Field Performance of Corrugated Pipe Manufactured with Recycled Polyethylene Content (2018)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

71 Based on the research results summarized in Chapter 2, pro- posals were established for AASHTO material and design stan- dards for pipes manufactured with recycled materials. A new test method, the UCLS test, was developed to assess the stress- crack initiation and propagation properties of HDPE materials containing recycled content. Additionally, an AASHTO Stan- dard Practice was developed for predicting the service life of corrugated HDPE pipes manufactured with recycled materials. This Standard Practice also details a method for establishing minimum UCLS requirements to ensure service life require- ments are met for a given set of service conditions (i.e., design stress and temperature). This provides the basis for a perfor- mance-based specification that was validated in the research. AASHTO material specification and design methodology proposals are detailed in this chapter. 3.1 AASHTO Material Specification Proposal for Pipes Manufactured with Recycled Materials As discussed previously and shown in the research, the property that governs the service life of corrugated HDPE pipes manufactured with recycled materials is resistance to Stage II brittle stress cracking. Standard notched tests are not sufficient for assessing the stress-crack resistance of pipes manufactured with recycled materials since they do not address the time to crack initiation from a void or contaminant, which is particularly critical for materials containing recycled content. Therefore, a new test method, the UCLS test—based on the BFF test from NCHRP Report 696, was developed and published as an ASTM standard (5). This test method provides the basis for a performance- based specification, as it was shown in Chapter 2 that the results of the UCLS test conducted at multiple temperatures and stresses can be used to accurately predict the service life of corrugated HDPE pipes manufactured with recycled materials. To establish minimum UCLS requirements for corrugated HDPE pipes manufactured with recycled materials, it is help- ful to review the process that was used to predict the service life of these pipes based on the UCLS test. The process can be summarized by the following steps: 1. Prepare compression-molded UCLS plaques according to the procedure outlined in ASTM F3181. The plaques may be prepared from resin blends or from chips taken directly from the pipe wall. It is important they are properly homogenized. 2. Prepare at least 15 UCLS test specimens from the plaques in accordance to the dimensions and procedures outlined in ASTM F3181. 3. Conduct the UCLS test in accordance with ASTM F3181 on five specimens at each of, at a minimum, three test condi- tions. Each specimen must be taken to failure. Record the individual and average failure times of the five specimens at each condition, as well as the coefficient of variation (the standard deviation of the failure times divided by the aver- age failure time). The average must be calculated on a Log basis. The minimum suggested test conditions are as follows: a. Condition I: 80°C, 4.48 MPa (650 psi) stress b. Condition II: 80°C, 3.10 MPa (450 psi) stress c. Condition III: 70°C, 4.48 MPa (650 psi) stress 4. Use the PSM multiplication factors shown in Equations 3.1 and 3.2 to shift the elevated temperature (T2,°C) average failure times (determined on a log basis) from Step 3 to the projected failure times at the desired in-ground ser- vice temperature (T1,°C). A conservative assumption for T1 is 23°C, though lower temperatures may be used for northern climates. Stress Shift Factor (3.1)0.0116 T2–T1e= ( ) Time Shift Factor (3.2)0.109 T2–T1e= ( ) 5. Plot the resulting three (or more, if additional conditions were evaluated) shifted average data points on a log-log C H A P T E R 3 Material Specification and Design Methodology Proposals

72 scale, with log time on the x-axis and log stress on the y-axis. Determine the best-fit curve for the data points, which should be linear on a log-log scale. 6. Calculate the 95% lower confidence limit of each of the shifted average data points by using the Student’s t-distribution as shown in Equation 3.3. Use the largest COV from the three (or more) data sets obtained in Step 3 for determination of the LCL. ( )= − ( )−LCL (3.3)95% 1X t COV X nn  where LCL95% = Lower 95% Confidence Limit X – = Log-based average of five test specimens t(n–1) = Student’s t value at (n–1) degrees of freedom = 2.132 COV = Maximum coefficient of variation of five test specimens n = Number of test specimens at each condition = 5 7. Determine the best-fit curve for the three (or more) LCL data points. 8. Extrapolate the LCL curve to the desired factored service stress condition to determine the predicted service life relative to Stage II brittle cracking. An illustration of the average and LCL curves for Pipe 4 (49% PCR) is shown in Figure 3-1. In this example, the predicted service life relative to Stage II brittle cracking is 234 years. Note that the RPM rather than the PSM could also be used to predict the service life of pipes manufactured with PCR materials based on UCLS failure data. However, since the data presented in Chapter 2 indicated that PSM is more conservative and accurate than the RPM for the majority of the materials evaluated, the PSM is suggested. 3.1.1 Minimum UCLS Requirement for Pipes Manufactured with Recycled Materials The UCLS test requirement can be determined by back- calculating from the desired performance at the service conditions, as shown in the following equations. Similar methodology has been previously employed by Hsuan to determine the minimum junction test performance require- ments to ensure 100-year service life for virgin HDPE pipes for FDOT (44), though the statistical adjustments and test specimens are different in this analysis. The calculations require determination of the slope of the brittle failure curve, represented generically by Equation 3.4. [ ][ ]= − − (3.4)1 2 1 2m Y Y X X where m = Slope of brittle failure curve at desired service temperature (X1, Y1) and (X2, Y2) = Any two points on the curve Avg: y = -0.2175x + 4.1196 R² = 0.9936 LCL: y = -0.2175x + 4.0719 R² = 0.9936 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 Lo g St re ss (p si) Log Time (h) Mastercurve at 23 deg. C 2.05E6 h = 234 yrs. 500 psi stress Figure 3-1. Illustration of procedure for predicting service life of Pipe 4 (49% PCR) at service conditions of 23°C and 3.4 MPa (500 psi) stress, based on application of PSM multiplication factors to UCLS failure data.

73 If the slope of the curve and one data point are known (or fixed), one can solve for any other data point on the curve. Choosing (X1, Y1) to correspond to a point on the shifted curve generated from PSM-shifted test values and (X2, Y2) to correspond to a point on the curve based on the service conditions, one can rewrite Equation 3.4 as shown in Equa- tion 3.5. The minimum required failure time at a given test condition can then be calculated by solving for tT, as shown in Equation 3.6. [ ] [ ] ( ) ( ) ( ) ( )= σ − σ − σlog log log log (3.5) SVC SVC m SF SF t t T t T   = 10 (3.6)t SFT C t where tT = Minimum required average failure time at test condition, h m = Slope of brittle failure curve SFσ = PSM stress shift factor from Equation 3.1 SFt = PSM time shift factor from Equation 3.2 σT = Stress at test condition, psi σSVC = Stress at service condition, psi tSVC = Required service life at service conditions, h log log log (3.7) SVC SVCC SF m t T( ) ( ) ( )= σ − σ   + σ The PSM stress and shift factors are dependent on the ser- vice and test temperatures; thus, if multiple test temperatures are used (e.g., 70°C and 80°C, as suggested in Step 3 above), the shift factors will differ for each condition. To obtain 95% confidence that the minimum average failure times calculated from Equation 3.6 will result in the desired projected service life requirements, these failure times must be statistically adjusted to account for the scatter in the data. This is done by setting the minimum failure time calcu- lated from Equation 3.6 equal to the LCL as shown in Equa- tion 3.8, and solving for X – 95% as shown in Equation 3.9. ( )= = − ( )−LCL (3.8)95% 95% 1 95%t X t COV X nT n  1 (3.9)95% 1 X t t COV n T n( )= − ( )− where LCL95% = Lower 95% Confidence Limit X – 95% = Average failure time needed for 95% confidence, h tT = Minimum required average failure time from Equation 3.6, h t(n–1) = Student’s t value at (n−1) degrees of freedom = 2.132 COV = Typical coefficient of variation of test data = 0.5 n = Number of test specimens at each condition = 5 Based on Equations 3.6 through 3.9, minimum UCLS performance requirements can be established for any given service condition (i.e., service temperature and wall stress). 3.1.1.1 Example Calculations for Minimum UCLS Criteria For purposes of illustration, the service conditions required by FDOT for 100-year service life applications of corrugated HDPE pipes manufactured with virgin materials will be consid- ered. FDOT requires corrugated HDPE pipes to last 100 years at a wall stress of 3.4 MPa (500 psi) and an underground tempera- ture of 23°C (44). Note that this wall stress, based on a factored wall strain of 2.25% and long-term modulus of 20,000 psi (23), exceeds the tensile wall stresses present in typical installations. Additionally, most states have underground temperatures lower than 23°C. Nonetheless, these design assumptions can be con- sidered as a conservative requirement in the absence of other design data. To complete the calculations, it is necessary to know the slope of the brittle failure curve as well as the COV in the UCLS data sets. From Table 2-7, the average slope of the brittle failure curves determined by the PSM for the six test pipes was −0.196. From Table 2-4, the average COV of 21 UCLS tests (six pipes tested at three conditions and three pipes tested at one condition) was 0.366. Based on these values, a brittle slope of –0.20 and a COV of 0.50 are conservatively appropriate for the calculations. To determine the minimum UCLS requirements at Condi- tion I [80°C, 4.48 MPa (650 psi) stress], based on a desired ser- vice life of 100 years (876,000 hours) at 3.4 MPa (500 psi) stress and 23°C, the following calculations are performed: 1. Using Equations 3.1 and 3.2, calculate the stress and time shift factors: Stress Shift Factor = SFσ = e0.0116 (80–23) = 1.937 Time Shift Factor = SFt = e 0.109 (80–23) = 499.2 2. Using Equation 3.7, calculate C:   C log log log log 1.937 650 log 500 0.20 log 876,000 C 3.937 SVC SVC SF m t T( ) ( ) ( ) ( ) ( ) ( ) = σ − σ    + = − −     + = σ

74 3. Using Equation 3.6, calculate the minimum average failure time, tT,650,80, at the 80°C, 4.48 MPa (650 psi) stress condition: = = = 10 10 499.2 17.33 h,650,80 3.937 t SFT C t 4. To account for the scatter in the data, the calculation in Step 3 should be considered the LCL for the 80°C, 4.48 MPa (650 psi) condition. To ensure 95% confidence that the average failure time meets this requirement, use Equa- tion 3.9 to statistically adjust this number to determine the actual test requirement for the minimum average failure time: ( ) ( )= −  = −  = ( )− 33.1 h 1 COV 17.33 1 2.132 0.5 5 95%,650,80 ,650,80 1 X t t n T n   Similarly, to determine the minimum UCLS requirements at Condition II [80°C, 3.10 MPa (450 psi stress)], based on a desired service life of 100 years (876,000 hours) at 3.4 MPa (500 psi) stress and 23°C, the following calculations are performed: 1. Using Equations 3.1 and 3.2, calculate the stress and time shift factors: Stress Shift Factor = SFσ = e0.0116 (80–23) = 1.937 Time Shift Factor = SFt = e 0.109 (80−23) = 499.2 2. Using Equation 3.7, calculate C:   log log log log 1.937 450 log 500 0.20 log 876,000 4.736 SVC SVCC SF m t C T( ) ( ) ( ) ( ) ( ) ( ) = σ − σ    + = − −     + = σ 3. Using Equation 3.6, calculate the minimum average fail- ure time, tT,450,80, at the 80°C, 3.10 MPa (450 psi) stress condition: = = = 10 10 499.2 109.0 h,450,80 4.736 t SFT C t 4. To account for the scatter in the data, the calculation in Step 3 should be considered the LCL for the 80°C, 3.10 MPa (450 psi) condition. To ensure 95% confidence that the average failure time meets this requirement, use Equation 3.9 to statistically adjust this number to deter- mine the actual test requirement for the minimum average failure time:   208.3 h 1 COV 109.0 1 2.132 0.5 5 95%,450,80 ,450,80 1 X t t n T n( ) ( )= −  = −  = ( )− Finally, to determine the minimum UCLS requirements at Condition III [70°C, 4.48 MPa (650 psi stress)], based on a desired service life of 100 years (876,000 hours) at 3.4 MPa (500 psi) stress and 23°C, the following calculations are performed: 1. Using Equations 3.1 and 3.2, calculate the stress and time shift factors: Stress Shift Factor = SFσ = e0.0116 (70–23) = 1.725 Time Shift Factor = SFt = e0.109 (70–23) = 167.8 2. Using Equation 3.7, calculate C:   log log log log 1.725 650 log 500 0.20 log 876,000 4.189 SVC SVCC SF m t C T( ) ( ) ( ) ( ) ( ) ( ) = σ − σ    + = − −     + = σ 3. Using Equation 3.6, calculate the minimum average fail- ure time, tT,650,70, at the 70°C, 4.48 MPa (650 psi) stress condition: = = = 10 10 167.8 92.1 h,650,70 3.937 t SFT C t 4. To account for the scatter in the data, the calculation in Step 3 should be considered the LCL for the 70°C, 4.48 MPa (650 psi) condition. To ensure 95% confidence that the average failure time meets this requirement, use Equation 3.9 to statistically adjust this number to deter- mine the actual test requirement for the minimum aver- age failure time: ( ) ( )= −  = −  = ( )− 175.9 h 1 COV 92.1 1 2.132 0.5 5 95%,650,70 ,650,70 1 X t t n T n   3.1.1.2 Simplified Method for Establishing Criteria for UCLS Test Since the slopes of the brittle Popelar shifted mastercurves presented in Chapter 2 were consistently around −0.20, it can be argued that it is sufficient to conduct UCLS testing at only one condition to establish performance requirements

75 for pipes containing recycled materials. Furthermore, since Condition I [80°C / 4.48 MPa (650 psi) stress] is the most severe condition, it is reasonable to incorporate this condi- tion as the minimum test requirement into the AASHTO M 294 material specifications and to include it as a quality assurance check to be conducted on a basis to be established by industry and AASHTO’s National Transportation Product Evaluation Program (NTPEP). Also, since the long-term performance of the corrugated HDPE pipes relative to Stage II cracking is dependent on tem- perature and stress, it may be prudent to consider a variable performance requirement for the UCLS test at this condition. Table 3-1 shows the minimum suggested UCLS values for the 80°C / 4.48 MPa (650 psi) test condition for a range of field stress conditions at underground service temperatures of both 20°C and 23°C. These minimum requirements assume the slope of the brittle failure curve is −0.20, a reasonable assumption based on the data presented in Table 2-7. They are also adjusted statistically as shown in the previous section, based on the Student’s t-distribution assuming a conservative COV of 0.50, which was greater than any COV on the test pipes in this study. It is interesting to note that higher service tensile stresses are more likely to occur in more poorly controlled instal- lations (e.g., agricultural installations) than in highway or railroad applications, where quality backfill materials and post-installation inspection procedures are specified. For example, pipes in agricultural applications are typically installed in accordance to ASTM F449, “Standard Practice for Subsurface Installation of Corrugated Polyethylene Pipe for Agricultural Drainage or Water Table Control,” which does not require structural backfill materials for the majority of installations (45). Also, there is not a 5% in-field deflection limit on these pipes, as is the case in highway and railroad appli- cations, so wall strains will be greater in agricultural drainage pipes. Therefore, it may be prudent to establish higher UCLS (and NCLS, for that matter) performance requirements for pipes installed with poorer backfill materials or lesser quality installation and inspection procedures than those used for railroad or highway applications, where deflections and back- fill materials are more tightly controlled. This is contrary to current practice, as pipes for railroad and highway appli- cations typically have stricter material requirements than those for agricultural or other land drainage applications. As an illustration of the lower stresses resulting from qual- ity installations, recall from Chapter 2 that the peak measured strains in the field test pipes from the SEPTA pilot study were around −4000 microstrain (compression), and no tensile strains were measured. Based on a long-term modulus of 138 MPa (20,000 psi), the maximum long-term stress in the pipe wall is 0.55 MPa (80 psi) compression. The UCLS test values at 80°C and 4.48 MPa (650 psi) were 1271 hours for Pipe 1 (virgin materials) and 99 hours for Pipe 2 (49% PCR materials). Both test values are significantly higher than the minimum suggested values from Table 3-1, even for service stress conditions of up to 4.14 MPa (600 psi), so both pipes are expected to last well beyond 100 years. For simplicity, design engineers and specifiers may choose to conservatively assume a factored design stress of 3.45 MPa (500 psi) and an under-surface temperature of 23°C (73°F), the conditions assumed by FDOT for its 100-year service applications. If this is the case, the minimum five-specimen average UCLS value should be 34 hours, and no single speci- men shall fail in less than 18 hours. 3.1.2 Other Requirements for Pipes Manufactured with Recycled Materials While the incorporation of the UCLS test is the most significant part of the proposed materials specification for pipes manufactured with recycled materials, there are some other suggested guidelines, as detailed in the following subsections. 3.1.2.1 NCLS Requirement Since the UCLS test addresses both the crack initiation and propagation phases of the SCG mechanism, it could be argued that the NCLS test, which primarily addresses crack Service Stress, MPa (psi) 20°C Service Temperature 23°C Service Temperature Avg. UCLS (h) No SpecimenLess Than (h) Avg. UCLS (h) No Specimen Less Than (h) 1.38 (200) N/A N/A N/A N/A 2.07 (300) N/A N/A 3 2 2.76 (400) 7 4 11 6 3.45 (500) 20 11 34 18 Table 3-1. Minimum UCLS requirements for the 80°C, 650 psi stress condition, based on a brittle curve slope of -0.20 and a maximum COV in the data of 0.50, for 100-year service applications.

76 propagation, is not necessary for pipes manufactured with recycled materials and should only be specified for virgin materials. However, given the good historical field perfor- mance of current AASHTO M 294 pipes relative to stress cracking, it is suggested to carry over the same requirements as specified for virgin pipes to pipes containing recycled content. Specifically, the pipes should have an NCLS value of 24 hours, with no specimen having less than 17 hours, when tested in accordance with ASTM F2136. It is suggested that the test only be conducted on specimens taken from compression-molded plaques made from chips from the pipe wall and/or liner, as it offers a better assessment of material crack propagation since it eliminates the effects of residual stresses that can occur when testing directly from the pipe wall. It should be noted that six of the seven pipes that contained recycled materials and were evaluated in this research project did not meet this 24-hour NCLS requirement (see Figure 2-2). 3.1.2.2 Oxidation Induction Time Requirement The current AASHTO M 294 standard does not have a requirement for oxidation induction time (OIT). Because the recycled materials used for pipes have experienced a prior heat history, it is prudent to incorporate a minimum OIT to ensure manufacturers add some antioxidants to their formulation to prevent premature Stage III chemical failures of the product. An OIT of 20 minutes when tested in accordance with ASTM D3895 is suggested, as it is a typical value used for solid-wall pressure pipe applications and can be achieved with the addi- tion of relatively common antioxidant packages. Of the seven pipes evaluated for OIT in this research project, only three met this suggested 20-minute OIT requirement (Pipes 5, 8 and 9). 3.1.2.3 Tests for Contaminants In NCHRP Report 696, it was suggested to include a break strain requirement of 150% on pipe materials when tested in accordance with ASTM D638 (1). The break strain test was shown to be a relatively quick check for contaminants. Since the UCLS test is also a contaminant test and is already being proposed in the materials specification, the break strain test is redundant. However, while the UCLS test is a good quality assurance test, the break strain test is more appropriate for quality control as it is a shorter test to conduct. Therefore, it is suggested to include a break strain requirement of 150% for pipe materials, as proposed in NCHRP Report 696. Rather than set maximum limits for various types of con- taminants (e.g., polypropylene, rubbers, labels), it is pro- posed to rely on the break strain and UCLS tests to govern the performance of the materials. Manufacturers should set frequencies for conducting these tests that will ensure the required performance limits are met on a consistent basis. 3.1.2.4 Cell Classification The cell classification for the final blend of materials used in the manufacturing of corrugated HDPE pipes contain- ing recycled content should be the same as currently speci- fied in AASHTO M 294 for virgin pipes. Specifically, the cell classification should be 435400C when tested in accor- dance with ASTM D3350. Because colorants and inorganic materials may be present in the recycled materials blends, it is suggested to use the ultrasound technique detailed in ASTM D4883 to determine base resin density since the inorganic materials have little or no effect on ultrasonic density. 3.1.2.5 Other Properties All the structural and performance properties currently specified in AASHTO M 294 for pipes manufactured with virgin materials (e.g., pipe stiffness, flattening performance, impact resistance) are proposed for pipes manufactured with recycled materials as well. 3.1.2.6 Testing Frequencies While this project did not specifically address the frequency of testing necessary for pipes manufactured with recycled materials, the frequencies currently specified in AASHTO M 294 and the NTPEP are proposed as a minimum. 3.1.3 Draft Specification Based on the previously discussed details, a draft of the proposed revisions to AASHTO M 294 was developed and is shown in Appendix J. 3.2 AASHTO Design Methodology Proposal for Pipes Manufactured with Recycled Materials Because pipes containing recycled materials are manufac- tured with the same molds and production processes as virgin materials, the current AASHTO design method is appropriate for these pipes. Furthermore, because the cell classification for pipes manufactured with recycled materials is proposed to be the same as for pipes manufactured with all virgin materials, the short- and long-term material design properties speci- fied in Table 12.12.3.3-1 of the AASHTO LRFD Bridge Design Specification (37, Section 12) are proposed to be the same. They are summarized in Table 3-2. Note that current AASHTO design methodology assumes that the long-term properties (i.e., modulus of elasticity and

77 tensile strength) of HDPE are stress independent. Preliminary evaluation of data compiled during this and previous NCHRP projects indicates that this may not be the case. Specifically, it appears that the long-term properties are dependent on the initial stress in the pipe wall. This is very important to under- stand, as it affects the long-term structural design criteria for corrugated HDPE pipes manufactured both with and without recycled materials. However, this was beyond the scope of this project and therefore was not fully understood or evaluated at the time of this report. It is suggested that further evalua- tion is needed regarding the dependence of long-term tensile strength and modulus on initial applied stress for both recycled and virgin HDPE materials to more accurately determine the appropriate long-term design values for these pipes. Proposed modifications to the AASHTO LRFD Bridge and Construction specification are shown in Appendix K. 3.3 AASHTO Standard Recommended Practice for Service Life Determination An AASHTO Standard Recommended Practice was devel- oped for determining the service life of corrugated HDPE pipes manufactured with recycled materials. This practice also provides guidance for determining minimum UCLS require- ments to ensure a desired service life at any service condition (i.e., service temperature and design stress). The Standard Recommended Practice is shown in Appendix L. Property Corrugated HDPE Pipes from Virgin Materials Corrugated HDPE Pipes with Recycled Content Tensile strength, Fu – initial 3,000 psi 3,000 psi Tensile strength, Fu – 50 year 900 psi 900 psi Tensile strength, Fu – 75 year 900 psi 900 psi Modulus, E – initial 110,000 psi 110,000 psi Modulus, E – 50 year 22,000 psi 22,000 psi Modulus, E – 75 year 21,000 psi 21,000 psi Factored compressive strain limit, yc 4.1% 4.1% Service long-term tension strain limit, yt 5.0% 5.0% Table 3-2. Summary of design properties for corrugated HDPE pipes manufactured with and without recycled materials.