Swale Filter Drain System: Inflow-discharge relation

by Erik Donkers

MSc thesis report

In urban water management for a long time it was the practice to discharge the storm water in a centralized manner. For instance the peak loads of the sewer systems and the wastewater treatment plants in times of heavy rainfall and the needs of water retention in urban areas nowadays ask for a decentralized discharge of the precipitation. The runoff is not discharged as fast as possible, but it is retained and (partly) discharged to the surrounding soil and the surface water. For this purpose among others there is made use of swale filter drain systems (SFDSs).

The principle of a SFDS is as follows: The precipitation in the urban area is transported to the SFDS by means of a storm water sewer system or a gully. The water ends up in a deepened grassplot and infiltrates into a soil improvement underneath the grass. Within the soil improvement a drain is installed which leads to the surface water or a storm water sewer system nearby. The water is partly infiltrated into the original surrounding soil and partly discharged by the drain.

SFDSs have both qualitative and quantitative features. The research described in this thesis focuses on the quantitative features. From this perspective the SFDS provides retention, infiltration and delayed discharge of the runoff through the drain. This research has focused on the quantitative relation between the inflow to the SFDS and the discharge through the drain.

The research question for this study is: ‘What is the inflow – discharge relation of Swale Filter Drain Systems, with respect to the total discharge reduction, the discharge peak reduction and the peak delay?’.

The purpose of this research is to determine the characteristics that have effect to the inflow – discharge relation. How these characteristics affect the relation and if the knowledge about this is usable for predicting the relation in practice. For this purpose measurements are carried out and a model is used.

By monitoring a SFDS in Leidsche Rijn, a district of Utrecht, the actual working of the system in practice is examined. With these measurements the effect of different inflow characteristics is determined. The studied inflow characteristics are the intensity, the peak inflow, the duration of the inflow and the total volume of the inflow. Hydrus 2D is used to make a two dimensional numerical model of a SFDS. This model gives the opportunity to investigate the effect of the change of some characteristics of the SFDS and its surrounding on the inflow – discharge relation. The studied characteristics are the drainage depth, the groundwater level, the shape of the trench (the soil improvements below the grass) and the width of the trench.

The main findings of the effects of the characteristics to the inflow – discharge relation are presented here point by point. There are more outflow characteristics analyzed which can be found in the conclusions as well.

• No clear inflow – discharge relation is found. This is caused by the different initial conditions for the measured events. Furthermore the distribution of precipitation events differ. Therefore during the analyses a rough distinction is made between all the measurements and those with a shallow initial water level in the trench. A distinction between short duration and long duration events is made as well. The short events have a duration between 77 minutes and 385 minutes. The long events have a duration of around 1000 minutes.
• For all measured events the peak delay has a range between 10 minutes and 108 minutes. The peak reduction has a range between 40% and 100%. The total volume reduction is between -8% and 100%.
• The events with a shallow initial water level in the trench give some different values. The peak delay is between 10 minutes and 41 minutes. The peak reduction is between 40% and 89%. The total volume reduction values are between -8% and 89%.
• The difference between the start of the inflow and the start of the outflow is between 1 minute and 100 minutes for all measurements. For events with a shallow initial water level in the trench the values are between 1 minute and 28 minutes.
• There is found a minor effect for some studied characteristics in case the peak delay is the most important outflow characteristic. Only for short events there is an effect of the duration of the event and the total inflow volume of the event. This means that during the design period this may be taken into account. When a certain peak delay is found during modeling this peak delay is different with for instance another duration of the event. Although the effect is very small.
• The inflow characteristics have no or not much effect on the peak reduction. When a large peak reduction is purposed one may use a shallow drainage level. Furthermore it is recommended to use a t-shaped trench instead of a rectangular shaped trench. The groundwater level has a certain effect on the peak reduction as well. When the groundwater level is deeper the peak reduction is larger. This may be important for calculations during the year. The groundwater level may change during the year.
• When the total volume reduction has the priority during the design phase the intensity and the peak inflow of the event have an effect to it. A larger intensity and peak inflow causes a smaller total volume reduction. This means that when the design is modeled with a certain event the results are different when the design is modeled with an event with a larger intensity. This needs to be taken into account during the design phase to prevent for overestimation of the volume reduction by the system. The total inflow volume has a minor effect to the total volume reduction. During the design phase it still may be taken into account.
• In case a large total volume reduction is purposed a shallow drainage level may be used. The groundwater level has an effect as well. This means that the fluctuation of the groundwater level during the year needs to be taken into account. By enlarging the volume of the soil improvement the total volume reduction is increased as well.
• For all outflow characteristics it must be noted that the emptying time of an SFDS for the native soils with a small hydraulic conductivity is relatively large. Because of this it takes more time after a precipitation event to get back to the maximum storage capacity of the SFDS. Because of this simulations with a sequence of events may be used. This will simulate the designed SFDS with different initial conditions caused by the former events.