In addressing the issues of low efficiency in bridge maintenance decision making, the inaccurate estimation of maintenance costs, and the lack of specificity in decision making regarding maintenance measures for specific defects, this study utilizes data from regular bridge inspections. It employs a two-parameter Weibull distribution to model the duration variables of the states of bridge elements, thereby enabling the prediction of the duration time of bridge elements in various states. Referring to existing bridge maintenance and repair regulations, the estimation process of maintenance costs is streamlined. Taking into account the specific types and development state of bridge defects, as well as considering the adequacy of maintenance and the restorative effects of maintenance measures, an intelligent agent sequential decision-making model for bridge maintenance decisions is established. The model utilizes dynamic programming algorithms to determine the optimal maintenance and repair measures for elements in various states. The decision results are precise, all the way down to the specific bridge elements and maintenance measures for individual defects. In using the case of the regular inspection data of 222 bridges along a highway loop, this study further validates the effectiveness of the proposed research methods. By constructing an intelligent agent sequential decision-making model for bridge element maintenance, the optimal maintenance measures for 21 bridge elements in different states are obtained, thereby significantly enhancing the efficiency of actual bridge maintenance and the practicality of decision results.

Bridge health assessment and deterioration prediction are essential for traffic safety and maintenance planning. However, conventional bridge evaluation still relies heavily on expert judgment and heuristic rules, limiting objectivity and long-term forecasting capability. This study proposes a physics-constrained multi-task learning–semi-Markov framework for bridge condition assessment, multi-year deterioration forecasting, and risk-aware maintenance prioritization. Guided by the Highway Bridge Technical Condition Rating Code, the proposed model jointly predicts health grade, defect type, and severity from inspection item/defect entry records, improving assessment robustness through cross-task information sharing. The predicted health states are then incorporated into a physics-constrained semi-Markov model to forecast bridge deterioration over a three-year horizon, and the current and predicted states are further used for maintenance prioritization. Experiments on real bridge inspection data show that the proposed model achieves test accuracies of 95.20%, 99.67%, and 99.72% for health grade, defect type, and severity, respectively, with a Cohen’s kappa of 0.872, while the proposed semi-Markov model achieves a three-year prediction accuracy of 95.4% and a weighted accuracy of 99.1%. Comparative and ablation studies further demonstrate the effectiveness of the proposed framework for bridge lifecycle management.

This paper proposes a Bayesian networks (BNs) method that combines Conditional Probability Table (CPT) compression with an improved Variable Elimination (VE) inference algorithm to improve the efficiency of connectivity reliability assessment for bridge networks. First, a Continuous-Time Markov Chain (CTMC) is used to model the deterioration process of individual bridges, and a series equivalence strategy merges multiple bridges within the same road segment into an "equivalent bridge", thereby obtaining the time-varying failure probability of each edge. Second, a two-layer BNs model with a component layer and a system layer is constructed, and a depth-first search (DFS) connectivity check is performed row by row on the edge state combinations in the CPT to generate the state column of the system-layer node. Third, a hybrid compression algorithm integrating run-length encoding (RLE) and phrase encoding (PE) is proposed to compress the state column into run or phrase tuples, generating a compression entry table, a companion dictionary, and row index lists. Finally, an improved VE inference algorithm is designed, in which mapping rules for variable elimination are defined based on the parity of the start row number and the tuple length, and variable elimination is performed directly on the compressed results. The method is verified through a 5-node 7-edge example and a regional highway bridge network in Shandong, China. Results show that, while maintaining accuracy, the proposed method has significant advantages over the pgmpy method in both inference time and memory usage.

Modal tracking and condition evaluation are two important procedures in modal parameters based bridge condition evaluation, yet they always face the risk of failure due to the spatial aliasing caused by the damage of the sensors and the variation from the noise corruption. This paper proposed an automated modal tracking method and a variation removal method to get precise condition evaluation results. An index defined as modal similarity criterion (MSC) is first proposed to distinguish the modes when spatial aliasing occurs. Then, based on this index, a fully automated modal tracking method is established. An automated baseline mode setting strategy and an adaptive threshold updating strategy are proposed to make the modal tracking process fully automated and adapt to variation and spatial aliasing, even for closely spaced modes. The adaptive bandwidth kernel density estimation (ABKDE) is applied to the identified frequencies to reduce the variation from a statistical perspective to obtain more precise temperature–frequency regression and more precise condition evaluation results. The novelties and feasibilities of the proposed index and modal tracking method are validated through a theoretical analysis and real application to a benchmark bridge with closely spaced modes. The feasibility of the variation removal and condition evaluation method is validated through application to a suspension bridge with 15 years of continuous monitoring. The results indicate that the proposed index and modal tracking method are able to distinguish and automatically track the modes even when the mode shapes are not distinguishable (spatial aliasing occurs). The ABKDE is capable of reducing the variation in frequencies and helps to produce more precise regression and condition evaluation results.