Technical Overview

CalME has been developed by Caltrans in collaboration with the UCPRC to enable Caltrans to design flexible pavements in California using a mechanistic-empirical method.  CalME shares the general framework with other M-E design methods but differs from them in many important ways in the implementation. Below is a list of the key differences:

•CalME uses an incremental-recursive performance prediction process in which the predicted damages causes predicted pavement responses to increase over time. This allows the calibration of damage models using data from accelerated pavement testing, in which the pavement responses (such as deflections) increases as the test section is damaged by trafficking.

•CalME introduces the concepts of within-project variability and between-project variability to clearly distinguish and account for uncertainties from different sources. This allows for proper considerations of performance related construction specifications, which essentially is a way of reducing uncertainties in performance of as-built materials.

•CalME introduces a new framework for calibrations. The damage models are calibrated using well-controlled accelerated pavement testing data and the transfer functions are calibrated using field observations throughout the whole California highway network over time.

•CalME includes a standard materials library that allows addition of new materials whenever laboratory test data are available. This allows quick introduction of innovative materials.

Also, having its own M-E program allows the Caltrans/UCPRC team to optimize some of the key components to make the program run faster, which in turn make the design process much shorter. Here are some examples of these optimizations:

•A built-in temperature profile solver based on pre-calculated pavement surface temperature history.

•A built-in multi-layer elastic theory analysis program that is highly optimized to run millions of analyses on a structure that has layer stiffness changes but no layer thickness change.

With these optimizations, CalME typically finishes estimation of the performance of a given design in about one minute.

The next subsections explain in more details the  incremental-recursive procedure, how to account for uncertainties and how all the pieces fit together.