Performance evaluation and construction design of concrete overlays
Peter C. Taylor
Concrete overlays extend the service life of existing pavement and are potentially one of the most cost-effective maintenance and rehabilitation strategies for pavement systems. While concrete overlays are not new, the long-term performance of various types of concrete overlays has not been fully investigated because there has been insufficient performance data available to support such evaluation. The Iowa Pavement Management Program (IPMP), the Iowa Concrete Paving Association (ICPA), and other agencies have created a complete concrete overlays historical performance database, and this historical performance database includes the Pavement Condition Index (PCI), the International Roughness Index (IRI), overlay type, construction year, overlay thickness, joint spacing, traffic, and other construction and design-related data over a 30-year period. This study included more than 300 overlay projects based on more than 1,400 miles of roadway to evaluate the long-term performance of concrete overlays.
The main purpose of this study is to evaluate the long-term performance trend concrete overlays. The effects of overlay type and design features (thickness and joint spacing) on long-term performance were also identified. The effects of structural design alternatives on concrete overlay performance have been identified using the latest version of AASHTOWare Pavement ME Design (Version 2.3.1). Furthermore, to develop and practical ANN model for predicting concrete overlay performance based on historical performance database. In addition, investigating differences in behavior between shorter joint spacing and conventional joint spacing for optimize concrete overlays joint spacing size.
Long-term performance trends can be evaluated by studying PCI and IRI (two measures representative of pavement performance) changes during pavement service life. Performance data dating back to 1998 for all in-service Iowa concrete overlays constructed over the last 38 years were collected and evaluated. To date, since concrete overlays do not reflect new technology, and concrete overlay design procedures still follow empirical methods, this study applied both mechanistic-empirical design software and machine-learning techniques (i.e. AASHTOWare Pavement ME Design (Version 2.3.1) and Artificial Neural Networks (ANN) model) to identify the effects of various design parameters and help in predicting concrete overlay service life. AASHTOWare Pavement ME Design (Version 2.3.1) is a powerful software package able to simulate alternative joint spacing design options on various types of concrete overlays, and it provides theoretical insights for developing recommendations for pavement design. An ANN model is a machine learning tool that has been successfully used in the field of pavement design and analysis. Compared with other statistical techniques, since the deterioration of pavement performance is a non-linear function, the ANN model has shown superior accuracy for pavement management systems. Four different groups (distress data, construction design data, traffic data, and climate data) of input variables were used to predict pavement performance in the ANN model. Non-destructive testing (NDT) is another method for identifying the effects of various design parameters in concrete overlay systems. In this study, ultrasonic low-frequency tomography (MIRA) proved effective in detecting whether a saw-cut was activated. By comparing joint activation results with slab length values and radius of relative stiffness ratio (L/ℓ), recommendations on joint spacing for Iowa concrete overlays were developed.
Results from a summary of long-term performance showed that concrete overlays can extend service life of existing pavement by at least 20 years. After a comprehensive review of concrete overlay performance data, the adequate and substandard performance data were identified, showing differences between adequate and substandard performance over a 10-year service life, and indicating that improving construction quality to eliminate premature failure can also increase concrete overlay service life. In additional, compared to historical performance-related data, the Pavement ME Design software results is conservative in predicting concrete overlay service life. On the other hands, according to the concrete overlays prediction models results, the ANN model resulted in a root mean squared error (RMSE) of less than 10% of the range of IRI values, indicating that the ANN model was successful in predicting Iowa concrete overlay performance. Compared with MIRA evaluate rates of joint activation results and slab length values and radius of relative stiffness ratio (L/ℓ), joint spacing should be based on L/ℓ value between 4 and 7.