Artificial Intelligence in the Water Industry: Myth or Reality?

Water suppliers are constantly seeking techniques to improve the quality of their services and reduce operational costs. This is traditionally done through ensuring that the water infrastructure is maintained regularly by performing routine maintenance and responding to faults within the infrastructure. Operators would usually discover a fault, analyse the data and respond accordingly, in a “reactive” manner. Although these techniques work, they mainly rely on human intervention which can sometimes be inefficient, slow and potentially costly. In addition, water suppliers generally follow a reactive approach to energy consumption, where potential savings are lost due to not taking external factors into account.

Artificial intelligence (AI) systems utilise real-time and historic data to optimise operations in terms of cost and quality, potentially making the water industry more “proactive”. Current AI systems can significantly reduce the response time to unusual events or faults, predict faults and learn how to respond in future occurrences. These systems can also produce maintenance schedules and assign their priority based on

predicted outcomes.

Another major benefit is the ability to optimise operations to save on energy costs based on various sources of data such as weather reports and electricity suppliers’ charges, creating the ability to estimate the optimal balance between reducing energy consumption and maintaining the required storage volumes for peak demand. Such systems would significantly reduce the cost of operations whilst ensuring a sufficient reserve.

This paper will investigate existing applications of AI in the water industry, specifically those that are scalable, practical and require minimum change to existing systems. In addition, real-world examples of AI implementation in the water industry will be discussed with the emphasis on lessons learnt. Finally, this paper will showcase platforms data collected by SCADA and historian systems to be converted into “smart data”, enabling water suppliers to utilise the latest AI technology at relatively small costs.