Statistical Analysis of Weigh-In-Motion Data for Bridge Design in Vermont.

摘要

This study investigates the suitability of the HL-93 live load model recommended by AASHTO LRFD Specifications for its use in the analysis and design of bridges in Vermont. The method of approach consists in performing a statistical analysis of weigh-in-motion (WIM) data collected between the years 2000-2012 at 12 stations across the state of Vermont. In total 36,754,819 individual WIM events were analyzed in this study. We compared the statistics of the lane moment and shear induced by the WIM data to the corresponding lane moment and shear induced by the HL-93 live load model. This analysis was performed on two types of very common bridge decks: (1) steel girders and concrete slabs and (2) concrete girders with concrete slabs. In all cases the decks were considered to be acting fully composite. We considered span lengths in the range of 5-60 meters (approximately 16-200 ft). The mains findings of this study are: (1) The probability that the lane moment and shear induced by the WIM data exceeds the corresponding values induced by the HL-93 model, decreases with span length. Averaged over all years considered in this study, the largest probability of exceedance was found to be approximately 1%. (2) For span lengths exceeding 10 m, the annual probability of failure induced by the WIM data analysis did not exceed the annualized AASHTO target probability of failure. This indicates that for typical bridge decks with span length exceeding 10 meters, the HL-93 live load model is adequate for its use in Vermont. (3) We propose that a more detailed study be carried out for short span structures such as culverts. Evidence from our study suggests that for very short spans (< 10 m), the HL-93 live load model might not be conservative.