Abstract

Stormwater runoff from urban catchments is affected by the changing climate and rapid urban development. Intensity of rainstorms is expected to increase in Northern Europe, and sealing off surfaces reduces natural stormwater management. Both trends increase stormwater peak runoff volume that urban stormwater systems (UDS) have to tackle. Pipeline systems have typically limited capacity, therefore measures must be foreseen to reduce runoff from new developed areas to existing UDS in order to avoid surcharge. There are several solutions available to tackle this challenge, e.g. low impact development (LID), best management practices (BMP) or stormwater real time control measures (RTC). In our study, a new concept of a smart in-line storage system is developed and evaluated on the background of traditional in-line and off-line detention solutions. The system is operated by real time controlled actuators with an ability to predict rainfall dynamics. This solution does not need an advanced and expensive centralised control system; it is easy to implement and install. The concept has been successfully tested in a 12.5 ha urban development area in Tallinn, the Estonian capital. Our analysis results show a significant potential and economic feasibility in the reduction of peak flow from dense urban areas with limited free construction space.

INTRODUCTION

Changing climate and densification of built-up areas will have a considerable impact on urban areas. One of the effects of this trend on urban areas in Northern Europe is the increase of stormwater peak intensities during rain events (Madsen et al. 2014). Current urban drainage systems (UDS) are usually not designed to cope with such extremes. In these cases, intense rainfall will cause the system to become surcharged, which will consequently trigger pluvial floods. Rapid urbanisation that is disrupting the natural stormwater cycle is accelerating the problem even more. As a result of these trends, urban areas are considered highly vulnerable to climate change (Tapia et al. 2017).

There are usually limited financial resources available to enlarge the UDS to handle higher flow rates. Therefore, special attention has to be paid if new development districts are planned to connect to the existing system. There are several options available to alleviate the pressure on existing UDS and therefore reduce the risk of surcharge at downstream. According to Fletcher et al. (2015), these can be broadly divided into structural and non-structural measures, both of which are underpinned with mitigation of changes in flow regime and improvement of water quality. Low impact development (LID) is considered by many authors as one of the efficient structural methods, while best management practices (BMP) contribute to the non-structural category (Joksimovic & Alam 2014; Saraswat et al. 2016).

LID can also be characterised as a small scale stormwater treatment facility located near the source (Fletcher et al. 2015). These techniques are usually divided into two groups: (1) green solutions, e.g. above ground bioretention systems and (2) grey solutions, e.g. underground concrete structures. Green solutions, also referred to as sustainable urban design systems (SUDS), attempt to restore a natural hydrologic budget (Joksimovic & Alam 2014) while underground detention facilities aim to accumulate the peak flow and discharge this into the system with a certain time lag (Andrés-Doménech et al. 2012). In this study, we focused on the improvement of underground LID structures, e.g. grey infrastructure, considering green solutions as a valuable additional measure in stormwater management.

Underground LID solutions aim to increase spare capacity in the system to accumulate excessive flow rates. The extra volume can be achieved by adding off-line or in-line underground storage into the UDS. Typical off-line facilities are detention tanks, having a connection to the mains, while free capacity of the drainage network has often been considered as in-line storage. Although effectiveness of off-line facilities is the objective of many recent studies (Lim et al. 2014; Thomas et al. 2016; Wang et al. 2017), in-line solutions are mainly considered as a possibility of utilising the excess capacity of the pipeline, not specially designed for storage (Garofalo et al. 2017). In this study, we intend to change this paradigm to include storage and flow control into the design as an additional objective.

Real time control (RTC) methodology that emerged with the development of information and communication technology (ICT) aims to bridge the structural and non-structural measures into one comprehensive solution (Beeneken et al. 2013). This is achieved by installing active network elements, e.g. weirs and valves, into the UDS. These actuators will be automatically adjusted on the basis of the data from the network sensors and thus allow UDS to be adaptable for different loading conditions. Therefore, RTC is seen as a key technology to improve the operation of UDS (García et al. 2015).

The main objective of our work is to find the most feasible solution to reduce stormwater peak runoff volume from compact real-estate development areas situated within a highly urbanised catchment. The paper is based on the results obtained in Kändler et al. (2018) with significant improvements to the overall analysis. The advantages of LID in-line reservoirs are coupled with RTC architecture to create a smart and sustainable solution for peak flow reduction from newly developed dense urban areas.

METHODOLOGY

Redevelopment of obsolete areas into new living and business districts is a constant process in every city. As these places are typically surrounded by highly urbanised catchments, constraints have to be imposed for the stormwater runoff into the existing UDS in order to avoid network surcharge.

In many cases, due to limited free space available, it is not possible to choose full-scale SUDS for flow mitigation, since these require notably more space than grey solutions (Fletcher et al. 2015). Free construction space is usually scarce in these areas because other underground communications and developers attempt to gain profit by maximising the building footprints.

Underground storage containing either off-line detention tanks or enlarged pipe sections for in-line storage are considered a feasible option to alleviate the stormwater runoff problem (Piro et al. 2010). Off-line detention tanks are usually cylinder shaped plastic or concrete underground barrels installed in the network with a connection to the UDS (Figure 1(a)). A diversion chamber is used to direct water either to the pipeline downstream or to the detention tank. After being filled, the tank will empty if the hydraulic grade line (HGL) in the system has lowered below a water level in the barrel. It was assumed in our study that the volume of the tanks will be compiled from 50 m3 plastic cylinders, as these have manageable dimensions for transportation and installation. It would be technically challenging to install larger barrels below the street area because of the limited free space. It was also assumed that the tanks will be filled and emptied only by gravity flow.

Figure 1

Underground stormwater detention solutions: (a) off-line detention tank; (b) in-line detention tank with a static orifice.

Figure 1

Underground stormwater detention solutions: (a) off-line detention tank; (b) in-line detention tank with a static orifice.

In-line storage facilities are typically enlarged pipe sections designed to be part of the network (Figure 1(b)). Outflow from these sections is restricted by a fixed orifice, i.e. a pipe section with a smaller diameter. During a rainstorm, these pipe sections will fill with water, hence reducing the peak flow at the outlet. Distinct from off-line storage solutions, water is always flowing through the system. In our study, it was assumed that the maximum diameter of these enlarged sections should not exceed 1 m, as larger conduits would significantly hinder installation of other communications in the street area.

An in-line storage system was coupled with an RTC solution to create a smart in-line storage unit. An RTC weir is designed to close the flow completely in order to harness the useful feature of off-line tanks. The system consists of an RTC manhole, a control algorithm, an optimisation and telemetry system which are described in more detail in the following paragraphs.

RTC manhole

A stormwater manhole with an adjustable weir is the key part of the developed system. The position of the weir changes the cross-section of the free opening and thus regulates the flow through the manhole (Figure 2). The position of the weir is changed by a threaded rod turned by an electric motor. The manhole also has an emergency overflow for handling extreme weather events and ensuring system operation during any malfunction.

Figure 2

In-line detention tank with an RTC weir.

Figure 2

In-line detention tank with an RTC weir.

An enlarged pipe section upstream of the manhole allows water accumulation. The length of the section can be determined on the basis of the terrain slope. Also, the number of RTC manholes in the system depends on the slope of the terrain and can be determined through simple optimisation if the hydraulic model of the system is available. If the terrain, i.e. the pipe slope, is significant, it is advisable to install control manholes in sequence in order to maximise the storage capacity. The RTC weir is equipped with a communication unit to exchange information with adjacent manholes and update the control curve. As stormwater systems will be built during the development of the area, simultaneously, trench works can be easily supplied with electricity from the grid system as the connection cables are installed.

Control algorithm

Operation of the system is based on the on-line data received from the water level sensors installed at the outlet and at the upstream section of each control manhole. The distributed mode predictive control (DMPC) algorithm is used to create the control curve for each weir-wall. DMPC is an optimisation-based solution capable of predicting a system status, i.e. relationship between the water level and the gate opening. This is first needed to eliminate measurement errors of the level sensors and second, to make the system's response to a rainfall hydrograph proactive.

Figure 3 presents the principal operational scheme of the control system. It can be seen from the scheme that no central control unit is needed and therefore each control manhole can operate independently, exchanging the information about the water levels and gate positions. Level sensors are used instead of flow measurements to reduce errors caused by sediments in the pipeline and reduce the cost of the system.

Figure 3

Architecture of the control algorithm.

Figure 3

Architecture of the control algorithm.

The following implementation strategy was used:

  1. The level measurement sensor SO is activating the system if the determined threshold level, i.e. flow rate out from the area, is achieved.

  2. The control manhole closest to SO starts to function, measuring the water height S1 at the upstream. This information x1 is fed to the DMPC1 unit and the respective operational curve u1 is calculated and sent to the actuator.

  3. If the weir-wall has reached a lower position but the water level at S1 is still increasing, the second control manhole DMPC2 is activated upstream. For that, DMPC1 is exchanging the status information w1. Information x2 about the water level upstream of the manhole is fed to the DMPC2 unit to calculate the necessary adjustments u2 for the actuator.

  4. As the flow through the gate depends on the water levels upstream (S1Sn), the position, i.e. the control curve of the weir-walls, is calculated by DMPCn units for each time step.

  5. When the rainfall is over, weir-walls will open in reverse order from activation. Weir-wall DMPCn at the upstream is sending information wn−1 to the downstream unit. Water levels are constantly measured to avoid exceeding the required peak outflow at SO.

Cascading operational action is designed for optimal system usage, as in the case of smaller rainfalls, only a few manholes are needed to be activated. This setup also reduces a risk of malfunction and avoids problems typically prevailing in centrally controlled RTC systems (van Daal et al. 2017).

Optimisation

Distributed model predictive control is an optimisation based control algorithm. Therefore, manipulated variables have to be optimised at each time step, taking into account the system prediction in order to have the model parameters fit to the actual targets, i.e. weir-wall positions. MS Excel spreadsheets with Solver Add-In module were utilised to code DMPC algorithm and perform necessary optimisations.

Optimisation results were tested and evaluated with EPA SWMM 5.1 software through a step-by-step process as presented in Figure 4. Water levels were calculated by the software and a new control curve was developed for each time step, i.e. runoff moment using the DMPC algorithm. Weir-walls were modelled as transverse weirs with a discharge coefficient of 1.84. The control curve calculated by the DMPC algorithm was imported to the modelling software to set the height of the weir-wall. The following optimisation constraints were imposed: (i) outflow should not exceed the threshold level at SO; (ii) maximum level of HGL shall be at peak moment not less than 0.5 m from the ground level; (iii) for safety reasons, no weir-walls should be allowed to be fully closed.

Figure 4

The process scheme of the optimisation process.

Figure 4

The process scheme of the optimisation process.

Telemetry system

Reliable communication is a key issue to ensure the smooth operation of the system. As the stormwater system will be built in conjunction with the area development, both cable-wired and wireless communication could be used. We chose the latter method in our study.

The weakening of an electromagnetic signal in soils with high moisture content and different structure layers (asphalt, gravel, sand) is so significant that communication with underground structures has been considered infeasible for decades. It has been found by Malik et al. (2018) that new low-power long-range communication technologies such as narrowband internet of things (NB-IoT) are opening new possibilities for underground data connections and are therefore especially suitable for urban drainage systems. These technologies have sufficiently low energy consumption, wide signal coverage and a suitable information rate for the described RTC system.

Economic aspects

The developed RTC in-line detention was compared with the traditional methods, such as off-line tanks and in-line detention, in economic terms to highlight the advantages of the solution. All three options were evaluated in reference to the scenario with no flow reduction measures implemented.

Two cost components were used to find the feasible solution (Table 1). The first component is the investment cost (IC) that includes the infrastructure cost (pipes, manholes, equipment, etc.) and the cost for the construction works.

Table 1

Cost components for the economic analysis

ComponentSpecificationCost
Investment Cost (IC)
Pipeline Diameter 0.1–0.4 m 230–350 EUR/m 
Pipeline Diameter 0.5–1.0 m 400–750 EUR/m 
Detention tank 50 m3 58,720 EUR/pcs 
Weir manhole Adjustable weir with communication unit 40,000 EUR/pcs 
RTC Sensors and control unit 50,000 EUR/pcs 
Penalties 
Exceeding the target flow per 1 m3 895 EUR/m3 
Replacement of the downstream pipeline Diameter >1.0 m 800 EUR/m 
ComponentSpecificationCost
Investment Cost (IC)
Pipeline Diameter 0.1–0.4 m 230–350 EUR/m 
Pipeline Diameter 0.5–1.0 m 400–750 EUR/m 
Detention tank 50 m3 58,720 EUR/pcs 
Weir manhole Adjustable weir with communication unit 40,000 EUR/pcs 
RTC Sensors and control unit 50,000 EUR/pcs 
Penalties 
Exceeding the target flow per 1 m3 895 EUR/m3 
Replacement of the downstream pipeline Diameter >1.0 m 800 EUR/m 

The second component, named as ‘penalty’, was imposed to consider the case when the technical solution appeared does not fully meet the design constraints and therefore extra investment is needed outside the street area, e.g. adding additional retention tanks on the plots. Penalty for the base scenario, i.e. with no LID used, takes into account the replacement of the pipeline outside the development area in order to allow higher volumes to pass.

Each system needs maintenance, which can be expressed as a cost per year. For example, stormwater pipelines need yearly cleaning from sediments and regular service is required for the weir-wall components. In addition, technical components have their unique lifetime, i.e. they need replacement after a certain period. Our consideration was that electronic components are expected to last up to 10 years, the pipeline, including manholes, need some restoration after 20 years and detention tanks need some repairs after 30 years of operation.

To take into account this variability, the net present value (NPV) for evaluated options was calculated as follows: 
formula
(1)

This method converts any future values to the present, thus providing an adequate answer about the feasibility of the investment. For that future payment Fn for each period n is summarised and subtracted with an interest rate i of this specific period n. Several assumptions were made for the calculation of NPV:

  1. Calculation period a was taken 50 years.

  2. Theoretical natural annual real interest rate was estimated to be around 3%.

  3. No price changes of investments were considered, i.e. inflation of the investment product is zero and therefore nominal and real interest rates are the same.

RESULTS

The developed RTC in-line detention system was tested and evaluated on the background of traditional off-line and in-line solutions in a 12.5-ha modern urban development area in Tallinn, Estonian capital (Figure 5). A base scenario with no LID imposed was taken as a reference to the other three options.

Figure 5

Urban development area in Tallinn with the base scenario (no flow reduction applied).

Figure 5

Urban development area in Tallinn with the base scenario (no flow reduction applied).

During the development, the old obsolete industrial territory will be turned into modern city environment with 1,600 apartments and offices. The build-up ratio of the plots is merely 100%, comprising both living blocks and underground parking garages. Taking into account street areas, including some pedestrian streets with minor greenery, the rate of impermeable surfaces is merely 90%. The infiltration and evaporation in the catchment areas were neglected to be in line with the cold climate conditions. A constant flow of 50 L/s, representing the infiltration from the upstream catchment areas, was applied to point A (Figure 5).

Standard design rainfall with a return period of two years, intensity of 28 mm/h and duration of 20 minutes was applied for the study on the basis of Estonian Design Standard (2013). As the catchment area is compact, unified rain intensity was used for the whole area.

Due to the limited capacity of the existing downstream UDS, the water utility has imposed the maximum limit of 300 L/s for the peak runoff. Calculations showed that under the base scenario, i.e. no LID foreseen, the peak runoff from the catchment will be 670 L/s, which is more than two times higher than that allowed.

Simple optimisation was performed to set up technical details for the three LID scenarios. The following optimisation constraints were considered for placing off-line and in-line storage units: (i) availability of free space for the construction and technical applicability, as described in the methodology; (ii) minimum IC; (iii) highest resilience, i.e. distance between the ground level and HGL is >0.5 m at the peak flow.

Nine off-line storage tanks with a total volume of 450 m3 were placed at the locations shown in Figure 6(a). As the storage tanks will cut off some of the peak flow rate, it was possible to use a smaller pipeline than in the base scenario presented in Figure 5. For the scenario of in-line storage, 0.9 km of accumulation pipeline was designed with a total volume of 480 m3 (Figure 6(b)). A fixed orifice with a diameter of 300 mm was installed at the outflow to restrict the flow. For the fourth scenario the section was replaced with the RTC weir and one additional weir was installed to the middle of the main conduit section, as presented in Figure 6(b).

Figure 6

Technical solutions for the peak flow reduction: (a) off-line detention tanks; (b) in-line detention with RTC.

Figure 6

Technical solutions for the peak flow reduction: (a) off-line detention tanks; (b) in-line detention with RTC.

The results of the analysis of the four scenarios are presented in Figure 7. It can be seen that only the fourth scenario, i.e. in-line detention with RTC, fully satisfies the outflow constraint (300 L/s). It reduces the peak flow by 57%. For the other three, some LID facilities had to be foreseen to provide an additional cut, resulting in penalties. It is important to note that in-line detention without RTC (option two), on the contrary, has the lowest effect on the peak flow reduction (26%). Traditional off-line detention reduced the peak flow by 39%, which correlates with the numbers presented in previous studies (Lim et al. 2014; Wang et al. 2017).

Figure 7

Peak flow graph of a 3-hour period for the different scenarios (300 L/s threshold presented with dashed line).

Figure 7

Peak flow graph of a 3-hour period for the different scenarios (300 L/s threshold presented with dashed line).

Results from the economic analysis are presented in Figure 8. It is quite obvious that the base scenario has the lowest IC, as this solution has no costly facilities conceived for flow detention. The IC of the other three options is on average 47% higher, which may lead to the decision that any flow detention is infeasible because of economic reasons. However, this conclusion is misleading since adding penalty costs to IC changes the sequence of total costs (TC) significantly. As can be seen from Figure 8, the base scenario has actually the highest penalty costs, which leads to the highest TC. High penalty cost stems from the need to rebuild about 0.9 km of the drainage collector downstream of the UDS to reduce the risk of network surcharge. For the other two options, the penalty cost was calculated on the basis of additional LID facilities needed on the plots. As RTC in-line storage is meeting the flow constraints, no penalty costs were applied.

Figure 8

Investment costs (IC), penalties and NPV of the analysed options.

Figure 8

Investment costs (IC), penalties and NPV of the analysed options.

It can be seen from the results that in-line tanks without RTC and off-line tanks have similar total costs. The penalty cost of the in-line tanks is the highest compared to other two options because of the substantial investments needed for an additional LID on the plots to cut the peak flow below the target level (300 L/s). It is also important to note that as the RTC in-line detention has no penalties, its total cost becomes, on average, 12% lower than that of the other two options, making the solution the most feasible.

Analysis of NPV of all four options showed that the base scenario has the highest investment value, and the in-line tanks with RTC the lowest, over a 50-year period. Although the in-line detention with RTC requires electronics that have a shorter lifespan, it reduces accumulation of sediments to be removed periodically in the other three cases. As a result, NPV of RTC in-line detention is 26% lower than in the base scenario and 5% lower than the in-line solution with no RTC.

DISCUSSION

Although it may seem that the traditional UDS are the most cost effective solution for new developments in dense urban areas, the results of our study show that if the penalty costs are taken into account, this option is actually the most costly. Upgrading the downstream UDS network needs much effort and adds additional economic burden to the option. Therefore, in terms of hydraulics and cost efficiency, the most feasible options are traditional off-line storage facilities and in-line detention with RTC. The TC of these solutions is merely a quarter lower than the traditional UDS (base scenario). It should be noted that the traditional in-line detention system showed a relatively low impact on the peak flow reduction and is therefore falling behind the other two with a comparatively high penalty cost.

The in-line system with RTC was also designed to mimic the features of off-line tanks, i.e. capability to temporarily cut the water flux during the filling period. However, this feature was not utilised in the case study because the system was capable of reducing the peak flow below the target limit even without completely closing the RTC weir-wall. It is important to note that this may not be the case in other case studies.

The footprint of the off-line tanks and in-line detention is relatively similar (240 m2 for the case study analysed), but there is a clear advantage in the latter option because the area is evenly distributed along the whole system. This facilitates the installation of the other communications in the street area, e.g. water supply, gas and electricity lines. Moreover, as the off-line tanks analysed in the study have a diameter of 2.4 m, it is not always possible to have them installed at the same level as the invert of the inflow pipeline. This may result in the accumulation of sediments that diminish the capacity of the tank and increase maintenance costs. Sediment transport and deposition have complex behaviour that is not easily characterised with conventional methods and understanding the influence on hydraulic systems requires advanced numerical simulations (Kaur et al. 2017). For that perspective, in-line storage with RTC is considered as the most feasible option to reduce peak runoff from this type of development area. The NPV of the solution is eventually 26% lower compared to the base scenario. This is a clear advantage in a long-term perspective.

CONCLUSION

A new decentralised real time controlled in-line detention system was developed to reduce the peak runoff from the urban catchment to the existing UDS downstream. The system has independently operated control manholes that are capable of exchanging information between each other and pro-actively adapt to the changing stormwater runoff hydrograph. The solution is easy to apply as neither a central control unit nor extra personnel are needed. As the system is situated completely underground, it is suitable for the locations of cold climate, especially for Northern Europe with moderate terrain conditions and smaller catchment areas where the infiltration capacity of the soil is non-existent.

The RTC in-line detention system was compared with two traditional solutions – off-line storage tanks and in-line detention facilities, and the base scenario – a system with no runoff control. Penalty rules were used to analyse actual indirect costs.

The methodology was tested in a 12.5-ha development area in Tallinn, Estonian capital. It was found that the in-line detention with RTC has the highest cut in the peak flow, the lowest total cost and the lowest investment costs over a 50-year period. Apart from that, the solution also has many other advantages, including better sedimentation removal and smaller construction footprint, which is particularly important in dense urban areas.

Future research will focus on more advanced optimisation methods, coding the solution in Python language and analysis of the different precipitation intensities, i.e. operation under extreme weather conditions.

ACKNOWLEDGEMENTS

This research was supported by the Institutional Research Funding IUT19-17 at Tallinn University of Technology and by the European Union (European Regional Development Fund) Interreg Baltic Sea Region Programme under Grant #R093. The authors thank Merko Ehitus Eesti AS for the permission to use the data of the development area in Tallinn for the case study.

REFERENCES

REFERENCES
Andrés-Doménech
I.
Montanari
A.
Marco
J. B.
2012
Efficiency of storm detention tanks for urban drainage systems under climate variability
.
J. Water Resour. Plann. Manage.
138
,
36
46
.
Beeneken
T.
Erbe
V.
Messmer
A.
Reder
C.
Rohlfing
R.
Scheer
M.
Schuetze
M.
Schumacher
B.
Weilandt
M.
Weyand
M.
2013
Real time control (RTC) of urban drainage systems – a discussion of the additional efforts compared to conventionally operated systems
.
Urban Water J.
10
(
5
),
293
299
.
Estonian Design Standard
2013
EVS 848:2013 Sewer Systems Outside Buildings. Estonian Centre for Standardization. Available from: www.evs.ee/tooted/evs-848-2013 (accessed 31 January 2018)
.
Fletcher
T. D.
Shuster
W.
Hunt
W. F.
Ashley
R.
Butler
D.
Arthur
S.
Trowsdale
S.
Barraud
S.
Semadeni-Davies
A.
Bertrand-Krajewski
J.
Mikkelsen
P. S.
Rivard
G.
Uhl
M.
Dagenais
D.
Viklander
M.
2015
SUDS, LID, BMPs, WSUD and more – The evolution and application of terminology surrounding urban drainage
.
Urban Water J.
12
(
7
),
525
542
.
García
L.
Barreiro-Gomez
J.
Escobar
E.
Téllez
D.
Quijano
N.
Ocampo-Martinez
C.
2015
Modeling and real-time control of urban drainage systems: a review
.
Adv. Water Resour.
85
,
120
132
.
Garofalo
G.
Giordano
A.
Piro
P.
Spezzano
G.
Vinci
A.
2017
A distributed real-time approach for mitigating CSO and flooding in urban drainage systems
.
J. Network Comput. Appl.
78
,
30
42
.
Kändler
N.
Annus
I.
Vassiljev
A.
Puust
R.
Kaur
K.
2018
Smart in-line storage facilities in urban drainage networks
.
Proceedings
2
(
11
),
631
.
Lim
S.
Ho
V. H.
Lee
S. Y.
Yoo
D. G.
Kim
J. H.
2014
Determination of optimal location and capacity of detention facilities
.
Procedia. Eng.
70
,
1037
1045
.
Madsen
H.
Lawrence
D.
Lang
M.
Martinkova
M.
Kjeldsen
T. R.
2014
Review of trend analysis and climate change projections of extreme precipitation and floods in Europe
.
J. Hydrol.
519
,
3634
3650
.
Malik
H.
Kändler
N.
Alam
M. M.
Annus
I.
Moullec
Y. L.
Kuusik
A.
2018
IEEE international environmental engineering conference
. In:
Evaluation of Low Power Wide Area Network Technologies for Smart Urban Drainage Systems. Paper Presented at the Conference 2018
,
March 12–14
,
Milan
.
Piro
P.
Carbone
M.
Garofalo
G.
2010
Distributed vs. concentrated storage options for controlling CSO volumes and pollutant loads
.
Water Pract. Technol.
5
(
3
),
1
15
.
wpt2010071-wpt2010071
.
Tapia
C.
Abajo
B.
Feliu
E.
Mendizabal
M.
Martinez
J. A.
Fernández
J. G.
Laburu
T.
Lejarazu
A.
2017
Profiling urban vulnerabilities to climate change: an indicator-based vulnerability assessment for European cities
.
Ecol. Indic.
78
,
142
155
.
van Daal
P.
Gruber
G.
Langeveld
J.
Muschalla
D.
Clemens
F.
2017
Performance evaluation of real time control in urban wastewater systems in practice: review and perspective
.
Environ. Modell. Software
95
,
90
101
.