Tank

Fabricating RP Tank

Classical engineering stress analysis shows that hoop stress (stress trying to push out the ends of the tank) is twice that of longitudinal stress. To build a tank of conventional materials (steel, aluminum, etc.) requires the designer to use sufficient materials to resist the hoop stresses that results in unused strength in the longitudinal direction. In RP, however, the designer specifies a laminate that has twice as many fibers in the hoop direction as in the longitudinal direction.

An example is a tank 0.9 m (3 ft) in diameter and 1.8 m (6 ft) long with semi-spherical ends. Such a tank's stress calculations (excluding the weight of both the product contained in it, and the support for the tank) are represented by the formulas:

where s = stress, p = pressure, d = diameter, and t = thickness.

Tensile stresses are critical in tank design. The designer can assume the pressure in this application will not exceed 100 psi (700 Pa) and selects a safety factor of 5. The stress must be known so that the thickness can be determined. The stress or the strength of the final laminate is derived from the makeup and proportions of the resin, mat, and continuous fibers in the RP composite material.

Representative panels must be made and tested, with the developed tensile stress values dien used in the formula. Thus, the calculated tank thickness and method of lay-up or construction can be determined based on:

and:

s = pd/4t for the end and longitudinal stresses

where:

20 x 103

2X 5

ti = 1/2 fh = 0.225 in. (or the same thickness with half the load or stress th = hoop thickness fi = longitudinal thickness sh = hoop stress s/ = longitudinal stress sh = 20 x 103 psi (140 MPa)

safety factor = 5

If the stress values had been developed from a laminate of alternating plies of woven roving and mat, the lay-up plan would include sufficient plies to make 1 cm (0.40 in.) or about four plies of woven roving and three plies of 460 g/m2 (11/2 oz) mat. However, the laminate would be too strong axially. To achieve a laminate with 2 to 1 hoop to axial strength, one would have to carefully specify the fibers in those two right angle directions, or filament wind the tank so that the vector sum of the helical wraps would give a value of 2 (hoop) and 1 (axial), or wrap of approximately 54° from the axial.

Another alternative would be to select a special fabric whose weave is 2 to 1, wrap to fill, and circumferentially wrap the cylindrical sections to the proper thickness, thus getting the required hoop and axial strengths with no extra, unnecessary strength in the axial (longitudinal) direction, as would inevitably be the case with a homogeneous metal tank.

As can be seen from the above, the design of RP products, while essentially similar to conventional design, does differ in that the materials are combined when the product is manufactured. The RP designer must consider how the load-bearing fibers are placed and ensure that they stay in the proper position during fabrication.

Underground Storage Tank

GFRP (glass fiber-TS polyester RP) underground tanks for storing gasoline and other materials have been in use worldwide since at least the 1950s. Experience with them initiated many tank standards for different materials. RPs provided much longer life than their steel counterparts. In fact, steel tanks previously had no "real life" or no requirement standards until the RPs entered the market. It has been estimated that more than 200,000 GFRP tanks were installed in the USA from 1960-1990. A previous study by the Steel Tank Institute (Lake Zurich, IL) reported 61% were of steel and 39% of GFRP. At present, at least 50% of all tanks are GFRP. This RP vs. steel debate escalated when the EPA gave service stations and fleet refueling areas 10 years to remove steel tanks that leaked.

Historically a Chicago service station documented the long life of RPs. A May 1963 installation remained leak tight and structurally sound when unearthed in May 1988. After testing the vessel, engineers buried it at another gas station. This tank was one of sixty developed by Amoco Chemical Co. It was fabricated in two semi-cylinder sections of glass fiber woven roving and chopped strand mat impregnated by an unsaturated isophthalic TS-polyester resin selected for its superior resistance to acids, alkalis, aromatics, solvents, and hydrocarbons. Two sections were bonded to each other and to end caps with RP lap joints. Today, the tanks are fabricated by using chopped glass fiber mixed with the isophthalic resin. This mixture is dropped from above onto a rotating steel mandrel. The glass-resin mix is sprayed to make the end caps.

Demand for this type of petroleum storage tank has grown rapidly as environmental regulations have become more stringent. Marina installation have taken advantage of these RP tanks. They permit for boat owners to purchase gasoline at the pier. Before they were installed, gasoline either had to be carried to the marina or purchased elsewhere, because of corrosive conditions underground for metal or other tanks, particularly ones next to salt water.

Standards require that today's underground tanks must last thirty or more years without undue maintenance. To meet these criteria, they must be able to maintain structural integrity and resist the corrosive effects of soil and gasoline, including gasoline that has been contaminated by moisture and soil. The tank just mentioned that was removed in 1991 met these requirements, but two steel tanks unearthed from the same site at that time failed to meet them. One was dusted with white metal oxide and the other showed signs of corrosion at the weld line. Rust had weakened this joint so much that it could be scraped away with a pocketknife. Tests and evaluations were conducted on the RP tank that had been in the ground for 25 years; tests were also conducted on similarly constructed tanks unearthed at 51 and 71 years that showed the RP tanks could more than meet the service requirements.

Prior to the development of the GFRP tanks, no standards were required for buried tanks such as loads or loading conditions, minimum depths of earth cover, or structural safety factors were available. At that time, sizes were 22,704 to 45,408 liter (6,000 to 12,000 gal), with a nominal width and height of 2.44 m (8 ft) for truck shipments to local gasoline stations. Standards have developed listing requirements for stored fluid type, environmental resistance, minimum earth cover, ground water submerged limits, and surface wheel load over tank. Increasing acceptance of buried GFRP tanks has widened the size range from 2,081 to 181,632 liter (550 to at least 48,000 gal) and the range in typical diameters from about 1.22 to 3.35 m (4 ft to at least 11 ft).

The tank configuradon is basically cylindrical, in order to provide the required design volumes within the established envelope of heights and widths. Length ranges are from 5.5 to 11 m (18 to 36 ft); they are well within pracdcal truck shipment limits. A circular shape is required to support the substantial internal and external fluid and earth pressures with good structural efficiency. Other considerations in selection of an efficient configuration are used.

With a vertical axis, tank underlay requires that much less land area than with axis horizontal, but very deep excavation is required where expensive ledge or ground water conditions will frequently be encountered. Both internal and external pressures are large, requiring a substantial increase in wall thickness and rib stiffness, compared with a horizontally-placed tank. With axis horizontal, maximum external and internal pressures do not vary with size (length), and can be resisted with economically feasible wall thickness and rib proportions. Tanks underlay a larger ground area. Uniform bedding is more difficult to attain.

Hemispherical shells, low rise dished-shaped heads and flat plate closures were all considered. Hemispherical shells were found structurally very efficient because of good buckling resistances under external fluid and earth pressure, good strength under internal pressure, no requirement for edge ring, and low discontinuity stresses at the junction with the cylinder. Flat end closures result in excessive deflections and large edge bending moments on the cylinder. Sandwich construction could be used to improve structural efficiency of flat ends. Sandwich wall construction was also investigated for attaining necessary buckling resistance of spherical shell end closures and found to be feasible but less cost effective.

Use of rib stiffening was required. It was found necessary to stiffen the cylindrical shell against buckling under external pressure from ground water and dimpling from local earth pressure due to surface wheel loads. Sandwich wall construction was investigated as an alternative to use of stiffening ribs and found to be feasible but less cost effective.

Shape-wise, a hollow trapezoid provides efficient bending strength and stiffness, a wide base for proper spacing of cylinder shell support against

Minimum practical total RP thickness is established as 4.8 mm (3/i6 in.) for the combined spray-up liquid seal and filament wound structural layers and 6.4 mm (V4 in.) for an all-chopped fiber spray-up laminate with sand filler. The choice for any construction is made on the basis of comparative design thickness, weight, and fabrication costs. The all-chopped fiber reinforced construction using somewhat greater wall thickness than the composite filament wound-chopped fiber wall is determined to provide the lowest tank cost; filament winding provides lower weight.

Hopper Rail Car Tank

In the past (1973) a severe shortage of railroad covered hopper cars for the transportation of grain developed. Cargill, Inc. provided a contract to Structural Composites, Inc. for determine feasibility studies on the potential of using RP in the design and fabrication of these cars. Test results showed structural deficient existed. By 1978 an acceptable design resulted fabricating the Glasshopper (registered name). It was used in rail service March 1981. Cargill Inc., Southern Pacific Transportation Co., and ACF Industries, Inc. (Fig. 4.18). It was larger and lighter in weight than the conventional steel covered hopper car resulting in being able to deliver more commodity per fuel dollar. Other advantages included corrosion resistance, and lower maintenance costs.

The first to be built was Glasshopper I. It successfully passed all of the

RP railroad covered hopper car

required American Association of Railroads (AAR) tests including the 454,000 kg (1,000,000 lb) static end compression test and the 568,000 kg (1,250,000 lb) coupler force impact test in the laboratory, and then successfully completed a round trip between St. Louis, MO and Oakland, CA [9700 km (6000 mile)].

From outward appearance, the RP designs were very similar to the standard ACF steel-covered hopper car. The first RP prototype, Glasshopper I that was in grain service, had four compartments. The car had a total capacity of 142 m3 (5000 ft3) and an overall length of about 16 m (53 ft). Its basic specifications are shown in Table 4.7.

lable 4.1 Glasshopper I basic specifications

Length inside

50 ft 3V2 in

Length over end sills

51 ft 55/8 in

Length over strikers

52 ft 11 in

Length over coupler pulling face

55 ft 6'/2 in

Length over running boards

53 ft 7/s in

Length between truck centres

42 ft 3 in

Extreme width

10 ft 8 in

Height, rail to top of running boards

15 ft 1 27/32 in

Height, rail to bottom of outlet

12 in

Extreme height, rail to top of hatch bumper

15 ft 6 in

Number of discharge outlets

4

Roof hatch opening, continuous

20 in x 44 ft 73/4 in

Curve negotiability, uncoupled

150 ft

Cubic capacity

500 ft3

Tare weight

59,000 lb

Gross rail load

263,000 lb

MR clearance diagram

Plate 'C

The second prototype car Glasshopper II that was later put into service had three compartments. The tare weight of the second car was 24,600 kg (54,200 lb), which was 4000 kg (8800 lb) lighter than a standard steel car weight of 28,600 kg (63,000 lb).

Construction details for Glasshopper I consist of a filament wound (FW) RP car body, RP/balsawood core sandwich panel bulkheads and slope sheets, steel side sills and shear plates, steel bolster webs, and RP hatch covers. Standard running gear and safety appliances were utilized, as were standard gravity oudets. Several changes in construction details such as the use of single laminate slope sheets were made in die design of Glasshopper II to reduce weight and manufacturing costs.

Table 4.8, shows the weight percentages of steel or RP materials. A significant amount (30 wt%) of the RP car structure is fabricated using RP materials. By subtracting the trucks steel weight, the remaining structure is RP. This construction allows the significant weight reduction to be possible. Finite element analysis (FEA) modeling was used throughout the design stages of the program to aid the structural analysis effort. The structural response in both static and dynamic loading conditions was characterized prior to initiation of the car construction.

: Glasshopper I component weight summary

Table 4.8, shows the weight percentages of steel or RP materials. A significant amount (30 wt%) of the RP car structure is fabricated using RP materials. By subtracting the trucks steel weight, the remaining structure is RP. This construction allows the significant weight reduction to be possible. Finite element analysis (FEA) modeling was used throughout the design stages of the program to aid the structural analysis effort. The structural response in both static and dynamic loading conditions was characterized prior to initiation of the car construction.

: Glasshopper I component weight summary

Component

Material

Weight lbs

Component weight ^ ^ Total weight

Car body

RP

6800

11.5

Sandwich panels

RP

4410

7.5

Wide flange beams

RP

640

1.1

Stiffeners

RP

2910

4.9

Top sill

RP

460

0.8

Roof/side angles

RP

510

0.9

Adhesive/bonding strip

RP

1070

1.8

Hatches

RP

860

1.5

Outlets

Steel

1800

3.0

End arrangement

Steel

9640

16.3

Side sills

Steel

4420

7.5

Running boards/safety appln.

Steel

1650

2.8

Brake system

Steel

1570

2.7

Misc. hardware

Steel

1060

1.8

Trucks

Steel

21200

35.9

Total weight

59000

100.0

£ RP component percentages = 30 £ Steel component percentages = 70

£ RP component percentages = 30 £ Steel component percentages = 70

FW process was used to fabricate both Glasshopper car bodies. It was determined that this process afforded the best mechanical properties for the lowest cost. Fabricating processes exist that can be highly automated which would help towards having RP-covered hopper cars compete economically with conventional steel-covered hopper cars in the marketplace.

Resin matrix material system chosen for fabrication of the car bodies was a proprietary isophthalic polyester resin system developed by Cargill specifically for the Glasshopper project. PPG, Certain Teed, and

OCF supplied the reinforcement of E-glass rovings. Both OCF 450 and 675 yield glass was used successfully in conjunction with the Cargill resin during FW operation. To provide adequate mechanical properties in the directions required to withstand the externally applied service loads, the FW apparatus was programmed to provide the multi-axial filament directional orientation capability.

Secondary bonding operations involving the attachment of stiffeners, etc., for the first RP car, used Hysol's epoxy adhesive (EA 919). This same adhesive was used in joints where both bonding and bolting with mechanical fasteners were employed. Lord Corp.'s acrylic adhesive system (TS 3929-70) was used successfully for Glasshopper II. Hat section stiffeners and wide flange beams were fabricated using the hand lay-up and pultrusion processes, respectively. The material used in the construction of the hat stiffeners included l'/2 oz mat, 24 oz woven rovings, 22'/2 oz unidirectional fabric, and the isophthalic polyester resin. Pultrusions were purchased finished, and were fabricated using standard pultrusion processes.

In order to demonstrate structural adequacy, Glasshopper I was tested in the ACF test laboratory located in St. Charles, MO. The test program was designed to show that the car meets and exceeds all requirements as specified by the AAR. Both static and dynamic tests were included in the testing. To determine the car's structural response under various applied loading conditions, Glasshopper I was instrumented with a total of 224 strain gauges, located at various areas determined through structural analysis to be of greatest importance and to provide maximum information. Glasshopper II, instrumented with 310 strain gauges, successfully completed the test program in 1983.

A series of six different static tests were successfully passed by Glasshopper I, including end compression, draft, vertical coupler-up, vertical coupler-down, coupler shank, and torsional jacking. The end compression test consisted of "squeezing" the car, while empty, with a hydraulic ram until a coupler force of 1,000,000 lb was measured. The draft test was conducted on the loaded car (105.9 tons) and consisted of pulling on the coupler until a force of 630,000 lb was experimentally observed. The remaining static tests were all conducted on the loaded car and involved using calibrated hydraulic rams to:

1. Jack the car upward with a vertical force of 22,700 kg (50,000 lb) applied at the coupler pulling face.

2. Jack the car downward with a vertical force of 22,700 kg (50,000 lb) applied at the coupler pulling face.

3. Lift the car free of the truck bolster by jacking at the coupler shank, a vertical force of 50,400 kg (111,0001b) was required.

4. Lift the car free of the truck bolster by jacking at the lifting lug/jacking pad assembly to verify torsional rigidity and stability, a verdcal force of 31,780 kg (70,000 lb) was required.

An analysis of test results show the experimentally observed strains to be very close to those predicted using FF.A techniques and "hand" calculations. This fact made it possible to use these techniques to further optimize the Glasshopper II design.

After successfully passing all required static tests, Glasshopper I was subjected to a series of impact tests. For these tests, the car that was fully loaded was pulled by cable up an inclined ramp and released to impact another fully loaded standing car that had its brakes released. Velocity of the car at impact was controlled by its height on the ramp at the time of release. Car's velocity was incrementally increased until an experimentally measured coupler force of 113,500 kg (250,000 lb) was developed during the impact.

Velocity of 14.9 km/h (9.24 m/h) was required to obtain the AAR specified load. It is noted that this velocity is significandy higher than the velocity required to reach the specified force with conventional steel covered hopper cars, which is about 12.1 km/h (7.5 m/h). Glasshopper I, with modified bulkhead joints, successfully passed the AAR impact test required and was subsequendy prepared for the extended road test.

Following completion of laboratory testing, Glasshopper I was tested over a 9700 km (6000 m) route on the Southern Pacific system. Fully loaded car with 9,600 kg (211,000 lb) made the trip from St. Charles, MO to Oakland, CA and back to Houston, TX. The car was unloaded at the Cargill export grain terminal in the Houston area and then returned empty to the facility in St. Charles, MO. The car was accompanied on the trip by the fully instrumented ACF test car used for data acquisition that monitored key strain gauges and load cells throughout the trip. All test results and visual observations showed the car performed well, and as predicted. During certain segments of the testing, speeds of 113 km/h (70 m/h) were reached with no dynamic problems (flutter, hunting, etc.) being observed.

It was determined that two major advantages of the RP covered hopper rail car are its tare weight and its corrosion resistance. As a result of its significandy lower weight and large size, the car is capable of carrying more payload per fuel dollar. This fact is extremely important in today's conditions of escalating and high fuelled prices. Glasshopper is able to carry many highly corrosive commodities (salts, potash, fertilizer, ore, etc.) without the need for expensive linings and with significandy reduced car maintenance costs. Also, the car's service lifetime would be gready extended in these severe service environments. Other advantages include the potential to eliminate painting requirements, reduced labor costs in manufacturing, lower center of gravity in the unloaded condition, ability to easily adapt to internal pressure designs, and rapid production changeover to alternate capacity cars.

Highway Tank

RP tanks on firm ground have been holding corrosive materials safely since the 1940s. Later the same technology, with some enhancements material-wise and design-wise, was applied to over-the-road tankers. Some tankers fabricated by Comptank Corp., Bothwell, Ontario, Canada have been on the road in the USA and Canada car rying a wide range of corrosive and hazardous liquids. These RP tank trailers are coded 312 for hauling corrosive and hazardous materials; special designed models haul acids or other corrosive chemicals; they unload by pressure, vacuum, or gravity.

The tankers are filament wound using E-glass rovings with polyester resin (Reichhold Atlac 4010 AC) and surfacing veil. RP moldings are integrated parts of the shell that is usually 15.88 mm (0.625 in.) thick. These parts include external rings/ribs, covers for steel rollover guards, spill dam, catwalk, hose trays, etc.

Very Large Tank

Large filament wound 150,000 gal (568 m3) tank has been fabricated by the Rucker Co. for Aerojet-General (1966). An example is shown schematically in Fig. 4.19 of the so-called racetrack-fabricating machine. A fabricated 56 m (22 ft) high, 152 m (60 ft) wide, 318 m (125 ft) long that weighed 32 ton, of all RP, large tank is shown in Fig. 4.20.

Just the mandrel for this FW machine weighed 100 ton all of metal. Total weight of the steel-constructed machine was 200 ton. The tank contained about 251 million km (158 million miles) of glass fiber, used 8 ton of textile creel containing 60 spools of glass fiber moving up to 7.24 km/h (4'/2 mph), and took three weeks to manufacture the epoxy-glass fiber RP tank in the Todd shipyard in Los Angeles, California.

Corrosive Resistant Tank

Chemical and corrosive resistant property of many plastics make them useful to contain different liquids ranging from water to acids. They are used extensively in water treatment plants and piping to handle

Tank's fabricating process

RP 150,000 gal tank

■ g 1 Large water filtration tank with 6 ft opening

drainage, sewage, and water supply. Glass fiber TS polyester RP water filtration tank is shown in Fig. 4.21. It is a 20 ft diameter, 32 ft high structure made in sections by a low pressure RP fabricating method. This bonded, assembled tank was shipped on a water barge to its destination. Structural shapes such as this tank for use under corrosive conditions often takes advantage of the properties of RPs and other plasdcs.

0 0

Post a comment