Browse Projects > Detailed View

Use of Field Measured Data For Improving Pavement Mechanistic-Empirical Design and Analysis Procedure


With advance computational tools and techniques, it is possible to estimate pavement responses due to mixed traffic, understand temperature and moisture sensitivities of different pavement materials, and predict different pavement distress over time. Pavement Mechanistic-Empirical (ME) analysis procedure helped in determining pavement responses mechanistically and relating those responses to pavement performances empirically. Layer Elastic Analysis (LEA) or Finite Element Method (FEM) is commonly used for mechanistic analysis, whereas the field measured data (e.g., Long Term Pavement Performance Study (LTPP) Study) coupled with Accelerated Pavement Studies (APT) are used for empirical analysis. A combination of these mechanistic and empirical approaches contributed to developing a robust Pavement ME model. However, there are some drawbacks in the existing ME models. When performing mechanistic analysis, the pavement layers are assumed to be perfectly elastic and material is isotropic and homogeneous. However, such conditions never exist in reality. Hence, the predicted stress/strain responses at the different pavement layers may not be accurate. As the damage models (a.k.a. transfer functions) are based on the predicted stress/strain responses, it affects the pavement performance predictions. Therefore, there is a need for improving the mechanistic model by utilizing field measured sub-surface pavement responses.


The current ME design algorithm uses sophisticated tools (e.g., numerical models based on LEA or FEM) for estimating “mechanistic” pavement response due to vehicle loadings by considering mix traffic scenarios. Although the vehicle load magnitude directly influences the mechanistic responses, other factors such as variations in climatic conditions (e.g., temperature, moisture) and material properties (e.g., temperature and loading rate dependency of asphaltic material; stress-dependency of unbound materials) also have significant influence. The pavement models usually consist of “elastic” asphalt and aggregate base/subbase layers over subgrade “half-space.” The assumed “elastic” behavior never exists due to heterogeneous material mix and temperature- and stress-dependencies. Besides, a wide range of loading patterns (varying loading frequency and magnitude) due to mixed traffic increases complexity of “mechanistic analysis.” The rutting damage model estimates the total surface deformation computed from vertical permanent deformations accumulated in each pavement layer. Such vertical deformation in each layer is predicted based on the mechanistic analysis. Similarly, the bottom-up asphalt cracking model is depended upon the horizontal strain responses. Therefore, it is required to predict layer-specific deformation (or strain) accurately for improving the reliability of the ME model in performance prediction.

The primary objective of this study is two-fold:

  • Measurement of stress and strain responses in pavement layers under various conditions such as different layer configurations, different subgrade conditions (soft and firm foundations), different material types (asphalt mix and aggregate base course mix), varying climatic conditions, and mixed traffic.
  • Comparison of field-measured stress/strain data and numerical model predicted stress/strain responses, and improvement in the mechanistic model by reducing the bias between the field measurement and model prediction.

FHWA, FAA, USCOE, State DOTs, and other transportation agencies already realized the need for accurate pavement modeling and predicting pavement performance. Therefore, most of these agencies have their Pavement ME design tool or similar tools. The basic principle of those tools is related to finding stress/strain at different pavement layers.

It is now well recognized that no two projects are the same. Due to variations in mechanical properties of pavement material and foundation conditions, construction practices, each pavement construction project has unique challenges. As such, these factors increase the uncertainty in pavement performance prediction. In addition, agencies are leaning toward using recycled material for reducing construction cost as well as developing sustainable pavements. In other words, new or materials with unknown characteristics are being used for pavement construction, and it is challenging to develop numerical models for such materials.

Beside the use of this research for improving pavement models, sub-surface stress/strain measurements also can be used for the early detection of many pavement distresses, because many pavement distresses originate from underneath the paved surface. Eventually agencies may have the opportunity to save budget costs allocated for pavement maintenance and rehabilitation with early detection. However, this is not in the scope of work, and a separate RNS should be developed for looking at this aspect in detail.


In order to achieve the objectives, the following tasks should be completed:

Task 1: A comprehensive literature review of previous related research and their findings; In-depth study of challenges and issues pointed out by past research; detailed study of recommendations made.

Task 2: Selection of Pavement ME design procedure (e.g., AASHTOWare Pavement ME, FAA’s Panda-AP, FHWA’s FlexPave); Determination of inputs required for the numerical model; Development of instrumentation plan to measure such input.

Task 3: Selection of sensors in cooperation with DOTs; Seek expert advice from APT owners and users of pavement instrumentation.

Task 4: Selection of sensor installation location and projects; It is recommended that multiple agencies should conduct this research so that the data could be collected under different climatic zones, and different subgrade types.

Task 5: Deployment and integration of sensors and data collection systems. It is required to collect traffic data as well (preferable Weigh in motion (WIM) data).

Task 6: Coordinating with the numerical model developers for comparison between measured and predicted pavement responses.

Task 7: Refining the numerical model and improving the pavement ME design procedure. The researcher should compare measured stress/strain responses from instrumented pavement sections with the ME model predicted stress/strain responses. Adjustment should be made in the predicted stress/strain responses, and the results should reflect on the damage models. In other words, the damage models should predict pavement performances accurately.

Task 8: Report preparation; Make suggestions for improving material characterization inputs and refining numerical models.


AKG60 (previously known as AFS20) survey results showed that there is a high need for measuring stress/strain response in pavement layers

Sponsoring Committee:AKG60, Geotechnical Instrumentation and Modeling
Research Period:Longer than 36 months
Research Priority:High
RNS Developer:Prajwol Tamrakar & Alessandra Bianchini
Date Posted:04/06/2020
Date Modified:02/10/2021
Index Terms:Field studies, Pavement design, Pavements, Mechanistic-empirical pavement design,
Cosponsoring Committees: 

Please click here if you wish to share information or are aware of any research underway that addresses issues in this research needs statement. The information may be helpful to the sponsoring committee in keeping the statement up-to-date.