Digital water:
From smart water networks to AI

Digitalization can affect practically every process that a utility uses. Planning, monitoring and management tasks, especially, are often already partly or fully digitalized.

There is a range of opportunities, as diverse as the technologies on the market. Today, it is relatively simple to connect fire hydrants to the utility's cloud, or to minimize leakage loss by using AI models.

Given the challenges awaiting the water sector in the coming decades, the speed and effectiveness of the digital transformation is an important topic.

Not everything can be digitalized, of course. But with advanced technologies such as digital twins, water utilities can come close to doing just that. One major precondition for this is a supply of precise and current data.

Modern, connected sensors are a basic requirement for the digitalization in the water sector. They also provide more precise and up-to-date data than older generations of sensor, let alone manual meter reading. The stream of remote data allows utilities to simplify a number of common management tasks and business processes - or even to automate them completely. 

Smart water meters, for instance, register even tiny flows, hamper manipulation and make meter reading superfluous. This gives suppliers a much better idea of how much water is used where in a given supply zone, and when. 

This is useful for better decision-making, but also allows precise, transparent billing. Utilities can enable their customers to track their own water use nearly in real time. There is another aspect to this. If the water provider can warn customers when water use suddenly jumps, they don't have to learn about it from the water bill, or damage to their home.

What makes this possible are not just advances in modern sensor technology, but especially in the development of IoT. This has resulted in extremely efficient devices that don’t just collect data for years on a single battery charge, but also transfer them using mobile networks.

An additional driver of this development is the spread of 5G and of power-saving communication standards, such as NB-IoT or LoRaWAN. The smart city trend is another enabler.

This makes IoT system very useful for the remote monitoring of water supply networks. Use cases include:

• Water level and flow
• Water meters
• Flooding alerts
• Temperature and pressure
• Monitoring of valves, e.g., control valves or air valves
• Access monitoring, e.g., with a door sensor
• Leak detection using noise loggers
• Water quality monitoring with a spectrophotometer
• Hydrant monitoring.

The assets of water utilities are literally all over the map. This makes geographical data particularly interesting for them. A popular application are interactive maps - with the added benefit of being accessible everywhere.

Web-based and mobile map apps, such as the Hawle.live APP, make it easy to visualize entire pipe networks. This lets operators see the position and condition of hydrants, control valves, gate valves and other important infrastructure at any time.

Besides maps, smartphones and mobile apps offer a number of other useful functionalities. Before leaving and while on the road, service staff can access maintenance records and product information (e.g., the maximum torque for a gate valve), record test results or access data from Hawle.live BOX sensors.

One particular use case are fire hydrants. They are part of the emergency infrastructure and connect to the drinking water network. Thus, they have to be ready for use at all times, but can't become a source of contamination. 

They are usually very accessible. This makes them somewhat liable to being hit by cars - and vulnerable to water thieves. Illegal hydrant use can damage them and contaminate the water, as can accidents. Both theft and damage may not be spotted immediately. The answer: IoT-enabled smart hydrants. Even existing hydrants can easily be upgraded with a specially designed cap. Either way, the hydrants can now be connected to the utility's management software, along with metadata such as GPS location or serial number.

Data on assets are not just useful in the field, such as when an air valve proves hard to find without precise coordinates. Whether piping, pump or water tank: There is a trend towards detailed records, real-time monitoring and modeling. 

To fulfill this need, manufacturers offer digital 3D models of their products, which planners can embed in their CAD or BIM systems. This makes designing facilities or pipe networks easier.

The information included in such a product model can range from the pure geometry over complex material characteristics to simulations of its function. If a pump is described in enough detail, then it’s possible to automatically dimension intake pipes, control valves or even mounts.

Infrastructure in the utility sectors are typically very long-lived. Thus, digitalization often means upgrading and retrofitting existing assets.

Luckily, thanks to increasingly mature technology, this is becoming a plug & play exercise. Ultrasound flow meters can easily be clamped onto pipes. Valves are automated with power-saving electric motors. Battery-powered noise loggers are dropped into existing shafts.

Digitalization depends on data and generates huge amounts of it. Thanks to machine learning and AI, utilities can put it to good use. 

By feeding ever larger volumes of data into machine learning models, operators get ever more realistic simulations of facilities and processes, all the way to digital twins. This technology is particularly useful for managing complex and critically important facilities such as water treatment plants. 

Built and trained on real data, digital twins behave exactly like their physical counterparts: from the digital water over realistic sensor behavior to the simulated control devices. 

One key advantage of digital twins is that they can be used to make very reliable predictions. Additionally, they allow operators to try experiments that would not be possible in the live facility. Digital twins are also very useful for staff training. 

Simpler models of this kind can be used to predict what condition the various parts of a water supply network are in. This allows operators to replace fixed maintenance intervals with more efficient predictive maintenance. 

The idea is to detect problems before they become obvious. At the same time, this prevents maintenance or costly replacements where they are not needed. For example, these models can predict which sections of a pipe network are the most likely to suffer breaks - and which pipes, though elderly, won't need replacing for years to come.

We are witnessing the digital transformation of the water sector: Digital processes and tools are now firmly established, whether they serve to create a better customer experience, enable pinpoint-precise maintenance or simplify facility planning. 

The use of IoT sensor technology, GIS applications, mobile apps and AI models is increasing: Water utilities are preparing for the future with a growing arsenal of smart technologies. The next trend may already be around the corner.