LTPP Data Analysis: Influence of Design and Construction Features on the Response and Performance of New Flexible and Rigid Pavements (2005)

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.

390 CHAPTER 8 - CONCLUSIONS AND RECOMMENDATIONS 8.1 INTRODUCTION This chapter presents a summary of findings from a comprehensive evaluation of SPS-1, SPS-2 and SPS-8 experiments, based on mid-term performance trends [Release 17 (Level-E) of DataPave]. The current status of the experiment, construction quality, and data availability are also briefly discussed in this chapter for each experiment. Finally, the limitations of the findings from this research and recommendations for future data collection and research are presented. A detailed description of the experiment designs and the current status of SPS-1, SPS-2 and SPS-8 experiments were presented in Chapter 2. The SPS-1 and SPS-2 experiments are fractional factorial experiments that were aimed at finding the relative influence of design and construction features on performance of new flexible and rigid pavements, respectively. Each site within the SPS-1 and SPS-2 experiments has twelve pavement test sections with each section representing a different structural design. There are eighteen sites in the SPS-1 experiment and fourteen sites in the SPS-2 experiment; these sites are distributed throughout the United States by climatic zones (wet-freeze, wet-no-freeze, dry-freeze and dry-no-freeze) and subgrade type (fine and coarse-grained). The SPS-1 experiment is designed to investigate the effects of HMA layer thickness, base type, base thickness, and drainage on flexible pavement performance, while the SPS-2 experiment is aimed at studying the effect of PCC slab thickness, base type, PCC flexural strength, drainage, and lane width on rigid pavement performance. The current status (details in Chapter 2) of the experiments indicates some deviations from the intended experiment design for both SPS-1 and SPS-2 experiments. The most important deviation from the experiment design is the distribution of sites by climatic zone, which caused an unbalance in the number of sites per climatic zone. The average age of test sections in the SPS-1 experiment is 7 years with a range of 3 to 11 years, while the average age of sections in the SPS-2 experiment is 7 years with a range of 5 to 12 years. It may thus be said that the pavements are “fairly young”, and high occurrence and levels of distresses may not be expected at this point in time. Thus, all conclusions from the analyses presented in this report should only be interpreted as “mid-term” performance findings. The SPS-8 experiment was designed to study the effects of the environment on pavement performance, in the absence of heavy traffic. The experimental factors include climate and

391 subgrade soil type. A total of 32 flexible pavement sections in 15 sites and 14 rigid pavement sections in 5 sites were constructed for the experiment. The average age of the flexible pavement sections in SPS-8 is 6 years with a range of 3 to 10 years. In the case of rigid pavements, the average age is 6 years with a range of 4 to 10 years. A summary of data availability was presented for SPS-1, SPS-2 and SPS-8 experiments in Chapter 3. The extent and occurrence of distresses in the test sections within each of the experiments were also presented in Chapter 3. Experimental factors (design features and site factors) were compared to the as-constructed details obtained from the LTPP database and from construction reports. The deviations observed were reported in the chapter. Based on the extent and occurrence of distresses, different methods of analysis were employed. A brief description of each of the methods used for this research is in Chapter 4. The majority of results from these analyses should be interpreted with caution in light of the “low” occurrence of distresses in the test sections within the SPS-1, SPS-2, and SPS-8 experiments. This is especially true for the SPS-2 and SPS-8 experiments. Also, it is suggested that the conclusions be considered while keeping in view the other limitations of the data, as explained later in this chapter. A synopsis of the salient findings from all the analyses presented in previous chapters (Chapters 5, 6, and 7), for each experiment, is presented in this chapter. This summary is intended to give the reader a brief overview of the effects of design and construction features on pavement performance and response. The findings are presented by each performance or response measure, and not by design factor, as most of the experimental factors (design and site factors) are interacting among each other.

392 8.2 EFFECTS OF STRUCTURAL FACTORS FOR FLEXIBLE PAVEMENTS — SPS-1 EXPERIMENT The SPS-1 experiment, entitled Strategic Study of Structural Factors for Flexible Pavements, is one of the nine special pavement studies in the LTPP program. The effects of the experimental factors on fatigue cracking, structural rutting, roughness, transverse cracking, and longitudinal cracking (WP and NWP) are discussed below. It should be noted that the effects presented herein are statistically significant and of practical significance unless mentioned otherwise. 8.2.1 Effect of Design and Site Factors on Pavement Performance The effects of the experimental factors on each performance measure are discussed below, one performance measure at a time. Fatigue Cracking • All the experimental factors were found to be affecting fatigue cracking, though not at the same level. On the whole, pavements with “thin” 102 mm (4-inch) HMA surface layer have shown more fatigue cracking than those with “thick” 178 mm (7-inch) HMA surface layer. Also pavements constructed with only DGAB have shown more fatigue cracking than those with ATB-over-DGAB, and those with ATB base only, with the latter base type showing the best performance. The main effect of base thickness is not statistically significant. However, on average, pavements with 16- inch (406 mm) base thickness have shown slightly better fatigue performance than those with 203 mm (8-inch) or 305 mm (12-inch) base thickness. It should be noted that only pavement sections with drainage have a 406 mm (16-inch) base thickness according to the SPS-1 experiment design; therefore, it is unclear whether this effect is caused by the increased base thickness or by drainage provided with PATB. In this regard, the frequency-based analyses did show that pavements with drainage have significantly lower chances of cracking than those without drainage. • In general, pavement sections built on fine-grained soils have more fatigue cracking than those built on coarse-grained soils. Also pavements located WF zone have shown more fatigue cracking than those located in WNF zone.

393 • Among un-drained pavements, on average, an increase in HMA surface thickness from 102 mm (4-inch) to 178 mm (7-inch) has a slightly higher effect on fatigue cracking for pavements with DGAB than for pavements with ATB. However, this effect is not statistically significant. • The main effect of HMA surface thickness is more significant for sections built on coarse-grained soils. • Among pavements built on fine-grained soils, the effect of drainage is seen only in those sections with DGAB; i.e., those with drainage have less fatigue cracking than those without drainage. For drained pavements built on fine-grained soils, those with 203 mm (8-inch) base have more cracking than those with 305 mm (12-inch) and 406 mm (16-inch) base. Hence, for pavements built on fine-grained soils, drainage helps improve fatigue performance for those with DGAB while thicker base helps improve fatigue performance for drained pavements (irrespective of base type). • The main effect of HMA thickness, discussed above, is mainly seen among sections located in WNF zone. This may be an indication that an increase of HMA thickness from 102 mm (4-inch) to 178 mm (7-inch) is not sufficient in resisting fatigue cracking for pavements in WF zone as compared to WNF zone. • Among sections located in the WF zone, those with DGAB have shown the highest amount of cracking while those with ATB have the least cracking. In addition, those with 406 mm (16-inch) drained base have the least amount of fatigue cracking. This suggests that among pavements located in WF zone, “thick” 406 mm (16-inch) treated bases with drainage are less prone to cracking. The effects of HMA thickness and base thickness discussed above imply that, among sections located in WF zone, an increase in base thickness to 16-inch (with drainage) has a greater impact than an increase in HMA thickness from 102 mm (4-inch) to 178 mm (7-inch), suggesting that a thicker base and drainage helps in reducing frost effects.

394 Structural Rutting The extent of structural rutting among the test sections in the SPS-1 experiment is 6.5 mm, on average, with a standard deviation of 2.4 mm. Their average age is about 7 years with a range between 4.5 and 10 years. The amount of rutting for the majority of these sections is within the normal range at this point in time. Therefore, the results at this point may only show initial trends and may not be of much practical significance. • Marginal main effects of drainage, HMA thickness, and base thickness on structural rutting were observed. Pavements with “thin” [102 mm (4-inch)] HMA surface layer have shown slightly more rutting than those with “thick” [178 mm (7-inch)] HMA surface layer. Also, on average, pavements with 406 mm (16-inch) drained base have shown somewhat better rut performance than those with 203 mm (8-inch) and 305 mm (12-inch) base. However, these effects of HMA surface thickness and base thickness were not found to be statistically significant. Pavements with drainage have less rutting than those without drainage. The effect is not of practical significance, at this point in time. • In general, pavement sections built on fine-grained subgrade have shown more rutting than those built on coarse-grained subgrade. On the other hand, there is no apparent effect of climate (WF vs. WNF) on structural rutting. • Among the pavements built on coarse-grained soils, those with 178 mm (7-inch) HMA surface have shown slightly less rutting than those with 102 mm (4-inch) HMA surface. However, this effect is not operationally significant at this point. • The above suggests that for sections built on fine-grained soils an increase in HMA thickness from 102 mm (4-inch) to 178 mm (7-inch) may not be sufficient in reducing the amount of rutting. Among pavements built on fine-grained soils, a marginal positive effect of drainage is seen in sections with ATB. • Among drained pavements located in WF zone, those with DGAB have shown more rutting than those with ATB. Also, among sections located in WF zone and built with ATB, those with drainage have shown significantly less rutting than those without drainage. This implies that, among pavements located in WF zone, those with ATB and drainage perform better than those with other combinations of base type and drainage.

395 • Among un-drained sections located in WNF zone, those with 305 mm (12-inch) base have less rutting than those with 203 mm (8-inch) base. For sections built on DGAB and located in WNF zone, those with drainage have shown slightly less rutting than those without drainage. This effect was found to be marginally significant. These early trends imply that the importance of drainage among pavements with DGAB is considerable in improving rut performance among sections located in WNF zone. On the other hand an increase in base thickness from 203 mm (8-inch) to 305 mm (12- inch) improves rut performance for un-drained sections, irrespective of base type. Roughness • All the experimental factors were found to be affecting roughness, though not at the same level. Pavements with “thin” [102 mm (4-inch)] HMA surface layer have higher change in IRI (∆IRI) than those with “thick” [178 mm (7-inch)] HMA surface layer. This effect is not of practical significance at this point in time. Also, pavements constructed with DGAB have higher ∆IRI than those with ATB/DGAB and ATB, while pavements with ATB have the best performance for roughness. Pavements with thicker bases have lower ∆IRI. Also pavements with drainage have lower ∆IRI than un-drained pavements. • In general, pavements built on fine-grained soils have shown higher ∆IRI than those built on coarse-grained soils, especially among sections in WF zone. Also, the change in roughness among sections located in WF zone is significantly higher than those in WNF zone. • Among pavements built on fine-grained soils, an increase in HMA thickness from 102 mm (4-inch) to 178 mm (7-inch) has a significant positive effect on change in roughness. Also for un-drained pavements, those with ATB have significantly lower ∆IRI than those with DGAB. Finally the effect of drainage is significant only for sections with DGAB. These effects suggest that, for pavements built on fine-grained soils, higher HMA thickness and/or treated base will help inhibit the increase in roughness. Also, drainage appears to be more effective in preventing an increase in roughness for sections with DGAB, especially among those located in WF zone.

396 • For un-drained pavements built on coarse-grained soils, an increase in base thickness from 203 mm (8-inch) to 305 mm (12-inch) causes lower ∆IRI. However, this effect is marginally significant and is not of practical significance at this point in time. Transverse Cracking • The effect of base thickness on transverse cracking is insignificant, at this point. Pavements constructed with DGAB have more transverse cracking than those with ATB/DGAB and ATB, while pavements with ATB have shown the least amount of cracking. However the effect is not of practical significance at this point in time. Slightly more cracking was observed on pavements with “thin” [102 mm (4-inch)] HMA surface layer. Also, pavements with drainage have shown slightly less cracking than un-drained pavements. However, these effects were not found to be statistically significant. • In general, pavements built on fine-grained soils have shown more transverse cracking than those built on coarse-grained soils. • Pavements located in WF zone have shown significantly more transverse cracking than those located in WNF zone. This confirms that transverse cracking occurs mainly in freezing environment. • Among drained pavements built on coarse-grained soils, those with ATB performed better than those with DGAB. • Among pavements with DGAB and built on fine-grained soils, those with drainage have shown significantly less transverse cracking than those without drainage. Longitudinal Cracking-WP • The effects of HMA and base thickness on longitudinal cracking-WP are insignificant at this point in time. Pavements with drainage have shown less cracking than un- drained pavements. The main effect of drainage is not of practical significance at this point. • In general, pavements built on fine-grained soils have shown more longitudinal cracking-WP than those built on coarse-grained soils.

397 • On average pavements in WF zone have shown higher levels of longitudinal cracking-WP than those in WNF, especially among pavements built on fine-grained subgrade. This effect was found to be only marginally significant. • Among pavements built on fine-grained soils, those built with DGAB have shown more longitudinal cracking-WP, and those built with ATB have shown the least amount of cracking. Also, drainage has a significant effect on longitudinal cracking, and this effect is more pronounced in pavements built with DGAB. This trend implies that if a pavement on fine-grained subgrade is constructed with a DGAB base, better performance (in terms of longitudinal cracking-WP) can be achieved by providing drainage. These effects are seen in both WF and. WNF zones. Longitudinal Cracking-NWP • The effects of HMA thickness, base thickness, and base type on longitudinal cracking-NWP are insignificant at this point in time. Pavements with drainage have shown slightly less cracking than un-drained pavements. However, the effect of drainage was found to be only marginally significant. • The effect of subgrade type was not found to be statistically significant. • In general, more longitudinal cracking-NWP was observed among sections located in “freeze” climate compared to those in “no-freeze” climate. • The effect of drainage is more pronounced (with marginal statistical significance) among pavements located in “freeze” climate. However, this effect is not of practical significance. These initial trends indicate that longitudinal cracking-NWP is caused by “freeze” climate (frost effects), and that pavements without drainage may be more prone to it. In summary, based on the above discussion for SPS-1 experiment, base type seems to be the most critical design factor for fatigue cracking, roughness (IRI), and longitudinal cracking- WP. This is not to say that the effect of HMA surface thickness is not significant. In fact, the effect of base type should be interpreted in light of the fact that a dense graded asphalt treated base effectively means thicker HMA layer. Drainage and base type, when combined also play an important role in improving flexible pavement performance, especially in terms of fatigue and

398 longitudinal cracking. Base thickness has secondary effects on performance, especially in the case of roughness and rutting. Subgrade soil type seems to be playing an important role in flexible pavement performance. In general, pavements built on fine-grained soils have shown worst performance, especially in the case of roughness. Also, climate is a critical factor in determining flexible pavement performance. Longitudinal cracking-NWP, transverse cracking, and longitudinal cracking-WP appear to be affected by climate. Longitudinal cracking-WP and transverse cracking seems to be associated with Wet Freeze environment, while longitudinal cracking-NWP seems to be the dominant in “freeze” climate.

399 8.2.2 Effect of Design and Site Factors on Pavement Response Three pavement response parameters were chosen for ANOVA−peak deflection under FWD load (d0), far-sensor deflection (d6), and AREA. A summary of the effects of design and site factors on each of the response parameters follows. Peak Deflection under FWD Load (d0): For pavement sections built on DGAB, those with 102 mm (4-inch) HMA surface thickness have higher d0 than those with 178 mm (7-inch) HMA surface thickness. Also pavements with thicker bases, irrespective of base type, have lower d0. This effect is more prominent in the case of sections with treated bases (ATB or ATB/DGAB). Also, pavement sections with PATB/DGAB have lower d0 than those with DGAB. In general, pavements built on fine-grained soils have shown significantly higher d0 as compared to those built on coarse-grained soils. This effect is more prominent on pavements located in WNF zone. Far Sensor Deflection (d6): Pavements built on fine-grained soils have higher d6 values as compared to those built on coarse-grained soils. This effect is more prominent on pavements located in WNF zone. Pavements built with DGAB have shown higher d6 values than those built on other base types. Pavements constructed on 203 mm (8-inch) bases have also shown significantly higher d6 values than those built on 305 mm (12-inch) or 406 mm (16-inch) bases. Furthermore, pavements with PATB/DGAB have smaller d6 values than those with DGAB. These effects of the design factors on d6 are based on statistical analyses only, and may or may not be of practical importance. AREA: For pavements built on DGAB, those with “thin” HMA surface layer have lower AREA values compared to those with “thick” HMA surface layer, implying that the upper layers of these pavements are “less stiff”. For pavements built on DGAB, increasing base thickness from 203 mm (8-inch) to 305 mm (12-inch) has not shown a significant effect on AREA; however a two-fold increase in base thickness [from 8 to16 inch (203 to 406 mm)] has shown a significant increase in AREA. Furthermore, pavements with PATB/DGAB have higher AREA values than those with DGAB. This may be an indication that the structural capacity of the PATB layer is somewhat higher than that of the DGAB.

400 8.2.3 Apparent Relationship between Response and Performance Two types of relations between flexible pavement response (FWD) and performance were explored for the SPS-1 pavements— explanatory and predictive. The salient findings are briefly presented below: Overall Analysis— Explanatory Relationship Following are findings regarding relationship between response and performance based on regression analysis, after adjusting for HMA surface thickness and pavement mid-depth temperature at the time of testing: • Older pavements have slightly lower deflections (d0) compared to younger pavements, which could be due to stiffening (aging) of the asphalt. • Pavements with “weaker” subgrade (higher d6) have higher d0 values. • Pavements with more cracking (fatigue cracking or longitudinal cracking) have a higher d0 values, compared to those with less cracking. Site Level Analysis— Predictive Relationships This section summarizes the observations regarding the predictive relationships between initial response and future pavement performance, based on site-level analysis. • In most of the sites, pavements with higher initial SCI or BDI, or lower initial AREA have higher fatigue cracking. • In most of the sites, pavements with higher initial BDI have higher IRI. • The deflection basin parameters have not shown a consistent relationship with rut depth for the various sites in the SPS-1 experiment. Overall Analysis— Predictive Relationships The main observations based on the analyses are as follows: • For pavements constructed on fine-grained soils, ones with higher SCI have shown more fatigue cracking, especially in WNF zone. • Stiffer pavements (higher AREA) built on fine-grained soils have shown more fatigue cracking, especially if located in WF climatic zone. • Higher longitudinal cracking-WP was observed for the pavement sections with higher AREA, especially among pavements located in WNF climatic zone.

401 • No apparent relation was observed between AREA and longitudinal cracking-NWP or transverse cracking, implying that these distresses could be independent of the pavement structural capacity. Dynamic Load Response for OH (39) test sections The observations based on analysis of DLR data from instrumented sections are: • In general, the strains in the longitudinal direction are higher than the strains in the transverse direction. • The sections that were observed to have higher initial strain values have shown worse fatigue performance. • The sections that were observed to have high initial stress at the top of the subgrade layer and those that were observed to have high initial surface deflection under the load have shown poor rut performance. 8.3 EFFECTS OF STRUCTURAL FACTORS FOR RIGID PAVEMENTS — SPS-2 EXPERIMENT The SPS-2 experiment, entitled Strategic Study of Structural Factors for Rigid Pavements, is one of nine special pavement studies in the LTPP program. The key conclusions regarding the influence of the experimental factors, based on this study, are summarized below. It should be noted that the effects presented herein are statistically significant unless mentioned otherwise; however, they may not be of practical significance at this point in time. 8.3.1 Effect of Design and Site Factors on Pavement Performance Transverse cracking: PCC slab thickness and base type seem to be the most important factors affecting the occurrence of transverse cracking, whereas, drainage seems to have a marginal effect. The effects of design and site features on transverse cracking are as follows: • The occurrence of transverse cracking among pavements with 203 mm (8-inch) PCC slab thickness is higher than that among those with 279 mm (11-inch) PCC slab thickness.

402 • The occurrence of transverse cracking among pavements constructed with LCB is higher than that among those with PATB/DGAB or with DGAB. Pavements with PATB/DGAB have shown the “best” performance (least occurrence of cracking). • Sections without drainage have slightly higher likelihood of cracking than sections with drainage. • On average, among sections built with LCB, those with 203 mm (8-inch) PCC slab have higher occurrence of cracking than those with 279 mm (11-inch) PCC slab. It is important to interpret these results in light of the construction issues, i.e. shrinkage cracking in LCB. • Pavements built on fine-grained soils have slightly higher chances for the occurrence of transverse cracking than those built on coarse-grained soils. Longitudinal Cracking: PCC slab thickness and base type seem to be the most important factors affecting the occurrence of longitudinal cracking. The effects of design and site features on longitudinal cracking are as follows: • The occurrence of longitudinal cracking among pavements with 203 mm (8-inch) PCC slab thickness is higher than among those with 279 mm (11-inch) PCC slab thickness. • The occurrence of longitudinal cracking among pavements constructed with LCB is higher than among those with PATB/DGAB or with DGAB. Pavements with PATB/DGAB have shown the “best” performance (least occurrence of cracking). • On average, among sections built with LCB, those with 203 mm (8-inch) PCC slab have higher occurrence of cracking than those with 279 mm (11-inch) PCC slab. It is important to interpret these results in light of the construction issues i.e. shrinkage cracking in LCB. Faulting: A majority of SPS-2 sections are exhibiting “good” performance with respect to joint faulting, at this point in time. Only 33% of the sections have 0 to 20% of the joints that faulted more than 1.0 mm, and just 5% of the sections have more than 20% of the joints that faulted more than 1.0 mm. Therefore, the results at this point may only indicate the initial trends/observations that may not be of much practical significance. Among all the design factors, lane width seems to be most important for faulting of PCC joints.

403 In general, pavements with standard lane [3.7 m (12 ft) wide lane] have shown higher faulting than those with widened lane [4.3 m (14 ft) wide lane]. The effect of lane width is more prominent among sections built on fine-grained soils than among those built on coarse-grained soils. Also, the effect is more pronounced among sections located in WF zone. Roughness: The initial roughness (smoothness) of the pavement sections in the experiment seems to be affected by the PCC slab thickness. Pavements with thicker slab (279 mm) were found to have more initial roughness compared to those with thinner slab (203 mm). Drainage and base type seem to be the most important factors affecting the growth in roughness, whereas, slab thickness seems to have a marginal effect. The effects of design and site features on change in IRI are as follows: • Pavements constructed with PATB have shown lower change in IRI (∆IRI) compared to those with DGAB or LCB, while pavements with DGAB have the highest change in roughness • Among pavements constructed with standard lane [3.7 m (12 ft) wide lane], sections with DGAB have shown higher ∆IRI than those with LCB or PATB. • Among pavements built on fine-grained soils, those with 203 mm (8-inch) PCC slab have higher ∆IRI than those with 279 mm (11-inch) PCC slab. This effect is more prominent among sections located in WF zone. • Among sections located in WF zone and built on fine-grained soils, those with drainage (i.e. sections with PATB) have shown lower ∆IRI compared to those without drainage. The above results suggest that the change in roughness can be inhibited by constructing pavements with PATB and drainage as compared to sections with DGAB or LCB, especially in the case of pavements built on fine-grained soils. Also, among pavements built on fine-grained soils, an increase in PCC slab thickness from 8” (203 mm) to 11” (279 mm) seems to help prevent an increase in pavement roughness. 8.3.2 Effect of Design and Site Factors on Pavement Response Analyses were performed on the peak deflection under the FWD load plate (do), the far sensor deflection at 60 inches (1524 mm) from the FWD load (d6), the “Area Factor” (AF), and

404 Effective Stiffness (ES) of the PCC slab. All the response parameters have been calculated using the midslab deflections. Peak Deflection under FWD load (do): Pavements constructed with DGAB have higher do values than the ones constructed with PATB. Also, pavements constructed on LCB have the least do values. Pavements with 203 mm (8-inch) thick slab have higher do values than the 279 mm (11-inch) thick slabs. In the Wet Freeze zone, the pavements built on fine subgrade soils have higher do values than those built on coarse subgrade soils. Similar results were obtained from the analysis of the latest (or final) do values. Far Sensor Deflection (d6): The pavements constructed on DGAB have higher d6 values than the ones constructed on PATB. The pavements constructed on LCB have the least d6 values. Also the pavements with 203 mm (8-inch) PCC slabs have higher d6 values than those with 279 mm (11-inch) PCC slabs. Similar results were obtained from the analysis of the latest (or final) do values. In the Wet Freeze zones, pavements built on fine subgrade soils have higher d6 values than those built on coarse subgrade soils. Area Factor (AF): Pavements with 279 mm (11-inch) PCC slab have higher AF than those with 203 mm (8-inch) slab. Among pavements with LCB, those constructed on coarse-grained subgrade have higher AF than those constructed on fine-grained subgrade soils. These effects were not significant for final survey AF values. Pavements with 6.2 MPa (900 psi) concrete have higher AF than those with 3.8 MPa (550 psi) concrete. Sections located in “wet” climate have higher AF than those in “dry” climate. Effective Stiffness (ES): The effect of PCC thickness on ES is more prominent among pavements with DGAB than among those with LCB. The effect of PCC flexural strength on ES is more apparent for pavements with DGAB or PATB than for sections with LCB. Also, pavements built on coarse-grained subgrade soil were stiffer than those built on fine-grained soil. The effects of PCC thickness and base type on ES from final survey were similar as in the case of initial ES. However, pavements built with drainage have higher ES than those

405 without drainage. Also, pavements with 6.2 MPa (900 psi) concrete have higher ES than those with 3.8 MPa (550 psi) concrete. 8.4 EFFECTS OF THE ENVIRONMENT IN THE ABSENCE OF HEAVY TRAFFIC FOR FLEXIBLE & RIGID PAVEMENTS — SPS-8 EXPERIMENT The SPS-8 experiment is entitled Strategic Study of Environmental Effects in the Absence of Heavy Loads for Flexible and Rigid Pavements. The study examines the effect of climate and subgrade type (active, fine, and coarse) on pavement sections incorporating different flexible and rigid pavements, which are subjected to very limited traffic as measured by ESAL accumulation. The SPS-8 pavements have “low” occurrence and extent of distresses, at this point. Most of the pavements in the experiment are performing at comparable levels. No formal statistical methods can be employed due to this. Therefore the observations presented here are just based on average performance of the distressed pavements. The observations, presented below, need to be considered as initial trends in light of these limitations. Flexible Pavements On average, pavements in WF zone have more fatigue cracking, longitudinal cracking- NWP, and roughness than pavements in other climates. Also, in general, pavements constructed on “active” subgrade (frost susceptible or expansive) soils have higher longitudinal cracking- NWP, transverse cracking, and fatigue cracking than pavements on “non-active” soils. Pavements located in “wet” climate, on average, have higher change in IRI than those in “dry” climate. Furthermore, pavements located in WF zone and those built on active soils have the higher changes in IRI. Rigid Pavements Longitudinal spalling, on average, was higher in sections located in “wet” climate. Spalling was not observed in any of the pavements located in the DF zone and in any of the pavements constructed on coarse-grained subgrade soil. Transverse cracking was not observed in any of the pavements constructed with thicker PCC slabs and in any of the pavements constructed on coarse-grained subgrade soils.

406 8.5 LIMITATIONS OF THE EXPERIMENTS AND ANALYSES All the above findings/observations on the effects of design and construction features on pavement performance and response should be considered in light of the limitations discussed herein. These limitations can be broadly classified under two categories— experiment-related and data-related. 8.5.1 Experiment-related issues • The SPS-1, SPS-2 and SPS-8 experiments, which are fractional-factorial designs, were rendered unbalanced because unequal numbers of sites were constructed in each zone- subgrade combination. This unbalanced design limits the “power” of the experiments. In the SPS-1 experiment only 2 sites each are located in DF and DNF zones, compared to 8 and 6 sites in WF and WNF zones. Moreover, both the sites of SPS-1 in the DF zone are located on coarse-grained soils. In the SPS-2 experiment, 7 are located in the WF zone compared to 2, 3 and 2 in WNF, DF, and DNF zones, respectively. Furthermore, in some of the sites not all sections were constructed on the same subgrade soil type [for example, KS (20), NV (32)]. • Initially, the SPS-1 and SPS-2 experiments were designed to have all 24 designs at a site. But later due to some implementation issues, 12 designs were constructed per site. Hence, the experiments do not have any “true” (statistical) replication of designs. In other words, though two sites (with 12 designs at each site) are located in a climate-subgrade soil combination, the traffic, age, and material-related properties vary between the sites. • The variation in age of the sites is considerably high for the experiments. If the sites were reasonably similar in age, the findings would be more reliable. It should be noted that age of the test sections was included as a covariate in all statistical analyses to address the above issue to some extent. • In both SPS-1 and SPS-2 experiments, in-pavement drainage was provided only for sections built with PATB. Moreover, all sections with PATB were provided with drainage. As a result of this, the effect of PATB and the effect of drainage are inseparable (confounded). • In the SPS-1 experiment, a 406 mm (16-inch) thick base was only provided for sections with drainage. In other words, none of the sections without drainage have a 406 mm (16-

407 inch) base thickness. Hence, the effect of a 406 mm (16-inch) thick base and the effect of drainage are also inseparable. • Among flexible pavement sections of SPS-1, all sections built with ATB-over-DGAB were not provided with drainage. Hence, any interaction effects of drainage and ATB- over-DGAB cannot be studied. • The sections with 203 mm (8-inch) thick PCC slabs have dowels of 32 mm (1.25-inch) diameter and sections with 279 mm (11-inch) thick PCC slabs have dowels of 38 mm (1.5-inch) diameter. The effect of PCC slab thickness, especially on faulting, is thus not “pure”. • Only the lower limit for traffic volume was specified for the SPS-1 and SPS-2 experiments. This resulted in considerable variability in traffic across sites. • In the SPS-8 experiment, the effects of HMA surface thickness and base thickness are confounded. Therefore, the “pure” effects of any of these factors cannot be studied. 8.5.2 Data-related issues • Reasonably accurate monitored traffic data is not available for all the sites in the experiments. This has further complicated the issue of controlling for traffic. • Large measurement variability was observed, over time, for some of the distresses (for example, longitudinal cracking in SPS-1 and faulting in SPS-2). This variability has made the time-series trends unclear for some of the performance measures. • The measurement variability discussed above is believed to be due to maintenance activity at some sites, which is a deviation from the experiment design. These activities tamper with the actual long-term field performance of the pavement sections. • The frequency of distress surveys is not uniform across sites, especially for SPS-1 and SPS-2 experiments. Wide gaps in distress surveys necessitate interpolation of performance, which may not always be accurate. • Pumping distress, in the case of rigid pavements (SPS-2), was not considered for analyses based on the recommendations by the project panel, as “the validity of related data is questionable”. • Though thorough construction guidelines were developed by the LTPP to minimize construction variability across sites, some deviations occurred. These deviations along with the material variability have added to the variability in performance across sites. If

408 material-related information were available for all the sections, the issues caused by performance variability could be better addressed. • Backcalculated layer moduli are unavailable for most of the sections in the SPS-1 and SPS-2 experiments. Some of the material-related issues could be dealt with if the data were present. • The data regarding the coefficient of thermal expansion (CTE) of concrete is not available in the DataPave IMS database (Release 17.0). Therefore, CTE of concrete could not be considered in the analyses. 8.6 RECOMMENDATIONS FOR FUTURE DATA COLLECTION AND RESEARCH Based on the above issues and the experience of the authors with the LTPP data, recommendations for future data collection and research are given below. 1) Reasonably accurate monitored traffic data should be made available for all the sites to allow for better adjustment of traffic loading variation across sites. 2) All the core sections of the experiments, especially SPS-1 and SPS-2, should be closely monitored until failure or to a stage when the long-term performance (at least 15-20 years) has been captured. 3) The core sections of both SPS-1 and SPS-2 experiments should be strictly supervised to prevent any maintenance activity, as per the experiment designs. This will ensure that the actual long-term performance of the pavements is observed. 4) Most of the test sections will soon enter a critical stage in their service life; in light of this, to reap maximum benefits from the experiments, the sections should be monitored at regular intervals and with greater accuracy. 5) Hall et al [1] have identified issues regarding in-pavement drainage for sections in SPS-1 and SPS-2 experiments. The findings from this study should be considered for inclusion in the DataPave IMS database. This may help study the effect of in-pavement drainage more accurately. 6) Some of the sections in the SPS-1 and SPS-2 experiments have shown premature “failure”. These sections should be considered for exclusion from DataPave, as they do not contribute to the study of long-term pavement performance.

409 7) Some of the sites in both SPS-1 and SPS-2 experiments are very close to the thresholds (regarding average annual precipitation and freeze index) defined for delineation of climatic zones. The definitions of the climatic zones may need reconsideration in light of this. 8) The definition of pumping, in the case of the SPS-2 experiment, should be revisited to allow for its inclusion in future studies. 9) Accurate material data should be made available for all the sections of the experiments to allow for addressing the variability in material quality across sites. Also, backcalculated layer moduli data should be made available for all sections in DataPave to help perform mechanistic analyses. 10) Most of the construction-related issues are available only from construction reports. These issues and/or deviations should be better highlighted well within the DataPave database. 11) The spatial location of some distresses (such as cracking) is sometimes important for research. It is practically cumbersome for the users to obtain distress maps and interpret the spatial location of the distresses. It is therefore recommended that each section be “discretized” (segmented) for data collection and the related data be made available in DataPave. This would greatly decrease the level of subjectivity in the data. 12) In general, the extent of distresses on the SPS-8 test sections is “low”, at this point in time. The performance data should thus be collected for sufficiently long time (15 to 20 years for all the sections in the experiment) to capture the effects of environment. A meaningful statistical analysis may then be performed to study the effects of environment. 13) The DLR instrumentation location (spatial location), alignment and designation details should be made more accurate in DataPave. 14) It is recommended that the complete actual traces of data from the instrumented DLR sections be considered for inclusion in the DataPave database (in addition to the “peak” and “valley” data) after proper quality checks. Also more data (i.e. more test series results) should be included in the DataPave. This data could be stored separately as in the case of profiles and FWD time histories.

410 When the long-term performance data is available for most of the sections of the SPS-1 and SPS-2 experiments, the methods employed in this research would be more “powerful” to study the effect of design and construction features. Methods that analyze the time-series data (such as survival analysis and ANOVA with repeated measures) can also be employed when the performance data for most of the sections is available for about 15 years.