WC130 MayJune2023 - Magazine - Page 10
INNOVATION
Here in Canada, applications currently being
explored within the water industry include uses
aimed at improving municipal decision-making with
respect to municipal asset management plans (AMP)
as well as infrastructure operation and maintenance.
Advances are being made in the use of ML models to
monitor the operational parameters of key infrastructure within water systems to allow for the scheduling
and planning of proactive maintenance, with the goal
of reducing the likelihood of unplanned shutdowns
due to equipment failure and the maintenance cost of
the infrastructure equipment. This would allow for a
more informed approach to asset management planning and a reduction in the operators’ risk associated
with equipment failure.
Well-considered
Consider the typical approach for municipal well
maintenance. Planning for the work to be done is
only started once a decline in the efficiency of the well
has been noted. Of course, this usually results in a
delay between when the asset should have rehabilita-
10
WATER C AN ADA • M AY/JUNE 2023
tion work initiated and when the work is completed.
A loss in efficiency of the well maintenance work will
occur during this delay due to a greater loss in well
efficiency, a portion of which can never be recaptured
during rehabilitation. There is currently an ongoing pilot program where artificial intelligence and
machine learning technology is utilized to monitor
the operation of a municipal water well with the goal
of predicting when well rehabilitation would need to
be completed so that arrangements can be made to
complete the work proactively before the decrease in
well efficiency has been observed. This will allow for
an increase in the useful operational life of the well by
reducing unnecessary losses in efficiency.
Flow forward
The usefulness of these models are not limited to
groundwater systems. Surface water features are
monitored extensively and an abundance of data is
collected. Conservation authorities in Ontario are
now starting to look at utilizing the data they have to
predict future trends in stream flow and water quality
WAT E R C A N A D A . N E T
