The pH values of pore solutions are of great significance for the durability of concrete, but the influencing factors and mechanisms of geopolymer pore solutions are still unclear and the composition of raw material elements has a great influence on the geological polymerization behavior of geopolymers. Therefore, we prepared geopolymers with different Al/Na and Si/Na molar ratios using metakaolin, and the pH and compressive strength values of the pore solutions were determined using solid–liquid extraction. Finally, the influencing mechanisms of sodium silica on the alkalinity and geological polymerization behavior of geopolymer pore solutions were also analyzed. The results showed that the pH values of the pore solutions decreased with an increase in the Al/Na ratio and increased with an increase in the Si/Na ratio. The compressive strength of the geopolymers first increased and then decreased with an increase in the Al/Na ratio and decreased with an increase in the Si/Na ratio. The exothermic rates of the geopolymers first increased and then slowed down with an increase in the Al/Na ratio, indicating that the reaction levels first increased and then decreased with an increase in the Al/Na ratio. The exothermic rates of the geopolymers gradually slowed down with an increase in the Si/Na ratio, indicating that an increase in the Si/Na ratio reduced the reaction levels. In addition, the results obtained from SEM, MIP, XRD and other test methods were consistent with the pH change laws of geopolymer pore solutions, i.e., the higher the reaction level, the denser the microstructure and the smaller the porosity, whereas the larger the pore size, the smaller the pH value of the pore solution.
To ensure the safety and durability of concrete structures, timely detection and classification of concrete cracks using a low-cost and high-efficiency method is necessary. In this study, a concrete surface crack damage detection method based on the ResNet-18 residual network was developed. This method was implemented by training a model with images to extract the cracks, where the image processing algorithms and deep learning were combined. The results show that the computational accuracy can meet the requirements by utilizing the established image dataset and appropriate model hyperparameters. The trained model had high recognition accuracy when the 256 × 256 resolution images were adopted, and the worst accuracy of crack recognition in the test set was over 90%. The average accuracy in the test set was 91.3% when considering environmental interference generated by processing the images with a brightness adjustment, salt-and-pepper noise, and localized interference. Then, it was demonstrated that the present model possesses good robustness for crack identification in different environments. The average recognition accuracy when dealing with images of a real bridge, which are outside the training dataset, was 99.7%. The residual network model developed in this study has the advantages of low cost, high efficiency, and practicality compared to traditional detection methods. Compared to the existing deep learning methods, the model created in this study requires less computational resources and storage space, and shows a faster training speed and higher accuracy.