Developing dry-weather flow and RDI/I models using regression analysis and unit hydrography

- By:

Courtesy of Innovyze

InfoWorks CS has been used to build an accurate sanitary sewer flow model for Jefferson Parish, Louisiana, despite there being gaps in the available input data. Reliable predictions of the dry-weather and rainfall-dependent flows were derived by using techniques including unit hydrography and regression analysis to determine the required parameters.

Jefferson Parish is close to New Orleans in the southern USA. The area has very flat topography and there is significant groundwater infiltration to the sewers as much of the area is at or close to sea level. The project team, consisting of Hartman Engineering, Inc., Digital Engineering & Imaging, Inc., and CDM, was asked to help in the better understanding of the sanitary sewer flows with a view to improving the infrastructure.

Flow parameters for input into InfoWorks were determined, refined and verified using SHAPE, a program that has been developed by CDM and which is due to be made available by the US Environmental Protection Agency. SHAPE was used to analyze the parameters for each storm and catchment area to produce median results for input into InfoWorks.    

A variety of information had to be obtained for each monitored area. The flow is made up of three components: a base volume that cycles over the course of the day and is made up of domestic, commercial and industrial wastewater; a constant volume of ground water infiltration; and rainfall-dependent infiltration and inflow (RDI/I) that rises to a peak in the aftermath of a storm and then tails away.

Several available methods for predicting RDI/I were considered before deciding on the triangular hydrograph. Research has shown that the triangular unit hydrograph yields results with a high degree of accuracy, even though it does not require very extensive or precise input data. This triangular graph provides a simplified but accurate representation of the relationship between a catchment’s flows and the time elapsed since the rainfall. The proportions of the triangle depend on the peak flow, the time taken for this to occur and the time for the flows to return to dry-weather values. The only data required concerns the contributing area, precipitation and corresponding sewer system flows.

There are flow monitors throughout the Jefferson Parish area and some work had already been carried out to establish which information was known to be reliable. RDI/I results from areas that had been checked could then be applied to other comparable catchments.

The dry weather flow for each of the monitored drainage areas was separated into its two components - the base wastewater flow and the groundwater infiltration. The best equation for predicting the flows was determined using regression analysis, which involved examining the relationship between the variables. Variables affecting the amount of groundwater infiltration included the population, the sewered area, the number of manholes and the length of pipe. The length of pipe proved to be the factor with the closest correlation to groundwater infiltration.

Key variables for the base wastewater flow were the area´s population and the land use. InfoWorks was supplied with details including the population of each subcatchment, its typical profile and the daily flows arising from residential, commercial and industrial users.

The methodology that was developed produced good information that proved easy to use within InfoWorks and enabled calibration of the model even in areas of limited data.

Customer comments

No comments were found for Developing dry-weather flow and RDI/I models using regression analysis and unit hydrography. Be the first to comment!