This study establishes an experimental database of shear failure for corroded RC columns, including 86 rectangular and 12 circular sections. An overall analysis method based on mean and standard deviation is conducted, and a combined analysis of key parameters (corrosion rate, shear span ratio, axial load ratio, effective stirrup characteristic value) is performed to systematically evaluate the predictive capabilities of 12 existing semi-empirical and semi-theoretical shear bearing capacity models. The results indicate that, from a holistic assessment perspective, for rectangular section corroded RC columns, the prediction accuracy of Wu Ranli’s formula is optimal (mean of 0.95, standard deviation of 0.35); For circular section corroded RC columns, Ngoc (2016) formula provides the best prediction (mean of 1.13, standard deviation of 0.18). The analysis of key parameters reveals significant differences in the predictive performance of each model. When the stirrup corrosion rate is in different ranges, Wu Ranli’s and Ngoc (2016) models are applicable for predicting the shear bearing capacity of rectangular section corroded RC columns for λ_v′ ≤ 0.1 and λ_v′ > 0.1, respectively. When the axial load ratio is in different ranges, Wu Ranli’s model is applicable for predicting the shear bearing capacity of rectangular section corroded RC columns in all cases.

With the continuous promotion of cross-regional water transfer projects, prestressed steel cylinder concrete pipe (Prestressed Concrete Cylinder Pipe, PCCP) has been widely used. However, due to the long-term erosion of Cl^- and $\text{S}{\text{O}}_4^{2-}$ in the soil, the external mortar expansion and cracking, internal steel wire corrosion, resulting in PCCP bursting have occurred repeatedly, so it is important to explore the different factors on the basic performance of the PCCP protective layer mortar, and erosion resistance of the law and to improve the anti-erosion performance of the PCCP protective layer mortar have important practical significance of the project. In this paper, the performance requirements and influencing factors of PCCP protective layer mortar are sorted out, and the improvement methods of erosion resistance of PCCP protective layer mortar are summarized. The results of the study show that the PCCP protective layer mortar mixing ratio, production process and maintenance conditions have a greater impact on its mechanical properties, water absorption and erosion resistance. In the PCCP mortar mixed with mineral admixture and mortar compacting agents and other additives can effectively improve the compactness of the mortar. Anti-corrosion coating on the surface of PCCP mortar protection layer can effectively prevent the intrusion of external erosion ions, thus improving the durability and service life of PCCP pipelines.

The accuracy of traditional prediction models for the shear bearing capacity of reinforced concrete(RC) columns is improved and existing experimental data is mined and utilized. Based on machine learning methods and the interpretable SHAP method, an artificial neural network model is established to predict the shear bearing capacity of reinforced concrete columns. Firstly, based on shear theory, 9 input features including longitudinal reinforcement ratio ρ_l, longitudinal reinforcement yield strength f_yl, and area shear reinforcement ratio ρ_sv are determined and their correlations are verified. With 441 sets of collected and organized experimental data on shear tests of reinforced concrete columns, the neural network model is compared with 5 machine learning models and 5 traditional semi-empirical and semi-theoretical formulas. The prediction results show that the neural network model established in this paper has better generalization and robustness, and its prediction results are more accurate(with R² reaching 0.99 and 0.92 on the training set and test set, respectively). In addition, the SHAP method is used to analyze the interpretability of the neural network model. The analysis results show that features such as section width b, axial compression force N, shear span ratio λ, effective section height h₀, and axial tensile strength f_t have significant influences on the shear performance of reinforced concrete columns. Moreover, the SHAP method also provides reasonably reliable analysis results for unknown samples. The study demonstrates that the data-driven and mechanism-driven neural network model and the SHAP interpretability method proposed in this paper can be applied to similar prediction problems of shear bearing capacity of reinforced concrete columns.

A shear performance experimental database containing 439 rectangular columns and 88 cylindrical columns was established, and the influence patterns of key parameters on the shear strength of RC columns were analyzed and discussed. The overall prediction results of 18 shear strength formulas are evaluated based on the database, and investigate the effect of the range difference of key parameters in the database on the prediction performance of 6 formulas, and the reasons are analyzed using the conclusions of the parameter effects. The results show that among the standardized formulas, the Canadian and Chinese formulas are the best predictors for rectangular and cylindrical columns, respectively; and among the scholarly formulas, Biskinis's and Sezen's formulas are the best predictors for rectangular and cylindrical columns, respectively.

Reinforced Concrete(RC) is susceptible to brittle shear failure, and diagonal cracking occurs and develops throughout the process, and cracking shear is one of the key indicators. In this paper, a database of 276 shear tests with cracking load information of RC deep beams was established, and the machine learning XGBoost ensemble algorithm was used to predict their cracking strength, and five statistical indicators were used to evaluate the prediction performance of the machine learning model. The prediction results of the proposed machine learning model were compared with five semi-empirical and semi-theoretical formulations according to the case of the deep beam without web reinforcement and with web reinforcement, indicating that the prediction accuracy of the proposed prediction model is high and the dispersion is small. The R² is 91%, the mean value of the ratio of predicted to tested values is 0.99, and the standard deviation is 0.27. In addition, SHAP(SHapley Additive exPlanations) interpretable method was used to globally and locally interpret the prediction results of the machine learning model, and the ranking of feature importance from important to general is: loading slab width, section height, and concrete compressive strength. The results show that the proposed model and interpretable method are consistent with the mechanism.