Evaluation of the CompsiSleeve TM Pipeline Repair System

Transcription

1 Evaluation of the CompsiSleeve TM Pipeline Repair System PN CRA Prepared for LMC Industrial Contractors, Inc. Avon, New York June 2011 Stress Engineering Services, Inc. Houston, Texas

2 Evaluation of the ComposiSleeve TM Pipeline Repair System PN CRA Prepared for LMC Industrial Contractors, Inc. Avon, New York Prepared by: Dr. Chris Alexander, P.E. Principal June 2011

3 EXECUTIVE SUMMARY Over the past three years Stress Engineering Services, Inc. has worked with Western Specialties and LMC Industrial Contractors, Inc. in evaluating the ComposiSleeve TM pipeline repair system. The ComposiSleeve TM system employs a hybrid technology combining advanced adhesives, steel half shells, and a water-activated urethane E- glass composite system. Each component in the system provides a vitally important role in reinforcing damaged sections of a pipeline. The main focus of the current report involved Stress Engineering s assessment of the repair of severe corrosion subjected to cyclic pressures and burst tests. The evaluation process involved full-scale testing where the ComposiSleeve TM repaired a test sample fabricated using inch x inch, Grade X42 pipe having a 75% deep corrosion section. The tests involved both burst testing and pressure cycle to failure testing. For the past five years Stress Engineering Services, Inc. (SES) has been the Principal Investigator in four research programs co-sponsored by the Pipeline Research Council International, Inc. (PRCI) performed in studying the performance of composite repair technology. In one program in particular, MATR-3-4, the focus is on evaluating the long-term performance of composite materials used to repair corroded pipelines. In this study strain gages were used to monitor strain in the steel beneath the composite materials. This permits SES to compare the relative performance of the competing composite technologies. These data are useful in comparing the relative performance of the ComposiSleeve TM system. During the burst tests phase of testing, the strain in the steel beneath the repair at 72% SMYS was measured to be 1,610 µε, whereas the average for all composites in the PRCI study was 4,497 µε. At 100% SMYS the ComposiSleeve TM and PRCI-average values were measured to be 2,370 µε and 5,692 µε, respectively. The ComposiSleeve TM sample failed at a pressure of 4,374 psi, a pressure value that is approximately 2.5 times the operating pressure (72% SMYS) of the pipeline. In the pressure tests a total of four repair configurations were tested, including different E- glass material thicknesses, as well as one sample having no composite material with

4 only the steel sleeves attached via adhesive to the pipe. Even this sample had significantly reduced strains when compared to the PRCI average values. In addition to the burst tests, SES performed a pressure cycle test on a sample having the same corrosion configuration as the burst test samples. This sample was pressure cycled from 36% to 72% SMYS until a failure occurred. SES has performed this type of test on eight different composite repair systems, with the cycles to failure for one sample being a low as 19,411 cycles. The ComposiSleeve TM failed after 767,816 pressure cycles had been applied, even though multiple failures occurred in the welds attaching the end caps to the test sample. This report also provides details on three other programs evaluating unique applications of the ComposiSleeve TM system in reinforcing various pipeline features and anomalies. These include the repair of leaking pipes, reinforcing pipes subjected to bending loads, and evaluating an increase in overburden capacity of a reinforced pipeline.

5 LIMITATIONS OF THIS REPORT The scope of this report is limited to the matters expressly covered. This report is prepared for the sole benefit of LMC Industrial Contractors, Inc. In preparing this report, Stress Engineering Services, Inc. has relied on information provided by the LMC Industrial Contractors, Inc. Stress Engineering Services, Inc. has made no independent investigation as to the accuracy or completeness of such information and has assumed that such information was accurate and complete. Further, Stress Engineering Services, Inc. is not able to direct or control the operation or maintenance of client s equipment or processes. All recommendations, findings and conclusions stated in this report are based upon facts and circumstances, as they existed at the time that this report was prepared. A change in any fact or circumstance upon which this report is based may adversely affect the recommendations, findings, and conclusions expressed in this report. NO IMPLIED WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE SHALL APPLY. STRESS ENGINEERING SERVICES, INC. MAKES NO REPRESENTATION OR WARRANTY THAT THE IMPLEMENTATION OR USE OF THE RECOMMENDATIONS, FINDINGS, OR CONCLUSIONS OF THIS REPORT WILL RESULT IN COMPLIANCE WITH APPLICABLE LAWS OR PERFECT RESULTS.

6 TABLE OF CONTENTS Page No. EXECUTIVE SUMMARY... 3 TABLE OF CONTENTS... 6 LIST OF FIGURES INTRODUCTION BACKGROUND REPAIR OF CORROSION: TESTING METHODS AND RESULTS Testing Methods Testing Results Burst Tests Pressure Cycle Fatigue Tests Bend Testing (November 2008) Leak Repair Testing (April 2009) Overburden Testing DISCUSSION CLOSING COMMENTS Appendix A Report on Bend Testing (Prior Work) Appendix B Report on Leak Repair (Prior Work) Appendix C Report on Overburden Testing (Prior Work)... 38

7 LIST OF FIGURES Page No. Figure 2.1 Schematic diagram of the ComposiSleeve TM system Figure 2.2 Western Specialties Composite Thickness Calculator (1/2) Figure 2.3 Western Specialties Composite Thickness Calculator (2/2) Figure 3.1 Schematic of test sample with machined corrosion Figure 3.2 Schematic showing location of strain gages on corrosion sample Figure 3.3 Photographs showing machined corrosion depths Figure 3.4 Photographs of test sample preparation efforts (Set #1) Figure 3.5 Photographs of test sample preparation efforts (Set #2) Figure 3.6 Photographs of test sample preparation efforts (Set #3) Figure 3.7 Photographs of test sample preparation efforts (Set #4) Figure 3.8 Photographs of three burst tests Figure 3.9 Hoop strain measured as a function of internal pressure Figure 3.10 Tabulated hoop strain data acquired during pressure testing Figure 3.11 Strain as a function of cycle number for 75% corroded sample Figure 3.12 Hoop and axial strain listed as a function of cycle number Figure 3.13 Service life based on pressure cycle range Figure 4.1 Schematic diagram of the bend test set-up Figure 4.2 Photographs of test sample having a thru-wall defect Figure 4.3 Test set-up using a 400,000 lbs load capacity frame... 33

8 1.0 INTRODUCTION The purpose of this report is to provide information on the testing program conducted by Stress Engineering Services, Inc. (SES) for Western Specialties in evaluating the ComposiSleeve TM system in repairing and reinforcing damaged pipelines. This system is a hybrid design that integrates steel half-shells combined with an E-glass composite having a water-activated urethane resin system. The approach employed by SES in evaluating the ComposiSleeve TM system is based on full-scale experimentation, where defects were machined into test samples and the performance of the system was evaluated by means of destructive testing. Much of this work is based on previous studies conducted by SES for the pipeline industry in evaluating competing composite technologies. Having access to data associated with prior studies as it permits a direct comparison of the ComposiSleeve TM system with other composite repair technologies. Of particular interest is the level of reinforcement provided by the reinforcing system to the damaged region of the pipe. Systems that are most effective are those systems that are able to successfully reduce strain in the damaged (i.e. corrosion and dents) to acceptable levels. As will be presented in this report, the ComposiSleeve TM system is effective in reducing strain in the corroded of a pipeline significantly below industry norms. In turn, the 767,816 cycles to failure measured for the pressure cycle fatigue sample is greater than any composite repair system tested to date. This report has been organized to provide the reader with a brief background on the ComposiSleeve TM system, including a schematic with descriptions of the system. A detailed discussion on the test program, including results, is presented. The primary means for evaluating the ComposiSleeve TM system is its ability to provide reinforcement to a 75% deep corrosion defect machined in a inch x inch, Grade X42 pipe sample. In addition to testing associated with evaluating the reinforcement of corroded pipelines, testing was also conducted to evaluate the ability of the ComposiSleeve TM system to repair leaking pipes and pipe sections subjected to bending loads.

9 Sections of this report include Background (with details on the repair of corrosion), Testing Methods and Results, Additional Testing: Methods and Results, Discussion, and Closing Comments.

10 2.0 BACKGROUND The ComposiSleeve TM system is a hybrid design that integrates steel half-shells combined with an E-glass composite having a water-activated urethane resin system. Unlike conventional composite repair systems that utilize fibers (i.e. typically E-glass and carbon) as the primary means for reinforcement, the ComposiSleeve TM system relies on the stiffness of steel and adhesive bonding between the outer surface of the pipe and inner surface of the steel half shell. Most of the composite repair systems currently on the market are either in compliance or seeking compliance with standards such as ASME PCC-2, Repair of Pressure Equipment and Piping, Part 4, Nonmetallic and Bonded Repairs. These standards were written primarily for wet wrap systems involving the use of either saturation in the field or pre-impregnated resins. Systems involving either pre-cured coils or half shell designs (ComposiSleeve TM system) are not explicitly addresses in the current standards. As a result, any assessment of these systems must rely on assessment via full-scale testing performance and not rely explicitly on calculations to determine the minimum required reinforcing thickness. Provided in Figure 2.1 is a schematic diagram showing the components of the ComposiSleeve TM system. The variables of interest include selection of the load transfer material, thickness of the steel, selection of the bonding adhesive between the pipe and steel sleeves, and thickness of the E-glass composite material. The preceding statement is predicated on the fact that the grade of steel and composite material selections have already been made.

11 Load transfer material material E-glass composite material Carbon steel half-shells Damaged pipe material Bonding adhesive (high shear strength) Keys to proper reinforcement of damaged pipelines: Shear strength of bonding adhesive attaching half-shells to pipe Load transfer material stiffness External composite wrap Figure 2.1 Schematic diagram of the ComposiSleeve TM system

12 At the present time there is no single published methodology that can be used to determine the minimum required thickness for a composite-based reinforcing sleeve such as the ComposiSleeve TM system; however, as a minimum, the following should be considered from a performance standpoint. Stresses must be reduced in the damaged section of pipe to an acceptable level Stresses in the reinforcing materials of the ComposiSleeveTM (steel, composite, and adhesive) must not exceed design stresses The repair must be able to withstand both static and cyclic pressure loading Long-term performance is an essential variable of interest in qualifying repair systems Western Specialties requested that SES develop a calculator that could be used to determine the thickness of the overwrapping E-glass composite material. Typically, the half shell thicknesses have not been less than 0.25 inches. In comparison to the relative low stiffness of the composite material, this thickness of the steel has been adequate for all testing completed to date (i.e. steel pipe thicknesses of inches). Screenshots for the calculator are presented in Figure 2.2 and Figure 2.3, along with the appropriate equations. The essential elements of this calculator include the following: Properties of the base pipe material including diameter, wall thickness, and grade Dimensions of the corrosion Aspects of the repair including length and adhesive lap shear strength Safety factors, especially with regards to the bonding adhesive The calculated required thickness of the composite material is based on the assumption that the composite material and the bonding adhesive are the two structural components that reinforce the damaged section of pipe, and resist that opposing force caused by the internal pressure. For conservatism, it is assumed that these two structural members resist a force caused by the product of the internal pressure and projected area of the corrosion. In other words, if the steel associated with the corroded region of the pipe were removed, the composite and bonding adhesive would be responsible for resisting the resulting force. While this condition would never exist, this approach provides a mechanics-based reference point on which to make calculations.

13 Note the following input data in Figure 2.2. Pipe: inch x inch, Grade X42 Corrosion geometry: 8 inches long x 6 inches wide Adhesive shear strength: 2,000 psi Composite tensile strength: 20,000 psi (long-term design strength) Length of repair: 18 inches Adhesive service factor: 0.2 (safety factor of 5) Composite service factor: 0.5 (safety factor of 2 on long-term strength) The design factors that contribute to the strength of the reinforcing system include the shear strength of the adhesive, tensile strength of the composite material, composite thickness, and the length of the repair. Assuming that material selection is not an option, the composite thickness and length of repair are the two critical variables.

14 Western Specialties Composite Thickness Calculator This calculator is used to estimate composite material thickness. No consideration of the ComposiSleeve steel "thickness" is considered (it is assumed that this material is sufficiently thick), although the adhesive bond strength is critical to the calculation process. Input values are as noted and include pipe dimensions, corrosion area, and material strengths. INPUT VALUES Measured yield strength of steel pipe or mill certification: s a := 43000psi Minimum Specified Yield Strength (SMYS) of steel pipe: S := psi Measured ultimate tensile strength of steel pipe or mill certification: s UTS := 76000psi Nominal wall thickness of test pipe: t s := in Outside diameter of test pipe: D := in Long-term strength of composite material: s lt := 20000psi Lap shear strength of the adhesive material: τ adh := 2000 psi Length of repair: L repair := 18 in Per layer thickness of composite:: t layer := in Length of corrosion: L corr := 8in Width of corrosion: w corr := 6in Service factor for composite material (ASME PCC-2 Table 5): f := 0.2 Service factor for adhesive material f adh := 0.2 Figure 2.2 Western Specialties Composite Thickness Calculator (1/2)

15 As observed in Figure 2.3, the resulting composite thickness is inches. If the shear strength of the adhesive, t adh, is varied, the following composite thickness, t c, values are calculated: t adh = 2,375 psi t adh = 2,000 psi t adh = 1,000 psi t adh = 500 psi t adh = 100 psi t c = inches t c = inches (default configuration) t c = inches t c = inches t c = inches What is interesting to note in the above calculations is the significant contribution that the adhesive makes to the overall strength of the reinforcement. As noted when the shear strength is assumed to be 2,375 psi, the need for composite reinforcement is practically non-existent. Of the three burst tests that were conducted, the test sample with no composite reinforcement did indeed demonstrate that that the adhesive-only case had adequate strength to resist the corrosion bulging and pipe expansion associated with the burst test. CALCULATED VALUES The thickness of the composite is determined by considering the level of restraint provided by the steel sleeve and the required containment pressure. 2S t s SMYS pressure: P := D P = 2529 psi 2s UTS t s Predicted burst pressure of pipe based on UTS: P burst := D P burst = 4471 psi 1 PL corr w corr t c τ 2f s lt L adh f π adh repair 2 w := corr t c = in t c Number of wraps based on per layer thickness of composite: N w := t layer N w = 7 wraps Figure 2.3 Western Specialties Composite Thickness Calculator (2/2)

16 3.0 REPAIR OF CORROSION: TESTING METHODS AND RESULTS Full-scale testing was performed to evaluate the performance of the ComposiSleeve TM system in repairing a corroded pipeline. While analytical calculations are useful for sizing the components within a system, it is essential that performance testing be conducted to validate the overall performance of a given composite reinforcement system. What is lacking in the calculations is the inter-dependant relationship that exists between the components in the system. For example, the load transfer material serves to ensure that load generated by the bulging corrosion is transferred into the composite fiber material. A filler material lacking the required rigidity will fail to engage the composite material, resulting in an excessive accumulation of strain in the reinforced steel material. The end result is a highly-loaded corrosion region that is susceptible to premature failure in the form of leaking, especially in the presence of cyclic pressure loading. This sections of this report that follow provide specific details associated with the testing methods and results used to evaluate the performance of the ComposiSleeve TM system in repairing a corroded pipeline. 3.1 Testing Methods Two types of testing were conducted. The first involved samples repaired to evaluate the increase in burst strength (i.e. limit state) of a corroded pipe section, while an additional test samples was prepared to evaluate the effects of cyclic pressure on the ComposiSleeve TM repair. The variable of interest in this study was the thickness of the E-glass water-activated urethane. An additional sample was also fabricated to determine what would happen if no composite materials were installed on the outside of the repair. The pipe test samples included a inch x inch, Grade X42 pipe with a corrosion region having a depth of 75% installed by machining. Figure 3.1 shows the schematic for this test sample. Strain gages were installed in the machined corrosion region, on the base pipe, and outside of the repair, as shown in Figure 3.2.

17 As stated previously, three types of tests were conducted, with details for the burst sample configurations listed below. The pressure cycle fatigue sample involved the steel plus 18 wraps configuration. Sample #1: Steel only Sample #2: Steel plus 18 wraps Sample #3: Steel plus 8 wraps Figure 3.3 provides a series of photographs showing actual measurements made after the machining was completed. The target machining depth was 75%, although as can be seen from these photographs, the machining depth averages ranged from 74% to 78%. After machining, end caps were welded to each sample. The samples were sandblasted to a near white metal finish (NACE 2).

18 8 feet (center machined area on sample) 8 inches long 0.75-inch radius (at least) Break corners (all around) inches 75% corrosion: remaining wall of inches NOTE: Perform all machining 180 degrees from longitudinal ERW seam. 6 inches Note uniform wall in machined region Measure wall thickness at 9 locations in the machined area using a UT meter. Details on machining (machined area is 8 inches long by 6 inches wide) Figure 3.1 Schematic of test sample with machined corrosion 3 Gages #4 & #5 on repair Strain gage locations Gage #1 Center of corrosion Gage #2 2-inch offset from center of corrosion Gage #3 12 inches from end of sample Gage #4 Outside of composite (on top of Gage #1) Gage #5 Outside of composite (on top of Gage #2) Figure 3.2 Schematic showing location of strain gages on corrosion sample

19 Sample # inches (78% corrosion) Sample # inches (74% corrosion) Sample # inches (76% corrosion) Sample # inches (74% corrosion) Figure 3.3 Photographs showing machined corrosion depths Figure 3.4 Photographs of test sample preparation efforts (Set #1)

20 Figure 3.5 Photographs of test sample preparation efforts (Set #2) Figure 3.6 Photographs of test sample preparation efforts (Set #3)

21 Figure 3.7 Photographs of test sample preparation efforts (Set #4) 3.2 Testing Results Results are presented in this section of the report detailing the measurements and final performance data associated with both the burst and pressure cycle fatigue tests. In addition to the actual performance data (i.e. burst pressure and number of cycles to failure), data were also generated from the strain gage measurements in the corroded steel region beneath repair. The strain measurements are a fundamental means for evaluating the ability of a repair system to reinforce a damaged section of pipe. With Stress Engineering Services, Inc. having performed corrosion repair tests on more than 15 different systems over the past 5 years, it is possible to evaluate the relative performance of a particular system to industry norms. A comparison of this type will be provided for the ComposiSleeve TM system.

22 3.2.1 Burst Tests The three burst tests were performed to determine the limit state capacity for each repair configuration. The recorded burst pressures for the three different repair configurations. Sample #1: Steel only P burst = 3,995 psi Sample #2: Steel plus 18 wraps P burst = 4,389 psi Sample #3: Steel plus 8 wraps P burst = 4,374 psi Figure 3.8 provides photographs of the three burst tests. Note that the failures occurred in the repaired region for all three samples; however, the occurred at pressure levels that would be expected for a perfect pipe. Sample #1: P burst = 3,995 psi (Steel only) Sample #2: P burst = 4,389 psi (Steel + 18 wraps) Sample #3: P burst = 4,374 psi (Steel + 8 wraps) Figure 3.8 Photographs of three burst tests

23 For the inch x inch, Grade X42 pipe, which has a minimum tensile strength of 66,000 psi, the estimated burst pressure is 3,529 psi. Even the minimum burst pressure of 3,995 psi (Sample #1 with steel only) is 113% of this value; while the maximum burst pressure associated with Sample #2 (i.e. 4,389 psi) is 124% of this value. Of the three burst tests that were conducted, Sample #1 provides a significant contribution to the overall study because prior to testing one of the questions to be addressed in the current scope of work what to identify the level of reinforcement provide by only the adhesive bonding the steel half shell to the outer pipe surface. While Samples #2 and #3 demonstrate that the composite material provides additional reinforcement, Sample #1 shows that the adhesive by itself provides a robust level of reinforcement. While the burst pressures are essential for understanding the limit state performance of a given repair, equally if not more important is measuring the level of strain beneath the repair. For a repair system to work properly, it must ensure that strain levels in the damaged section of pipe (i.e. corrosion in this particular study) are minimized. There is no current limitation on strain in the reinforced section specified in any of the current design codes such as ASME PCC-2; however, it is possible to quantify acceptable strain levels by considering performance of the repair in testing. Burst testing is one means of establishing performance, although a better means for quantifying the capability of a repair system for corrosion and dents is via pressure cycle fatigue testing. The section that follows, Pressure Cycle Fatigue Tests, provides details on the pressure cycle fatigue test that was conducted on the ComposiSleeve TM system. Presented in this section of the report are the strain gage results measured during burst testing of the three different repair configurations. The presentation includes plotted data, as well as tabulated data measured at pressure levels equal to 72% and 100% SMYS (1,780 psi and 2,470 psi, respectively). Figure 3.9 plots hoop strain as a function of internal pressure. Results are included for the following data sets: Leading E-glass technology

24 Base pipe (non-corroded section) Sample #1: Steel only Sample #2: Steel plus 18 wraps Sample #2: Steel plus 18 wraps (outside surface of repair) Sample #3: Steel plus 8 wraps Data point at 72% SMYS for Industry average for other competing E-glass systems 1 Data point at 100% SMYS for Industry average for other competing E-glass systems 1 There are several observations made in viewing this plot; however, the most noteworthy point is the significant reduction in strain provided by the ComposiSleeve TM system when compared to other E-glass technologies. The large number of fatigue cycles acquired during the pressure cycle test is testimony to the appreciable level of performance provided by the ComposiSleeve TM system. Also included are the tabulated data shown in Figure As observed at 72% SMYS the PRCI average for E-glass material was 4,497 µε, (where µε is microstrain with 10,000 µε corresponding to 1% strain). For the ComposiSleeve TM systems that were tested the minimum measured strain data was 1,610 µε (steel + 8 wraps), while the maximum measured of the three data sets was 2,220 µε. Results at the 100% SMYS pressure level demonstrated similar performance characteristics when compared to the PRCI data set, although the results for the ComposiSleeveTM are not as non-linear as the data for the PRCI E-glass systems. A detailed assessment of the results will be provided in the Discussion section of this report. 1 These data acquired as part of the long-term testing program (MATR-3-4) being co-sponsored by the Pipeline Research Council International, Inc. and 13 other composite repair systems. For additional details consult

25 Hoop Strain Versus Pressure for Composite Repair Burst test of inch x inch, Grade X42 pipe with 75 % corrosion. Gages beneath repair on steel. Test pressures: 72% SMYS = 1,780 psi SMYS = 2,470 psi Base pipe burst pressure = 3,936 psi 4,500 4,000 Internal Pressure (psi) 3,500 3,000 2,500 2,000 1,500 1, Leading E-glass technology Base pipe ComposiSleeve - Steel with 18 wraps (outside surface) ComposiSleeve - Steel only ComposiSleeve - Steel with 18 wraps ComposiSleeve - Steel with 8 wraps MAOP Industry E-glass Average (4,497 microstrain) SMYS Industry E-glass Average (5,692 microstrain) 0 0 2,000 4,000 6,000 8,000 10,000 12,000 Hoop Strain (microstrain) (10,000 microstrain is equal to 1 percent strain) Figure 3.9 Hoop strain measured as a function of internal pressure Strain at 72% SMYS (design pressure) PRCI average: 4,497 µε Steel only: 2,220 µε Steel + 8 wraps: 1,610 µε Steel + 18 wraps:1,910 µε Strain at 100% SMYS (yield pressure) PRCI average: 5,692 µε Steel only: 3,630 µε Steel + 8 wraps: 2,370 µε Steel + 18 wraps:2,970 µε Data from PRCI Test Program ALL MATERIALS PRCI average measured strain values for 75% corrosion MAOP 3,734 µε SMYS 4,905 µε MAOP min 1,828 µε MAOP max 8,852 µε SMYS min 2,250 µε SMYS max 8,791 µε E-Glass Material Only PRCI average measured strain values for 75% corrosion MAOP 4,497 µε SMYS 5,692 µε MAOP min 2,667 µε MAOP max 8,852 µε SMYS min 3,185 µε SMYS max 8,472 µε (actually higher, gage failed) Carbon Material Only PRCI average measured strain values for 75% corrosion MAOP 2,524 µε SMYS 3,292 µε MAOP min 1,828 µε MAOP max 3,087 µε SMYS min 2,250 µε SMYS max 4,106 µε At both the design and yield pressures, the ComposiSleeve TM strain is less than one-half of the industry average for E-glass repair materials when considering the steel + 8 wraps configuration. Figure 3.10 Tabulated hoop strain data acquired during pressure testing

26 3.2.2 Pressure Cycle Fatigue Tests For the pressure cycle fatigue tests the inch x inch, Grade X42 pipe having 75% corrosion was cycled from 36% to 72% SMYS until a leak developed in the corroded region beneath the repair. The ComposiSleeve TM system achieved 767,816 cycles before a leak failure developed in the repair. Of the 11 different repair configurations tested to date by Stress Engineering, the ComposiSleeve TM system achieved the greatest number of cycles to failure. Provided below are data for other competing technologies. E-glass system: 19,411 cycles to failure (MIN) E-glass system: 32,848 cycles to failure E-glass system: 129,406 cycles to failure E-glass system: 140,164 cycles to failure E-glass system: 165,127 cycles to failure Carbon system (Pipe #1): 212,888 cycles to failure Carbon system (Pipe #2): 256,344 cycles to failure Carbon system (Pipe #3): 202,903 cycles to failure E-glass system: 259,537 cycles to failure Carbon system (Pipe #4): 532,776 cycles (run out, no failure) ComposiSleeve TM system: 767,816 cycles to failure (MAX) It should be noted that in order to achieve this number of cycles, the girth welds joining the end caps to the ComposiSleeve TM test sample had to be repaired multiple times. As with the burst tests, the pressure cycle fatigue tested integrated strain gages installed in the corroded region beneath the repair. The strain gages measured hoop and axial strains at designated cycle periods during testing. Figure 3.11 plots the maximum hoop strain and hoop strain range for Gage #1 located beneath the repair up to 100,000 cycles, while Figure 3.12 lists hoop and axial strains at designated cycle periods. The strain range is the difference between the maximum and minimum strains.

27 Hoop Strain Beneath Repair (microstrain, 10,000 µe = 1 percent) Strain as a Function of Cycle Number Fatigue test of inch x inch, Grade X42 pipe with 75 % corrosion cyled from 890 to 1,780 psi (72% SMYS) with ComposiSleeve having E-glass thickness of 0.36 inches Maximum Strain Strain Range Low strain range critical to ensuring long-term performance (outside repair in base pipe with NO corrosion only Δε = 460 µε) Cycle Number Figure 3.11 Strain as a function of cycle number for 75% corroded sample Cycle Count ,000 2,000 5,000 10,000 20,000 50, ,000 Strain Type H3 A3 H1 A1 H2 A2 H4 A4 H5 A5 µe µe µe µe µe µe µe µe µe µe Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Maximum Minimum Range Figure 3.12 Hoop and axial strain listed as a function of cycle number (refer Figure 4.2 for location of strain gages)

28 In viewing the strain data from Figure 3.11 and Figure 3.12 there are several important observations. Note that the data plotted in Figure 3.11 is for Gage #1 located beneath the repair. The maximum measured strain beneath the repair (i.e. Gage #1) increase from 1,756 µε at start-up to 3,318 µε at 100,000 cycles. This is approximately a 90% increase over the designated cycle period. Over the 100,000 cycle recording period there was actually a decrease in the Gage #1 hoop strain range from 914 µε at start-up to 810 µε at 100,000 cycles. This represents a strain reduction on the order of 12%. The strain range in the base considering the 36% SMYS pressure differential was 460 µε. The average hoop strain range measured during the 100,000 cycle period was 811 µε, a value that is approximately 76% more than the nominal base pipe. The DOE B-curve is provided below, where ΔS represents strain range in psi. If the maximum hoop strain of 914 µε is considered, the corresponding design life is estimated to be 1,836,094 cycles. The actual fatigue failure data was 767,816 cycles. The likely explanation for the difference is rooted in the presence of a stress concentration factor (SCF) in the vicinity of the machined corrosion. A stress concentration of only 1.24 generates a cycle to failure condition of 776,000 cycles using the DOE B-curve. An SCF of this magnitude is completely consistent what could be expected for a machined corrosion region. σ N = 2.343E It is clear from the significant number of cycles that the ComposiSleeve TM system is uniquely effective in reducing strain in the reinforced region of the pipe. This statement is reinforced by the fact that the ComposiSleeve TM system was able to increase the fatigue life beyond other composite-based systems (although one carbon data point at 532,776 cycles is a run out condition). To be useful from a design standpoint the 767,816 cycles to failure must by adjusted to account for variations in fatigue data and to introduce an acceptable level of conservatism. In other words, even though the test ran for more than 750,000

29 cycles, it is not appropriate to assume that an actual repair can function for this many cycles. If a safety factor of 10 is employed, the number of design cycles for the ComposiSleeve TM system is 76,781 cycles (i.e. 767,816 cycles / 10). From a high pressure gas pipeline standpoint pipeline applications standpoint, the 76,781 design cycles for a pressure range of 36% SMYS corresponds conservatively to a design condition of 21 years for a very aggressive pressure cycle condition (3,683 cycles per year), 228 years for a moderately aggressive pressure cycle condition (337 cycles per year), and 600 years for a light pressure cycle condition (128 cycles per year). The data from Kiefner et al 2 in Figure 3.13 was used in generating the above service life values. Percent Very SMYS Aggressive Aggressive Moderate Light Total Single equivalent number of cycles with DP as noted 72% % 3, Figure 3.13 Service life based on pressure cycle range 2 Kiefner J. F. et al, Estimating Fatigue Life for Pipeline Integrity Management, Paper No. IPC , Presented at the International Pipeline Conference, Calgary, Canada, October 4 8, 2008.

30 4.0 ADDITIONAL TESTING: METHODS AND RESULTS In addition to the corrosion tests presented previously, Western Specialties conducted several other tests to evaluate the ability of the ComposiSleeve TM to reinforce pipelines subjected to bending loads, repair leaks, and also provide external reinforcement to pipeline subjected to excessive overburden loads. Provided in Appendix A, Appendix B, and Appendix C are the detailed reports issued previously on these respective testing efforts; however, a brief synopses of these test programs are included in this section of the report. 4.1 Bend Testing (November 2008) Testing was conducted to evaluate the ComposiSleeve TM system s ability to reinforce pipes subjected to bending loads. The evaluation involved conducting physical bend tests where inch x 0.25-inch, Grade X42 pipe samples were subjected to constant bending loads that might be experienced during flash floods and washouts in buried or partially-buried pipelines. The system was evaluated under two separate configurations: one configuration included orienting the half shell seams parallel to the bending loads, where the seams were at 0 and 180 this configuration is referred to as the Top Seam Reinforcement (TSR); the second configuration had the half shell seams oriented perpendicular to the bending loads at 90 and 270 this configuration is referred to as the Side Seam Reinforcement (SSR). A schematic of the test set-up is shown in Figure 4.1. All bend tests were performed with the samples pressurized to 72% SMYS. The final three point bend test results show the TSR configuration having approximately two times the bending capacity of the SSR for a pipe loaded with internal pressure of 72% SMYS. Furthermore, the SSR configuration displayed minimal advantage to a bare pipe loaded with internal pressure to 72% SMYS. Similar were observed in both the four point and three point bend tests. Finally, the design bending moment per the ASME B31.4 stress limit of 54% SMYS is 90 kip-ft. The TSR increases the bending capacity of the pipe to a level that is approximately two times the design limit. Refer to additional details and the full report in Appendix A.

31 I-beams (fixed by all-thread studs) I-beam Strong-back This configuration forces all of the deformation to take place within the reinforced region of the pipe. Vertical deformation and the hydraulic ram load were monitored during testing. Figure 4.1 Schematic diagram of the bend test set-up 4.2 Leak Repair Testing (April 2009) Two (2) repairs were made on two test samples, using inch x inch, Grade X42 pipe material, that each had a 60% deep corrosion defect. One of the test samples had a thru-wall hole. Prior to burst testing, each test sample was subjected to 22,503 pressure cycles from 890 psi to 1,780 psi, or from 50% to 100% of the maximum operating pressure (MAOP). After the pressure cycles, the samples were subjected to a static burst test. Sample #1 developed a noticeable leak at approximately 4,000 psi of internal pressure. The maximum pressure reached was 4,163 psi. Sample #1 did not burst due to a leak that developed at the high pressure. The leak occurred in the repair region. Sample #2 burst in the base pipe away from the repair where the maximum pressure applied to the sample was 4,033 psi. Figure 4.2 provides two photographs of the thru-wall defect test sample prior to repair. Refer to additional details and the full report in Appendix B.

32 Figure 4.2 Photographs of test sample having a thru-wall defect 4.3 Overburden Testing The testing program involved four pipe samples fabricated from 24-inch x 0.25-inch, Grade X42 pipe material. The first sample did not have any reinforcing materials, while the second sample was a composite-reinforcement only sample that included the installation the Pipe Wrap A+ material (0.64-inch thick). The third and fourth samples included the installation of the Composi-Sleeve TM pipeline reinforcement system having a steel thickness of 0.25 inches and Pipe Wrap A+ composite material thicknesses of 0.20 inches (10 wraps) and 0.30 inches (15 wraps), respectively. The set-up used by SES included support saddles placed on top and bottom of each test sample. Instrumentation included strain gages, displacement transducers, and pressure

33 transducers. A vertical load of 400,000 lbs was applied to each sample that was pressurized to 630 psi (72% SMYS). The primary observations from testing included the reduction in axial strain in the pipe provided by the Composi-Sleeve TM system, as well as the increased radial stiffness. As a point of reference, at a vertical load of 400,000 lbs the deflection in the unreinforced pipe was 2.35 inches, while the 15-wrap Composi- Sleeve TM system had a deflection of only 1.5 inches. As a point of comparison, the calculated stiffness values for these two different systems are lbs/in and lbs/in, respectively. Figure 4.3 are photographs showing the test set-up for the compression test involving a load frame having a 400,000 lbs capacity. Refer to additional details and the full report in Appendix C. Figure 4.3 Test set-up using a 400,000 lbs load capacity frame

34 5.0 DISCUSSION This report has provided details on testing conducted by Stress Engineering Services, Inc. to evaluate the ComposiSleeve TM pipeline repair system. The results demonstrate that the ComposiSleeve TM system can significantly reduce strain in corroded sections of pipe, the end results being the ability to restore pressure integrity and increase pressure cycle fatigue life. One of the challenges concerning the ability of composite-based repair systems to repair damaged pipelines concerns long-term performance. The issue primarily relates to degradation of the polymer-based resins and adhesives that are central components in every composite repair system. One of the techniques commonly used in the composite repair industry is to use large safety factors (i.e. between 5 and 10) on design stresses relative to long-term adhesive lap shear and composite tensile strengths. The ensures that even in the presence of material degradation performance levels of the composite repair system do not reach unacceptable levels. Additionally, as long as the composite system possesses adequate strength and stiffness, the strain in the damaged sections of the pipeline will not reach unacceptable levels. As observed in the detailed reports provided in the appendices, the ComposiSleeve TM system can be specifically designed for reinforcing a wide range if pipeline anomalies. As demonstrated in this report, performance-based testing is the ideal means for evaluating the ability of the ComposiSleeve TM system. The overall approach should, as a minimum, include identifying the pipeline anomaly, determining the required level of reinforcement, designing and installing the repair, and evaluating the overall performance of the repair system via full-scale destructive testing. Through monitoring strain gages beneath the repair and identifying the ultimate capacity of the repair, it is possible to estimate how the repair will perform in actual service.

35 6.0 CLOSING COMMENTS Composite materials are widely-accepted as a viable means for repairing damaged high pressure pipelines. The ComposiSleeve TM is a hybrid system that employs a bonding adhesive, a steel strong-back, and a composite overwrap. Results provided in this report demonstrate that the ComposiSleeve TM significantly reduces strain in the damaged section so that the integrity of even severely corroded pipeline are restored to levels similar to what would be expect in an undamaged section of pipe. The ComposiSleeve TM represents an engineered system that can be designed to meet the specific needs for a high pressure pipeline system.

36 Appendix A Report on Bend Testing (Prior Work)

37 BEND TESTING OF THE WESTERN SPECIALTIES ComposiSleeve TM REINFORCEMENT SYSTEM Prepared for Western Specialties, LLC Sedro-Woolley, WA November 2008 PN117847CRA Stress Engineering Services, Inc. Houston, Texas

38 Bend Testing of the Western Specialties Pipeline ComposiSleeve TM Reinforcement System SES PN CRA Prepared for Western Specialties, LLC Sedro-Woolley, WA Prepared By: Julian Bedoya Senior Analyst Chris Alexander, Ph. D. Principal November 2008

39 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 EXECUTIVE SUMMARY Stress Engineering Services, Inc. (SES) was contracted by Western Specialties, LLC (WS) to conduct testing to evaluate the Western Specialties Pipeline ComposiSleeve TM Reinforcement System (patent pending). The WS reinforcement system integrates steel half shells adhered to the pipeline using an epoxy mixture, and then reinforced with circumferentially wrapped composite materials. The evaluation involved conducting physical bend tests where inch x 0.25-inch X42 pipe samples were subjected to constant bending loads that might be experienced in flash floods and washouts in buried or partially-buried pipelines. The WS Pipeline Reinforcement System was evaluated under 2 separate configurations: one configuration included orienting the half shell seams parallel to the bending loads, where the seams were at 0 o and 180 o this configuration is referred to as the Top Seam Reinforcement (TSR); the second configuration had the half shell seams oriented perpendicular to the bending loads at 90 o and 270 o this configuration is referred to as the Side Seam Reinforcement (SSR). All bend tests were performed with the samples pressurized to 72% SMYS. The final three point bend test results show the TSR configuration having approximately two times the bending capacity of the SSR for a pipe loaded with internal pressure of 72% SMYS. Furthermore, the SSR configuration displayed minimal advantage to a bare pipe loaded with internal pressure to 72% SMYS. Similar were observed in both the four point and three point bend tests. Finally, the design bending moment per the ASME B31.4 stress limit of 54% SMYS is 90 kip-ft. Therefore, one can conclude that the TSR increases the bending capacity of the pipe to a level that is approximately two times the design limit. Stress Engineering Services, Inc. Page i SES Project No.:

40 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 LIMITATIONS OF THIS REPORT The scope of this report is limited to the matters expressly covered. This report is prepared for the sole benefit of Western Specialties, LLC. In preparing this report, Stress Engineering Services, Inc. (SES) has relied on information provided by Western Specialties, LLC. SES has made no independent investigation as to the accuracy or completeness of such information and has assumed that such information was accurate and complete. Further, SES is not able to direct or control the operation or maintenance of client s equipment or processes. All recommendations, findings and conclusions stated in this report are based upon facts and circumstances, as they existed at the time that this report was prepared. A change in any fact or circumstance upon which this report is based may adversely affect the recommendations, findings, and conclusions expressed in this report. NO IMPLIED WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE SHALL APPLY. STRESS ENGINEERING SERVICES, INC. MAKES NO REPRESENTATION OR WARRANTY THAT THE IMPLEMENTATION OR USE OF THE RECOMMENDATIONS, FINDINGS, OR CONCLUSIONS OF THIS REPORT WILL RESULT IN COMPLIANCE WITH APPLICABLE LAWS OR PERFECT RESULTS. Stress Engineering Services, Inc. Page ii SES Project No.:

41 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 TABLE OF CONTENTS Page No. EXECUTIVE SUMMARY... i TABLE OF CONTENTS... iii LIST OF FIGURES... iv 1.0 INTRODUCTION TESTING METHODS Sample Preparation Testing Protocols Testing Procedure RESULTS Four point bend tests Three Point Bend Tests Finite Element Analysis DISCUSSION CONCLUSIONS...19 Appendix A: Pipe Mill Test Report...20 Appendix B: Half Shell Mill Test Report...23 Stress Engineering Services, Inc. Page iii SES Project No.:

42 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 LIST OF FIGURES Page No. Figure 1. Strain gauge installation locations for both samples Figure 2. Picture of sample with the WS Pipeline Reinforcement System in place... 7 Figure 3. Strain gauge installation locations on the composite material Figure 4. Test set-up for the bending tests... 8 Figure 5. Picture of sample in four point bend test fixture Figure 6. Bending moment vs. microstrain for samples 1 and 2 during the 4-point bend tests Figure 7. Sketch of the three point bend test set-up Figure 8. Actual photograph of the three point bend tests Figure 9. Bending moment vs. center displacement for the three point bend tests Figure 10. General geometric features of the finite element models of the four point bend tests Figure 11. Finite element mesh used in the numerical models of the four point bend tests Figure 12. Von Mises stress for a pressurized (1,186 psi, 72% SMYS) reinforced and unreinforced pipe at the load indicated Figure 13. Von Mises stress for a pressurized (1,186 psi, 72% SMYS) reinforced pipe at the load indicated Figure 14. Moment vs. center deflection results of the modeled four point bend finite element analyses Stress Engineering Services, Inc. Page iv SES Project No.:

43 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November INTRODUCTION Western Specialties, LLC (WS) has been contracted by several pipeline companies to develop a system to reinforce pipelines that are in regions of flashfloods and washout. During a flashflood, buried or partially-buried pipelines may experience increased exposure due to washout. This results in increased unsupported lengths subject to higher bending loads and stresses. These resulting stresses have the potential of exceeding the allowable stresses per existing ASME pipeline standards and regulations. The WS reinforcement system integrates steel half shells adhered to the pipeline using an epoxy mixture, and then reinforced with circumferentially wrapped composite materials. Stress Engineering Services, Inc. (SES) was contracted by WS to conduct testing to evaluate the WS Pipeline Reinforcement System under bending loads. The testing program consisted of comparing the moment bearing capacity of an unreinforced pipeline relative to 2 different configurations of reinforced pipelines. To better represent inservice conditions, the pipelines were tested while pressurized to the maximum operating pressure (MAOP) or 72% of the minimum specified yield strength (SMYS) and then subjected to bending loads. A finite element model was also analyzed using the same loading conditions to compare to the test results. As a benchmark, a hand calculation predicts a plastic moment of 158 kip-ft for an unreinforced inch x 0.25-inch X42 pipe. The following sections of this report detail the test methods, set-up, and results of the tests conducted on the WS Pipeline Reinforcement System. Stress Engineering Services, Inc. Page 5 SES Project No.:

44 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November TESTING METHODS The sections that follow provide specific details on testing methods associated with the bending tests. 2.1 Sample Preparation The test samples were fabricated using a 30 foot long section of in x 0.25-in, Grade X42 pipe. Weld caps were welded to the ends of the samples, and the center 10 feet of the samples was sandblasted. Six strain gauges were installed as shown in Figure ft long sample (typical) Measure everything from center of sample 15 feet 1B, 0 o 5U, 90 o 4U, 0o 2B, 90 o Note: B refers to biaxial, and U refers to uniaxial strain gauges. 6U, 180 o 3B, 180 o 6 inches Figure 1. Strain gauge installation locations for both samples. WS proceeded to install the steel half shells (9 feet in total length two 3-foot sections and two 18-inch sections per sample) at the center of each sample. One sample had the half-shell seams oriented at 0 o and the other sample at 90 o to test the 2 possible extremes under which the reinforced sections might be loaded in. The half shells were installed on the pipe by coating it with epoxy and securing them on the pipe with ratchet straps for at least 60 minutes. Pipe Wrap A+, an E-glass water-activated urethane system, was then installed on top of the steel half shells at 7 locations, each 10 inches in length see Figure 2. Stress Engineering Services, Inc. Page 6 SES Project No.:

45 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Figure 2. Picture of sample with the WS Pipeline Reinforcement System in place. After curing for 24 hours, an additional 3 biaxial strain gauges were installed at the center of the reinforced region to measure the strain at the composite fibers during the bend tests, as shown in the schematic in Figure ft long sample (typical) Measure everything from center of sample 15 feet 7B, 0 o 8B, 90 o Note: B refers to biaxial, and U refers to uniaxial strain gauges. 6 inches 9B, 180 o Pipe wrap reinforcement and half shells (9ft total length). Figure 3. Strain gauge installation locations on the composite material. Stress Engineering Services, Inc. Page 7 SES Project No.:

46 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November Testing Protocols The sections that follow provide specific details on testing protocols associated with the burst and fatigue tests Testing Procedure Once the installation of the strain gauges in the samples was finished, they were placed in a four-point bend test set-up as shown in Figure 4. A four-point bend set-up ensures a constant bending moment is imparted to the reinforced section of the samples during the test located between the two hydraulic cylinders. Four Point Test Set-up Assembly Not To Scale! 30-ft long sample (typical) Measure everything from center of sample 29-in inches inches inches 42 inches 36-in 6 inches Reaction braces (2 required) 54-in Hydraulic cylinders (2 required) 0 o always in compression I-beam Strong-back Notes: 1. Both samples will be approximately 30-ft 2. Use cradle atop the half-cylinder fixtures to apply loading from hydraulic rams (omitted for clarity) 3. Applied bending moment is Ram Load (kip) times inches Measure deflection at these locations (5 places) Figure 4. Test set-up for the bending tests. Stress Engineering Services, Inc. Page 8 SES Project No.:

47 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Figure 5. Picture of sample in four point bend test fixture. Once the sample was securely placed in the fixture, strain gauges, displacement transducers, and pressure transducers were instrumented and calibrated. During the actual tests, data were recorded at a rate of 1 scan per second. Prior to applying the bending loads, the sample was pressurized to 72% SMYS, which corresponds to an internal pressure of 1,186 psi. During the application of the bending moment loads, the internal pressure was maintained at 1,186 psi +/- 50 psi. Table 1 shows a summary of the applied loads during the tests. Table 1. Summary of Applied Ram Loads to be held for 2 minutes. Applied Ram Percent Yield Moment Total Load To Induce %My Load While Sample is Pressurized 0% 0 kip 25% 6.34 kip 1.95 kip 50% 12.7 kip 8.28 kip 100% 25.4 kip kip 150% 38.1 kip kip Plastic Collapse Stress Engineering Services, Inc. Page 9 SES Project No.:

48 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November RESULTS The sections that follow provide specific details on results associated with the bend testing efforts. Four-point bend tests were conducted to evaluate the performance of the TSR and the SSR under bending loads. Three-point bend tests were conducted to determine the maximum bending capacity of the reinforced sections under both reinforcement extremes and to compare to the four-point bend tests. 3.1 Four point bend tests Sample 1 had the half-shell seams oriented parallel to the direction of loading 0 o (at the top and bottom of the reinforcement). Sample 1 reached a maximum bending moment of 204 ft-kip. The test was stopped at this load due to the displacement at the center of the sample exceeding 8 inches. The maximum axial strain in tension under the reinforced section was 0.15%. Sample 2 had the half-shell seams oriented perpendicular to the direction of loading (on the sides of the sample, 90 o ). Sample 2 reached a maximum bending moment of 203 kip-ft, but the maximum displacement was 9 inches. In addition, the magnitude of the axial strain was significant; the axial strain gauge in the reinforced region saturated at 3.5%, indicating axial strains in excess of 3.5% were experienced by the sample. Figure 6 shows bending moment as a function of strain for samples 1 and 2. According to ASME B31.4, there is an axial stress limit of 54% allowed in pipes during service 1. Sample 1 with the seam at the top, or 0 o Top Seam Reinforcement (TSR) reached this threshold at a moment of 95 ft-kip; sample 2 with the seam at 90 o Side Seam Reinforcement (SSR) reached this threshold at a moment of 40 ft-kip; and a bare pipe reached this threshold also at 40 ft-kip. All three samples were pressurized to 72% SMYS, or 1,186 psi. As can be appreciated, the SSR did not differ from the bare pipe results, suggesting no reinforcement advantage. However, the TSR showed a moment capacity that is 2.4 times of a bare pipe and a SSR. 1 It is stated in paragraph (d) of ASME B31.4 that The sum of the longitudinal stress due to pressure, weight and other sustained external loadings shall not exceed 0.75 S A. Furthermore, the Code also states The allowable stress range S A for unrestrained lines shall not exceed 72% of the SMYS of the pipe. Therefore, the axial stress limit is 0.75 * (0.72 * SMYS) which is 54% SMYS, or 22,500 psi for a Grade X42 pipe. Stress Engineering Services, Inc. Page 10 SES Project No.:

49 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Bending Moment (kip-ft) Bending Moment as a Function of Axial Strain Experimental results on the Western Specialties half-pipe bending reinforcement system with measurements taken by strain gages located beneath the reinforcement on the base pipe. 756 me elastic strain limit (54% SMYS per ASME B31.4) WS Repair w ith seam at TOP (Design Moment of 135 kip-ft) Axial Gage (Sample #1 - Top Seam) Axial Gage (Sample #2 - Side Seam) Double elastic slope line (Sample #1) Double elastic slope line (Sample #2) Bare pipe with no reinforcement WS Repair w ith seam at SIDE (Design Moment of 90 kip-ft) Strain (microstrain) (10,000 microstrain is equal to 1 percent strain) Figure 6. Bending moment vs. microstrain for samples 1 and 2 during the 4-point bend tests. 3.2 Three Point Bend Tests In an effort to confirm the trends observed in the TSR and SSR four-point bend tests and to determine the capacity of both reinforcement extremes, a three-point bend test of the same samples put through the four-point bend tests were conducted. The main difference between a four-point bend and a three-point bend is that in the former, there is a constant and pure bending moment load in a specified region of the sample (the region in between the hydraulic cylinders), whereas in the three point bend, the maximum bending moment only occurs in the center of the sample at the location of the single hydraulic cylinder. Because the hydraulic cylinder was placed in the center of the reinforcement, a three point bend forces the maximum deformation to occur in the reinforced region, whereas in a four point bend, there could be significant deformation and strain at the location where the pipe transitions from reinforced to unreinforced and thus the maximum deformation and strain will not necessarily occur at the center of the sample. Of secondary importance to these specific tests conducted for WS, is that there is a non-zero shear load in addition to the bending load in a three-point bend test. This shear load imparts stresses that are Stress Engineering Services, Inc. Page 11 SES Project No.:

50 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 roughly one order of magnitude smaller than the bending stresses for the sample size used for the WS tests. Figure 7 shows a sketch of the experimental set-up, and Figure 8 shows a photograph of the actual set-up. For simplicity, only bending moment, internal pressure, and center displacement of the samples were monitored during the three-point bend tests. I-beams (fixed by all-thread studs) I-beam Strong-back Figure 7. Sketch of the three point bend test set-up. Figure 8. Actual photograph of the three point bend tests. Stress Engineering Services, Inc. Page 12 SES Project No.:

51 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Figure 9 shows the bending moment as a function of maximum displacement for the three point bend tests. This test determined the true limit of the capacity associated with the reinforcement. The design bending moments based on the three point bend test for the TSR and SSR are 194 kip-ft and 97 kip-ft, respectively. Bending Moment (kip-feet) Bending Moment as a Function of Displacement Three point bend test conducted to determine maximum capacity of Western Specialties steelcomposite reinforced system. Design bend capacities for both seam orientations shown below. WS Repair with seam at TOP (Limit Load of 324 kip-ft) Design Moment = 0.6 * 324 kip-ft = 194 kip-ft) WS Repair with seam at SIDE (Design Moment of 162 kip-ft) Design Moment = 0.6 * 324 kip-ft = 97 kip-ft) Vertical Displacement (inches) Half-shell Side Seam Orientation Half-shell Top Seam Orientation Double Elastic Slope Line (Top Seam) Double Elastic Slope Line (Side Seam) Figure 9. Bending moment vs. center displacement for the three point bend tests. 3.3 Finite Element Analysis Finally, an effort was conducted to model the WS Pipeline Reinforcement System via finite element methods. The four-point bend tests were modeled including the bare pipe, the steel half shells, the composite wraps, and the epoxy. The saddles were also modeled; except they were modeled as being rigid (infinitely stiff) and point loads were applied at the same locations where the rams were applied in the actual tests. The finite element models also considered internal pressure and plasticity for the steel components. Figure 10 shows some of the details of the finite element modeled components. Figure 11 shows the finite element mesh used in the region of interest of the model (the center of the reinforced pipeline). Table 2 shows the mechanical properties used in the finite element models. Stress Engineering Services, Inc. Page 13 SES Project No.:

52 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Figure 10. General geometric features of the finite element models of the four point bend tests. Figure 11. Finite element mesh used in the numerical models of the four point bend tests. Table 2. Summary of material properties and thicknesses used in the finite element models. Component Young s Modulus (psi) Yield Strength (psi) Ultimate Tensile Strength (psi) Elongation % (2 Specimen) Thickness Modeled Pipe 30X ,000 60,000 20% 0.25 inches Sleeves 30X ,000 60,000 20% 0.25 inches Epoxy 1.5X10 5 Modeled as Elastic Modeled as Elastic Modeled as Elastic 0.25 inches Composite 3X10 6 Modeled as Elastic Modeled as Elastic Modeled as Elastic 0.25 inches Figure 12 shows the Von Mises stress in the bare pipe and in a reinforced pipe at a load of 125 kip-ft. Note that the deformation scale is the same for both results. There is significantly more deformation in the unreinforced model than in the reinforced model. Stress Engineering Services, Inc. Page 14 SES Project No.:

53 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Finite Element Results: Von Mises Stress at 14.2 Kips, 125 kip-ft, Deformation Scale=3.0 Bare Pipe Regions in Gray are above SMYS Repaired Pipe Figure 12. Von Mises stress for a pressurized (1,186 psi, 72% SMYS) reinforced and unreinforced pipe at the load indicated. Figure 13 shows the reinforced pressurized pipe at a moment of 181 kip-ft. Note how the regions where the reinforcement begins has significant stresses due to an abrupt change in stiffness in the system. The deformed shape is similar to that observed in the actual tests. Finite Element Results: Von Mises Stress at 20.6 Kips, 181 kip-ft, Deformation Scale=3.0 Regions in Gray are above SMYS Repaired Pipe Figure 13. Von Mises stress for a pressurized (1,186 psi, 72% SMYS) reinforced pipe at the load indicated. Stress Engineering Services, Inc. Page 15 SES Project No.:

54 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Finally, Figure 14 shows a plot of bending moment vs. center deflection for a modeled bare pipe, the TSR, and the SSR reinforcement configurations with 72% SMYS internal pressure. Moment vs. Deflection for a 72% SMYS Pressurized 12.75" x 0.25" Pipe With and Without Reinforcement BARE PIPE Seam At 90 Degrees Seam At 0 Degrees Moment (kip*ft) Center Deflection (in) Figure 14. Moment vs. center deflection results of the modeled four point bend finite element analyses. Stress Engineering Services, Inc. Page 16 SES Project No.:

55 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November DISCUSSION In this study the WS Pipeline Reinforcement System was analyzed in physical testing as well as a very basic numerical (finite element) analysis. The major points of discussion include the level of reinforcement achieved with each configuration (TSR vs. SSR). For the four-point bend tests, there was no discernible difference between the SSR configuration and a bare pipe. This means that if a reinforced pipe were to be subjected to loads in which the seams of the half shells were perpendicular to the direction of the bending moment, there would be no advantage of the pipe being reinforced. On the other hand, if the bending moment were to be parallel to the seams of the half-shells, the fourpoint bend results indicated a bending capacity of 2.4 times a bare pipe or a pipe with SSR reinforcement. In the three point bend tests, the total load capacities of the samples increased, however, the trends remained generally the same; the TSR configuration showed a design moment of 194 kip-ft whereas the SSR showed a design moment of 97 kip-ft. A hand calculation for a inch x 0.25-inch X42 pipe using minimum Code material properties indicates a yield moment of kip-ft (including a pressure end load that is 36% SMYS), again confirming that there is minimal advantage to being reinforced in the SSR configuration. Referring to ASME B31.4, the axial stress limit of 54% SMYS corresponds to a bending moment of 90 kip-ft. In contrast, the finite element results suggest there is a benefit of reinforcing the pipe with the WS Pipeline Reinforcement System regardless of the orientation of the seams of the steel half shells. However, the finite element model does not account for the reduction in effectiveness of the epoxy in shear or normal loads, and thus the model is inherently incapable to capture this important behavior. Mechanically, the half shells in the SSR configuration are much less capable in providing reinforcement to the pipe because the half shells have significantly less moment of inertia than the half shells in the TSR configuration. From the SSR results, it seems that the composite wrap material is not adding significant reinforcement to the system when installed in discrete sections, although it might be preventing the half shells from Stress Engineering Services, Inc. Page 17 SES Project No.:

56 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 separating from the pipe. It is thought that a continuous section of composite material could provide an improvement in the reinforcement. From a final assessment standpoint, it is suggested that the results from the three point bend test be used in developing the design basis for the ComposiSleeve TM system. Consider the following: The ASME B31.4 axial stress limit corresponds to a calculated design bending moment of 90 kip-ft The design moment capacity for the TSR configuration is 194 kip-ft The design moment capacity for the SSR configuration is 97 kip-ft Therefore, one can conclude that the ComposiSleeve TM in the TSR orientation increases the bending capacity of the ASME B31.4 bare pipe design by 2.15 times. The SSR orientation yields minimal increase over the stress limits based on the test results. It is possible with some modifications to the design that the SSR capacity could be increased. Stress Engineering Services, Inc. Page 18 SES Project No.:

57 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November CONCLUSIONS This report has provided technical details on the performance of the WS ComposiSleeve TM Reinforcement System under two reinforcement configurations; one where the seams of the half shells are oriented parallel to the applied bending load (TSR), and one where the half shell seams are perpendicular to the bending load. The testing results indicate that the TSR configuration reinforces a bare pipe on the order of 2.15 times the design moment per the ASME B31.4 axial stress limit of 54% SMYS. Furthermore, the results of the SSR configuration show that there is minimal benefit of the reinforcement system when loaded in this scenario. The finite element results show a benefit of reinforcement, but the simple model analyzed cannot capture the disbonding of the epoxy layer so there is no distinction between TSR and SSR. The finite element model can be further refined, however, the effort would be more significant and it would only capture what the physical tests have already confirmed. Stress Engineering Services, Inc. Page 19 SES Project No.:

58 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Appendix A: Pipe Mill Test Report Stress Engineering Services, Inc. Page 20 SES Project No.:

59 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Stress Engineering Services, Inc. Page 21 SES Project No.:

60 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Stress Engineering Services, Inc. Page 22 SES Project No.:

61 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Appendix B: Half Shell Mill Test Report Stress Engineering Services, Inc. Page 23 SES Project No.:

62 Western Specialties, LLC Bend Testing of the Western Specialties ComposiSleeve TM System November 2008 Stress Engineering Services, Inc. Page 24 SES Project No.:

63 Appendix B Report on Leak Repair (Prior Work)

64 SENIOR PRINCIPALS President Joe R. Fowler, Ph.D., P.E. Senior Vice President W. Thomas Asbill, P.E. Vice Presidents Ronald D. Young, Ph.D., P.E. Clinton A. Haynes Jack E. Miller, P.E. J. Randy Long, P.E. PRINCIPALS James W. Albert, P.E. Christopher Alexander, Ph.D. Claudio Allevato, Corp. LIII Kenneth K. Bhalla, Ph.D. Mark A. Bennett, P.E. Richard S. Boswell, P.E. Helen Chan, C.P.A. John F. Chappell, P.E. Kimberly O. Flesner, P.E. S. Allen Fox, P.E. David L. Garrett, Ph.D. Andreas T. Katsounas, P.E. Terry M. Lechinger Christopher Matice, Ph.D., P.E. Charles A. Miller, P.E. George R. Ross, Ph.D., P.E. Ramón I. San Pedro, P.E. Teri Shackelford Matthew J. Stahl, D.Eng., P.E. David A. Tekamp, P.E. Kurt D.Vandervort, Ph.D., P.E. Kenneth R. Waeber, P.E. Robert E. Wink, P.E. Bobby W. Wright, P.E. SENIOR ASSOCIATES/ STAFF CONSULTANTS Glenn A. Aucoin, P.E. Richard C. Biel, P.E. J. Kirk Brownlee, P.E. Michael J. Effenberger, P.E. Greg Garic, P.E. Robert B. Gordon, Ph.D., P.E. Lori C. Hasselbring, Ph.D., P.E. David P. Huey, P.E. Daniel Krzywicki, P.E., CSP Kenneth R. Riggs, Ph.D., P.E. SENIOR ASSOCIATES Rafik Boubenider, Ph.D., P.E. Roger D. Cordes, Ph.D., P.E. Donnie W. Curington Nripendu Dutta, Ph.D., P.E. Steven A. Garcia Mark E. Hamilton Brett A. Hormberg William A. Miller, P.E. John M. Moore Ronald A. Morrison, P.E. Thomas L. Power, Ph.D., P.E. Brian S. Royer, P.E. Mahmod Samman, Ph.D., P.E. Daniel A. Pitts, P.E. Lane E. Wilson Leo Vega STAFF CONSULTANTS Ray R. Ayers, Ph.D., P.E. Clinton H. (Clint) Britt, P.E. Joe W. Frey, P.E. Michael W. Guillot, Ph.D., P.E. Steve Hoysan, Ph.D., P.E. Joe Kintz, P.E. Paul J. Kovach, P.E. Ron Scrivner Jackie E. Smith, P.E. ASSOCIATES Lyle E. Breaux, P.E. Douglas E. Drahem Kenny T. Farrow, Ph.D. Stuart J. Harbert, Ph.D., P.E. Scot T. McNeill David F. Renzi Chad Searcy, Ph.D. Obaidullah Syed, P.E. Kevin Wang, Ph.D., P.E. SENIOR ANALYSTS Napoleon F. Douglas, Jr. Lixin Gong, Ph.D. Yun Han, Ph.D. Won Kim, Ph.D. Dilip Maniar, Ph.D. Bo Yang, Ph.D. Hong Zhou, Ph.D. ANALYSTS J. Julian Bedoya Jonathan Brewer Rhett L. Dotson Michael L. Ge Sachin V. Kholamkar Karen Lucio Brent Vyvial April 23, 2009 Mr. Carl Brooks Western Specialties, LLC 601 W. State Street Sedro-Woolley, WA Phone: (916) SUBJECT: Carl, Report on testing Pinhole Leak Composite Repair System PN117847CRA Stress Engineering Services, Inc. appreciates the opportunity to provide engineering testing services to you in evaluating the Western Specialties ComposiWrap composite leak repair system. Included in this document is a summary report detailing results from the cycling prior to the burst tests as well as the burst tests performed on April 6 th, 2009 in our Houston lab. Two (2) repairs were made on two test samples that each had one thru-wall hole. Each test sample was subjected to 22,503 pressure cycles from 890 psi to 1,780 psi, or from 50% to 100% of the maximum operating pressure (MAOP). After the pressure cycles, the samples were subjected to a static burst test. Sample #1 developed a noticeable leak at approximately 4,000 psi of internal pressure. The maximum pressure reached was 4,163 psi. Sample #1 did not burst due to the leak. The leak occurred in the repair region. Sample #2 burst in the base pipe away from the repair where the maximum pressure applied to the sample was 4,033 psi. We look forward to continuing our work with you. Please contact me if you have any questions. Regards, Chris Alexander chris.alexander@stress.com (281) (direct phone) (281) (cell phone) Report prepared by: Julian Bedoya Associate Julian.Bedoya@stress.com (281)

65 Summary Report Prepared for Western Specialties, LLC. Page 2 of 14 Prepared by Stress Engineering Services, Inc. SES PN Testing the W.S. ComposiWrap System Stress Engineering Services, Inc. (SES) performed two (2) pressure tests that included pressure cycling followed by two (2) burst tests on inch x inch Grade X42 pipes. Both samples had a simulated 60% external machined corrosion region measuring 8-inches longitudinally x 6-inches circumferentially, and a thru-wall drilled hole having a diameter of inches in the center of the corroded region see Figure 1. Western specialties then led the preparation of the samples for the installation of the pinhole repair system. Sample Preparation The sample preparation involved the following steps: 1. Sand the area to be repaired with 36 grit paper, including the inch diameter hole creating scuff marks for enhanced adhesion, and for debris and rust removal. * See Figure Sand 18-inch long steel half shells using the same technique as the repair region. See Figure Apply a bead of epoxy on the inch hole to cover it, and spread. See Figures 4 and Allow epoxy to cure. 5. Wipe the epoxy clean with a rag. 6. Apply epoxy filler material in machined area on pipe. Figure 6 shows the filler material being applied. 7. Sand high spots on epoxy surface to create a smooth contour. Figure 7 shows filler material being ground smooth. 8. Test metal sleeves over epoxy area for fit. 9. Measure and mark pipe to center sleeves on corroded region. 10. Wipe down surfaces to remove dust. 11. Place adhesive to inside surface of sleeves and install each sleeve over corroded region. Figure 8 shows the application of the epoxy to the inside of the steel half shells. 12. Ratchet strap the sleeves on pipe and let set for 90 minutes. See Figure 9. Figure 10 shows the steel half shells installed prior to installing the PipeWrap A+ product. 13. Apply thin layer of adhesive over sleeves and pipe. 14. Apply E-Glass water-activated urethane wrap (PipeWrap A+) 6 Rolls while spraying with water. See Figure Apply shrink wrap over E-Glass and perforate with holes to release gases. Figures 12 and 13 show Sample #1 and #2 with the complete installation. 16. Post mortem OD measurements were made for both samples. Sample #1 had 0.55-inches in thickness of the PipeWrapA+ material applied over the steel half shells and measuring 35 inches in length. Sample #2 had 1.01-inches in thickness of the PipeWrapA+ material applied over the steel half shells and measuring 18.5 inches in length. Fatigue Test Activities Once both repairs were complete, the following steps were carried out. 1. The samples were filled with water and placed in a trough for cycling and burst. 2. The samples were connected in series with the cyclic pump and with each other. See Figure The trough shields were shut for safety during cyclic pressurization. * The two samples were sandblasted back in September 2008, but had minor flash rust as the samples remained in SES test lab.

66 Summary Report Prepared for Western Specialties, LLC. Page 3 of 14 Prepared by Stress Engineering Services, Inc. SES PN Both samples were to be cycled simultaneously to a target of 20,000 cycles from 50% to 100% MAOP. (pressure range of 36% SMYS or psi) 5. The cyclic pump was then started, noting the cycle counter number. 6. After 22,503 cycles, the cyclic pump was shut down. Burst Test Activities Sample #1: 1. For the burst tests, only one sample was in the trough at a time. 2. High pressure fittings and autoclave tubing were used to connect the water-filled sample to the static water pump for burst testing. 3. The SES StrainDAQ data acquisition system was set-up for monitoring internal pressure of the sample. 4. After calibrating the pressure transducer, the testing of sample #1 began. 5. The pressure was increased at a rate of approximately 10 psi/second. 6. At approximately 1,778 psi, or 100% of the MAOP, the pressure was held constant for 2 minutes. 7. The pressure was then increased to the value corresponding to a hoop stress equal to the specified minimum yield strength (SMYS for the Grade X42 pipe material), or 2,470 psi of internal pressure (42,000 psi of hoop stress). The pressure was held constant for 5 minutes. 8. The pressure was then increased with the intent of bursting the sample. 9. At approximately 4,000 psi, a decrease in pressure was noticed as increases in pressure could not be achieved with the pumping unit. 10. A few minutes later, a visible water drip was noticed from the outside of the trough shields. 11. At this point in time, it was not clear if the leak existed in the tubing or in the repair region because the safety shields were in place. 12. An effort was made to try to burst the sample by continuing to increase the pressure in the sample. 13. The pressure in the sample continued to rise very slowly in part due to the leak, and also because the air pressure driving the pump was at the highest possible pressure. After 70 minutes of testing and the noticeable water puddle, the shut off valve in the static water pump was closed, and the sample lost pressure due to the existing leak. A maximum pressure of 4,163 psi was achieved. 14. Once the pressure reduced to a small value, the safety shields were opened, and the fitting connections were examined for leak. No evidence of leakage was found in any of the fittings (see Figures 15 17); however, the repair region in the sample was wet. Burst Test Activities Sample #2: 1. Steps 1 8 for Sample #1 were followed for Sample #2 with the exception that both holding periods were 5 minutes. 2. After 35 minutes of testing, the pressure in the sample reached 4,000 psi, and the rate of pressure increase was very slow. 3. After 55 minutes of testing the sample burst outside the repair region at 4,033 psi. See Figure 18. Test Results: Figure 19 plots the pressure vs. time recorded for each of the samples. The maximum pressures for each sample are noted below. Sample #1 reached a maximum pressure of 4,163 psi. Sample #1 began leaking at approximately 4,000 psi. Sample #2 reached a maximum pressure of 4,033 psi. Sample #2 burst outside the repair region and did not leak.

67 Summary Report Prepared for Western Specialties, LLC. Page 4 of 14 Prepared by Stress Engineering Services, Inc. SES PN inch x inch, Grade X42 pipe (8-feet long) 8 feet (center machined area on sample) 8 inches long 0.75-inch radius (at least) Break corners (all around) inches 60% corrosion: remaining wall of inches NOTE: Perform all machining 180 degrees from longitudinal ERW seam. 6 inches Note uniform wall in machined region Thru-wall hole Details on machining (machined area is 8 inches long by 6 inches wide) Figure 1 Sketch of machined corrosion and thru-wall hole. Figure 2 Sanding of sample.

68 Summary Report Prepared for Western Specialties, LLC. Page 5 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure 3 Sanding of steel half shells Figure 4 Plugging of inch hole with epoxy.

69 Summary Report Prepared for Western Specialties, LLC. Page 6 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure inch hole with epoxy spread out. Figure 6 Application of epoxy filler material to machined area.

70 Summary Report Prepared for Western Specialties, LLC. Page 7 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure 7 Grinding of hard epoxy prior to applying steel half shells. Figure 8 Application of epoxy to steel half shells prior to application.

71 Summary Report Prepared for Western Specialties, LLC. Page 8 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure 9 Secured steel half shells while curing. Figure 10 Cured steel half shells prior to applying PipeWrap A+ product.

72 Summary Report Prepared for Western Specialties, LLC. Page 9 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure 11 Application of PipeWrapA+ product. Figure 12 Completion of Sample #1 installation.

73 Summary Report Prepared for Western Specialties, LLC. Page 10 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure 13 Completion of Sample #2 installation. Figure 14 Placement of samples 1 and 2 in trough prior to starting fatigue testing.

74 Summary Report Prepared for Western Specialties, LLC. Page 11 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure 15 Dry fitting in sample 1 after shields were removed to inspect for leaks. Figure 16 Dry fitting in sample 1 after shields were removed to inspect for leaks.

75 Summary Report Prepared for Western Specialties, LLC. Page 12 of 14 Prepared by Stress Engineering Services, Inc. SES PN Figure 17 Wet repair region in sample 1 after shields removed to inspect for leaks. Figure 18 Sample #2 after burst test.

76 Summary Report Prepared for Western Specialties, LLC. Page 13 of 14 Prepared by Stress Engineering Services, Inc. SES PN ,500 4,000 3,500 Internal Pressure vs. Time For a inch X inch X42 pipe with External Machined Corrosion of 60% and a inch Hole in the Center of the Corrosion Repaired with Steel Half Shells, and Composite with 22,503 Cycles From 50% to 100% MAOP. 4,000 psi Leak Pressure 4,033 psi Burst Pressure Internal Pressure (psi) 3,000 2,500 2,000 1, pinhole-burst-sample sample2-pinhole-burst 1, ,000 2,000 3,000 4,000 5,000 6,000 Time (Seconds) Figure 19 Pressure vs. time for Samples #1 and #2.

77 Summary Report Prepared for Western Specialties, LLC. Page 14 of 14 Prepared by Stress Engineering Services, Inc. SES PN Conclusions The results of these tests demonstrate that the Western Specialties ComposiWrap repair system can reinforce thru-wall defects in inch x inch pipe with 60% simulated external corrosion with a inch drilled hole in the center of the corrosion patch. Considering the fact that both samples were subjected to over 22,000 50% to 100% MAOP pressure cycles prior to the burst tests is a significant result for this repair system. The fact that the failure in Sample #2 occurred outside of the repaired region is also an important consideration for this repair system used to reinforce damaged pipelines. As far as Sample #1 is concerned, it is estimated that the leak occurred at approximately 4,000 psi; significantly above the MAOP of the pipe and close to the tensile strength of the base pipe. Additionally, it is worth noting that the burst pressure of Sample #2 was close to the leak pressure in Sample #1. What has not been specifically addressed in this program is the long-term performance of the leak repair system. This can be achieved in two ways. 1. The adhesive that is to come into contact with the internal pipeline product (e.g. natural gas, ethanol, gasoline, crude oil, etc.) must not deteriorate over time due to extended exposure. 2. The externally-applied composite repair system must have long-term performance and be stressed to a level that is within the acceptable design levels. Based on prior communication with Western Specialties, both Lord and PipeWrap, LLC have communicated that they can provide the required documentation to address the above longterm performance elements. A final document can be prepared by SES that includes this required documentation.

78 Appendix C Report on Overburden Testing (Prior Work)

79 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System PN CRA Prepared for LMC Industrial Contractors, Inc. Avon, New York March 2010 Stress Engineering Services, Inc. Houston, Texas

80 Evaluation of the Composi-Sleeve TM Pipeline Reinforcement System PN CRA Prepared for LMC Industrial Contractors, Inc. Avon, New York Prepared by: Chris Alexander, Ph.D. Principal Testing work supervised by Mr. Brent Vyvial, Associate Finite element analysis work performed by Mr. Rhett Dotson, P.E., Senior Analyst March 2010

81 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page i Prepared by Stress Engineering Services, Inc. March 2010 EXECUTIVE SUMMARY Stress Engineering Services, Inc. (SES) was contracted by LMC Industrial Contractors, Inc. (LMC) to perform an assessment of the Composi-Sleeve TM pipeline reinforcement system that has been designed to provide structural rigidity to pipelines having overburden loads. The primary focus of this study was full-scale testing, although numerical modeling of pipeline overburden loads was accomplished using finite element analysis. One observation from the current study is that many of the closed-form equations used by industry fail to account for the stiffening contribution of soil and internal pressure in minimizing deformation of the pipe. As a result, the predicted deformation might be greater than will actually exist in a real pipeline. The testing program involved four pipe samples fabricated from 24-inch x 0.25-inch, Grade X42 pipe material. The first sample did not have any reinforcing materials, while the second sample was a composite-reinforcement only sample that included the installation the Pipe Wrap A+ material (0.64-inch thick). The third and fourth samples included the installation of the Composi- Sleeve TM pipeline reinforcement system having a steel thickness of 0.25 inches and Pipe Wrap A+ composite material thicknesses of 0.20 inches (10 wraps) and 0.30 inches (15 wraps), respectively. The set-up used by SES included support saddles placed on top and bottom of each test sample. Instrumentation included strain gages, displacement transducers, and pressure transducers. A vertical load of 400,000 lbs was applied to each sample that was pressurized to 630 psi (72% SMYS). The primary observations from testing included the reduction in axial strain in the pipe provided by the Composi-Sleeve TM system, as well as the increased radial stiffness. As a point of reference, at a vertical load of 400,000 lbs the deflection in the unreinforced pipe was 2.35 inches, while the 15-wrap Composi-Sleeve TM system had a deflection of only 1.5 inches. As a point of comparison, the calculated stiffness values for these two different systems are lbs/in and lbs/in, respectively. From an absorption standpoint the imparted energy to the test samples was on the order of 25,000 ft-lbs. This is equivalent to dropping a 1,000 lbs object from 25 feet (potential energy), or a 500 lbs rock traveling at 40 miles per hour (kinetic energy).

82 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page ii Prepared by Stress Engineering Services, Inc. March 2010 One can conclude from the test results of this program that the Composi-Sleeve TM provides reinforcement to the pressurized pipeline subjected to vertical loading. As observed from testing, one of the more significant contributions is reducing the amount of lateral expansion that takes place in a pipeline that is subjected to overburden loads.

83 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page iii Prepared by Stress Engineering Services, Inc. March 2010 LIMITATIONS OF THIS REPORT The scope of this report is limited to the matters expressly covered. This report is prepared for the sole benefit of LMC Industrial Contractors, Inc. In preparing this report, Stress Engineering Services, Inc. has relied on information provided by the LMC Industrial Contractors, Inc. Stress Engineering Services, Inc. has made no independent investigation as to the accuracy or completeness of such information and has assumed that such information was accurate and complete. Further, Stress Engineering Services, Inc. is not able to direct or control the operation or maintenance of client s equipment or processes. All recommendations, findings and conclusions stated in this report are based upon facts and circumstances, as they existed at the time that this report was prepared. A change in any fact or circumstance upon which this report is based may adversely affect the recommendations, findings, and conclusions expressed in this report. NO IMPLIED WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE SHALL APPLY. STRESS ENGINEERING SERVICES, INC. MAKES NO REPRESENTATION OR WARRANTY THAT THE IMPLEMENTATION OR USE OF THE RECOMMENDATIONS, FINDINGS, OR CONCLUSIONS OF THIS REPORT WILL RESULT IN COMPLIANCE WITH APPLICABLE LAWS OR PERFECT RESULTS.

84 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page iv Prepared by Stress Engineering Services, Inc. March 2010 TABLE OF CONTENTS Page No. EXECUTIVE SUMMARY... i TABLE OF CONTENTS...iv LIST OF FIGURES... v 1.0 INTRODUCTION ANALYTICAL METHODS AND RESULTS Description of Finite Element Analysis Simulating Actual Buried Pipelines Simulating the Experimental Test Set-up EXPERIMENTAL METHODS AND RESULTS Sample Fabrication Test Set-up Testing Results CLOSING COMMENTS... 25

85 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page v Prepared by Stress Engineering Services, Inc. March 2010 LIST OF FIGURES Page No. Figure 2.1 Example contour plot showing stress in dent... 3 Figure 2.2 Example contour plot showing stress in wrinkle bend... 3 Figure 2.3 Soil support model showing position of spring elements... 5 Figure 2.4 Stresses in soil support model including overburden and pressure... 5 Figure 2.5 Plot showing vertical deflection of soil support model... 6 Figure 2.6 FEA model used to simulate experimental test set-up... 8 Figure 2.7 Boundary conditions and loading on FEA test model... 8 Figure 2.8 End view of FEA test model showing vertical displacement... 8 Figure 2.9 Contour plot for FEA test model... 9 Figure 2.10 Comparison of analysis to testing results... 9 Figure 3.1 Steel half-shells provided by LMC Figure 3.2 Installation of adhesive on inner surface of steel half-shells Figure 3.3 Preparing to install steel half-shell on pipe surface Figure 3.4 Positioning restraining straps on outer surface of half-shells Figure 3.5 Half shells installed on pipe sample prior to the composite installation Figure 3.6 Installation of composite materials on the outer surface of the half shells Figure 3.7 Installation of stricture wrap on the outer surface of the composite Figure 3.8 Schematic diagram showing strain gage installation Figure 3.9 Basic diagram showing general loading scheme Figure 3.10 Photograph showing isometric view of the test set-up Figure 3.11 Photograph showing end view of the test set-up Figure 3.12 Close-up view showing the upper and lower saddles with rubber inserts Figure 3.13 Side view of the test set-up showing the vertical displacement transducer Figure Force versus vertical deflection of pipe Figure Force versus hoop strain in pipe beneath reinforcement Figure Force versus axial strain in pipe beneath reinforcement Figure Calculation of stiffness shown graphically... 24

86 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 1 Prepared by Stress Engineering Services, Inc. March INTRODUCTION This report provides details on the analysis and testing work performed by Stress Engineering Services, Inc. (SES) for LMC Industrial Contractors, Inc. (LMC) to evaluate the performance of the Composi-Sleeve TM pipeline reinforcement system to provide structural rigidity to pipelines subjected to vertical overburden loads. Full scale testing was performed using 24-inch diameter pipes, as well as numerical modeling via finite element analysis, to evaluate stresses and deflections in the pipe due to different loading conditions. For this particular application, the principal aim of the Composi-Sleeve TM system is to ensure that adequate reinforcement is provided to a pressurized pipeline subjected to vertical compression loading. From a mechanics standpoint the Composi-Sleeve TM system increases the structural rigidity and stiffness of the reinforced pipe so that when it is loaded it provides a greater resistance to deformation. The primary components of the Composi-Sleeve TM system include the bonding adhesive, the steel half shells, and the outer composite wrap. The variables of interest include length of the repair, thickness of the steel shells, and thickness of the composite material. This program has demonstrated that it is possible to estimate the level of reinforcement provided by the Composi-Sleeve TM system. Subsequent pipeline projects requiring additional strength can use the information gained in this study to properly design a reinforcing system. The sections of this report that follow include details on the finite element modeling effort incorporating analysis methods and results. Also included are details on the testing program that incorporate a presentation on the Composi-Sleeve TM installation process, along with documentation on the test program including set-up techniques and results. A Closing Comments section provides general thoughts on the overall performance of the reinforcement system.

87 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 2 Prepared by Stress Engineering Services, Inc. March ANALYTICAL METHODS AND RESULTS This report provides details on the analysis and testing work performed by Stress Engineering, where this particular section of the report deals specifically with the analysis efforts. Two types of pipeline models were analyzed. The first involved simulating actual buried conditions including soil support with loading from above the pipeline. The second model replicated actual test conditions in an effort to calibrate the model to match the test results. The benefit in performing analysis (numerical modeling) is that different conditions can be evaluated without having to run multiple tests, the latter of which can be expensive. The sections that follow provide a brief description of the analysis work that was done where results are presented as load-deflection and stress contour plots. 2.1 Description of Finite Element Analysis Finite element analysis is a tool used by engineers to solve problems using computers. Models are built of a particular system, structure, or component and loaded to represent actual conditions. In much the same way as we conduct experimental investigations, finite element analysis (FEA) allows us to estimate the effects of different loads on a system. Example results include calculating stresses, temperatures, and deflections. Also, FEA permits multiple loads to be imparted at the same time to determine the combined effects. For the problem at hand, we can consider the effects of soil surrounding the pipe, internal pressure, and what deflections in the pipe occur when imposing a vertical load on top of the soil. Figure 2.1 and Figure 2.2 are example contour plots where stresses in a dent and wrinkle bend were calculated considering a nominal internal pressure load. One of the contributions of finite element analysis to pipelines is the ability to determine either the ultimate capacity or safe operating conditions of an actual pipeline. While experimental investigations are a necessary part of pipeline engineering, FEA permits engineers to evaluate the effects of different scenarios on stresses in the pipeline. An ideal condition is to calibrate an FEA model using test results and then perform subsequent evaluations using the FEA model. This is the suggested scenario for the Composi-Sleeve TM system. Future projects requiring pipeline overburden reinforcement can use insights gained in the study to numerically simulate the response of the Composi-Sleeve TM system.

88 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 3 Prepared by Stress Engineering Services, Inc. March 2010 Figure 2.1 Example contour plot showing stress in dent Figure 2.2 Example contour plot showing stress in wrinkle bend

89 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 4 Prepared by Stress Engineering Services, Inc. March Simulating Actual Buried Pipelines Stress Engineering Services, Inc. constructed a model to simulate conditions associated with an actual pipeline that included soil around the pipe, internal pressure, and vertical loading on top of the pipe due to weight of the soil and external loads. Figure 2.3 is a general schematic showing the arrangement of the model including the soil springs. As noted in this figure, the vertical and horizontal springs were used to model the soil around the pipe. These soil stiffness values used by Stress Engineering Services, Inc. are representative of a clay-type soil and have been used previously in FEA pipeline analysis work. Stresses and displacements are generated in the pipeline model by internal pressure, as well as the overburden loading. Figure 2.4 shows the general state of stress in a 24-inch x 0.25-inch, Grade X42 pipeline subjected to 630 psi (72% SMYS, Specified Minimum Yield Strength) and a non-specified overburden load. Figure 2.5 specifically plots deflections in the pipe as a function of vertical loading. The vertical load is assumed to have distribution on a per foot basis. There are four data sets included in Figure 2.5 that include: 24-inch x 0.25-inch, 630 psi 24-inch x 0.25-inch, 0 psi 24-inch x 0.50-inch, 630 psi 24-inch x 0.50-inch, 0 psi There are several noteworthy observations made in reviewing Figure 2.5. The first is that for the two models including internal pressure, there is a deflection that occurs with no vertical loading. The displacements that are shown are absolute in value so that when internal pressure is applied there is radial expansion of the pipe. For example, at 630 psi the strain in the pipe is that corresponds to a radial expansion of approximately inches for the 24-inch diameter pipe; a value that is observed in the plotted data. Another important observation is that the presence of internal pressure acts to stiffen the pipe and reduce the overall deflection due to external forces acting on the pipe. Some of the equations used by industry in evaluating stresses in buried pipelines fail to take this condition into account. A third observation concerns the effects of increased wall thickness. As expected, as the wall thickness of the pipe is increased, structural rigidity is also increased. As an example, Figure 2.5 shows that with no internal pressure and a vertical load of 50,000 lbs the deflection for the 0.50-inch thick pipe is 0.45 inches, while for the 0.25-inch thick pipe the deflection is approximately 0.68 inches.

90 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 5 Prepared by Stress Engineering Services, Inc. March 2010 Vertical Pressure Internal Pressure Horizontal Soil Springs (Compression Only) K=58,118 lb/ft/ft Vertical Soil Springs (Compression Only) K=58,702 lb/ft/ft Figure 2.3 Soil support model showing position of spring elements Units of psi Overburden Load with 630 psi Internal Pressure (24-inch x 0.25-inch with soil reaction) Figure 2.4 Stresses in soil support model including overburden and pressure

91 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 6 Prepared by Stress Engineering Services, Inc. March Vertical Deflection of Pipe (180 Degree Crown Location) Results from finite element model for 24-inch x 0.25-inch, Grade X42 pipe including the effects of internal pressure and soil support in the form of vertical and lateral springs (500 lbs/in/in). 24 x 0.25, 0 psi, With Support 24 x 0.25, 630 psi, With Support 24 x 0.50, 0 psi, With Support x 0.50, 630 psi, With Support 180 (Pipe Top) Deflection (in) Z 270 (Pipe Side) X 90 (Pipe Bottom) (Pipe Bottom) Vertical Load (lbs) Figure 2.5 Plot showing vertical deflection of soil support model 2.3 Simulating the Experimental Test Set-up In addition to the analysis work used to model the behavior of an actual pipeline, a series of finite element models were used to simulate the response of the pipe subjected to experimental loading. The model replicated the test fixtures used in the current program to actually test the Composi-Sleeve TM reinforcement system. Loading included pressure end loading (i.e. to represent the sample end caps), internal pressure, and vertical overburden compression loading. Figure 2.6 shows the set-up for the FEA model including the top and bottom saddle supports. The saddles each spanned 120 degrees. A side view of the model is shown in Figure 2.7 including the pressure end load, boundary conditions, and axial position of the saddle. Figure 2.8 is an end view of FEA test model showing the vertical displacement configuration. Four basic models were evaluated in this study that included the following: 24-inch x 0.25-inch with no internal pressure 24-inch x 0.50-inch with no internal pressure

92 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 7 Prepared by Stress Engineering Services, Inc. March inch x 0.25-inch with 630 psi internal pressure 24-inch x inch with 630 psi internal pressure and fixed end A contour plot for the FEA test model considering the 24-inch x 0.25-inch geometry with 630 psi internal pressure and fixed end is shown in Figure 2.9. The fixed end condition best simulated the actual behavior of the experimental results. In addition to the contour plot, a graph is provided in Figure 2.10 showing the load-deflection response for the four models. Also plotted are the experimental results for the unreinforced test sample that included 630 psi. What is observed in this plot is that the FEA model simulating the 24-inch x inch pipe with 630 psi internal pressure and fixed ends best simulates the actual test set-up. In actuality the test sample s wall thickness was likely greater than 0.25 inches and based on the Mill Test Report the actual yield strength was 49,300 psi (i.e. greater than the 42,000 psi yield strength used in the model). The data plotted in Figure 2.10 demonstrate that it is possible to develop a finite element model that simulates the general behavior of the test set-up.

93 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 8 Prepared by Stress Engineering Services, Inc. March 2010 Top Saddle Radius = 15-inches Saddle Length: 2 feet Inner Radius: 15 inches ABAQUS Rigid Surface 24-inch OD Pipe Pipe Length: 10 feet Pipe OD: 24 inches Wall Thickness 0.25 or 0.50 inches ABAQUS Element Type: S4R Elevation View Bottom Saddle Radius = 15-inches Figure 2.6 FEA model used to simulate experimental test set-up 285 kips Pressure End Load Top & Bottom Saddles Fixed Boundary Conditions Figure 2.7 Boundary conditions and loading on FEA test model Top Saddle Displaced 3-inches Down Load steps: Internal pressure: 630 psi Top saddle pressed down 3 inches Bottom Saddle Fixed Figure 2.8 End view of FEA test model showing vertical displacement

94 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 9 Prepared by Stress Engineering Services, Inc. March 2010 End View Fixed Boundary Condition Results shown for case with inch wall thickness Figure 2.9 Contour plot for FEA test model Force versus Deflection for 24-inch Grade X42 Pipe Finite element analysis results using a 15-inch radius saddle that spanned 120 degrees on top and bottom of pipe. Loading conditions as noted in legend inch WT (no pressure) inch WT (no pressure) 0.25-inch WT with 630 psi Load (lb) inch WT with 630 psi and Fixed End Experimental data (0.25-inch with 630 psi) Differences in results likely due to rubber used in experimental work Deflection (in) Figure 2.10 Comparison of analysis to testing results

95 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 10 Prepared by Stress Engineering Services, Inc. March EXPERIMENTAL METHODS AND RESULTS The preceding section provided details on the analysis methods used in this study and the associated results. This section provides details on the experimental work done to evaluate the performance of the Composi-Sleeve TM reinforcement system. The primary components of the Composi-Sleeve TM system include the following: Bonding adhesive between the pipe surface and rolled half-shell steel material Rolled steel half-shell material (0.250 inch thick material used in this program) Composite outer wrap material (e.g. Pipe Wrap A+ system) The intent of the test program was to evaluate the level of reinforcement provided by the Composi-Sleeve TM reinforcement system relative to an unreinforced pipe. A total of four samples using 24.-inch x 0.25-inch, Grade X42 pipe material were fabricated and tested. These four samples included the following configurations. Unreinforced pipe section (referred to as the base pipe test case) Composite only (Pipe Wrap A+ with 32 layers: approximately 0.64 inches thick) Composi-Sleeve TM reinforcement system using Pipe Wrap A+ with 15 layers Composi-Sleeve TM reinforcement system using Pipe Wrap A+ with 10 layers The sections that follow provide specific details on the sample fabrication and installation activities including photographs take at Stress Engineering s Houston test lab. A section provides details on the test set-up including discussions on instrumentation. A final section presents results with an emphasis on differentiating performance between the four test samples. 3.1 Sample Fabrication Stress Engineering Services, Inc. had four test samples fabricated using 24-inch x 0.25-inch, Grade X42 pipe material (ERW weld seam). According to the Mill Test Report (MTR) the yield and ultimate tensile strengths for this material are 49,300 psi and 73,950 psi, respectively. End caps were welded to all test samples and 1-inch NPT thread-o-let bossets were installed for filling the samples with water and pressurization during testing. The test samples were sandblasted to near white metal as requested by LMC.

96 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 11 Prepared by Stress Engineering Services, Inc. March 2010 The figures that follow provide pictorial descriptions of the Composi-Sleeve TM installation process. Figure 3.1 Steel half-shells provided by LMC Figure 3.2 Installation of adhesive on inner surface of steel half-shells Figure 3.3 Preparing to install steel half-shell on pipe surface Figure 3.4 Positioning restraining straps on outer surface of half-shells Figure 3.5 Half shells installed on pipe sample prior to the composite installation Figure 3.6 Installation of composite materials on the outer surface of the half shells Figure 3.7 Installation of stricture wrap on the outer surface of the composite After the steel half-shells of the Composi-Sleeve TM reinforcement systems were installed, the outer composite materials were installed and allowed to cure over a 48-hour period. 3.2 Test Set-up Before the Composi-Sleeve TM systems were installed on the test samples, strain gages were installed on select locations beneath the reinforcements. Figure 3.8 is a schematic diagram showing the position of the strain gages on the pipe beneath the reinforcement and on the outside surfaces of the Composi-Sleeve TM. In setting up for the testing work, Stress Engineering modified one of its dent installation rigs having a load capacity in excess of 400,000 lbs. Figure 3.9 is a basic diagram showing the general loading scheme of the compression test. Figure 3.10 and Figure 3.11 are photographs showing isometric and end views of the test set-up, respectively. Figure 3.12 is a close-up view showing the upper and lower saddles. The saddles each spanned 120 degrees. Note in this photograph the positioning of the 1-inch thick rubber between the saddles and the test sample. Figure 3.13 is a side view of the test set-up showing the displacement transducer that was used to measure the vertical displacement of the upper platen of the load frame during testing. A hydraulic cylinder generated the vertical forces required to compress the test sample. During testing, data were recorded at a rate of 1 scan per second. The measurements included strain, vertical displacement, vertical load, and internal pressure. A specific sequence of testing was developed to ensure uniformity of measurements and assessment of performance. The

97 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 12 Prepared by Stress Engineering Services, Inc. March 2010 intention in developing four different and unique test samples was to evaluate the effects of different variables on the overall rigidity of the test samples. For example, the base pipe with no reinforcement serves as the primary reference point against which all other tests are measured. Of particular interest was the level of deflection that occurred as a function of load. One of the unknowns in testing was the level of damage would be imparted to the strain gages. Every practical effort was made to ensure their protection. Provided below is the protocol used in this test program. Fabricate four (4) 24-inch x 0.25-inch, Grade X42 pipe samples (0.375-inch thick end caps). Install strain gages on pipe (to be located outside of and beneath repair on each sample as shown in Figure 3.8). Make repairs as prescribed below to fabricate samples for the test matrix: o Base pipe (unreinforced with NO repair materials) o Composite only (Pipe Wrap A+ with 32 layers: approximately 0.64 inches thick) o Composi-Sleeve TM reinforcement system using Pipe Wrap A+ with 15 layers o Composi-Sleeve TM reinforcement system using Pipe Wrap A+ with 10 layers Place test sample in SES dent fixture (one test performed at a time) Apply internal pressure to pipe (72% SMYS: 630 psi). Force saddle indenter into pipe as shown in Figure 3.9. Record the following data at a rate of one scan per second: internal pressure, strain, vertical pipe deflection, and indenter load. Continue to increase the compression load applied to the pipe until gross plasticity develops in the test sample (impose a maximum load of 400,000 lbs).

98 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 13 Prepared by Stress Engineering Services, Inc. March 2010 Figure 3.1 Steel half-shells provided by LMC Figure 3.2 Installation of adhesive on inner surface of steel half-shells

99 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 14 Prepared by Stress Engineering Services, Inc. March 2010 Figure 3.3 Preparing to install steel half-shell on pipe surface Figure 3.4 Positioning restraining straps on outer surface of half-shells

100 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 15 Prepared by Stress Engineering Services, Inc. March 2010 Figure 3.5 Half shells installed on pipe sample prior to the composite installation Figure 3.6 Installation of composite materials on the outer surface of the half shells

101 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 16 Prepared by Stress Engineering Services, Inc. March 2010 Figure 3.7 Installation of stricture wrap on the outer surface of the composite 24-inch x 0.25-inch, Grade X42 pipe 18-inch 10-ft length 5-feet #5 #2 Gages on pipe #4 End gages #6 #1 #3 #7 Center gages Gages on composite (outside surface) 12:00 and 9:00 12:00, 3:00, and 9:00 All strain gages to be hoop-oriented except Gages #1, #4, and #6 which are the bi-axial (hoop and axial) Figure 3.8 Schematic diagram showing strain gage installation

102 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 17 Prepared by Stress Engineering Services, Inc. March 2010 Original Test Set-up Deformed Test Configuration Figure 3.9 Basic diagram showing general loading scheme

103 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 18 Prepared by Stress Engineering Services, Inc. March 2010 Figure 3.10 Photograph showing isometric view of the test set-up Figure 3.11 Photograph showing end view of the test set-up

104 Evaluation of the Compsi-Sleeve TM Pipeline Reinforcement System Page 19 Prepared by Stress Engineering Services, Inc. March 2010 Figure 3.12 Close-up view showing the upper and lower saddles with rubber inserts Figure 3.13 Side view of the test set-up showing the vertical displacement transducer