Warren R. True
Pipeline/Gas Processing Editor
Commissioning began last month near Columbus, Ohio, on the last of four elements that comprise the Gas Research Institute's pipeline simulation and test facility.
The commissioning will continue through mid-1995 to test and define the operating capabilities of a 4,700 ft, 24-in. OD pipeline test-flow loop that is the major component of the simulation facility.
The project is part of a nondestructive evaluation (NDE) program being implemented by the Gas Research Institute, Chicago. The loop, located about 18 miles west of Columbus at Battelle Memorial Institute's West Jefferson site, simulates an operating pipeline and will be used for research on pipeline inspection, operations, maintenance, and rehabilitation.
It incorporates features that reflect, says Battelle, an actual cross country pipeline such as bends, road crossings, and underground sections.
The line will operate under pressures of 200-1,000 psi.
NDE PROGRAM
The flow loop resembles several vendor loops around the U.S. as well as the metering research facility at the Southwest Research Institute, San Antonio.
The main differences, says Battelle, are the wide range of pressure and flow conditions that can be simulated while an inspection tool is running, the variety of pipeline features and defects included in the loop piping, and the ability to run inspection tools continuously around the loop without stopping and starting.
In addition, the SwRI metering facility was built for research and development on gas flow measurements.
It provides close controls on temperature, pressure, and flow rates but cannot accommodate inspection tools.
GRI's nondestructive evaluation program was established in 1987 to advance the state of the art of inspection technologies for gas-transmission pipelines.
The program's goal, says GRI, is to develop technologies that will help gas-pipeline companies maintain the physical integrity of their transmission systems, prevent pipeline shutdowns, and reduce maintenance costs.
It consists of facilities development and research on current and future inspection technologies.
GRI's research on current inspection technologies aims at improving current magnetic-flux leakage (MFL) analysis techniques and developing new and more powerful ones.
These studies are being used to improve defect detection and characterization analysis for corrosion and third-party damage and to develop next-generation analysis techniques, says GRI.
Research on future inspection technologies centers on stress-corrosion crack detection and characterization and on techniques for inspecting hard-to-inspect the lines.
The cracking program evaluates and incorporates worldwide technological advances to develop improvements that can be used by in-line inspection.
Facilities that support the program are based at Battelle.
The first is a nondestructive evaluation laboratory. Battelle says the lab allows detailed analytical modeling and carefully controlled experiments.
-The second element is an outdoor pull rig which consists of four 300-ft lengths of pipe (Fig. 1).
Test-bed vehicles, or inspection tools, can be pulled through these pipe lengths each of which contains controlled defects.
The third element is the flow loop which will be discussed in detail presently.
The fourth element is a set of test-bed vehicles that are modular systems upon which inspection equipment can be mounted and used in the pull rig or flow loop.
Fig. 2 illustrates how the four parts of the facility relate to each other.
OTHER FACILITIES
The NDE laboratory contains a linear test rig (Fig. 3) which is a movable platform on which sensor systems can be installed and used to take detailed measurements.
The rig was designed with the assistance of Pipetronix Ltd., Toronto, and built by Mecon Industries Ltd., Scarborough, Ont.
The test rig can measure the effects of velocity on defect-signal relationships and can support virtually any nondestructive evaluation sensor technology, according to Battelle.
The rig contains an inspection platform that can be pulled by a cable along with a 24-ft guide rail either under a partial-diameter pipe section or through a full-diameter pipe.
The speed of the inspection platform remains constant at up to 8 mph in the middle 10 ft of pipe with a magnetic load and brush friction and at speeds up to 10 mph without a magnetic load and friction.
The speed control is -0.1 mph with a maximum acceleration distance of 8 ft.
In the second element of the pipeline simulation facility, the outdoor pull rig, fundamental studies can be conducted to determine the relationship between defects and signals.
Battelle says the pull rig bridges the gap between the laboratory setting of the linear test rig and full-scale testing in the flow loop and actual application in pipelines.
It provides a full pipe-diameter section and full vehicle configuration for development of technologies while the test environment remains closely controlled.
An inspection tool or test-bed vehicle can be passed by several known defects and a large amount of data generated quickly and cost effectively.
The pull rig is a set of pipe runs with removable defect sections through which in-line inspection tools and test bed vehicles can be pulled.
The pull rig contains:
- Four unpressurized 300-ft pipe runs of 12, 24, 30, and 36 in. OD, each
- Three 80-ft removable defect pipe sections on each pull run
- A pull winch, which can generate pull forces up to 56,000 lb and velocities of up to 25 mph
- Instrumentation to collect such data as pulling force and tool velocity.
The pull rig also contains two metal-loss defect sets and a stress-corrosion cracking set. Plans call for other defect sets, including one to indicate third-party damage, to be added.
The metal-loss research and development defect sets have been installed in the removable 80-ft segments of the 24 in. pull run.
The defects represent pipelines under normal conditions, says Battelle. They were selected to allow study of the effects of defect parameters on the amplitude and shape of MFL signals and the study of defect-detection thresholds.
The stress-corrosion cracking defect set was installed in the 30-in. pipe run. This set was assembled from pipe with cracks that had been removed from service.
Before installation, the pipe samples were stripped of coating and lightly grit blasted to remove any residual coating and rust scale.
Each pipe sample in the stress-corrosion cracking set was inspected for cracks with two methods:
- The first was a continuous wet fluorescent magnetic-particle inspection. The magnetic field was oriented to highlight longitudinal cracks.
- The second method was an eddy current technique.
Many of the cracks in the pipe samples are shallow: 120 of the 134 crack colonies are less than 20% deep. Seven crack colonies are between 20 and 30% deep, and seven more are greater than 30% deep.
The fourth element in the simulation facility consists of two test-bed vehicles built as test platforms for use in the pull rig and flow loop. These vehicles remove the need to develop a test platform for each research scenario.
Both test-bed vehicles were designed to make it easier for independent researchers to attain field realism, which is something only vendor laboratories had been able to do previously.
Both vehicles were built for use in 24-in. pipe.
The first test-bed vehicle is called the advanced sensor vehicle. It was designed and fabricated by Southwest Research Institute, San Antonio, and T.D. Williamson, Tulsa, to support development and evaluation of ultrasonic and other nondestructive testing methods.
The vehicle can be used for experiments involving ultrasonics, eddy currents, electromagnetic acoustic transducers, or multiple inspection technologies.
Battelle says that the versatility of the advanced sensor vehicle pen-nits a wide range of experiments because an experimenter can essentially replace any one of the vehicle sections with one of his own design.
The MFL test-bed vehicle was designed and built by Pipetronix and Battelle. The main difference between the MFL test-bed vehicle and the advanced sensor vehicle lies in the data acquisition system and the sensor module.
The sensor module contains full magnetizing and sensing equipment. The MFL vehicle uses slower acquisition rates than the advanced sensor vehicle; slower data acquisition allows more channels of data to be taken.
Also, although the design of the vehicle was oriented toward research in magnetic flux leakage, its use is not limited. The test bed vehicle can be used for any research which does not require extremely fast data-acquisition rates.
FLOW LOOP
The 4,700 ft, 24-in. flow loop, in which both pressure and flow velocity can be controlled to desired setpoints, incorporates pipeline features (bends, road crossing, underwater sections, etc.) that provide realism in a closed loop system.
JOINT DEVELOPMENT
The design requirements for the flow loop were developed in a series of meetings between GRI, various industry advisory groups, and Battelle between 1988 and 1990. The design was completed by Fluor Daniel Williams Bros., Tulsa.
- Design operating pressure range for the loop of 200-1,000 psig was selected to span typical operating pressures in existing natural-gas pipelines.
- Design tool speed range of 2-10 mph encompasses the most commonly used inspection speeds. Because higher gas velocities are typical in operating pipelines, the design calls for future expansion capability to 25 mph.
- Pressurizing medium may be either air or natural gas.
Natural gas is obviously the normal condition in an operating pipeline, but there are no expected or significant differences in tool performance between air and natural gas.
Air will therefore be the medium for initial operation.
- Flow may eventually be bidirectional (reversible) with addition of appropriate valves and piping.
This capability will allow inspection tools to approach various loop features from either direction (for example, a tool moving up or down through a vertical section).
It will also allow the effect of a reversed magnetic field on the residual magnetism to be studied.
- Continuous operation allows large amounts of data to be generated without retrieving and relaunching the tool, thereby facilitating durability and wear studies.
- Pressure differential across a tool can be at least 10 psi under all operating conditions and 50 psi under at least some conditions.
The loop (Fig. 4) is constructed of 24 in. x 0.344-in. W.T., X-70 pipe. It is divided into two main segments by full-opening valves in the loop piping. These valves are used to provide continuous looping capabilities.
Removable defect and bend test areas in the flow loop include 320 ft of replaceable straight pipe (Fig. 5) and one removable bend (elbow; Fig. 6). The replaceable straight pipe can be mounted in sections as short as 20 ft and as long as 320 ft.
In addition, in the removable straight region (Fig. 4), an expansion loop can be mounted. The removable bend section allows bend radiuses as small as 1.5D (3 ft; D = 1 pipe diameter) or as long as 10D (20 ft). The removable bends are all 90.
Fixed-bend radiis are included in the loop along with the bend-radius test section.
The fixed-bend radii are nominally 10 D (20 ft), although one radius is 20 D (40 ft) and two are 75 D (15 ft). The major fixed bends are between 55 and 90; all bends were fabricated by Naptech Inc., Clearfield, Utah.
One set of bends is mounted vertically. These two bends are also situated in a compound arrangement, in which the plane of the second bend is not the same as that of the first bend. In addition, both right hand and left hand bends are included.
Additionally, the loop's design called for a road crossing consisting of shorted eccentric and unshorted underground casings. One heavy-wall road crossing has been installed.
Pipeline crossings and adjacent lines are part of the loop's design.
In the two adjacent line sections, one simulates a pipeline near other pipe lines; the other a pipeline between other lines. In addition, a pipeline crossing is included.
For liquid slugs, fill and drain points have been located near the launching and receiving areas. In addition the section scrubber and filter have been sized to receive water slugs.
Special features included in the design were added at the request of the industry advisory group or because natural conditions necessitated their use,
An underwater section simulates a water crossing with set-on and clamp-on weights (Fig. 7). A natural marsh area simulates a swamp crossing with screw anchors and a concrete coated section.
In addition to these features, the flow-loop includes instrumentation for measuring the pressure and temperature of the gas in the loop and the pressure drop across the compressor system.
The design included pig detectors for monitoring the tool in critical areas.
OPERATING, CONTINUOUS LOOPING
Attainable tool speeds are a function of the pressure differential across the tool and vice versa (Table 1).
With air, the compressor system can propel a tool with a pressure differential of 10 psi with speeds of 15.4 and 11.3 mph for operating pressures of 200 and 1,000 psi, respectively.
For natural gas, the velocities are 24 and 13.5 mph at 200 and 1,000 psi, respectively.
All values exceed the minimum specified value of 10 mph. The natural-gas velocities approach the upper limit of 25 mph for the design.
At 200 psi, the compressor system can provide a maximum pressure differential of 53-54 psi for all normal operating velocities.
For air at 1,000 psi, the maximum pressure differential is between 47 and 89 psi. For natural gas, the maximum pressure differentials are between 36 and 74 psi for normal operating pressures and velocities.
These values exceed the minimum design value of 10 psi for all conditions, and as specified they provide a 50-psi capacity for at least some of the operating conditions.
The flow loop's design allows true continuous operation without relaunching. That is, a tool can be repeatedly propelled around the loop without stopping and restarting.
To provide continuous operation, six gate valves are used to control the suction and discharge of the flow compressor to the loop. These valves cycle one time for each passage around the flow loop and will therefore experience rugged service.
The valve positions used to achieve continuous looping are shown in Fig. 8.
Valves used for continuous operation are valves A, B, C, D, E, and F. The following paragraphs discuss how the valves are opened and closed to achieve continuous looping.
Fig. 8a shows the tool moving in a counterclockwise direction near the start of a run. Valves A, D, and E are closed, and Valves B, C, and D are open. The discharge into the loop is through Valve B and the suction is through Valve F.
As the tool moves up and around the top half of the loop, Valve A is opened and Valves B and C are dosed. This operation switches the discharge to Valve A (Fig. 8b). Suction is maintained at Valve F.
As the tool continues to move around the loop Valves D and E are opened and Valve F is closed. This operation switches the suction to Valve E (Fig. 8c). Now, the valve positions are exactly opposite those shown earlier in Fig. 8a.
A similar sequence of opening and closing a set of three valves is repeated as the tool moves around the bottom half of the loop. in this manner, the discharge is switched back and forth between Valves A and B, and the suction is switched between Valves E and F.
By continuation of this process, a tool or test-bed vehicle can be pushed around the loop repeatedly.
LINE PIPE, COMPRESSORS
The line pipe was selected on the basis of a design operating stress level of 50% of the specified minimum yield stress. A 50% design basis is commonly used for compressor stations and for areas of moderate to high personnel exposure.
Several combinations of yield strength grade and wall thickness could have been used (Table 2). Most of the wall thicknesses exceed those used on many gas transmission pipelines.
Because the flow loop was meant to simulate typical gas transmission lines, a material grade of X-70 was chosen, which requires a wall thickness of 0.344 in.
An X-80 material grade was rejected because the gas-transmission industry has had little experience using X-80 materials.
Line pipe was purchased from Confab, Brazil, in late 1990 and delivered to New Orleans in July of 1991. The pipe received an external coating of fusion-bonded epoxy (FBE) applied by Bayou Pipe Coating.
The specification called for cleaning, coating with 12 mils of FBE, inspection, and testing. The cleaning was specified as an acid wash with a water rinse prior to coating. Inspection consisted of testing for holidays with 1,500 y (hot end) and 2,000 y (cold end).
All main loop valves were supplied by Daniel Valve Co. and specified for service up to 1,100 psi. The sequencing valve operators are specified to open fully each valve in 18 sec with enough torque to overcome a 250 psi pressure differential.
Electric operators were selected rather than pneumatic or pneumatic/hydraulic on the basis of reliability, simplicity, and ease of interfacing with the programmable logic center.
Design of the flow loop includes two field compressors that were donated by Enron Corp., Houston.
The compressors are 330 hp Ariel four-stage reciprocating compressors driven by Caterpillar engines. They contain their own interstage coolers and scrubbers.
The flow loop requires a low-head, high-flow compressor system. The loop will have the full flow of a 24-in. transmission line, but it will have the pressure drop of less than 2 miles (equivalent length) of main line pipe.
This 2-mile distance is less than 5% of the normal distance between stations.
Both reciprocating and centrifugal compressors were considered for the low-head, high-flow requirements. Both can theoretically meet the flow and head conditions.
Reciprocating compressors use the pressure energy of the flowing gas stream to open the compressor valves, however. This use of pressure energy is a net loss of energy from the system.
Under low-head conditions, the net loss is a considerable fraction of the total power requirements.
Centrifugal compressors do not have these losses. Under comparable conditions, reciprocating compressors need several times the horsepower of centrifugal compressors.
The flow compressor was provided by Solar Compressors Inc., San Diego. The 1,300-hp unit is used to provide the flow rates discussed earlier.
MID-YEAR OPENING
Dedication of the flow loop is set for May 1995.
At or about that time, the entire facility will be open for use by the pipeline industry for product development, trials, and demonstrations conducted under safe and controlled conditions, says Battelle, without the necessity of interfering with an operating pipeline.
Copyright 1994 Oil & Gas Journal. All Rights Reserved.
Issue date: 12/26/94
