The Problem: Ineffective Water Treatment
The American Society of Civil Engineers released a report card that grades America’s infrastructure. In that report, drinking water and wastewater infrastructure earned grades of D and D+, respectively.
Of particular note is that, out of 16,000 wastewater treatment plants in the U.S., on average, they’re using about 81% of their design capacities, while a mere 15% have either reached or exceeded it. As a whole, they’re not functioning optimally.
The same is happening with drinking water, with about 6 billion gallons of treated water lost in the U.S. per day because of water main breaks. Those breaks occur in the up to 2.2 million miles of underground piping that delivers fresh water to the country’s residents. The water lost could also fill over 9,000 swimming pools.
That’s a lot of water a lot of waste – and a big problem that needs to be addressed. The question is, how do we identify and fix these issues, and where do we get the manual power to do so?
The Solution: Adopting AI in Industrial Water Treatment
Without delving too far into the basics, AI and machine learning technologies use data, by ingesting and processing the information it contains, to take action. That action is empowered by trends and patterns that exist within the data, which can be scoured to reveal useful and practical insights about an operation, process, event, or task.
For example, in water treatment, an IoT-enabled sensor (connected to the internet or a local network) can detect when there is a leak or water main break and shut off the water until it can be addressed. The system would learn based on prior leakage events, eventually knowing what to identify, like water pressure changes, temperature changes and so on.
It would also involve several interconnected technologies that allow the sensor to communicate with the control system so the water shuts off. The final stage is automatically flagging the problem by sending alerts and notifications to the appropriate parties, including information about precisely where the leak occurred.
Even from this quick example, you can start to see how AI and machine learning technologies – with the help of IoT sensors – can vastly improve water-based infrastructure. And these are not conceptual ideas. Products such as the water leak sensor in the aforementioned example already exist. Some consumer-grade examples include Flo by Moen, D-Link’s DCH-S161 Wi-Fi water sensor, Flume 2, Phyn’s smart water sensor, and many more.
While these devices probably wouldn’t be as successful in industrial settings, it shows that it is completely possible to create these systems – and we are already fully capable of doing so.
How Do We Make It Work?
All of that sounds great in theory, but how do we make AI in industrial water treatment a reality? What kinds of things can the technology do for the field?
The first step is to identify the true cause of the problem, which in this case is due to highly underdeveloped infrastructure or just dated developments, in general. When most of the U.S. drinking and wastewater infrastructure was built, we didn’t have the technologies we do now. Of course, the industry has evolved over the years, and there are certainly some advanced implementations out there in the wild, and already fully operational, but they are not the norm, which is the major issue.
That can be remedied now by implementing these technologies and principles in modern infrastructure. Honoring the core principles of eco-engineering is a great start and fits wastewater treatment processes perfectly. But we can also do better by improving regular operations with smarter, more contextually aware solutions.
AI in Industrial Water Treatment: Changing the Game
Implementing the type of predictive maintenance and smart data solutions that AI and machine learning provide, within water and wastewater plants, would change the game. The possibilities include preventive maintenance before major events or disruptions, fully automated and intelligent control systems, minimal worker injuries and improved safety, smarter and more optimized operations, and cost savings across the board – especially when waste is reduced.
Reducing waste is one of the major selling points of this technology. Whether it involves identifying and addressing leaks early for freshwater solutions or more effective treatments for wastewater, AI-driven technologies can make it happen. The technology can monitor and detect toxic chemicals and organisms by adjusting treatments and optimizing flow cytometry.
AI tools can also be used for infrastructure cleaning and maintenance tasks, such as removing small silica particles from pipes. When silica builds up inside pipes, it can have adverse effects on a host of elements, including machinery, water treatment performance, and so on.
In boilers, a single millimeter of silica can increase the fuel consumption of machines and equipment, in turn elevating costs. Just imagine what it can do to wastewater or freshwater treatment operations. Smart pipe sensors can identify it, adjust pipe routes accordingly so it can be cleaned, or even flag infrastructure that needs to be replaced.
Industry on the Precipice
The final piece of the puzzle is to discuss when AI in industrial water treatment solutions might be accessible to operators in the industry.
We’re already there. Besides the consumer-grade devices used to detect leaks and other anomalies in home infrastructure, these technologies are already being adopted within real-world industrial operations.
Severn Trent, in the U.K., will soon leverage AI and machine learning to automate water catchments. Academic studies are exploring the idea of real-time modeling and predictive controls in wastewater treatment using machine learning. Another wastewater treatment plant in Germany began using a proprietary AI platform to analyze the plant’s data via their existing SCADA system, which led to a reduction in aeration energy usage by as much as 30%.
The ultimate goal is to upgrade industrial wastewater treatment operations and facilities to reduce waste, lower operating costs, and improve results in nearly every way – including by getting fresh water to those who need it most.
