Wouter C. Brink

Sensitivity of Input Variables for Flexible Pavement Rehabilitation Strategies in the M-E PDG

The Mechanistic-Empirical Pavement Design Guide (M-E PDG) can be used to design different flexible rehabilitation options; namely HMA over HMA (overlay), HMA over JPCP (composite), and HMA over JPCP fractured (rubblized). Each rehabilitation option has different input variables for new and existing layers. The type and magnitude of the input parameters have significant impact on the predicted flexible pavement performance. In order to evaluate the impact of various design inputs on the predicted performance for the rehabilitation options, one-variable-at-atime (OAT), and detailed sensitivity analyses were performed. While the OAT sensitivity analysis investigated the input variables specific to the existing pavement layers, it also determined the variable effects on the overall pavement performance. The OAT results identified the significant input variables of the existing pavement layers. Subsequently, based on the results of OAT sensitivity analysis, a detailed sensitivity analysis was performed. Full factorial matrices were designed to determine statistically significant main and interaction effects. This paper highlights the process of identifying the most sensitive input variables for flexible rehabilitation design options. The results of the sensitivity analyses show that the existing pavement condition rating and existing thickness for HMA over HMA overlays is critical for all performance measures. On the other hand, existing PCC modulus and thickness are important in determining the performance of HMA overlay over intact and rubblized JPCP. Keyword: Pavement rehabilitation, M-E PDG, Sensitivity analysis, Input variables

Local calibration of rigid pavement performance models using resampling methods

Abstract The performance prediction models in the Pavement-ME design software are nationally calibrated using in-service pavement material properties, pavement structure, climate and truck loadings, and performance data obtained from the Long-Term Pavement Performance programme. The nationally calibrated models may not perform well if the inputs and performance data used to calibrate those do not represent the local design and construction practices. Therefore, before implementing the new M-E design procedure, each state highway agency (SHA) should evaluate how well the nationally calibrated performance models predict the measured field performance. The local calibrations of the Pavement-ME performance models are recommended to improve the performance prediction capabilities to reflect the unique conditions and design practices. During the local calibration process, the traditional calibration techniques (split sampling) may not necessarily provide adequate results when limited number of pavement sections are available. Consequently, there is a need to employ statistical and resampling methodologies that are more efficient and robust for model calibrations given the data related challenges encountered by SHAs. The main objectives of the paper are to demonstrate the local calibration of rigid pavement performance models and compare the calibration results based on different resampling techniques. The bootstrap is a non-parametric and robust resampling technique for estimating standard errors and confidence intervals of a statistic. The main advantage of bootstrapping is that model parameters estimation is possible without making distribution assumptions. This paper presents the use of bootstrapping and jackknifing to locally calibrate the transverse cracking and IRI performance models for newly constructed and rehabilitated rigid pavements. The results of the calibration show that the standard error of estimate and bias are lower compared to the traditional sampling methods. In addition, the validation statistics are similar to that of the locally calibrated model, especially for the IRI model, which indicates robustness of the local model coefficients.

Process and Data Needs for Local Calibration of Performance Models in the AASHTOWARE Pavement ME Software

Local calibration of the performance models in the AASHTOWARE Pavement ME software (Pavement ME) is a challenging task, especially if the data are limited. This paper summarizes the local calibration process for flexible and rigid pavements in Michigan. Other agencies can learn from the steps needed to accomplish a more streamlined local calibration. The local calibration process includes several sequential steps. An adequate number of pavement sections needs to be identified from the pavement management system (PMS) database on the basis of pavement type, age, geographical location, and number of collection cycles for performance data. The selection of the final set of pavement sections is based on distress magnitude over time. The selected pavement sections must be categorized on the basis of measured distresses because the local calibrated models are typically used to predict normal pavement performance at the design stage. For the selected pavement sections, the as-constructed input variables are collected from construction records. However, when such input information is unavailable, the best estimates are used to represent the pavement design and construction practices of the Michigan Department of Transportation. Finally, the typical steps for local calibration by using various resampling techniques are demonstrated for the rutting (flexible) and transverse cracking (rigid) models. The techniques are compared through use of the standard error of the estimate (SEE). The SEE of a technique shows how much variance is explained by the model. The main advantage of using resampling is to quantify the variability associated with the model predictions and parameters. The quantification of the variability will also help in determining more robust design reliability in the Pavement ME.

Updates to Hourly Climate Data for Use in AASHTOWare Pavement Mechanistic–Empirical Design

The AASHTOWare Pavement Mechanistic–Empirical Design software requires hourly temperature, wind speed, percentage sunshine, precipitation, and relative humidity to properly calculate pavement damage and distresses. Actual or measured values, which vary hourly throughout a day for a given site, are required to properly capture the damage caused by environmental loadings. Currently the mechanistic–empirical design hourly climatic data contain approximately 1,200 U.S. and 300 Canadian stations. The U.S. stations typically contain data from 1995 through 2005, and data from the Canadian stations vary in length from 10 to 50 years, with the exception of some weather stations. Some agencies expanded their historical weather data to include longer periods of time. This paper documents the process and data sources that were used to update the current set of climate stations with climate data dating back to 1979 using the North American Regional Reanalysis (NARR) database. The results of the comparison between new climate files and the existing older climate data files for use in pavement design are presented. Overall, the NARR-generated climate data showed a very good comparison. The paper details the background of the NARR and its limitations and compares the performance predictions made by using the old and new climate data. The results indicate there is no systematic bias between the two climate data sets.

Local Calibration of Rigid Pavement Cracking Model in the New Mechanistic- Empirical Pavement Design Guide using Bootstrapping

The local calibration of the performance models in the new mechanistic-empirical pavement design guide is a challenging task, especially due to the lack of needed data. For the selected set of pavement sections for local calibration, the data requirements include: (a) a wide range of inputs related to traffic, climate, design and material characterization, and (b) a reasonable extent and occurrence of observed performance data over time. In addition, to data limitations, the conventional statistical methods of split sampling for model calibration and validation further add to these complications. In the traditional approach about 70% of the data set is used for calibration and remaining 30% is utilized for validation. However, most of the States have a limited number of identified pavement sections for local calibration. Therefore, there is a need to employ statistical methodologies that are more efficient and robust for model calibrations given the data related challenges encountered by state highway agencies. In this paper, the rigid pavement cracking model was calibrated using the traditional and advanced statistical resampling approaches like jackknifing and bootstrapping. Jackknifing and bootstrapping methods provide more reliable assessment of the model prediction accuracy than the alternative methods. While traditional split sample approach uses a two-step process for calibration and validation, advance approaches can simultaneously consider both steps. Moreover, the goodness-of-fit statistics are based on predictions rather than on data used for fitting the model parameters. The efficiency and robustness of such approaches become more important when the sample size is small. The results in the paper show the effect of different approaches on model parameter estimations and compare the role of such parameters in reducing the bias and the standard error of the cracking model.

Issues Related to Pavement Rehabilitation Design Options in the New Mechanistic- Empirical Pavement Design Guide

Since the DARWin-ME is expected to be implemented by State DOTs for rehabilitation and new pavements analysis and design in the near future, it is important to highlight the practical issues related to the execution of these options. Several issues were encountered while executing the DARWin-ME rehabilitation options. These concerns are related to structural and material characterization of the existing pavement conditions. More specifically, the issues are related to: (a) characterizing the existing Portland cement concrete (PCC) pavement structure for unbonded overlays, (b) modeling of the HMA interlayer for unbonded overlays, (c) using backcalculated subgrade modulus of resilience (MR), and (d) utilizing Falling Weight Deflectometer (FWD) testing. Typical PCC elastic moduli for existing PCC layers range between 3 and 6 million psi. For unbonded overlays, the elastic modulus is used to characterize the existing PCC condition. However, a limiting value of 3 million psi is suggested by the developers of the DARWin-ME. Counter-intuitive results were observed when higher values of elastic modulus were used to represent field conditions. It was also found that the asphalt interlayer structural and material properties have insignificant effect on the equivalent PCC slab thickness. Furthermore, the current performance models in the DARWin-ME are calibrated using backcalculated subgrade MR values. Generally, the values are much greater than ones used for AASHTO 93 designs. In addition, DOTs cannot use the AASHTO 93 MR values in the DARWin-ME directly because the values are internally reduced by the software. Finally, because of critical impact of existing pavement conditions on the rehabilitation design, use of FWD to characterize the existing pavement structure has become more important. This paper documents the above mentioned issues and their impacts on pavement performance. Moreover, remedies are discussed to ensure reliable and accurate results by using the DARWin-ME overlay options.