Using an SVG simulation tool to design an urban stormwater harvesting system for the city of Salisbury
Date
2007
Authors
Thomas, P.
Howlett, P.G.
Boland, J.W.
Piantadosi, J.
Editors
Oxley, L.
Kulasiri, D.
Kulasiri, D.
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Conference paper
Citation
MODSIM07: land, water and environmental management : integrated systems for sustainability, 2007 / Oxley, L., Kulasiri, D. (ed./s), pp.219-225
Statement of Responsibility
Conference Name
International Congress on Modelling and Simulation (MODSIM07) (10 Dec 2007 - 13 Dec 2007 : New Zealand)
Abstract
We describe development of a Scalable Vector Graphics (SVG) simulation tool for design and optimal management of urban stormwater harvesting systems. The system is modelled on the Helps Road Drain in the City of Salisbury, South Australia and consists of a stormwater supply channel in an urban catchment with a series of in-line storage dams. It is used as a temporary water storage for flood mitigation and control of environmental flows, and to capture, hold and harvest stormwater for supply to consumers or for aquifer storage and recovery (ASR). The purpose of the simulation is to analyse the operation of a typical urban stormwater harvesting system and to study the movement of water through a series of connected dams. Our model uses a system of ordinary differential equations to describe the flows. The differential equations relate the rates of change in storage volumes to rates of inflow, rates of drainage outflow and rates of harvesting. We used experimental simulations to test various different configurations and to gain insight into the operation of the system. Each storage unit consists of a permanent storage component below the level of the outflow pipe or weir and a temporary storage component above this level. During “wet” periods water flow is controlled by filling of the temporary storage component above the outflow level. This water continually drains away and passes on to the next downstream storage. During “dry” periods the water level is often below the outflow level. At all times water can be harvested by pumping from the dams for direct supply to consumers or for aquifer storage. Our work can be used to determine effective design parameters for the various storage units and to simulate the operation of the system using appropriate management policies. The simulation will provide water managers with an easy-to-use computer-based management tool. The model will incorporate generations of simulated rainfall and run-off scenarios to enable a realistic assessment of system capacity and performance and will allow managers to visualise limiting factors and assist in understanding the interdependence of different system components. The simulation tool can be used to investigate the effectiveness of various water management policies by evaluating the Conditional Value-at-Risk (CVaR) for failure of the stormwater supply system. In this application CVaR is the expected volume of shortfall in supply to consumers given that the shortfall exceeds some significant threshold known as the Value-at-Risk (VaR). CVaR was introduced originally in financial applications (Rockafellar and Uryasev, 2000). Simulations on configurations with identical dams found that a system with low storage capacity was unable to satisfy demand on a regular basis and that upstream dams were likely to overflow. The term overflow is used to describe water that escapes from the system and may cause flooding. Systems with greater storage capacity were better able to satisfy demand but the downstream dams were often empty. The simulation includes a comparison mode in which identical systems with different initial states are subject to the same inflow patterns and the same consumer demands. The simulation shows that two such systems, one initially empty and the other initially full, eventually converge to the same state. The average time taken for a full system and an empty system to converge to the same state defines an intrinsic time-scale for each fixed configuration. We found that systems with smaller storage capacity converged more quickly and that systems with greater storage capacity took longer to converge. Systems with very large capacity and very long time-scales may be difficult to manage. The simulations suggest that upstream dams should be designed primarily for flood mitigation with large temporary storage capacities and high outflow rates and that downstream dams should be smaller and designed primarily for permanent storage. A system of this type can be more easily managed and should allow evenly spread demand with less overflow upstream and less drying out downstream. Although this is a prototype simulation modelled on a particular system the model could be adapted and applied to other stormwater harvesting systems, both locally and nationally.
School/Discipline
Dissertation Note
Provenance
Description
Access Status
Rights
Copyright status unknown