Neural network algorithms for the correction of concrete slab stresses from linear elastic layered programs

Elastic layered programs (ELPs) are currently being used in mechanistic-based pavement design procedures for the analysis of jointed portland cement concrete pavements. Corrections must be made to stresses obtained from ELP solutions to account for the effects of finite slab size, load location on the slab, and load transfer efficiency of the joints. A preliminary artificial neural network (ANN) model is trained and used as a tool to predict the results of a finite-element analysis program for a standard pavement section. Under identical loading conditions, the trained neural network produces stresses within 1.2 percent of those obtained from finite-element analyses. The trained ANN model is found to be very effective for correcting ELP stresses, practically in the blink of an eye, with no requirements of complicated finite-element inputs. The preliminary ANN algorithm is currently being expanded to handle more general input conditions covering a wide range of slab sizes, slab thicknesses, subgrade supports, and loading conditions. Design curves created from these ANN algorithms will eventually enable pavement engineers to easily incorporate sophisticated state-of-the-art technology into routine practical design.