Reinforced concrete (RC) is the result of a combination of steel reinforcing rods (which have high tensile) and concrete (which has high compressive strength). Additionally, the prediction of long-term deformations of RC flexural structures and the magnitude of the influence of the relevant material and geometric parameters are important for evaluating their serviceability and safety throughout their life cycles. Empirical methods for predicting the long-term deformation of RC structures are limited due to the difficulty of considering all the influencing factors. In this study, four popular surrogate models, i.e., polynomial chaos expansion (PCE), support vector regression (SVR), Kriging, and radial basis function (RBF), are used to predict the long-term deformation of RC structures. The surrogate models were developed and evaluated using RC simply supported beam examples, and experimental datasets were collected for comparison with common machine learning models (back propagation neural network (BP), multilayer perceptron (MLP), decision tree (DT) and linear regression (LR)). The models were tested using the statistical metrics ${\mathrm{R}}^2$, $\mathrm{R}\mathrm{A}\mathrm{A}\mathrm{E}$, $\mathrm{R}\mathrm{M}\mathrm{A}\mathrm{E}$, $\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{E}, \mathrm{V}\mathrm{A}\mathrm{F}, \mathrm{P}\mathrm{I}, {\mathrm{A}}_{10-\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{x}} \mathrm{a}\mathrm{n}\mathrm{d} {\mathrm{U}}_{95}$. The results show that all four proposed models can effectively predict the deformation of RC structures, with PCE and SVR having the best accuracy, followed by the Kriging model and RBF. Moreover, the prediction accuracy of the surrogate model is much lower than that of the empirical method and the machine learning model in terms of the RMSE. Furthermore, a global sensitivity analysis of the material and geometric parameters affecting structural deflection using PCE is proposed. It was found that the geometric parameters are more influential than the material parameters. Additionally, there is a coupling effect between material and geometric parameters that works together to influence the long-term deflection of RC structures.

In order to reveal the non-uniform distribution of grain size in thick direction for engineering heavy plate, microstructure of 40 mm-thick Q345 steel was observed and measured under different short-term high temperature environments formed by fire. Moreover, the influence of the short-term high temperature environment was revealed on the distribution of ferrite grain size in the Q345 steel. Under different fire service environments, there was a log-normal distribution relationship between the distribution parameter Nf (number of ferrite grains) and df (average grain diameter), as well as ρAf (area fraction density) and df, at different positions along the thickness direction. However, the statistical results are greatly affected by the length of the statistical interval. When df is about 4 to 6 times the length of the statistical interval, the statistical accuracy is higher. By using nonlinear fitting method, multiple non-uniform distribution empirical models including Nf-df empirical formulas and ρAf-df empirical formulas were established at different positions along thick direction under various fire environments. Furthermore, the interrelationships between fire temperature T and Nf , T and ρAf , fire duration t and Nf , t and ρAf were revealed, respectively.

To reveal the distribution characteristics of the non-uniform temperature field of structural steel under fire conditions, Q355B steel for grain storage was taken as the research object. A multi-channel temperature detection instrument was used to monitor the temperature distribution of Q355B test plate in real time under gas fire, and the distribution characteristics of the non-uniform temperature field of the steel were studied under different heat inputs. The results showed that the temperature changes within 600 mm from the flame point were significant, while the temperature changes beyond 600 mm were not sensitive. The peak temperature Tmax and the heating rate vr at the fire source points of each test plate increased linearly with the increase of heat input Q at the fire source. With the increase of distance d, the variation trend of the time tmax for each test plate to reach the peak temperature in the rolling direction, was the same. As d increased, tmax gradually increased and tended to stabilize. The average cooling rates vd1 and vd2 of each test plate decreased with the increase of d. The temperature distribution patterns of each test plate were the same. The Tmax-d curve of each test plate followed the Boltzmann function distribution, the Q-∆T followed a linear relationship, and the Q-d0 followed an exponential function relationship. By combining the three, a non-uniform distribution model of Tmax-Q temperature field under gas fire was obtained.