Roadway Shallow Water Flow Modeling By Velocity Distribution

Manning's equation is a widely used method for determining flow discharge in a channel with unconfined gravity flow. The roughness value, n-value, is a critical factor for Manning equation to obtain the accurate amount of flow. A precise estimation of Manning "n" is difficult to obtain and varies by investigator justification and experience. Flow on a roadway is a type of open channel flow normally determined by Manning's equation. To ensure reliability and highway safety, the hydraulic geometry dimensions such as spread, depth and discharge must be accurately estimated. A Texas Department of Transportation Manning n-value research project collected data on surface roughness and estimate n-value of four different types of roadway sections; asphalt, treatment, smooth (worn) concrete and TxDOT's standard concrete surface. This research used full scale roadway sections with the varied flows and longitudinal and transverse slopes. This study focused on the estimating n-values for the entire roadway flow width. A velocity distribution method is used as an alternative method to study the flow characteristic and estimate n-values of each roadway cross-section. The velocity distribution equations use basic geometry data from the TxDOT research. The data for the four types of roadways from TxDOT Manning n-value research were used as input for the velocity distribution modeling. The percent accuracy of model simulation is estimated from a comparison of result, discharge and n-value, between the velocity method result and the original TxDOT research data. The modeling utilizes theoretical survey, statistical-analysis, numerical-analysis and flow methods to simulate roadway flow. Statistical analysis such as normality, data cleaning, and outlier detection, were used to improve results. The results indicate velocity distribution equations are potentially a good method for estimating discharge and n-values for a roadway section. It shows comparable discharge volumes and average n-values to the original TxDOT laboratory result with an acceptable percent of error.