Within WRc’s Technology area we are developing a Portfolio collaborative project for January 2014 to facilitate and accelerate the move to true, real-time hydraulic/water quality models and decision support systems. Drawing on academic and other work, it will pull together:
- Model calibration and validation; use of advanced machine learning techniques;
- Demand prediction including use of real-time data from smart meters/AMI;
- Optimum spatial and temporal resolution for sensors; use of virtual sensors to improve model accuracy;
- Data quality and automatic sensor data validation;
- Technical architecture; use of computational steering as a control system allowing end users to interact with simulations;
- Accompanying changes to business processes and organisation/people.
The project will use an advanced distributed computing architecture and prototype software developed at the University of Manchester (UoM). High performance computing will be used to provide simulation results in near-real-time, including running multiple simulations in parallel to model uncertainty. The use of commercial on-demand cloud services (e.g. Windows Azure) will also be demonstrated and assessed. The modelling work will be based on EPANET and also the newer CWSnet object-oriented library.
A key factor in achieving successful outcomes will be augmentation of real-time data collected from planned events such as hydrant and valve operations, in conjunction with automated network calibration algorithms to interpret the whole of the data captured. This will provide a control room embedded capability for the resolution of legacy anomalies within water mains networks such as long-term unknown closed valve locations and location of historically persistent leakage hotspots. It will also provide enhanced problem solving capability for the real world network for the resolution of unplanned events and deteriorating network performance as it occurs.
A large-scale practical trial will be carried out with one or more participating water companies. Participants will also be able to trial the use of real-time models and DSTs for their own networks using prototype software on mobile devices. Use of a dedicated test bed and software prototyping will allow different approaches to be tested and evaluated, and issues exposed by the work to be addressed within the project.
The work will be supported by a Knowledge Transfer Partnership with UoM. The TSB-funded scheme provides for academic support and a full-time associate to work on the project, and normally attracts a 67% government grant.
For further information contact John Brown on 01793 865016 or email@example.com.
Return to Newsletter