| Abstrakt: |
The accurate assessment of water performance in concrete structures requires a comprehensive understanding of material properties and structural connections. Various methodologies and formulas based on established codes have been developed to evaluate the water pressure on highway bridge piers. This research article focuses on developing data-driven models to calculate the water pressure on highway bridge piers and assess their effectiveness and applicability. Computational fluid dynamics analysis was used to obtain research data samples related to water pressure on highway bridge piers. Poisson regression models were created using this data to estimate the water pressure on the piers. Additionally, two conventional code-based equations were employed to compute the water pressures of highway bridges for comparison purposes. The accuracy and effectiveness of these data-driven models were verified by comparing the predicted water pressures from the models with the computational fluid dynamics results. Various statistical evaluation metrics, such as mean square error, mean absolute deviation, mean absolute percentage error, coefficient of determination, and root mean square error, were considered to assess the performance of these models. The results indicated that all of these models successfully forecasted the water pressure on highway bridge piers within reasonable limits. In comparison to the two standard code-based equations, the Poisson regression models built on actual data demonstrated greater precision and reliability. To minimise errors and provide a more robust mathematical equation, a univariate regression-based technique was employed to propose a unique relationship between water pressure and velocity, considering the dependency on velocity. |
Full text from SpringerLink