Life needs water to survive, as history continually shows us. For millennia, human settlements have sprung up where water is present, with agricultural cycles and advancements virtually dictated by the resource and its reliability, supply, security, and cleanliness. Entire civilisations prospered, crumbled, or moved based on water supply.
But while water management has very much remained the same with reservoirs and piping, management tools have not. Now, water utilities are increasingly turning to data-driven technologies that give them access to information on water demand and supply considered unattainable before, allowing them to realise sustainable water supply and management for the future, according to Trevor Hill.
And one of the newer tools – on top of the Internet of Things (IoT) and Industrial Internet of Things (IIoT) – is artificial intelligence (AI). When it was first introduced, the software decision support systems originally used hard-programmed algorithms based on certain rules to derive decisions or outputs by whittling down the available choices. But while they were functional, the systems very quickly became too cumbersome when faced with the perennially dynamic environment of the water sector along with increasingly stiff competition.
But now, the newest tools of AI are completely changing the game. With pattern recognition applications with a series of inputs, weighing factors and many other functions that are – by definition – autonomously and dynamically updated as new information is consistently presented.
In more ways than one, AI replicates how humans learn. During programming, it “learns” when input data is correlated to known outputs, and allowing the algorithms to learn over time. In the operational phase, the programme will begin to make sense of patterns as new data is brought in. And because AI is adaptable and capable of processing large amounts of data in real-time, it is ideal for water management in a sector that is always changing and evolving, allowing water utilities to maximise revenue and plan for the future.
On top of allowing better planning and management, municipalities can also better understand water loss in real-time, and operate water distribution networks more efficiently. And there are more benefits – with AI, utilities can engage customers in an engaging, informative, and personalised method, with real-time water consumption, bills, and information on the conditions of the water supply. Moreover, water utilities would be able to educate customers with the big data provided on conservation, as well as slowly changing behavioural patterns.
Unfortunately, according to Trevor Hill, AI is only as good as the data and how well the output is understood. And if the results are to be optimal, AI would need a validated data set that is adequately sized with enough information to categorise the issue as well as test and train the network – and it is not absolutely failsafe.
Although AI is able to perform basic analyses and more of the computational grunt work, it is not a full solution. In fact, it does abstract the issue, and takes away some insight from humans. Utility managers would need to be able to critically interpret the outcome AI provides. Machines and technology can work at a faster rate with fewer unforced errors, but ultimately, human interaction is still required. Human traits that are utterly unique much like investigating and questioning are still vital, and experienced and knowledgeable employees would still be needed to be an essential and critical part of delivering potable water to customers.
AI has the potential to completely and fundamentally change how water utilities operate and manage water in an environment that is becoming increasingly dynamic and volatile. But the main challenge now is to structure data and management services in such a way that AI can be maximised. From there, the sector can then turn its concentration on developing new models that optimise operations.
Source: Water Finance and Management