Teaching your treatment
works to be smarter
The Talk: interview
As the industry drives towards ever greater efficiencies, could artificial intelligence
(AI) and neural networks help organisations cut waste and costs, while reducing their
carbon footprints? Ruth Clarke (RC) and Dr Jöerg Gebhardt (JG) from Xylem outline the
opportunities for AI in the wastewater industry.
10 | MARCH 2020 | WWT | www.wwtonline.co.uk
Q. Why do we need intelligent
decision solutions like neural
networks?
RC: "Across the industry, especially in the
UK, there is a drive to do more with less.
Utilities want to cut leakage and pollu-
tion incidents, as well as reduce costs.
Optimising the operation of treatment
works is one key area where energy and
chemical consumption can be reduced
substantially. Intelligent decision solu-
tions, such as artificial neural networks,
can help wastewater facilities operate
more efficiently, while ensuring that they
tions and have a single point of optimal
operation, our sector continually juggles
changing circumstances.
"Neural networks can reduce inef-
ficiencies through improved prediction.
Consisting of a series of algorithms
programmed to recognise underlying
relationships in data sets, such as those
produced by a plant over the course of a
year, they are essentially pattern recogni-
tion systems.
"And because they adapt to changing
input, they can generate the best pos-
sible result without the need to redesign
An aerial shot of a wastewater treatment plant.
still meet, or even improve, effluent qual-
ity levels.
"Put simply, AI in wastewater is about
using the reams of data that are already
at operators' fingertips to learn the plant's
patterns – and use that information to
move from a reactive to a proactive way of
working."
Q. What are artificial neural net-
works?
JG: "Running a wastewater treatment
plant is challenging. Unlike most indus-
tries that operate under stable condi-
