Real-time monitoring of the microbiological quality of drinking water is opening up new possibilities for water suppliers. Meanwhile, enhanced data capabilities are transforming utility management. Torben Lund Skovhus and Bo Højris, authors of a new book on microbiological sensors, call for a rethink on the opportunities for drinking water quality monitoring as part of the digital future of water utilities.
All drinking water suppliers aim to provide clean, safe drinking water. However, they face an ongoing battle against pollution and degrading water quality. Suppliers must combat this with measures such as source protection, continuous risk assessment, and monitoring.
Unfortunately, people still do become ill from drinking water that was supposedly safe, and often this occurs before suppliers are aware of any problem. This happens for two main reasons: laboratory analyses of routine spot samples are time consuming, and many microorganisms cannot be cultured in the laboratory despite being quite viable in their natural habitat. This means it takes longer to obtain the results of analyses than it does for water to pass through the drinking water distribution system.
So utilities often have to rely on the consumer to report illness, and today that can mean the public raising concerns on social media.
What utilities need is fast and accurate online methods for real-time monitoring, providing data to help identify changes in water quality. In fact, while they may not be well known yet, such methods are already available.
The rise of sensor technology
Microbiological sensor technology is evolving rapidly, and the practice of real-time microbiological water quality monitoring is therefore becoming more common, particularly in the case of bacteria. Viruses, and parasites such as cryptosporidium, present challenges for application of sensors because, for example, of the small size of viruses and the low incidence of parasites required for infection to occur. These parameters, and the use of sensors, therefore fit within a wider approach to water quality management.
Even so, a gap is emerging between traditional approaches to microbiological water quality monitoring, operations and regulation – which have significant limitations – and what is now possible. Not only this, but the range of locations where sensor technologies can potentially be applied is also developing, all the way from the source, through the point of abstraction, production and the supply network, to the point of use.
This evolution is taking place against a backdrop of changes for water suppliers. For example, treatment processes are being developed to use less energy and fewer chemicals and still achieve the required water quality. Operating these processes requires close monitoring of the source waters and the treatment itself.
Another key issue is that ageing water distribution systems in industrialised countries require retrofitting and management. This needs to be informed by monitoring of the actual performance. Also, centralised systems are being augmented with diverse decentralised systems producing potable and non-potable water, as a result of which microbial water safety must be addressed.
Current developments in online monitoring of microbiological water quality are adding new aspects to operators’ improvement of safety and operations by integrating monitoring into vulnerability assessments and Water Safety Plans. These are part of a trend towards viewing the supply network holistically, as the complexity of managing and operating water production and treatment grows.
Alongside all of this, there is a growing realisation that water utilities need to develop digital strategies. So there is a need for an industry wide exchange of experience and thinking to pave the way for wider use and integration of microbiological sensor technology.
Utilities face other issues when looking to apply sensor technologies. For example, more challenging sources of water are being used in situations where resources are scarce or stressed. Such sources require the use of more advanced treatment and closer process control.
The task of supplying water is also becoming increasingly complex. Meeting consumer expectations for safe drinking water means developing a robust, multibarrier approach where monitoring will be crucial. Water quality and public health risks need to be analysed, managed and minimised, which means understanding raw water quality and the catchment, as well as treatment processes. Online monitoring enables utilities to provide safety and respond to incidents, both accidental and intentional.
There are also sensor related challenges: the market is not yet mature, and each sensor has its own advantages and disadvantages. Total cost of ownership depends on sampling frequency, the amount of data produced, and the use of chemicals. Also, sensors create vast amounts of data. This is not easy to interpret – and this is even before considering how to feed it into decision-making and to implement steps needed to mitigate a problem.
Sensors have the potential to follow water quality changes through a distribution network, but this will only become viable when relatively cheap and near maintenance-free sensors are available. Sensors can currently detect specific microorganisms; measuring different microorganisms online and simultaneously at very low levels is desirable, but not yet possible.
It is also difficult to compare the output of one sensor with another, even when they target the same parameter. This is because of the differing measurement principles used. The challenge of comparing results with laboratory methods also makes it difficult for sensors to be accepted as a replacement.
Achieving progress in the age of big data
The rise of big data, including analytics, is a key parallel development. Wide use of online sensors will generate much larger volumes of data. That data will need to be processed. Aspects such as regulatory requirements, safety issues, event detection, and operational control and improvement will all shape that processing. So implementing a successful monitoring strategy will depend on how well specified objectives and aims are addressed.
There are various questions that must be answered when deciding how to move to online systems and connect what the sensors provide with the interpretation and decisions supported by big data. These include the role of parameters such as turbidity and conductivity, and indicator microorganisms. Turbidity and conductivity can be measured in seconds, while trials of microbiological sensors have found a trade-off between specificity and speed. So, if a sensor uses particle counts as a proxy for microbial contamination, it is important to understand the conditions for which this proxy is valid, and when it is not.
Machine learning approaches (artificial intelligence) will also be able to help flag outliers in datasets or to identify unexpected behaviour. This is needed to create trustworthy data streams and databases. However, such an application requires an understanding both of processes and sensors.
Initially, there is potential for a general approach, monitoring total microbial activity in a treatment works, where sensor values can be linked to data from a SCADA system. Automated batch sampling can now deliver total bacteria counts in minutes, and can be installed, connected and compared directly online through a common interface.
Ultimately the best solution may not involve a single sensor that meets every requirement; instead, it might make sense to have a division of duties into event alarm, automatic sampling, and laboratory characterisation. In such a system, a sensor, or two sensors based on different principles, would provide the event alarm, while advanced automatic samplers combined with laboratory analyses would give further characterisation. With this diverse approach, all contamination events can be detected, whether sudden or slow, and threats mitigated to secure consumer water quality.
Sensors by themselves will not be enough to improve water systems – this will require innovative engineers, responsible operators, and engaged regulators to jointly explore and advance future water systems. Combining sensors with big data analysis will help to make progress possible.
Education has an important role to play in this progress too, ensuring, for example, that staff at water supply and wastewater utilities appreciate the potential for microbiological contamination of supplies – as a result of operations and maintenance activities. The addition of sensors can help by providing monitoring. Appropriately trained staff can then complement this by being better able to interpret information gathered through the sensors.
The time has come to share the progress that is being made in the field of microbiological sensors for use in drinking water quality applications, and to develop a common understanding of – and approach to – their use in the increasingly digital future of water utilities.
Dr Torben Lund Skovhus is associate professor and project manager at VIA University College, Horsens, Denmark, in the Centre of Applied Research and Development in Building, Energy, Water and Climate.
Dr Bo Højris is a senior researcher at Grundfos Holding A/S, Bjerringbro, Denmark, in the Global Research and Technology Department.
Sensors support security in the Antwerp region
Provinciale en Intercommunale Drinkwater-maatschappij der Provincie Antwerpen (Pidpa) is one of the biggest Flemish drinking water companies. It treats deep groundwater for almost 500,000 customers – more than one million people – in the Antwerp region of Belgium. The utility considers microbiological sensors to be complementary to traditional approaches, because, currently, one type of sensor cannot fulfil all the legislative requirements.
Pidpa uses a combination of chlorine and UV disinfection to provide a disinfection barrier to make drinking water safe. However, the disinfection dose was so low that it would not solve a major contamination event, and gave a false sense of security, masking potential failures in the production and distribution systems.
After removing chlorine dosing units, Pidpa decided to install online sensors for total particle and bacterial counts at the entrance to its 12,000km distribution network and several of its 62 water towers, to detect possible treatment process failures.
In this example, sensors provide operational control and early warning to ensure operational excellence in future.
New from IWA Publishing
Microbiological sensors for the drinking water industry
Torben Lund Skovhus and Bo Højris (Eds)
This 300-page book, published by IWA Publishing, contains 16 chapters, including the history of water monitoring and early sensors, the current landscape of microbiological sensors and measuring concepts, and today’s water needs. It also describes several state of the art technologies in detail and gives case studies and examples of online data collection and handling. The book looks at microbiological sensors applied in education and industry workshops, and regulatory aspects of their introduction.
This book is aimed at water utility managers, technical staff working on drinking water safety, educators, and regulators around the world. Four chapters are available open source.