Sustainable urban drainage systems in Ljubljana

Challenge:

The city of Ljubljana, Slovenia, faced challenges related to uncontrolled stormwater flow, which can lead to urban floods and excessive combined sewer overflows. The need for effective stormwater control measures (SCMs) to manage rainfall-runoff (RR) and achieve desired hydrological responses in urban catchments was critical. Traditional methods for designing and calibrating SCMs were time-consuming and required iterative expert-driven processes.

Approach:

The researchers from the University of Ljubljana and the Jožef Stefan Institute developed an automated modelling approach using the Process-Based Modelling Tool (ProBMoT) to optimise stormwater control measures (SCM) design. They created a new knowledge library for modelling rainfall-runoff (RR) and SCMs, which allowed for the automated discovery of optimal RR models and SCM designs. The approach involved encoding RR models and SCM processes in the modelling library, elaborating conceptual models of the case study and SCM scenarios, and using ProBMoT to discover the optimal model structure and parameters. The EPA Storm Water Management Model (SWMM) was used for RR and hydraulic modelling of urban drainage, incorporating several SCM elements. Precipitation data were provided by the Slovenian Forestry Institute, collected using a Rain Gauge Smart Sensor, while flow measurement data was provided by the local public utility company, collected using non-contact radar velocity measurements and ultrasonic water level measurements. Design rain events were based on IDF curves for the Ljubljana–Bežigrad weather station. Nine alternative structures of RR models were explored, calibrated, and validated against measured data. Six SCM scenarios (detention pond, storage tank, bio-retention cell, infiltration trench, rain garden, and green roof) were designed and simulated. The SCMs were evaluated for their ability to achieve target outflows using the Nash–Sutcliffe efficiency coefficient (NSE), peak flow ratio (PFR), and total volume ratio (TVR). The best-performing RR model was used to simulate catchment outflow for rainfall events with different return periods. The performance of SCMs was evaluated through simulations and adjustments to the conceptual models.

Outcomes:

The automated modelling approach proved to be efficient, with the RR models achieving "very good" performance according to the Nash–Sutcliffe efficiency criteria. The SCM scenarios were evaluated for their ability to achieve target outflows, with detention ponds and storage tanks showing near-perfect matches to the target outflow. Infiltration trenches also performed well but required more space. Rain gardens and bio-retention cells were effective for less intense design events but were limited by soil infiltration rates. Green roofs showed potential by replacing impervious areas and reducing outflows. The study demonstrated that the automated approach could effectively design SCMs and optimise urban drainage systems.

Data Sources:

Data for the study was sourced from the Slovenian Forestry Institute, local public utility company, and the EPA Storm Water Management Model (SWMM). Precipitation data, flow measurements, and catchment characteristics such as land use, topography, and soil properties were used to calibrate and validate the models.

Additional Information:

• Design and Simulation of Stormwater Control Measures Using Automated Modeling
•  

Point of Contact:

• infoijs [dot] si (info[at]ijs[dot]si)