A Condition Assessment Method for Bridges at Network Level Based on Multi-Source Information

Abstract:

This paper studied the commonality and personality of bridges at network level, summaried, and generalized the intrinsic rules according to their similar deterioration patterns. A network-level assessment and prediction method was proposed, for the first time, based on multi-source data. For large quantities of inspection reports, monitoring reports and bridge design drawings, the data integration and regulation process was conducted in order to extract key parameters and data sets applying to neural network of structures and components. A series of deterioration models were generated incorporating maintenance effects, based on which the condition assessment and deterioration prediction was realized for bridge populations. Based on the highway network in Shijiazhuang, an example was presented to verify the systematization and effectiveness of the proposed method for the purpose of bridge network management and decision-making.