DETENTION DRAIN DESIGN WITH DYNAMIC ROUTING

Objective and methodology #

In this section, we will examine a Storm Water Management Facilities Report prepared by one of our users, Alpha Consultant (Ir Kong Chee Vui). The report includes a hydraulic assessment of a detention drain system designed to regulate post-development stormwater discharge from the project site before it enters the existing downstream drainage network, thereby minimizing the risk of flooding.

The detention drain layout presented in the report will be reconstructed in MiTS 3 to establish a “calibrated model”. Using the Dynamic Wave Routing method, the simulation will be carried out to evaluate and compare the resulting discharge flow (Q), velocity (V), and water depth with the corresponding values from the original analysis. Later on this calibrated model will be used to compare the effect of CCF, and the adaptation of MSMA 2 design for detention drain. We will draw certain important conclusions from it.

File Sample can be found below:

1. MiTS 3 Dynamic Routing (Analysis Mode)
2. MSMA 2 Approach
3. MiTS 2/ MiTS 3 Steady State Approach (Design Mode)

Site information #

Detention drain is selected to be a storm water detention facility instead of on-site detention (OSD) or detention pond, since the catchment area is less than 0.1 ha, which is more practical to be constructed on a smaller project area.

Alpha Consultant’s Hydraulic Report Overview /  Methodology #

The hydraulic analysis was carried out using the Laurenson Non-Linear Reservoir Method, a rainfall–runoff routing approach embedded in the XPSWMM hydrologic and hydraulic modeling software. This method simulates the transformation of rainfall into runoff by representing each sub-catchment as a series of non-linear storage reservoirs. The method internally applies a storage delay parameter rather than explicit depression storage, allowing realistic simulation of time-varying flow from both pervious and impervious surfaces.

Hydrological parameters in this hydraulic report were derived from the Manual Saliran Mesra Alam Malaysia (MSMA), including infiltration losses modeled by the Horton Infiltration Model, with design rainfall intensities based on Kota Kinabalu IDF curves for 5-year (minor) and 50-year (major) ARI storms.

Alpha Consultant’s Hydraulic Modelling #

The generated hydrographs from each sub-catchment were routed through a detailed network of reinforced concrete trapezoidal drains, modeled in XPSWMM’s dynamic flow environment. The system included a detention drain and a discharge control sump equipped with orifice and weir structures to restrict post-development peak discharge (Qpost) to levels equal to or lower than pre-development flow (Qpre).

Input Data from Alpha Kong’s Hydraulic Report #

From the hydraulic report, the relevant details are summarized as follows:

Note:
Under the steady state condition (Q = CiA/360), the system assumes there is no storage or attenuation within the drain. In other words, any flow entering the drain will exit at the same rate (Qinflow = Qoutflow).

MiTS 3 Dynamic Routing Comparison #

The requirement of MSMA 2 is that the Qpost at the last drain before discharging into another drain/water body  must be less than Qpre. The last drain of both report and MiTS 3 model is compared below:

Design Storm ARI: 5years #

Storm Duration: 50min
Drain name: LOut01 (Last Drain)

Design Storm ARI: 50years #

Storm Duration: 50min
Drain Name: LOut01 (Last Drain)

Discussion #

The results don’t fully agree between Alpha Kong’s report and ours is to be expected because he is using MSMA 1 and we are using MSMA 2 hydrological data. Thus the purpose of comparing MiTS StormSync analysis with the Alpha Kong’s report is to calibrate our results, so that we can ensure that the modelling in MiTS is correct (e.g.: with orifice element and drain size properly set). This will set the stage for the more important comparison later on, namely the comparison with the Steady State analysis for drain, and MSMA module detention drain calculation in MiTS 2.

Context of Different Routing Methods #

In hydraulic modelling, the selection of routing methods, (e.g.: Laurenson, Kinematic Wave, Dynamic Wave or etc) typically has only a minor effect on the final results (inflow, outflow, water depth etc). The key nuances lie primarily in the hydrology, where the rainfall is transformed into runoff. The hydrological component governs the inflow and ultimately exerts the greatest influence on the model’s outcome/ result.

In detention pond analysis, two main processes control the results; the hydrology and the hydraulic routing.

The hydrology defines how much water flows into the pond based on rainfall data and catchment characteristics. In this study, the inflow hydrograph is generated using the MSMA 2 methodology with Climate Change Factor (CCF), which generally produces higher and longer runoff peaks compared to the older MSMA 1 method due to updated rainfall intensity data.

The hydraulic routing method, on the other hand, determines how water moves through and out of the conveyance and storage facility. The Laurenson and Dynamic Wave methods both simulate water level changes over time, but Dynamic Wave applies a more detailed hydraulic approach that can account for effects such as backwater, surcharge, or tailwater conditions.

However, in many urban stormwater cases, the difference between Laurenson and Dynamic Wave routing is relatively small compared to the effect of using different hydrology inputs. Therefore, the major variation in pond storage results between this study and the reference report is mainly due to the updated hydrology (MSMA 2 + CCF) rather than the routing method itself.

Result/ Discussions #

We first establish that the Dynamic Routing simulation in MiTS 3 produces results consistent with Alpha Consultant’s hydraulic report, with only minor deviations attributable to modelling method differences. The detention drain provides adequate flow attenuation and capacity under both minor and major storm conditions, validating its hydraulic performance and confirming compliance with design criteria. This is just another validation of our software. We have also performed validations of MiTS 3 StormSync hydraulic analysis with other hydraulic software, such as EPA SWMM. You are welcome to browse through benchmarks on our website here.

More importantly, we compare against the common practice of doing detention drain, either via modified MSMA 2 table method (chapter 7) or just peak flow static analysis of drain (CiA/360). We found that dynamic routing is more accurate and also, more economical in terms of detention drain design, by a very, very wide margin.

Gotchas! #

The significance of Routing Steps parameter #

When running Analysis Mode, one might question why the graph output is not how you expected it to be? Sometimes you see sharp spikes, sudden drop or even weird oscillations that don’t make physical sense. You might assume that the graph should depict a rising water level or flow rate corresponding to the rainfall input, but it turns out that is not how it simulates in the graph. 

Observe these 2 hydrographs below. Why is the trend line different when we select the same drain mark and storm duration? You might get an absurd hydrograph like the first image, in which you might question the validity of the analysis.

 

Why does this happen? #

This is actually related to a numerical solver engine; a computational core that carries out all the calculations behind the flow simulation. When you perform any flow routing method, the numerical solver engine calculates how water moves from one point to another over time. The physics equations involved are time series equations, involving the second order differential equations with time. So to solve them on a computer, one would need to discretize them, and develop them into time difference equations. Then we select appropriate time steps to run the equations, one time step at a time to trace out the evolution of the fluid in terms of its characteristics like velocity, flow, water depth and so on. 

The routing step (e.g. 60s or 30s) that tells the solver how often to update the calculation, is a very prominent parameter here.

You may think of the routing steps like FPS (frame per second) in video, but for water analysis. Here’s the analogy:

In a video, FPS means how many frames (images) are shown per second.

1. Higher FPS (e.g. 60 fps) → smoother, more accurate motion.
2. Lower FPS (e.g. 10 fps) → jumpy, less accurate, you miss some movements

 

In a drainage simulation, the routing step is like the time between frames.

1. A bigger step (60s) means the solver engine only checks what’s happening every 60 seconds.
2. A smaller step (30s) means it performs calculations every 30 seconds, twice as often.

 

Because the simulations are too hard to solve using one big formula, the solver engine breaks them into many small pieces and solves them one by one, step by step. It then combines all those small answers to get the full result. In other words, smaller routing steps produce more “frames” of the simulation, capturing more detail of how the flow changes over time and resulting in a smoother, more accurate hydrograph.

How to calibrate/ and check the validity of the graph? #

You can adjust the routing steps parameter to a smaller slice of seconds and compare the output graph.

But of course, the drawback of using a smaller routing step will increase computational time when running the simulation, as the solver performs more calculations during the analysis!

Now, can you see the difference when we reduce the routing steps? 

So, what is the correct time step we can use? There is no universal formula. But I have a rule of thumb. First you try to use a bigger time step to get a feel on how the order of magnitude of the result. Then you rerun it using a smaller time step to ensure that the result still looks about the same. The moment you find that a smaller time step doesn’t yield big changes in the time series result, that means that your time step is already good enough.

And of course, if you see spikes in the time series results, you should always rerun it with a smaller time step; most likely this is just an artifact of an inappropriate time step.

Significance of CCF inclusion in routing method #

What happens if you include the Climate Change Factor (CCF)  in your analysis? 

The CCF will be applied to the rainfall intensity to account for the potential increase in rainfall due to future climate change conditions. In this section, we will look into how CCF values affect the dynamic routing analysis.

Refer article here on how to toggle CCF inclusion in drainage and MSMA calculation.

Inflow Hydrograph #

Outflow Hydrograph #

Water depth #

Table comparison #

CCF value = 1.20

Discussion/ Conclusion #

From the graph and the table above, it is observed that with the CCF inclusion in the hydraulic analysis, we can see that the inflow (rainfall intensity) increased about 20%. As a result, the inflow hydrograph will become higher, reflecting the greater volume of runoff expected from more intense rainfall events, but take note that the water depth is only marginally increased.

The inclusion of CCF will generally make your drain bigger, but as shown in this case, in dynamic routing, it doesn’t increase the capacity requirement linearly, like how steady state is. This is just another indication that instead state analysis will tend to overdesign the conveyance and the storage units. See more below

Why detention drain should be designed in the StormSync Module instead of MiTS 2 (Steady-State Analysis) or MSMA Module #

In this section, we will compare detention drain design using steady-state calculation (Q=CiA/360) and MSMA 2.0 guideline with respect to MiTS 3 Dynamic Wave routing analysis.

To match with the peak flow from Alpha Kong’s hydraulic report, certain parameters such as rainfall intensity and catchment area will be calibrated to have an apple to apple comparison.

From the image above, for detention drain length of approximately 48m, the drain depth and width required to store and convey the runoff is 2m and 1.2m respectively. 

This results in a highly conservative design that not only increases construction cost but also leads to an inefficient use of space and materials.

Also a limitation on the above model. Take note that this only applies on a single long section of the detention drain. What if the detention drain system is a Y shape drain system, namely 2 drains flow into a single drain, or other complicated topologies with multiple branch connections? The above methodology– which is just built by assuming that the detention drain is a single long drain will break down. 

We need to have a more versatile way of analyzing detention drains.

Another shortcoming of the above MSMA 2 detention drain methodology is that we are designing detention drain purely as a storage unit, but however, detention drain shall function as both the storage unit and conveyance, which means that the methodology is not considering the actual function of detention drain adequately.

MiTS 3 Analysis (Dynamic Wave Routing) #

Result/ Discussion #

Observe the steady state calculation. Although the peak flow is lower than the drain’s capacity for the selected size, the post-development peak flow (Qpost) still exceeds the pre-development flow (Qpre) of the catchment area (0.063 > 0.042 m³/s). In certain cases, the peak flow can exceed the drain capacity (Qprov < Qreq), even when the drain size is already relatively large. If a larger drain size is proposed, the clients may require justification for the selection. 

Therefore, the use of steady flow calculation is not practical for designing a detention facility, since it assumes no storage system and the maximum peak flow discharged through each conveyance at a single point in time, which is unrealistic under actual hydraulic conditions. 

Now look at the MSMA design. Notice that the drain size required to abide with MSMA guideline is excessively large for a detention drain of 48m long. In practice, no engineer would typically propose such an oversized drain for a small project area. Since the orifice diameter of 100mm will discharge 0.42m3/s, we can’t opt for a bigger orifice diameter because it will eventually increase the flow discharge and will exceed Qpre,50 = 0.042m3/s. We also cannot opt for a smaller drain size than 1200 x 2000 because it will cause the drain to be overflowed under this MSMA design. This is about the most optimum design we can get under such an approach.

 However, when dynamic wave routing is used for hydraulic analysis, the peak detention drain outflow (Qpost) can be reduced ( see column 3), since it accounts for both storage and conveyance units in one drainage system. This enables the use of smaller and more efficient drain sizes compared to the previous two methods.

Conclusion #

First, an existing report from a user is used to help calibrate our detention drain model in MiTS 3 using dynamic routing analysis. Then, we use the model to compare the dynamic routing analysis in detention drain design with the conventional drain steady state design and detention drain design according to the philosophy of MSMA 2 storage design. We also compare the effect of CCF and without the CCF.

We found that the drain outflow doesn’t increase linearly with the increase in peak inflow due to the nature of routing. Unlike the drain steady state design. This means that the effect of CCF is less in dynamic routing than the steady state design.

We also found that designing detention drains directly using the MSMA Module often leads to overly conservative results. This is because the MSMA method does not properly capture the combination of conveyance and storage characteristics, hence resulting in overdesign. Similarly, detention drains should not be designed using simple static drain formulas, as these methods also neglect the effects of storage and flow attenuation.

In contrast, performing detention drain analysis in the StormSync Module (Dynamic Routing) of MiTS 3 provides advantages and more realistic simulation. The dynamic routing method considers the changing inflow, outflow, and storage conditions over time, allowing the drain to be optimized based on actual hydraulic performance rather than a single peak flow value. Not only that, it’s robust – applicable on any configuration involving conveyance and storage units in any combination under any situation. Overall, dynamic routing produces more reliable and defensible results that reflect real-world conditions and support better engineering judgement in stormwater design.