ABSTRACT
Bridge Weigh-In-Motion systems can be installed and serviced on highway bridges to provide measurements of truck traffic without interrupting traffic. This paper presents the application of Gaussian Process Modeling to BWIM. The Kriging metamodeling approach, allows for the strategic selection of training vehicles to generate a BWIM metamodel. The proposed approach extends the BWIM concept of influence area from response time histories to geometric input spaces from the strain response wave. The simulated dynamic response of simply supported bridges in the US Interstate Highway System were performed using Finite Element Modeling of moving loads, generating response waves. Geometric properties of the response waves are the input to train the Kriging metamodel of a predefined training truck group. The performance of the prediction of the GVW is evaluated with the ASTM WIM Standard.
