U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
|This report is an archived publication and may contain dated technical, contact, and link information|
Publication Number: FHWA-HRT-05-079
Date: May 2006
Optimization of Traffic Data Collection for Specific Pavement Design Applications
Chapter 2. Identify Scenarios And Knowledge Gaps
Methodology For Identifying Key Pavement Design Scenarios
Pavement design requirements are a function of the importance of a roadway facility. It is traditionally defined in terms of functional classification (Interstate, U.S. highway, State highway, or secondary road), which to a large extent reflects the traffic volumes and axle loads that need to be accommodated. Importance, in turn, defines the acceptable reliability in the pavement design of a facility and, hence, dictates the required quality of input data for both materials and traffic. Reliability is defined as the probability that a pavement section will not fail before the end of the analysis period is reached. Table 7 gives the pavement design reliability levels established by the 1993 edition of the AASHTO design guide, which provides a guideline for establishing reliability levels in the current study. (14) If the variation in the NCHRP 1-37A design guide output is known, appropriate levels of reliability of input can be selected. For traffic data input, this reliability will define the type of monitoring equipment and the length of data coverage required. This is the methodology that will be followed in establishing the traffic data collection scenarios required for specific pavement design applications.
Results based on a survey of the AASHTO Pavement Design Task Force.
As described later, the traffic data collection scenarios will be identified by expanding the four levels of traffic input defined by the NCHRP 1-37A design guide (table 4) to account for varying time lengths of coverage in the site-specific traffic data.
Knowledge Gaps In The Sensitivity Of The Pavement Design Process To Traffic Input
The majority of literature in this area treats traffic input in terms of cumulative ESALs and concerns the performance-based design process of the 1993 AASHTO design guide.(14) This process treated uncertainty in predicting performance as the present serviceability index (PSI) by artificially increasing the estimated number of ESALs. This was done by adding to the logarithm of the estimated ESALs, the product of the standard normal deviate corresponding to the desired reliability multiplied by the standard error in predicting PSI. This increased significantly the number of ESALs input to the empirical performance equations (e.g., for 85 percent confidence and a standard error in predicting PSI of 0.5, the logarithm of ESALs was increased by 1.037 x 0.5 = 0.5185, which arithmetically is a factor of 3.3). Although this is not directly applicable to the design philosophy of the new NCHRP 1-37A design guide, it does reflect the significant uncertainties in quantifying traffic loading.
There are few exceptions in the literature where axle-load spectra are used directly in damage calculations, such as the rigid pavement design procedure developed by the Portland Cement Association.(15) This procedure uses axle-load spectra and computes the resulting slab fatigue damage and joint erosion through a Miner's hypothesis-type accumulation algorithm. Experience with this method shows that:
The third fact reemphasizes the need for performing any pavement design sensitivity analysis of traffic loads by considering the thickness of the layers involved.
In summary, the knowledge gap in this area is considerable. Little is known about the sensitivity of the new NCHRP 1-37A design guide design process to traffic input. For a particular pavement type and combination of layer thicknesses, there is a need to study the extent of variation in pavement life predictions by distress type, in response to variations in traffic input.
Knowledge Gaps In Data Variation From Different Traffic Collection Scenarios
As summarized in the literature review, considerable work has been done to analyze the effects of various traffic data collection scenarios on the accuracy of traffic volume estimates, such as AADT and AADTT, as well as cumulative pavement damage, such as ESALs.(2,11-12) The common method used in these studies is simulating traffic data collection scenarios from continuous traffic records and comparing the traffic estimates to the ground truth, thereby establishing accuracy levels. There has been little work, however, on the accuracy in axle-load distribution estimates from short-term WIM data and particularly on the ability to capture the few high axle loads that cause disproportionately high pavement damage. Hence, there is some literature related primarily to the first four traffic input components to the NCHRP 1-37A design guide (table 6), but little is available on the fifth traffic input component, the distribution of axle loads. As described later in "LTPP Data Analysis" (Chapter 4), extended-coverage WIM data from LTPP sites will be used to simulate the effects of various sampling scenarios on the traffic data input components.
Two main knowledge gaps were identified in selecting a traffic data collection effort for particular NCHRP 1-37A design guide applications:
These knowledge gaps will be filled by analyzing data from the LTPP database and conducting a sensitivity analysis of the new NCHRP 1-37A design guide software output with respect to traffic input.