Analysis and Modeling of Flooding in Urban Drainag 2004
Analysis and Modeling of Flooding in Urban Drainag 2004
Analysis and Modeling of Flooding in Urban Drainag 2004
net/publication/222412347
CITATIONS READS
294 2,901
3 authors, including:
Some of the authors of this publication are also working on these related projects:
Sustainability Assessment for the Long-Term Transition of Urban Water Supply Systems View project
All content following this page was uploaded by Theo Schmitt on 16 November 2021.
Abstract
The European research project in the EUREKA framework, RisUrSim (S!2255) is presented. The project consortium
includes industrial mathematics and water engineering research institutes, municipal drainage works as well as an insurance
company. The overall objective has been the development of an integrated planning and management tool to allow cost
effective management for urban drainage systems. The paper outlines the regulatory background of European Standard EN 752
defining flood frequency as the one hydraulic performance criterion. The phenomenon of urban flooding caused by surcharged
sewer systems in urban drainage systems is analyzed leading to the necessity of dual drainage modeling. A detailed dual
drainage simulation model is described based upon hydraulic flow routing procedures for surface flow and pipe flow. Special
consideration is given to the interaction between surface and sewer flow in order to most accurately compute water levels above
ground as a basis for further assessment of possible damage costs. The model application is presented for small case study in
terms of data needs, model verification and first simulation results.
q 2004 Elsevier B.V. All rights reserved.
Keywords: Urban drainage; Flooding; Dual drainage modeling; Hydraulic surface flow simulation; Dynamic sewer flow routing
Table 1
Recommended design frequencies in EN 752
flooding caused by system surcharge will be described conditions may eventually lead to a rise in the water
afterwards. Its application and data need is demon- level above surface where water either escapes from
strated in a case study. The European Standard EN 752 the sewer system or prevents surface water from
‘External Drain and Sewer Systems’ applies to entering the sewer system. Fig. 1 describes different
drainage systems designed essentially for gravity stages of surcharge. Fig. 2 illustrates the phenomenon
flow. Hydraulic performance criteria established in of surface flooding.
EN 752 for urban drainage systems in terms of design
storm frequencies and design flooding frequencies 2.2. Analysis of flooding phenomena
differentiating between rural, residential and indus-
trial/commercial areas, and city centers are listed in Flooding in urban drainage systems as defined
Table 1. For larger developments and existing drainage above may occur at different stages of hydraulic
systems with complex hydraulic flow patterns (e.g. surcharge depending on the drainage system (separate
with loops, backwater effects, etc.), direct assessment or combined sewers), general drainage design charac-
of hydraulic performance by sewer flow simulation teristics as well as specific local constraints.
models is recommended thus checking flooding When private sewage drains are directly connected
frequencies in accordance with Table 1. Implicated to the public sewer system without backwater valves,
demands for a more detailed simulation approach and the possible effects of hydraulic surcharge depend on
an extended data base will be discussed below. the levels of the lowest sewage inlet inside the house
(basement), the sewer line and the water level during
surcharge, respectively. Whenever the water level in
2. Flooding in urban drainage systems the public sewer exceeds the level of gravity inlets in
the house below street level, flooding inside the house
2.1. Hydraulic surcharge and flooding will occur due to backwater effects. In such a case
flooding is possible without experiencing surface
Linking drainage system hydraulic performance flooding. In the same way, hydraulic surcharge in the
requirements directly with the frequency of flooding sewer system might produce flooding on private
demands a clear definition of flooding and a distinc- properties via storm drains, when their inlet level is
tion from the state—or different stages—of surcharge. below the water level of the surcharged storm or
According to EN 752 ‘flooding’ describes a ‘condition combined sewer.
where wastewater and/or surface water escapes from In both cases, the occurrence of flooding, being
or cannot enter a drain or sewer system and either linked directly to the level of inlets versus water level
remains on the surface or enters buildings’. (pressure height) in the sewer can be ‘easily’ predicted
Distinct from flooding, the term ‘surcharge’ is by hydrodynamic sewer flow simulations, assuming
defined as a ‘condition in which wastewater and/or the availability of physical data of the private drains
surface water is held under pressure within a gravity and the public sewer system.
drain or sewer system, but does not escape to the Distinct from the situations described above, the
surface to cause flooding’. Extended surcharge occurrence and possible effects of surface flooding
302 T.G. Schmitt et al. / Journal of Hydrology 299 (2004) 300–311
depend much more on local constraints and surface street flooding can limit or completely hinder
characteristics, e.g. street gradient, sidewalks and curb the functioning of traffic systems and has indirect
height. These characteristics, however, are much consequences such as loss of business and opportu-
more difficult described physically, and these data nity. The expected total damage—direct and indirect
are usually not available in practice. In addition, monetary damage costs as well as possible social
today’s simulation models are not fully adequate to consequences—is related to the physical properties of
simulate the relevant hydraulic phenomena associated the flood, i.e. the water level above ground level, the
with surface flooding and surface flow along distinct extend of flooding in terms of water volume escaping
flow paths. from or not being entering the drainage system, and
Due to these deficiencies, the German standard the duration of flooding. With sloped surfaces even
ATV-A 118 ‘Hydraulic design and simulation of the flow velocity on the surface might have an impact
drainage systems’ considers flood frequency to be on potential flood damage.
inappropriate for direct computational assessment
(ATV, 1999; Schmitt, 2001). The ‘surcharge
frequency’ is established as an additional criterion of
hydraulic performance defined as the rise of (maxi-
mum) water level at manholes up to ground level. This
borderline case of surcharge—the transition from
pressurized pipe flow to surface flooding—can be
accurately described by nowadays dynamic sewer
flow simulation models (Schmitt and Thomas, 2000).
3. Modeling urban flooding - single drainage areas (roofs, streets, parking lots,
yards etc.), where rainfall is transformed into
In regard of the distinct stages and processes of effective runoff depending on surface character-
surcharged sewer system and urban flooding as istics (slope, roughness, vegetation, paved/un-
described above, simulation models for flood risk paved surface area etc.);
analysis are required to accurately describe the - distinct surface drainage components, e.g. street
hydraulic phenomena of surcharged and flooded gutters, that lead surface runoff to the underground
sewer systems, particularly sewer system via inlets;
- surface areas, where surface flow might occur in
- the transition from free surface flow to pressure case of surface flooding (e.g. street surface)
flow in the sewer pipes - closed underground sewers forming the sewer
- the rise of water level above ground level with network (including manholes, control structures
water escaping from the sewer system and outlets).
- the occurrence of surface flow during surface
flooding, and The single areas are connected to the sewer
- the interaction between surface flow and pressur- systems via gutters and/or inlets followed by closed
ized sewer flow. pipes. In simulation models the single areas are
mostly comprised to sub-catchments that are linked to
The consideration of distinct surface flow and its distinct input elements of the sewer network,
interaction with sewer flow in surcharged sewer generally to the manholes being represented as system
systems is denoted as ‘dual drainage modeling’ with nodes. In general, the distinct surface drainage
flow components on the surface and underground, first components are not represented by runoff models.
described by Djordjevic et al. (1999) and illustrated The sum of all sub-catchments form the overall
in Fig. 3. catchment area.
For dual drainage simulation single areas need to
be further distinguished as follows:
3.1. System analysis in view of dual drainage
modeling. (a) in regard to their connection to the sewer system
as
In regard to dual drainage modeling, urban - areas linked completely via closed drains (e.g.
drainage systems comprise of roofs)
- areas linked via surface inlets and closed (uni-directional). The module RisoSurf computes
drains (e.g. parking lots) surface flow based upon a simplified representation
- areas on private sites draining to the street or of the shallow-water-equations using GIS-based sur-
side-walk surface face data, e.g. street area and slope (longitudinal and
- areas not connected to the sewer system (not lateral), gutters, culvert height, buildings and other
runoff-relevant) features relevant for surface flow patterns. Dynamic
(b) in regard to possible surface flow as sewer flow routing is applied for all underground
- areas not subject to flooding (no interaction drainage elements in the module HamokaRis. The
between surface and sewer flow; e.g. all roof model allows bi-directional exchange of flow volume
areas between surface flow module RisoSurf and sewer flow
- areas where surface flow occurs and is module HamokaRis at defined exchange nodes. The
simulated during flooding, or bi-directional exchange is realized by interpreting
- areas not to be considered in surface flow inlets to the sewer system as possible sinks or sources
simulation. in the mathematical model of both, surface and sewer
flow simulation.
the terms of inertia in the momentum equation, leads momentum equation and instantaneous water levels
to a simplified mathematical representation, expressed at the nodes at the end of the last time step. In the
as Eq. (1) next step of the dynamic flow routing procedure
The application of this detailed hydraulic method the flow volume is balanced at each node, taking
would be restricted to small areas only. Therefore, it into account inlets from house drains and all surface
only served as a benchmark for a further simplified inlets connected, as well as inflows and outflows
two-dimensional approach, where Manning’s from sewers connected at the nodes. The resulting
equation provides a closed-form expression for the change of volume is drawn to free water surface
two-dimensional velocity vector (u,v) that is used in ‘available’ at the node, thus producing a change of
the momentum Eq. (1) water level at the node. In order to improve
numerical stability, the two phases are applied in a
vh vuh vvh half-step–full-step procedure during each time step
C C Z Sp (1)
vt vx vy as described in Roesner et al. (1988) and Schmitt
(1986).
The underground sewer system is represented by a
t time variable,
network of nodes and conduits (sewer segment
x,y spatial variable,
between nodes). In contrast to conventional modeling,
h water level,
not only manholes but also street inlets and house
u depth averaged velocity in x direction,
drains are considered as extra nodes to fully achieve
v depth averaged velocity in y direction,
the connection of surface and underground drainage
Sp sink/source term as the exchange value with the
system at all locations where interaction between
pipe flow model,
surface and sewer flow and potentially flooding might
occur. This will be further discussed in context with
The RisoSurf approach will be described in greater
the case study below.
detail in Ettrich et al. (2004).
From a mathematical point of view, the new and
crucial point in this approach of hydraulic surface 3.2.5. Modeling interaction of surface and sewer flow
flow simulation is the coupling of the shallow water The simulation of the interaction between surface
equation model of surface flow with the dynamic and sewer flow is based upon the definition of
sewer flow model. Sink/source term Sp, being the exchange locations. Each runoff area is allocated
primary term of exchange with the pipe flow model, to one specified exchange location as illustrated in
requires particular consideration of numeric stab- Fig. 5. Here, all relevant information for surface and
ility. This crucial point of coupled hydraulic flow sewer flow simulation (instantaneous runoff, water
routing procedure is further discussed in the level, exchange volume) is available at the beginning
paragraph ‘Coupling modules RisoSurf and Hamo- of each time step for all simulation modules and is
kaRis’ below. renewed at the end of the time step in the following
way:
3.2.4. Dynamic sewer flow modeling (1) The hydrologic runoff model RisoReff only
Sewer flow is simulated applying fully dynamic supports uni-directional flow and is applied to
flow routing of unsteady, gradually varied flow and all areas not considered for surface flow.
solving Saint-Venant-Equations numerically in an Computed runoff from those ‘hydrologic areas’
explicit difference scheme. The explicit difference is passed to the single exchange location to
scheme is applied in variable time steps that are which the area is connected. The exchange
permanently adjusted to the COURANT-criterion, volume would be the runoff volume in the
guaranteeing numerical stability (Schmitt, 1986). according time step.
At each time step, the procedure of dynamic flow (2) Areas simulated with the hydrologic model
routing starts by computing flow values for each approach can be connected to the underground
conduit (sewer segment between nodes) based upon drainage system in two alternative ways:
306 T.G. Schmitt et al. / Journal of Hydrology 299 (2004) 300–311
data during rainfall events for model calibration depression storage and infiltration.
under surcharge conditions. This, however, has not
been successful as during the period of monitoring (1) Imperviously (2) Previously (3) Unpaved
not a single surcharge or even flooding event paved areas paved areas areas
occurred.
1.1Streets and 2.1 Streets 3.1 Green
other traffic areas and other traf- roofs
4.1. Data needs fic areas
1.2 Span-roofs 2.2 Yards, 3.2 Lawns,
Data needs for detailed simulation of urban (O10% grade) private park- garden area
flooding are related to a distinct representation of ing lots etc.
the runoff areas, the surface characteristics and 1.3Flat roofs 3.3 Garden
local constraints that possibly influence surface (except green area
flow patterns, and the underground sewer system roofs)
itself. 1.4Yards, private
parking lots etc.
4.1.1. Runoff areas
In order to most accurately describe runoff their
behavior, single runoff areas within the sub-catchment 4.1.2. Surface characteristics
have been classified and sized according as with A Digital Terrain Model (DTM) has been set up
distinct model parameters to quantify interception, in a high resolution in order to apply detailed
T.G. Schmitt et al. / Journal of Hydrology 299 (2004) 300–311 309
Fig. 9. Mathematical representation of the street surface in a triangular network in surface flow module RisoSurf.
hydraulic surface flow simulation module RisoSurf. - private house drains, represented as closed pipes
The DTM includes distinct levels of street cross- uni-directional input node
sections, location and height of manholes and street - street inlets, represented by a bi-directional
inlets, side-walks and street-curbs as well as the exchange location, followed by a closed pipe.
border-line between public (street, side-walk) and
private space (building sites). Fig. 9 shows the These inlet elements have been connected to the
mathematical representation of the street surface public sewer system either via existing adjacent
area in a triangular network for the hydraulic manholes or by introducing extra system nodes
surface flow module RisoSurf. Adjustments had to (‘virtual manholes’). This is done automatically by a
be made with regard to the representation of street- software routine depending on the location of the
curbs by a double polygon describing the location single inlets in relation to the sewer line.
and different heights of street gutter and side-walk.
For practical reasons it was decided to exclude
areas on private sites from surface flow simulation 4.2. Simulation results
in regard to the lack of detailed physical surface
data. Due to the fact that no surcharge or flooding event
could be monitored, the RisUrSim Software has been
4.1.3. Detailed sewer system applied to a variety of test scenarios using synthetic
The sewer network, normally represented by design storms. These applications have been done to
manholes nodes and sewers only, has been further verify the most crucial model features of hydraulic
detailed by including surface flow simulation and the interaction between
310 T.G. Schmitt et al. / Journal of Hydrology 299 (2004) 300–311
Fig. 10. Water level distribution as simulated for the test case system KL-Erzhuetten during a synthetical design storm after 15 and 25 min.
surface flow and sewer flow under surcharge and The simulation results of the real-case system
flooding conditions. Erzhuetten are shown the Fig. 10 in terms of water
Tests simulations have been done with single level distribution along the street surface at 15 and
modules using simple structures as input data too 25 min simulation time, respectively. In this system
complex systems with data from ‘real-world’ sys- the surface elevation decreases from north-east (right)
tems. Since the demand on data to achieve realistic to south-west (left) while the sewer-system flow
simulations results esp. to model the surface structure direction is oriented in the opposite direction. This has
is very high compared to ‘traditional’ hydrologic let to problems with flooding in this area in the past.
surface representation in sewer-system simulation The representation of simulated water levels in the
even the real-world tests can cover only small manholes and on the street surface illustrates the
regions. surface flow pattern from surcharged, flooded
T.G. Schmitt et al. / Journal of Hydrology 299 (2004) 300–311 311
manholes to street areas with lower surface levels (on and Trondheim as well as from the insurance
the left side of the graph). This proves that with the company Deutsche Rück. Special thanks are given
RisUrSim Software the surface flooding could be to the project partners in Norway (Prof. Wolfgang
reproduced realistically. Schilling, Dr Sveinung. Sægrov) and the research co-
workers Christoph Garth and Michael Hilden.
5. Conclusions