Asset Optimisation: Building data and intelligence strategy for AMP7

As we move towards Water Company final determinations for AMP7 one thing is consistent. For the problems that matter to customers, there is going to have to be a significant change in approach, behaviour and technology to meet the challenges being targeted by Water Companies and set by Ofwat.

From a wastewater perspective, the common performance commitments to reduce pollution and internal flooding have all attracted significant percentage reductions by the end of 2024/5.

An example is the 20% reduction in Pollution Incidents targeted by United Utilities and a 24% reduction in Internal Flooding incidents targeted by South West Water (both Fast Track companies). These targets are then stretched further with long term aspirations to, in some cases, eliminate these critical service failures. The scale of these reductions will force innovation as continuing with the same approaches will lead to penalty.

What is clear is that an improved understanding of how systems operate in all conditions is a fundamental part of achieving these targets. Managing wet weather performance will play its part, however understanding the risk to our wastewater networks of dry weather performance will be a critical factor, particularly when trying to mitigate the risk of internal (and external) flooding.

Obviously, data will play its part, but the critical element is making intelligent decisions based on this data. AMP6 saw the roll out of Event Duration Monitoring (EDM) at Combined Sewer Overflows (CSO). Whilst these sensors are intended to support the Storm Overflow Assessment Framework (SOAF) and enable Water Companies and the Environment Agency to target high spilling overflows for intervention, if set up right, they can also provide early warning of operational issues which may cause dry weather operation and a likely pollution incident. Understanding the ‘normal’ operating conditions at these locations will enable timely and targeted intervention. At RPS, we have been working with Water Companies using Machine Learning principles along with historical information to develop normal performance bands and alerting models where conditions deviate from ‘normal’. This has allowed early identification of operation issues likely to cause blockage and rectification of the problems on site.

This is a principal that can be taken through to any monitored site, so as monitoring programmes expand (another commitment for many Water Companies in AMP7) there is a real opportunity to utilise Machine Learning and Artificial Intelligence (AI) principles to better understand networks and improve serviceability. Again, RPS is working with Water Companies to utilise these techniques, looking at blockage/flooding performance but using the same principles to assess longer term impacts of new development.

It should be noted that data does not necessarily mean intelligence. Placing 5,000 monitors into a network is not going to reap the level of rewards that this level of investment should unless there is strategic intelligence as to why that location needs to be monitored. The EDM monitors at CSOs implemented in AMP6 provide instant value as they are the primary location for interaction with watercourses and sources of pollution incidents. Monitoring the most beneficial locations for flooding and pollution from pipe assets needs a depth of understanding of the consequence of failure. Once you understand what happens to a network should every asset fail, you can then build the understanding of likelihood of failure across a catchment and therefore risk. Whether this is achieved through deterioration modelling, historical performance analysis or the use of Machine Learning and AI to prioritise clusters of assets, these locations then become the targets for serviceability monitoring. This does enable us to use these innovative processes, supported by engineering judgement and knowledge to position monitors to provide the highest serviceability benefit.

With thousands of monitors comes gigabytes of data. A real big data challenge. And it is not just the scale of the task that is a challenge, it is the co-ordination of information coming from a multitude of places. Monitors will live stream data but are potentially on different platforms accessed in different ways. All of these data sources are separate from the intelligence of the geo-spatial data and modelling information held by Water Companies. This poses an important question; how do you pull all of this together to inform a coherent strategy?

This is where RPS has been able to build on experience from the clean water side of our business, where extensive sensor and monitor distribution has been in place for years. From this experience, RPS has developed the WaterNet Pro platform, (previously known as Wastewaternet), which enables all of these disparate datasets and live streams to be visualised, assessed and actioned in real time, supported by Machine Learning and AI where required to drive the serviceability improvement programmes. We are currently working with Water Companies to implement this system to start to achieve these strategic changes.

But as we become more intelligent customers, our behavioural impact also needs to play a part. Can we change serviceability performance based on a few key changes in our behaviour? In the future property level sensors will also play a part and the IOT will become part of the solution. Will we have a system that can incorporate this scale of data, link to existing Water Company systems? Provide analysis, intelligence and smart decision making in real time? We feel that all of this is possible, and we need to gear up our systems and tools to enable this. WaterNet Pro will be instrumental in achieving this.

Finally, can we deliver a shift change in serviceability to meet the targets of AMP7? At RPS we certainly are on the journey to accomplishing this.

For more information please contact James Hale, Technical Director, RPS, james.hale@rpsgroup.com