Finding sense amongst the noise: SiW 2019 write up

October 17th, 2019

The Water Industry is going through its “Digital Transformation” and at the moment there is a lot of “noise” around what sensing technologies are the right fit. In addition, innovations such as the Internet of Things and a wide variety of technologies bring the data and information together to allow the industry to work as efficiently as it can.

Recently the Water Industry gathered to “Find sense amongst the noise,” at the biennial Sensing in Water Conference. It was certainly an apt strapline for this conference. Throughout the 18 presentations by 20 speakers there was an underlying question: How do we convert data into information to provide real insight into the water industry? Pieces of the jigsaw making up the answer were to be found in virtually every presentation that was made at the conference as outlined below.

Rhys Stevens of Dwr Cymru Welsh Water

Catchment monitoring was the first area of focus and the areas of discussion were very much about preventing problems at source. Rhys Stevens of Welsh Water spoke about various areas of challenges that sensing technology can help with although collaborative projects are needed to address areas of concern. One case study presented was from the Cantref reservoir in the Brecon Beacons where landslides can cause huge increases in turbidity in the water. Advanced warning would allow Welsh Water to put contingency plans in place which can help with the resilience of the water treatment works. Operational resilience was one topic presented by the keynote speaker, Alison Fergusson from Ofwat, when she talked about the drivers that keep the industry moving forward. There were other challenges raised by Rhys Stephens including Taste & Odour Issues, Pollution Incidents affecting water sources and draw-off & intake management of raw water supplies.

The pollution of water sources can of course come from the strangest of places and the last presentation of the session concentrated on a project which tracked cattle movements and the impacts that they have on small watercourses. This sandwiched presentations on using surrogates for traditional measurement including UV254 and Suspended Solids as a surrogate for phosphorus measurements and peCOD as a surrogate for measuring disinfectant by-product formation.

Phil Tomlinson of Metasphere (a SWIG Director) chairing the panel session on Catchment monitoring.

The next session concentrated on the wastewater network. This began with a presentation by Phil Hulme & Rob Whittaker of the Environment Agency and concentrated on how the industry can work to improve the environment, its ultimate aim, through concentrating on the management of flows through the wastewater system. This started a number of years ago with the ministerial direction to monitor the vast majority of Combined Storm Overflows. The aim was to prevent spills outside of storm events by first monitoring when they are happening and then taking investigative action to identify when remedial action needs to be taken. This is being followed up in AMP7 of the water industry with a flow to full treatment programme to prevent spills from wastewater treatment works. Together these programmes aim to produce a much wider environmental benefit by controlling flow through the wastewater system and to ensure everything that should be treated is treated. This has similarities to the thoughts surrounding OPRA-PBC which seek to put Process Based Control in place in the wastewater system to achieve very similar aims. There of course has been work on this over the years and this was highlighted by Pete Skipworth when he talked about the various network control systems including the CENTAUR system which has been used in Europe to provide flood protection by using automated systems to protect assets.

It was good to hear the next presentation by Debbie Bell of Anglian Water and Robin Corbie of Technolog who are working together to produce a tool for the management of wastewater collection systems by using a combination of sensors and the ICM Live modelling system to provide an alert system for the control centres to use in order to detect problems within the network. Currently Anglian Water have ten models running on an hourly basis. This is one of the approaches that the water industry has taken in this area, following Southern Water setting up a comprehensive smart wastewater network a number of years ago as part of a project to protect the city of Portsmouth and its catchment area. This of course is the very start of the implementation of smart wastewater networks across the water industry as a whole, the benefits at the moment are largely unknown and the technologies are still to be developed in their entirety but the maturity of the industry in this area is assuredly growing.

Frank van der Kleij of Bristol Water

An area that has taken this step forward is that of the water distribution network and this is the area that started day 2 of Sensing in Water. The conference heard from Frank van der Kleij of Bristol Water and Derek Leslie of Severn Trent Water about the challenges that the Water Industry faces and how addressing these challenges is an opportunity for the supply chain. In particular the need for Advanced Network Monitoring over above the quantity monitoring that the industry conducts for leakage management. Bristol Water is going down the route of installing a large number of sensors on the water distribution network and creating a Digital Twin – a virtual model of the physical system that is regularly updated by the sensor network to provide insight into the operation of the network. One of the things that Bristol Water have not overlooked, which is unusual in the water industry, is the validation of sensor data to ensure that the data quality that they are collecting is actual right and assessing this data to ensure that maintenance requirements are accurately understood.

Derek Leslie of Severn Trent Water

So what benefits can all of these sensors bring? Presentations by John Gaffney of Siemens and Grigorious Kyritsakas of the University of Sheffield indirectly addressed this when they talked about discolouration events and Chlorine Decay within the distribution network. The work by John Gaffney looked at how results from turbidity monitoring can be converted using a calibration curve between turbidity and metal ions to create a solid flux model within the network itself with the theory that discolouration events are linked to this solid flux which in reality is the case.

This is in fact linked to the work that Grigorious Kyritsakas has done in looking at chlorine decay in his work with Sheffield University and Scottish Water which uses a combination of Machine Learning and Artificial Neural Networks to predict chlorine concentrations with the results thought to be beneficial to interventions on low chlorine concentration but also being useful to predict areas of high chlorine consumption as a surrogate for both disinfection by-product formation, discolouration and other such potential events within the system. In reality it raises the question as to whether the industry can use simple surrogate parameters such as temperature and turbidity combined with artificial intelligence to manage water quality within the distribution network.

The last session at this year’s conference was on data analytics and the use of data and in this session we heard some of the practical areas where there is a need for data within the wider water and environmental industries. The first example was that of the various requirements for water under abstraction licensing for agricultural uses. There is a burden of proof for agricultural abstractors to prove that they need their abstraction licence which can strongly affect the economics of farming. This has led the agricultural industry to look at a range of sources of water for agricultural purposes and a number of innovative projects for water resources. Chief amongst these projects is the Felixstowe Project which looks to abstract between 1500-2000 ML of water per annum from freshwater flow to the estuary. The problem being the lack of flow data that is available and also the management of the water resources between the different stakeholders. This is where concepts such as Blockchain can come in where specific units of any product be that product be it water resources between farmers or water companies or anything else from sludge to pollution credits can be “traded”. It is a concept that has been around industry that never fits in too well but Ana Pobrezenha explained it beautifully.

The areas considered were around

  • Water Security
  • Non-revenue water loss
  • Pollution
  • Sustainability
  • Sludge Trading

The case for sludge trading, although not active within the water industry at the current time is more than possible as within the next asset management period within England & Wales the water industry will be switching to an open market for bio-resources and there is increasing amount of anaerobic digesters within the wider environmental industry.

The second use of data is in customer influencing and this is an area that has been largely driven by a company called Advizzo in the UK Water Industry by using customer data to influence behaviour in water consumption. It is an area that looks at not just smart water data and advises customers as to their consumption patterns and uses household devices to advise customers when their water consumption is high. It is an area that exhibited some resistance and concerns as there is potential to compare yourself with the area average consumption data but in reality the level at which it acts does not infringe on people’s privacy.

Prof Guangtao Fu of the University of Exeter

Data and the Smart Water Industry has the potential to revolutionise the way that the industry works and this is the point that was emphasised at Sensing in Water as it was emphasised in the last presentation of the conference by Professor Gaungtao Fu of the University of Exeter when he talked about the use of AI-enabled Digitisation Pathways and the use of not only Artificial Neural Networks but Deep Learning which is a stage on from this. The case study used Deep Learning to predict bursts within the water distribution network as an alternative to more traditional and intensive methodologies.

In truth the undercurrent of the whole conference is that the water industry is heading towards a future where it is digitally transformed. It is going through stages of learning and picking up new concepts such as Artificial Intelligence and Machine Learning which is an area that the academic community are familiar with for a number of years now. In this way the water industry needs to collaborate in much closer ties with the academic community to understand what the “art of the possible” actually is. The academic community also need to collaborate with the water companies and the supply chains to understand not only the application but the sensing technology that is currently available. In short there is a need, if the Smart Water Industry is to be successful, for the Water Companies, Supply Chain and Academia to collaborate in much closes ways, something that SWIG was always designed to do.

Oliver Grievson, Deputy Chairman of SWIG and Technical Lead on the Water Industry at Z Tech Control Systems Ltd.