By Andrew Rowland, Technical Lead (Clean Water Networks), WSP
The concept of using smart technology to improve water management isn’t new, but examples of the fully-automated systems we imagine when we think of ‘smart’ are limited. One of the barriers is the perception that becoming smarter requires significant investment. But, by taking a step-by-step approach, we can all start to benefit.
Selecting targets for investment
Opportunities to apply smart technologies are proliferating so fast that choosing the most appropriate projects to support is a major headache for water companies. With so many immediate demands on funds, it can be difficult to look too far ahead.
Some companies are ‘testing the water’ – often quite literally – when it comes to assessing their options. We are working with several companies in the UK on projects that examine how smart strategies can be applied to various aspects of their business. It is a relatively low-risk route to developing a good understanding of the benefits that smart technologies can bring.
Data analytics in Ripponden
Yorkshire Water is amongst the water companies that are leading the way in smart analytics. We worked with their modelling and data team on a project to better understand the role that technology may be able to play in identifying and managing inflow and infiltration (I&I) flows.
We focused on the case study catchment of Ripponden, situated in the Pennines, as a test-bed for a smart strategy. In Ripponden, I&I from the surrounding hillside puts pressure on the sewerage system. By monitoring and modelling these issues, using a variety of sensors and bespoke data analytics, we could confidently develop options to reduce and manage I&I flows.
Water management and the IoT
Academic research into smart technologies for water management, particularly as it relates to optimising asset operations to meet climate change challenges, will help us prepare for what the future might bring. We participate in targeted research as part of our ‘Future Ready’ approach to seeing the future more clearly, and helping our clients design for it.
With the University of Surrey, and water industry research specialists PyTerra, we have been developing machine learning to predict flooding and optimise storage in catchments, in real-time. By using IoT-enabled sensors to monitor river water level data, and linking this to data on weather conditions and forecasts, we have devised an approach that promises accurate and early flood warning. This approach maximises natural catchment flood storage and offers communities vital time to prepare for floods and avert damage, by taking simple measures to protect their homes. The benefit is augmented if water companies choose to bolt on technologies such as active management of up-stream storage to reduce downstream flooding.
Our international team has already developed a system that uses IoT-enabled devices to pump stormwater away from the busy three-level stacked freeways of the Turcot Interchange in Montreal. The system processes various data with the aim of diverting water to nearby wetland, to reduce pressure on the storm sewers, and hence overflows to the St. Lawrence River.
What emerges most clearly from our work, is that implementing a smart approach doesn’t necessarily mean huge up-front expense. There is often a tiered set of options ranging from the relatively simple to the highly complex.
Digital twins in the West Midlands
Transferable knowledge and technologies can be harvested from other industries. We have been sharing with our water company clients some of the output from our work with Transport for the West Midlands (TFWM) on using ‘digital twins’ - models that fully describe a real or planned physical asset, such as a section of road, a sign-post, or lane-markings. TFWM is using the models we have built with them to test new vehicles, services and technologies in a virtual environment that closely replicates the real world.
The parallels with the water sector are evident: in a digital twin of an existing network, solutions to supply or sewerage problems could be identified and trialled before implementation, saving time and money and yielding an optimal outcome. Using the twin to examine failure events, as well as machine learning and AI to predict future failures, enables us to determine the best response plans to help realise better system resilience.
Given the many thousands of existing assets that a water company may have to map, developing a digital twin could be an extensive and expensive process, but it will be an important and worthwhile investment.
Working up, tier-by-tier
The approach water companies take to digital investments may well be tiered. Understanding what’s happening in a water infrastructure network using sensors and distributed network modelling is a stride in the right direction. Developing digital twins to aid decision-making could be a viable next step. Ultimately, completely smart solutions such as using robotics to identify and carry out repairs to buried infrastructure, may well be the goal. Only by starting out, however, can the journey truly begin.