Christopher M Dean

Effect of Frequency of Pavement Condition Data Collection on Performance Prediction

Monitoring pavement surface conditions over time is essential for pavement management and performance models. Time series distress data can be used to determine the remaining service life at the project level, which can then be used for all projects to assess the overall health of the pavement network. Observed field performance is also crucial for calibrating performance prediction models for pavement design purposes. Therefore, highway agencies collect pavement condition data to accomplish both policy and engineering objectives. However, differences exist among agencies between the monitoring frequency used for pavement surface distress (imaging) and that used for sensor-measured features. These differences pertain primarily to the relative difficulties in collecting and processing imaging data. Many agencies collect sensor data more frequently than images. Most highway agencies monitor pavement condition at 1-, 2-, or 3-year frequencies. Discrepancies between performance model predictions and observed field performance are conventionally attributed solely to errors in predicted pavement distresses. In fact, inherent uncertainty may also be present in measured pavement distresses due to spatial variability, sampling, and measurement errors. The frequency of distress data collection adds further uncertainty in performance prediction. This paper explores the effect of pavement condition monitoring frequency on pavement performance prediction. Analyses of observed pavement performance show that condition data collection frequency can significantly affect performance prediction. Therefore, more frequent data collection for image-based methods can reduce the associated risk in performance prediction and thus be more effective for better decision making for pavement management.

Impact of three state practices on effectiveness of hot-mix asphalt overlay

The effectiveness of pavement treatments is a function of several factors, including material quality, design, distribution of pavement condition states (conditions and rates of deterioration) before treatment, and construction quality. Because of these factors, state highway agencies have established estimates of treatment service lives with significantly wide ranges. These wide ranges make the calculation of treatment benefits or effectiveness a difficult task unless the role of each factor that affects service life is well understood. This study analyzed the impacts of the state of the practice of Colorado, Louisiana, and Washington on the effectiveness of thin (< 2.5 in.) hot-mix asphalt (HMA) overlay treatment. This paper reports that analysis of time-series pavement condition and distress data of previously treated pavement sections can be tabulated in a matrix format to demonstrate a snapshot examination of past practice. Such matrices are called treatment transition matrices (T2Ms). The data in the matrices express the probability that a pavement treatment will have certain effectiveness. T2Ms for pavement projects subjected to thin HMA overlays by three state highway agencies are discussed. Differences in treatment effectiveness were related to differences in the state of the practice in the three states. Selection of treatment timing and project boundaries significantly affected treatment effectiveness, and similar states of the practice yielded similar treatment effectiveness.

Defining Benefits from Pavement Rehabilitation and Preservation

Pavement service life can be defined as the estimated number of years between pavement construction or rehabilitation and when the pavement section reaches a given condition (rutting, cracking, roughness, etc.) threshold value. At that time, the pavement section is typically subjected to rehabilitation or preservation actions. The two overarching methodologies for determination of pavement fix benefits are: 1) the extension in the pavement service life due to a given fix and 2) the calculation of the area between the pavement performance curves and a given threshold value. The life extension method places importance in prolonging the service life of the pavement. On the other hand, the area under the performance curve method places importance on the level of pavement distress over time. Selection of the pavement fix type and time with the lowest cost to benefit ratio could be considered “optimal”.

Pavement Condition States Before and After Treatment

When a pavement section is subjected to preservation or rehabilitation treatment, its surface condition is transformed from one set of condition states—those before treatment—to another set—those after treatment. This transformation is a function of the pavement condition before treatment, the type and timing of treatment, the project boundaries, and the quality of construction. For most pavement sections, the surface condition before and after treatment varies substantially along the section and over time. This study sponsored by FHWA analyzed the distribution of pavement surface condition and distress before and after treatment along several flexible pavement projects in the state of Louisiana. The study showed that current practices regarding the selection of pavement treatment type, treatment time, and project boundaries were independent of pavement surface condition and distress before treatment. For all pavement projects that received certain types of treatment in the past, knowledge of the relationships between the before- and after-treatment distributions of the pavements surface condition and distress was crucial to the establishment of a future cost-effective strategy for pavement treatment. This paper shows that such relationships can be expressed by the probabilities of transforming before-treatment condition states to after-treatment condition states. These probabilities can be housed in one matrix format called the treatment transition matrix (TTM). For each treatment type, the TTM presents a snapshot of the state of the practice.

Modeling Pavement Conditions of Multilane Roads with Measured Driving Lane Data

In the past, highway authorities have selected pavement treatment type on the basis of the driving lane conditions and have applied the same treatment to all lanes. Recently, authorities started applying different treatments to the driving and passing lanes at different times. Because of costs, most highway authorities collect only the driving lane pavement condition data. Therefore, the passing lane pavement conditions need to be modeled with the driving lane data. To address this need, time series pavement condition and distress data along the driving and passing lanes of the Minnesota Road Research Project were obtained. The data were analyzed to develop and verify methodologies to estimate the passing lane conditions and distresses accurately with the use of the measured driving lane data. The methodologies are based on the fact that most pavement distresses are caused by material and environmental factors and by traffic loads. The effects of the first two factors are similar on adjacent driving and passing lanes. The differences in the conditions between adjacent lanes are mainly because of traffic distribution, which is expressed by the lane distribution factor. The developed methodologies use the measured pavement conditions along the driving lane and the lane distribution factors to estimate the time series conditions of the passing lane. For asphalt pavements, the time-dependent conditions of the passing lane can be accurately estimated (with low standard errors) with the measured conditions of the driving lane and the lane distribution factors. Procedural steps for implementing the methodologies are included.

The remaining service life a good pavement management tool

The remaining service life (RSL) of a given pavement section is the estimated number of years between now and the time when the pavement conditions reach certain amount of distress. The calculation of RSL is based on two steps; fitting the proper mathematical function to the time series condition data, and predicting the time at which the pavement reaches the threshold conditions. Time series pavement condition data of several miles of roads were obtained from four State Highway Agencies (SHAs). The data were analyzed and the RSL of each 0.1-mile section of each road was calculated based on the International Roughness Index, (IRI), rut depth and cracking data. Results of the analyses are presented in this paper. It is shown that, the RSL is a good communicating and strategy planning tool. It could be used to address engineers, the public, legislators and managers.

The Calculation of the Remaining Service Life Based on the Pavement Distress Data

The distress points, which are the numerical values assigned to pavement distress are not universal or standardized. Most agencies assign these points based on past experience with the intention to periodically calibrate them. The distress points significantly affect the distress indices, the life cycle cost, the remaining service life (RSL), and the accuracy of the pavement decisions. Distress data were obtained from four State Departments of Transportation. The data were analyzed and the RSL was calculated and equalized to the dollar values of the network and its rate of depreciation. Results of the analyses are presented in this paper. It is shown that the RSL could be calculated without the need to assign distress points. It is also shown that the weighted average RSL of the network reflects the dollar value of the network and its rate of depreciation. The RSL could be used to support accurate pavement decisions.

The Effect of Pavement Condition Data Sampling on Project Boundary Selection

Pavement condition data are either continuously collected or sampled. In sampling it is assumed that the time and costs of data collection can be reduced and the pavement condition of shorter segments represent the conditions of larger sections. In a study sponsored by the Federal Highway Administration, pavement distress data along several miles of roads were obtained from four State Highway Agencies. Each department collects and stores distress data on a continuous basis for each 0.1 mile section along the network. The continuous data were sampled and the impacts on the accuracy of pavement decisions were analyzed. Results of the analyses are discussed herein. It is shown that, for variable pavement conditions, ten percent sampling leads to inaccurate decisions. Pavement sections in need of repair are ignored whereas healthy sections are selected for repair. The costs incurred due to inaccurate decisions could be much higher than the saving incurred by sampling.