High-Performance Computing–Enabled Hydraulic and Flood Modeling for Urban Water Supply and Drainage Systems

Abstract

This study examined the effectiveness of high-performance computing–enabled hydraulic and flood modeling for urban water supply and drainage systems through a quantitative, quasi-experimental simulation-based approach. The research compared conventional hydraulic modeling frameworks with HPC-enabled models across multiple urban catchments, rainfall intensities, and drainage conditions to evaluate differences in computational efficiency and predictive accuracy. A total of 120 simulation scenarios were analyzed using high-resolution geospatial and hydrometeorological datasets, ensuring consistency in input conditions across both computational environments. The findings demonstrated that HPC-enabled models achieved a substantial reduction in simulation runtime, with mean execution time decreasing from 148.2 minutes in conventional models to 53.6 minutes, representing an efficiency improvement of approximately 63.8%. In addition to computational gains, HPC-enabled models showed enhanced predictive performance, with flood depth error reduced from 0.21 m to 0.12 m and inundation accuracy improving from 78.4% to 89.7%. Statistical analysis confirmed that these differences were significant, with p-values below 0.05 and large effect sizes observed for runtime (d = 2.85) and moderate to large effect sizes for accuracy-related metrics. Subgroup analysis revealed that the advantages of HPC were more pronounced in high-intensity rainfall scenarios, where runtime reductions reached up to 68% and prediction accuracy improved by approximately 15%, particularly in large and complex urban catchments. Sensitivity analysis further indicated that HPC-enabled models exhibited greater stability, with output variability reduced by over 20% compared to conventional approaches. Visual and spatial analyses supported these findings, showing closer alignment of HPC-generated inundation maps with observed flood patterns. Overall, the results provided strong empirical evidence that HPC integration significantly enhanced both the efficiency and reliability of hydraulic modeling, supporting its application in large-scale urban flood analysis and decision-making processes within water management systems.