An evaluation of alternative analysis methods for production load rating of culverts
Wood, Timothy Aaron
MetadataShow full item record
This study seeks to determine whether an increase in sophistication of production-oriented analytical models for culverts will produce more reliable load ratings. In this thesis, reliability means the coherence between predicted and observed events. Load-soil-structure interaction was the only portion of reliability investigated. Those aspects of reliability associated with method repeatability, actual behavior and acceptable conservatism were assumed to remain constant. Three analytical programs were evaluated: (1) CULV-5 â€“ a prepackaged, two-dimensional, culvert-specific structural frame program written by the Texas Department of Transportation (TxDOT), (2) RISA-2D with spring supports â€“ a two-dimensional, linear-elastic, structural frame program, and (3) RISA-2D with linear-elastic finite-elements (LEFE) â€“ a two-dimensional, linear-elastic, soil-structural model. These programs were used to calculate demand moments, shears and axial thrusts in order to load rate a sample of 100 culverts representing TxDOTâ€™s full population of reinforced concrete box culvert designs. Findings show that inventory load ratings from CULV-5 and RISA-2D with spring supports are similar for each culvert design analyzed. RISA-2D with linear-elastic finite-elements, the most sophisticated of the three programs, produced higher and ostensibly, more reliable, load ratings than CULV-5 and RISA-2D with springs for the case when the soil modulus was sufficiently high. For low soil modulus values, RISA-2D with LEFE produced lower load ratings than the less-sophisticated RISA-2D with springs and CULV-5.