out of place, shunts contacting the tank shell above the secondary seal, trash in the drain, out-of-place roof support and bleeder vent locks, and bird nests in foam chambers.

The initial cost of a floating-roof storage tank does not represent the true cost of operation, Gallagher notes. A life cycle cost analysis represents the true cost and should be the primary consideration when selecting a storage system. When properly designed and constructed, a floating-roof tank should provide from 15 to 30 years of service with a minimum level of preventive maintenance.

Though many floating-roof tanks provide long life, problems with the service history of internal floating tanks has prompted an American Petroleum Institute subcommittee on pressure vessels and tanks to revise the design standard for internal floating roofs (IFRs), says Robert Ferry, The TGB Partnership, Hillsborough, North Carolina.

“The recently approved revisions to the API 650 Appendix H standard will require all IFRs, including the skin-and-pontoon type, to be designed to support a uniformly distributed live load,” Ferry says. “This will effectively require the deck leg attachment to be designed to support a significantly greater load than the 500 pounds previously specified.”

At the 2000 Independent Liquid Terminals Association annual meeting, Ferry updated a presentation he made at a 1987 ILTA session suggesting that design specifications developed by Colonial Pipeline Company appeared to be adequate for an acceptable service life of a skin-and-pontoon IFR. The paper cited as a case history an IFR that had been in service for eight years at that time.

In the early 1960s, Colonial had steel external floating roofs in their tank farms, Ferry says. By 1978, Colonial had evaluated several roof replacement options and devised a replacement program. An aluminum geodesic dome roof with an aluminum internal floating deck was adopted as the company's standard roof design for roof replacements and new tanks.

Rigorous Tank Duty

One of the aluminum internal floating roof designs installed by Colonial early in the company's replacement program was a skin-and-pontoon type manufactured by Conservatek. Colonial installed the system in a 180-ft-diameter tank built in 1979 to store distillate products.

“This tank is filled and emptied through a 36-inch line opening into a center sump,” Ferry says. “A deflector plate mounted above the sump is designed to prevent the incoming flow from surging upward into the floating roof deck, but there is no diffuser to reduce the speed of the flow. Incoming liquid directed along the tank floor by the deflector plate rushes outward until it meets the tank shell. Thus, the deck support legs can be knocked about by the torrent of incoming product.”

Besides that abuse, the deck support legs also are subject to being pushed laterally when they land on the sloped bottom, he added. The slope of the floor is such that the legs near the center are nearly two feet longer than the legs near the tank shell.

In the 20 years since the tank was put in service, the floating roof has had 1,000 to 2,000 landings, Ferry says. The repeated exposure of the space under the landed floating roof to moist air created an extremely corrosive environment that eventually compromised the integrity of the tank.

Design Tests

In the summer of 1999, the tank was taken out of service for repairs. The annular ring of the tank bottom was to be replaced, and the entire bottom and the lower portion of the tank shell were to be painted with a protective coating. Colonial determined that these tank repairs could be performed without removing the internal floating roof.

“The floating roof has more than 100 pontoon end caps linking its structural frame, and the floating roof was inspected after the tank was taken out of service to repair the steel bottom. There was not a single crack or failure of any type in any of the pontoon end caps.”

Colonial specified that the deck leg attachment to the pontoon end cap be more durable than API 650 Appendix H required, Ferry concludes.

Puncture Test

Results of another case study on an internal floating roof come from Kenneth Erdmann, engineering manager of Matrix Service Inc, Tulsa, Oklahoma. The case study looks at a full-scale puncture test on the internal steel pontoon-floating roof. The floating roof consists of a single steel deck surrounded by a pontoon. The pontoon ring is divided into several compartments by radial bulkheads.

API 650 Appendix H requires any component of the floating roof in contact with liquid or vapor to be a minimum 3/16 inch thick, Erdmann notes. Steel not in contact with the liquid or vapor must be a minimum 0.095 inch thick.

A common regulatory instrument that prompts industry to use floating roofs is the Code of Federal Regulations, Title 40, Part 60, sub-part K, Ka, and Kb, he says. Sub-part K initially was enacted in June 1973 and regulates the use of floating roofs in tanks that contain 40,000 gallons or more of product with a true vapor pressure equal to or greater than 1.5 pounds per square inch atmosphere but no greater than 11.1 psia.

Sub-part Ka was introduced in May 1978. It requires either an external pontoon or double deck floating roof, a fixed roof with an internal floating roof cover, or a vapor recovery system that reduces emission by at least 95%. The floating roofs are required to have continuous closure devices around their circumference. External floating roofs must have a double rim seal.

Sub-part Kb came into effect in July 1984. It applies to tanks from 20,000 gallons to 40,000 gallons that contain product with a true vapor pressure from 4 psia to 11.1 psia. Sub-part Kb is similar to sub-part Ka, with a few additions.

Design Requirements

API 650 Appendix H describes the types of loads that internal floating roofs must be designed to withstand. All floating roofs must be capable of supporting the weight of two people (500 pounds) placed in one sq-ft area anywhere on the floating roof. All internal roofs are required to remain buoyant while supporting twice their dead weight.

The legs of internal floating roofs must support the roof's dead weight plus 12.5 pounds per square foot of live load unless means of draining accumulated liquid off the roof is available.

In addition, each type of floating roof must be capable of remaining buoyant in various punctured conditions. Bulkheaded, double-deck, and hybrid floating roofs must be capable of floating with any two compartments punctured. Pontoon floating roofs must remain afloat with the center deck and any two adjacent pontoon compartments punctured. “In many cases, these punctured compartment design requirements will be the controlling load condition that drives the design of the floating roof,” Erdmann says.

Two factors are critical to the ability of a pontoon-floating roof to withstand the punctured condition specified in API 650. First, the pontoon ring must be capable of supporting the loads being applied to it without being overstressed. Secondly, in addition to supporting the applied loads, the floating roof's deflection must be limited to prevent the product surface from entering other sealed pontoons, resulting in eventual flooding and sinking of the floating roof.

In the punctured condition, the pontoon-floating roof will deflect in three basic ways, Erdmann says. First, the center deck will deflect. It is no longer uniformly supported but is hanging from the pontoon ring around the perimeter. Secondly, the pontoon ring will list to one side as a result of uneven support when two adjacent pontoon compartments are punctured. Thirdly, the pontoon ring will undergo a deformation as a result of no longer having uniform buoyancy support.

Errors in the shape of the punctured deck will not affect the vertical loads applied to the pontoon ring, which must still support the weight of the deck, regardless of the deck's shape. The shape of the deck will affect the radial load that is applied to the pontoon, but this will primarily result in changes to the structural status of the pontoon and will have minimal effect on the vertical deflections of the pontoon ring.

“To properly design a pontoon floating roof, the various types of deflections must be closely modeled,” Erdmann says. “Errors in the calculation of these deflections will directly affect the accuracy of determining the surface level of the product with respect to the pontoon ring and pontoon manways. Since the only remaining buoyancy in the floating roof is in the pontoon ring, errors in the deflected shape of this ring could result in unacceptable product heights when in the punctured condition.”

Erdmann evaluates several methods of analyzing pontoon roof deflection. He compares these analysis techniques with the case study on a 123-ft-diameter pontoon-floating roof placed in the punctured condition.

Methods of Analysis

One method of analysis used in the petroleum industry is simply to require the pontoon ring to make up 30% of the area of the floating roof. The 30% method is based purely on area and would be insufficient for floating roofs that had corrosion allowances that would affect the weight of the roof. When compared to more precise analysis techniques, this method tends to provide non-conservative calculations for small tanks and very conservative results for large tanks.

Another simple approach is to assume the floating roof is rigid and all deflections are vertical (no tilting or local deflections). This method typically uses a 1.25 flexibility factor to account for additional deflection from the roof tilting and from local deformations. It provides a fairly consistent design, but when the floating roof has corrosion allowance applied the 1.25 flexibility factor is no longer adequate.

A third method assumes rigid planar tilting of the floating roof. This method typically assumes that local deflections are not significant. It appears to be over-conservative on smaller floating roofs and non-conservative on larger floating roofs, according to Erdmann.

Curved Beam Analyses

A popular analysis technique is finite element analysis (FEA). It models a floating-roof structure as a mesh of connected elements. STAAD III is the software used for FEA. “This method of analysis produced a deflected shape that was very close to the measured deflected shape of the full scale punctured pontoon roof that was tested,” Erdmann says. “However, the maximum deflection of the outer rim appeared to be slightly conservative.”

Another method of evaluating punctured pontoon roofs is to model the pontoon ring as a beam on an elastic foundation. This analysis was performed, and the results very closely matched the measurements of the tested pontoon roof. Though the deflected shape of the tested floating roof was closely matched, the maximum deflection was conservative.

“One primary consideration is how accurately does each method calculate the deflection of the pontoon ring in the punctured area,” he says. “The curved beam and the STAAD III FEA analysis methods were both slightly conservative at the point of maximum deflection. The curved beam provided the closest approximation of the deflected shape.

“The FEA and curved beam methods provide sufficient results. In the future, additional full-scale punctured tests should be performed to validate the results of these two methods of analysis over a wide range of tank diameters.”

First Floating-Roof Tanks

Since the first floating-roof storage tanks were developed in the 1920s, it has been the most widely used system for storage of volatile petroleum products. Many design changes and improvements have been made since then.

The first floating-roof storage tank was demonstrated by Chicago Bridge & Iron Company in 1923 as a means to reduce product evaporation, and that is still the reason floating-roof tanks are in use. Another obvious reason is safety. Crude and refined petroleum products are volatile and will readily evaporate at normal storage and handling conditions, producing vapors that are combustible over a range of concentrations with air.

“The floating roof originally was developed for use in open top tanks,” Gallagher says. “As product, weather, and environmental concerns became more of an issue, floating-roof designs were adapted for use in a tank with a fixed cover. The most advanced type of installation is used in applications that require absolute control of all emissions from the storage tank.”

New equipment developments have reduced overall environmental impact from a floating-roof tank. Thus, recent developments are important from a regulatory compliance standpoint.

A floating-roof design can range from a simple pan roof to the complicated structure of a double-deck floating roof. Design selection depends on many of the same factors used previously to select the tank configuration — product characteristics, site weather conditions, system operating requirements, storage capacity, and required through-put.

“The first floating-roof designs were stiffened pans,” Gallagher says. “In 1923 when the first floating roof was tested, the only application was as an external floating roof. The roof deck was sloped to the center for water drainage and to permit vapors to pass from under the deck to the perimeter rim space.”

The need for improved rim seals, roof drains, manual and automatic bleeder vents, rolling ladders, and other tank equipment has never stopped.

Floating roof systems are complex structures when it comes to design, analysis, and construction, Gallagher says. Many of the products currently in use were developed before the computer. Design work now may be completed using computer analysis and verified data obtained from the original field test programs. However, designs of a floating-roof structure still require a rigorous analysis.

“Design conditions for an external floating roof differ significantly from those that impact an internal floating roof,” Gallagher says. “A major difference is the added requirement that the external floating roof design must accommodate loads due to varying weather conditions. Other differences include floating roof access and the design of product emission control hardware. However, many of the basic design principles remain unchanged.”

Basic Principles

Gallagher goes on to list basic design principles for external and internal floating roofs: They must be designed to remain buoyant on liquid with a specific gravity of 0.7 under specified design operating conditions. They should maintain full liquid contact to minimize evaporation and reduce product side corrosion.

Floating roofs and their accessories must accept the full range of roof movements during unattended operation from low operating level to the maximum design liquid level. They should keep product emissions to a minimum. They should be designed to provide an extended service life with minimum maintenance.

Additional requirements for external floating roofs take weather conditions into consideration. API 650 Appendix C requires that an external floating roof must have sufficient buoyancy to remain floating on liquid with a specific gravity of 0.7 when the single-deck roof drain is inoperative and the roof is exposed to 10 inches (250 mm) rainfall in 24 hours. Double-deck roofs may be designed for a lesser amount provided the roof is equipped with emergency roof drains to prevent over-accumulation of water.

Detail Differences

Design and construction details used in double-deck floating roofs are important. As with the variations in design for single-deck roofs, different manufacturer details can make a difference in the mechanical stability and overall performance of a double-deck floating roof.

More design options are available for internal floating-roof tanks. However, design options for a fixed-roof tank are strictly limited. Internal floating roofs are available in conventional welded steel construction, but also may be constructed from lightweight aluminum or composite fiberglass. Specific design criteria for internal floating roofs are presented in API 650, Appendix H — Internal Floating-Roofs.

“External floating roofs can be — and have been — used as internal floating roofs,” Gallagher says. “In general, this occurs when a fixed roof is added to an existing external floating-roof tank. For new construction, the internal floating roof is designed as such and will have slight differences in construction details in response to reduced loading conditions.”

Roof Selection

Selection of a floating roof design from the various types available should be done carefully. Petroleum storage tanks serve many purposes in a refinery, marine terminal, or pipeline terminal environment. Tank operations, the type and quality of product being stored, and site-specific conditions must be considered.

“Selection of the floating roof design and associated equipment can ensure an excellent return on the investment, or it can provide a costly maintenance headache for many years,” Gallagher says.

Mechanical stability of the floating roof is probably the most important parameter of any floating roof. Any load acting on the floating roof that is not balanced will force the roof to float over to one side of the tank or, in the most severe cases, the roof will tilt due to unbalanced loads.

Product factors that should be considered when choosing a roof include product composition and chemical stability, specific gravity, viscosity, and wax content.

“Increased storage and handling risks are associated with new crude oils that are modified at the well or mine site to ensure that the end product can be safely transported,” Gallagher says. “All product characteristics can be affected by site ambient conditions. Variations in ambient temperature should be considered when reviewing the potential design impact on product true vapor pressure and product viscosity.”

Design Considerations

Product composition is important because of the potential effects certain chemicals will have on the service life of some of the equipment associated with the floating roof. The greatest potential for problems is with non-metallic materials used with the floating-roof perimeter seals. Some chemicals can destroy commonly used seal materials after only a few weeks of exposure.

Other considerations when choosing a floating roof are product specific gravity and true vapor pressure. In most cases, floating roofs are designed for a specific gravity of 0.7. It is unusual to find a product with a lighter specific gravity but it is possible, Gallagher says.

It is important that the roof designer know the minimum specific gravity since this will determine how far a given roof will float into the product. The distance from the product surface to the top of the floating roof deck is one measure of overall roof stability. This information also is used to check the relative position of the rim seal with respect to the liquid surface.

Product true vapor pressure (TVP) is the single most critical design parameter when selecting the type of floating roof. Most current environmental regulations limit the product TVP to less than 11.1 psia. As the TVP increases above 11 to 12 psia, daily heating of the product under a center deck will produce enough vapors to balloon the deck. It is common for these vapors to condense during the cooler evening hours, allowing the roof to resume a normal flat shape.

“It must be emphasized that if the tank is in an area of significant rainfall, ballooning of a single-deck roof may not permit normal water drainage to the primary roof drain,” Gallagher says. “An unbalanced load can quickly be developed that can sink the floating roof.”

A secondary consideration with the high vapor pressure products is that with increasing TVP, the overall effectiveness of any floating roof design is reduced. More evaporation will occur, and this vapor will escape to the atmosphere above the floating roof. Air pollution and the risk for a fire are increased under these conditions.

“If the TVP is going to be high, approaching 11.1 psia or greater, consideration should be given to using a double-deck floating roof,” he says. “A double deck will maintain its ability to drain water while containing some amount of product vapor. The double-deck roof can help reduce vapors from product heating due to the insulating effect the design provides to the product surface. Properly designed, the double-deck floating roof can be the most stable design available.”

Conditions upstream from the tank farm should be considered when selecting the type of floating roof. Process flow rates must be identified, as well as any process that might result in large quantities of vapor being mixed with the product fill stream.

“Each of these conditions can be addressed in the roof selection process and during the detailed design of the storage tank,” Gallagher says. “If the application is an internal floating roof and there is a possibility that significant quantities of vapor may be included in the fill stream, a pan or lightweight floating roof should not be considered. The minimum suggested roof design would be a pontoon single-deck roof, or in extreme cases, a double-deck roof equipped with surface discretionary vents.”

Floating-roof storage tanks and associated equipment must be kept up to date and a working knowledge of the equipment maintained by refiner and terminal operator. New roof and tank designs are being developed for more efficient product storage.

“Conventional floating-roof tanks operate with a low landing position of about three feet (less than one meter) and require five feet (1.5 meters) or more vertical clearance for the floating roof,” Gallagher says. “New designs and equipment options are available that would permit a floating roof to land almost on the tank bottom, increasing the storage and operating efficiency of the system. However, some of these options need further development and testing.”