Introduction To CalME
Historical Background on Mechanistic-Empirical Design in California and Calibration of CalME
The University of California Pavement Research Center (UCPRC) has been supporting the Caltrans effort to implement ME pavement design by working on a series of tasks since 2000. This work began under the technical guidance of the Pavement Standards Team, with the Division of Design in the lead. One of the tasks was to develop and calibrate ME flexible pavement design models, which were used for the design of the first Long Life Asphalt project on I-710 in Long Beach, constructed in 2003.
In 2005, the California Department of Transportation (Caltrans) approved an issue memo titled “Adoption of Mechanistic-Empirical (ME) Pavement Design Method,” which calls for the adoption of ME pavement design methodology to replace existing pavement design methods that have been in place since the early 1960s. Work on ME design for California pavements has continued under the direction of the Caltrans Office of Pavement, the successor the Pavement Standards Team.
The first step in a Mechanistic-Empirical (ME) pavement design or evaluation is to calculate pavement response - in terms of stresses, strains, and/or displacements - using a mathematical model. In the second step, the calculated response is used as a variable to predict structural damage (decrease in moduli and accumulation of permanent deformation). A third step then follows to estimate pavement distress based on the predicted damage. The first step is mechanistic, the second step has both mechanistic and empirical parts, while the third step is empirical. An introduction to M-E pavement design is included in a later part of this manual.
The first step must be reasonably correct. If the calculated response bears little resemblance to the pavement’s actual response, there is no point in trying to use the calculation to predict future damage to the pavement and later correlate the damage to pavement distress through empirical relationship. In other words, only if the calculated response is reasonably correct does it make sense to try to drive damage accumulation using the calculated pavement response.
The validation and calibration of the models in CalME was first performed using performance data from Heavy Vehicle Simulator (HVS) tests completed by the UCPRC between 1995 and 2004. The results of that work are documented in a report titled “Calibration of Incremental-Recursive Flexible Damage Models in CalME Using HVS Experiments", and focus on calibration of the strain and stress responses, damage models, and initial calibration of fatigue and reflective cracking, and rutting. The CalME damage, fatigue cracking, and rutting models were subsequently calibrated in 2006 using the materials, deflection, and condition survey data from the FHWA 1995-1996 WesTrack closed circuit track accelerated pavement testing experiment. New damage models for full-depth recycling were included in CalME in 2020, and damage models for cement stabilized base and subgrade were updated. The CalME empirical models relating asphalt fatigue and reflective cracking to predicted damages in the asphalt layers were recalibrated in 2021 using thousands of miles of condition survey, as-built and traffic data from 1978 to 2018 in the Caltrans pavement management system (PMS).
Report links:
-2006 WesTrack calibration report
-2021 Updates to CalME and Calibration of Cracking Models
CalME Versions 1 and 2
The first version of CalME, v1, was released in 2011, and the second version, v2, was released in 2014; these were both desktop applications written in Visual Basic. The desktop application was originally developed as a research tool that had a working user interface and workflow but it was very hard to maintain and enhance. The installation of the desktop application proved to be cumbersome to Caltrans since administrative privileges were required.
CalME Version 3
CalME v3 is a complete rewrite of the desktop version of CalME (v2). CalME v3 is a web-based application with a user interface written in ASP.NET to access a new, modern and easy-to-maintain compute engine, which uses Node.js to provide various web services and C++ to run the actual simulations to optimize performance. CalME v3 were tested against v2 to make sure the results are the same before incorporating various improvements in the models.
CalME v3 uses the same traffic data used by Caltrans' pavement management system, PaveM. This source of traffic data is more current than that used by v2 and it also uses traffic information from the Performance Measurement System (PeMS) in order to fine-tune traffic data. CalME v3 makes a suggestion on the traffic data (traffic volume and load spectrum) based on the location of the project on the highway system selected by the user. The user should compare this suggestion with the traffic data provided by the Caltrans traffic division. They should be close and the latter should be used, otherwise one should double-check if large discrepancy exists between the two.
CalME v3 also uses the same climate zone data used by PaveM and makes a suggestion for the best zone to use. Again, the user is able to select any climate zone.
CalME v3 also allows the import of backcalculation results generated by Caltrans' CalBack program via an export file. CalME v3 create a new project and a series of trials that represents the sections generated by CalBack.