Probabilistic Transient Propagation (PTP)

The supply of sufficient quantities of safe, clean water is a corner stone of modern civilised society. Despite this water is something that most people take for granted, being unaware of the huge scale of the water distribution infrastructure and the engineering challenges faced in ensuring day-to-day operation and the future of our supply.

Water Distribution Systems (WDS) are in a state of constant flux. The pressure of the water supplied and the velocity it travels at are continually changing, and these changes can occur extremely rapidly. These phenomena are known as hydraulic transients and are variations of pressure that spread out around the network in waves of alternating highs and lows. These changes can cause damage to the pipes and features of the network. The vision of this project is to change how hydraulic transients in WDS are modelled by taking into account, from the start, the inherent uncertainties we have in these systems. This probabilistic approach will provide water utilities and their engineers on the ground with the required knowledge to assess the risks that transients may pose to assets, allowing them to make decisions to minimise damage to their systems from extreme or cyclic transients, to be able to target investment in maintenance, repair or replacement and therefore ensure the long-term sustainable operation of these indispensable yet ageing networks.

Traditionally hydraulic transients have only been given real consideration in simple systems, as they were thought to occur for very short periods and to die away very quickly. This has been compounded by the lack of tools available to model, and the technology to measure, transients in real systems. Recent research has highlighted that transients occur far more often and are far more widespread than previously realised. Being able to make predictions of the size and shape of hydraulic transients would be an exceptionally useful tool for network operators to help them manage the risks they pose. Our current modelling tools give good results in well controlled lab environments but fail when applied to real, complex systems. The aim of this research project is to predict the travel of these transient waves in WDS and to propagate the uncertainties based on system parameters and boundary conditions.

The uncertainty being considered in this project is primarily due to the system properties, for instance the roughness of pipes (a critical factor affecting how much energy is needed to transmit water) or the speed that the transient waves travel (a factor that is influenced by the type of material and the level of deterioration of the pipes). As the model will take into account our uncertainty of the system it will not give us a single value for the prediction of a transient's properties at a given time and location; rather it will give us a range of possible results and the probability that each will occur. The project will first develop a robust but computationally expensive sampling approach, then explore new techniques to improve the efficiency of the modelling process to allow it to be applied to full scale systems and to ensure the wide uptake of the techniques by industry.

Hydraulic transients capture a huge amount of system information as they are modified by every feature of the system through which they pass. This information, if suitably decoded, can give access to vital knowledge of the condition and operation of the networks. The research undertaken in this first grant proposal will provide the foundation for future work to demonstrate the potential of combining simulation results with measured data; to reduce the uncertainty in actual system parameters and give WDS network operators the first viable widely implementable condition assessment tool.

Water utility companies will be the main initial beneficiaries as they have the responsibility to operate and maintain water distribution systems with an obligation to improve the efficiency of their operations. Customers of the water utilities will benefit by seeing a decrease in the cost of their water supply due to improvement in services. Companies specialising in the development or deployment of pressure and flow monitoring equipment will benefit through the increase in the use of their services and the intelligent procedures that allow the maximum information gain at least cost. In addition, commercial water distribution system modelling software companies will benefit by having access to new, robust and reliable methods for modelling transients in complex systems. Ultimately, the UK economy and the general public will be the key beneficiaries with improved service, effectiveness, efficiency and cost savings within the UK Water Industry. This will be achieved through the optimal use of existing infrastructure, giving them extended life and minimising targeted future investment. The new modelling tools developed and the experience and knowledge gained will have worldwide potential, and will ultimately lead to improved global water supply. Additionally many of the research outputs will be widely applicable to other academics working in the related fields of water informatics, uncertainty quantification, and health and condition monitoring.

To ensure the full impact of the project is achieved a number of dissemination processes will be undertaken. Firstly a steering panel of academics and industry partners will be convened; this panel will be primarily to direct the project but will provide the first means of disseminating to and involving the wider community in the research being undertaken. The PI and PDRA will attend and have a presence at industry meetings and events (for instance the Clean Water Modelling Advisory Group (CwMAG), SWIG and WATNET) to discuss the current use of modelling of transients and to promote the research outputs. To improve the visibility of the research a project specific website will be created and maintained throughout the course of the project. This website will be used as a hub to promote the project, but also as an open-source repository of outputs from the project. This will ensure greater discussion of the work in the academic community, cross pollinating research with other academics and increasing the overall reach of this project.

At the culmination of the project an industry dissemination event will be held at Sheffield. Here research outputs from the project will be presented and the event will be used to discuss the implementation of the research and to provide a spring board for future development and funding. Academic impacts will be maximised by the publication of project results in high impact journals and the running of a workshop on future transient modelling techniques.