Evaluation and Analysis of LTPP Pavement Layer Thickness Data

Previous | Table of Contents | Next

Summary and Conclusions

This study was conducted to assess quality and completeness of pavement layering information and layer thickness data and to provide recommendations for improvement of the data that are currently available in the LTPP database. Within-section layer thickness variability was characterized, and as-designed and as-constructed thicknesses were compared. Additionally, a Guide for LTPP Layer Thickness Data was developed.

Data Availability and Completeness

In the course of the study, layer thickness data available in the LTPP database were examined to evaluate quality and completeness using Levels A to E data. The layer thickness data availability assessment indicated that the TST_L05B and TST_L05A tables contain the most complete set of information about the representative layer structure and thickness for section-level analysis. Only 16 pavement structures from LTPP regular sections and 1 pavement structure from a supplemental section do not have any layer structure (including thickness) information in either TST_L05B or TST_L05A. Analysis of data completeness at QC Level E revealed 3,457 pavement layer structures in the EXPERIMENT_SECTION table. Some 3,240 of these structures (93.7 percent) had records in table TST_L05B, while 3,229 structures (93.4 percent) had records in table TST_L05A.

Layer Thickness Quality and Consistency

Following the data completeness evaluation, pavement layer thickness and other related data from different data sources were evaluated to determine consistency of layer functional description, material type, and thickness data between different data sources. In addition, layer thickness variability indicators, within-section material type consistency, and material type and thickness reasonableness were evaluated using selected tables where these parameters were available.

The results of the data consistency evaluation showed that the pavement layer functional descriptions are consistent between different LTPP tables for 93 percent of all cross-section layers evaluated in the study. Material type descriptions were found to be consistent between different tables for 79 percent of all section layers evaluated in the study. Evaluation of material type consistency was constrained by the absence of a unified material coding scheme. Representative layer thickness values were found consistent between different tables for 89 percent of all pavement cross-section layers evaluated in the study. In the cases where inconsistency in data from one or more data sources was identified, a layer was flagged for further review. Inconsistencies in pavement layering data were reviewed and reported to the LTPP data managers in the form of the data analysis/operations feedback reports along with recommendations for data anomaly resolution.

Additionally, reasonableness (or validity) of material type description was evaluated. The purpose of the reasonableness check was to evaluate whether the material description code for the layer is consistent with the layer functional description. While most of the records had valid material codes, 642 records out of 41,111 (1.56 percent) had erroneous material codes, and some records were missing material codes. The identified records were reported to the FHWA in a data analysis/operations feedback report.

Reasonableness of layer thickness data was evaluated using the representative layer thickness ranges specified in the SHRP-LTPP Lab Guide [3]. As a result of the layer thickness reasonableness evaluation, thickness values outside the representative thickness ranges were identified and reported to the FHWA.

Within-Section Thickness Variation

The variation in layer thickness data from SPS experiments obtained at different locations within sections was analyzed and characterized using theoretical statistical distributions. The analysis included layers with different material and functional types, including AC surface courses, combined AC surface and binder courses, AC binder courses, dense-graded aggregate bases, dense-graded AC-treated bases, permeable AC-treated bases, lean concrete bases, PCC surface layers, and PCC overlay layers. To assess layer thickness distribution characteristics, descriptive statistics such as mean, standard deviation, skewness, and kurtosis were computed for each section. A combined test for skewness and kurtosis was selected to test the normality of layer thickness distributions for 1,034 SPS layers. The statistical analysis results indicated that, for 84 percent of all layers, thickness variations within a section indicate a normal distribution. These results can serve as a very important input to pavement engineering applications involving reliability of pavement design and also for quality assurance construction specifications.

As-Designed versus As-Constructed Thickness Comparison

As-constructed core and elevation layer thickness measurements were compared to the design (or target) thickness values for newly constructed SPS layers. The data were evaluated to determine the percentage of the individual measurements either within or outside specific values from the target thickness.

Statistical analyses of the measured mean thickness values versus the designed values were performed using t-tests. Two sided t-tests with 95 percent confidence level were used for each section and layer to estimate whether the differences between as-designed and as-constructed thicknesses are significant. One-sided t-tests with 95 percent confidence level were used for each layer for the difference between as-designed thickness and the mean as-constructed thickness and for allowances of 6.35 mm (0.25 in), 12.7 mm (0.5 in), and 25.4 mm (1 in).

Based on the analysis of both data sources, the following conclusions can be made:

• The majority of the LC and PCC layers have constructed or measured thickness above the design values. This is particularly true for thin (76 mm [3 in] and 123 mm [5 in] thick) PCC bonded overlays of PCC slabs.
• Thin PCC and asphalt concrete surface and binder layers have the highest number of sections with mean as-constructed thickness that significantly deviates from the designed thicknesses.
• Elevation thickness measurements for PATB are more evenly distributed around the target value. However, the core measurements for PATB show that significant number of sections have thicknesses lower than the target thickness. It appears that for some cores the entire thickness of PATB layer was not obtained. The analysis shows the values currently stored in the database. A feedback report was submitted regarding these questionable data. In some cases the core thicknesses were below 25.4 mm (1 in), although target thickness is 102 mm (4 in).
• About 60 percent of section/layers have mean thickness within ±6.35 mm (0.25 in) from the target thickness. For the tolerance level of 25.4 mm (1 in) this percentage is above 90 for most layer types and target thicknesses.

A comparison between analysis results from the elevation and core thickness measurements shows that the percentage of measurements within tolerance limits for all three tolerance levels is approximately the same. However, the percentage of measurements lower than the target value is consistently higher for core measurements than for elevation measurements.

Based on elevation measurements, it is observed that more than 70 percent of sections with DGAB have as-constructed thicknesses within ±6.35 mm (0.25 in) from the design value.

Researcher's Guide for LTPP Layer Thickness-Related Data

One important product from this study is a Researcher's Guide for LTPP Layer Thickness Data. The main purpose of this researcher's guide is to provide guidance for selecting layer material type and thickness data from the LTPP database. The guide also contains a discussion about within-section layer thickness variability and a comparison between as-designed and as-constructed layer thicknesses. The researcher's guide is presented in a separate report.

Recommendations

Computed Quantity Data for Inclusion in the LTPP Database

Along-the-section variability of layer thickness is an essential input for reliability-based pavement design and performance modeling. This input is characterized by the statistical distribution attributes. During the evaluation of within-section layer thickness variability, comprehensive descriptive statistics were obtained from rod and level elevation measurement along the LTPP sections, for pavement structural layers (base and surface course):

• Mean
• Standard deviation
• Skewness
• Kurtosis

These data provide means for evaluating the distribution shape of layer thickness measurements observed along the LTPP sections. Tests of normality were carried out to identify sections and layers that have thickness values distributed normally. This valuable information provides statistical characteristics of the within-section variability in pavement layer thickness for different pavement layers and material types required for pavement engineering studies involving assessment of pavement design reliability, such as mechanistic-empirical pavement design procedures or pavement management procedures involving risk analysis. As such, we recommend including these statistics in the LTPP database as a new computed parameter tables (one table for each SPS experiment). The essential fields recommended for the new tables are:

• Layer type
• Mean
• Standard deviation
• Skewness
• Kurtosis
• Normality indicator

Researcher's Guide to LTPP Layer Thickness Data

Pavement layer material type and thickness data are very important for many types of pavement engineering analyses. The accuracy of layer thickness data has a great impact on the outcome of practically all analyses of pavement performance. As part of the LTPP program data collection effort, a large amount of data related to layer material type and thickness data have been collected from several sources. These data are stored in many different tables. Based on the analysis type, data from one or another table may be more appropriate.

To make the process of navigation through the LTPP layer thickness data more user-friendly, a Researcher's Guide for LTPP Layer Thickness Data was developed in this study. This guide discusses the field sampling, materials testing, and other layer thickness data collection activities utilized in LTPP. The layer thickness data that currently reside in the LTPP database are presented in relation to the data collection activities or data sources. The guide also explains how to search for the most appropriate thickness for different research purposes. Characterization of the within-section thickness variation and designed versus constructed or measured thickness data variation for the LTPP sections are also included in the guide. We recommend that this guide be used as a reference when selecting LTPP pavement layering data sources.

Improvement of LTPP Pavement Thickness Data Quantity and Quality

In an attempt to improve LTPP layer thickness data quality and quantity, an extensive review of layer thickness data available in the LTPP database was carried out in this study. As a result, several issues concerning questionable or anomalous data have been identified and reported to FHWA in a form of feedback reports. To improve the quality of existing layer thickness data and to fill in any identified data gaps, the reported data problems should be reviewed by the appropriate parties and, where warranted, the LTPP database should be updated and cleaned to remove anomalous data.

Previous | Table of Contents | Next