water temperature, turbidity, conductivity, colour, pH, etc.) is not currently
collected in a very consistent and automated manner across networks, and
these sensor technologies (and their price) have moved little in the last decade.
New spectroscopic methods are now coming into use (see chapter “Spectroscopic
Methods for Online Water Quality Monitoring”). Water quality measurements can
help to identify discolouration events and monitor chlorine residuals (see chapter
“Using Radial Basis Function for Water Quality Events Detection”). As the technology
improves and the whole life cost of ownership falls, water quality instruments could be
used for the provision of operational data and become as widespread as for flow
metering. Water quality data analytics are relatively undeveloped because there are
limited significant deployments of real-time water quality monitoring within networks
[22]. Future technologies promise enhanced sensors for natural biological/biochemical
markers. Interpretation and analysis across multiple online parameters is expected to
provide deeper understanding of water supply system state and asset condition.
The proliferation and diminishing costs of automated data transfer, such as
by GPRS, 4G/5G, Wi-Fi, LoRa/LoRaWAN and Sigfox systems, are allowing all
types of recorded data to be transferred from many disparate points on the networks.
New developments and products in the fields of micro-sensors, alternative power
supply and wireless telecoms all revolutionise the whole domain of real-time
monitoring and consequently real-time management. It is easy to anticipate that
the environment may quite soon be teeming with tens of thousands of small,
low-power, wireless sensors. Each of these devices will produce a stream of data,
and those streams will need to be monitored and combined to detect interesting
changes in the environment. The emergent properties of data from networks
of simple, low-cost sensors will be increasingly fruitfully explored.
IoT will be dominated in the coming years by the growth of edge computing
systems as roll-outs of the technology in the field become more complex and larger
scale. Edge computing represents cutting-edge hardware and software co-located
at IoT endpoints that make the technology far more efficient, scalable, secure and
manageable. The presence of on-board CPUs can form basic analytical functions
(e.g. which data to send to customer’s smartphone apps, which data to send to the
WSP and when it should be sent), combined with on-board data storage that would
really make devices smart, and not just a measurement and communication device.
It will also make IoT solutions much smarter by enabling the deployment
of machine learning and AI capabilities much nearer the point-of-use that will
enrich and optimise what is possible. Edge computing will enable AI applications
in scenarios where processing is better performed locally. Advances in sensors
and low-power computing architectures will enable edge computing with highperformance, real-time and increasingly complex AI solutions. This has promise
to yield huge cost savings for remote locations with limited or expensive Internet
connectivity and to reduce the overhead of centralised processing. Traditionally,
in IoT, the predictive analytics have been done by analysing the data in the cloud.
However, that may not always be possible with large amounts of streaming
data arriving from the edge devices. Making edge devices smarter will thus be
critical. The prediction rules may be discovered a priori in the cloud, and
these lightweight rules can then be deployed on the edge devices.
Data Science Trends and Opportunities for Smart Water Utilities
11
collected in a very consistent and automated manner across networks, and
these sensor technologies (and their price) have moved little in the last decade.
New spectroscopic methods are now coming into use (see chapter “Spectroscopic
Methods for Online Water Quality Monitoring”). Water quality measurements can
help to identify discolouration events and monitor chlorine residuals (see chapter
“Using Radial Basis Function for Water Quality Events Detection”). As the technology
improves and the whole life cost of ownership falls, water quality instruments could be
used for the provision of operational data and become as widespread as for flow
metering. Water quality data analytics are relatively undeveloped because there are
limited significant deployments of real-time water quality monitoring within networks
[22]. Future technologies promise enhanced sensors for natural biological/biochemical
markers. Interpretation and analysis across multiple online parameters is expected to
provide deeper understanding of water supply system state and asset condition.
The proliferation and diminishing costs of automated data transfer, such as
by GPRS, 4G/5G, Wi-Fi, LoRa/LoRaWAN and Sigfox systems, are allowing all
types of recorded data to be transferred from many disparate points on the networks.
New developments and products in the fields of micro-sensors, alternative power
supply and wireless telecoms all revolutionise the whole domain of real-time
monitoring and consequently real-time management. It is easy to anticipate that
the environment may quite soon be teeming with tens of thousands of small,
low-power, wireless sensors. Each of these devices will produce a stream of data,
and those streams will need to be monitored and combined to detect interesting
changes in the environment. The emergent properties of data from networks
of simple, low-cost sensors will be increasingly fruitfully explored.
IoT will be dominated in the coming years by the growth of edge computing
systems as roll-outs of the technology in the field become more complex and larger
scale. Edge computing represents cutting-edge hardware and software co-located
at IoT endpoints that make the technology far more efficient, scalable, secure and
manageable. The presence of on-board CPUs can form basic analytical functions
(e.g. which data to send to customer’s smartphone apps, which data to send to the
WSP and when it should be sent), combined with on-board data storage that would
really make devices smart, and not just a measurement and communication device.
It will also make IoT solutions much smarter by enabling the deployment
of machine learning and AI capabilities much nearer the point-of-use that will
enrich and optimise what is possible. Edge computing will enable AI applications
in scenarios where processing is better performed locally. Advances in sensors
and low-power computing architectures will enable edge computing with highperformance, real-time and increasingly complex AI solutions. This has promise
to yield huge cost savings for remote locations with limited or expensive Internet
connectivity and to reduce the overhead of centralised processing. Traditionally,
in IoT, the predictive analytics have been done by analysing the data in the cloud.
However, that may not always be possible with large amounts of streaming
data arriving from the edge devices. Making edge devices smarter will thus be
critical. The prediction rules may be discovered a priori in the cloud, and
these lightweight rules can then be deployed on the edge devices.
Data Science Trends and Opportunities for Smart Water Utilities
11