Artificial intelligence (AI) is the hot topic that no one can seem to escape right now. As AI technology has progressed in leaps and bounds within the past few years alone, it has unlocked new potential for its use, as AI’s ability to process and interpret data far exceeds what previous technologies have been able to do.

But don’t think AI applies only to the tech industry or start-ups. While some may be rightfully apprehensive about AI’s quick adoption, the water-treatment industry has already found efficient and cost-saving ways to utilize AI: Precision engineering programs, accurate facility-simulation projections, and many more applications are already in use.

What Is AI?
When looking for a concrete definition for this technology, I asked the source itself. I queried ChatGPT (chat.openai.com), a popular AI “chatbot,” for a short definition of AI. It said:

  • Artificial intelligence (AI) is the simulation of human intelligence processes by machines, including learning, reasoning, and problem-solving. It encompasses various technologies like machine learning and natural language processing, aiming to enable computers to perform tasks typically requiring human intelligence, advancing automation, and problem-solving capabilities.

Succinct and accurate. There are many methods for designing an AI program, but once each program is coded, it is then trained through an assortment of assessments that gauge its ability to perform its designated tasks. Many AI programs are trained to use complex algorithms to find patterns in large data sets and learn to make predictions based on the information. Applications like ChatGPT may be designed to respond to questions, or they may be utilized to synthesize more complex data.

AI in a Wide Variety of Water Applications
AI has proven to be helpful in water use, conservation, and distribution. Several municipalities and agencies use AI to monitor and predict maintenance needs before a catastrophe happens. Rifat Alam, business lead at Texas Water Resources, explained that AI can monitor sensors that notify agencies of a rise in contaminants or other dangers, such as an algae bloom.1 Floods can be more accurately predicted. For example, Google Research’s FloodHub model is the most talked-about flood prediction AI-based model.

Another way AI can be important is in the optimization of water usage. AI can be a critical tool in redistributing water from low-use to high-use areas, forecasting usage needs, or making irrigation more efficient. One example is PlutoShift’s AI platform, which uses real-time data to optimize water usage, anticipate and prepare for potential water shortages or excesses, reduce costs, and build sustainability.3

AI in Wastewater
In addition to monitoring for leaks and keeping tabs on water sensors, AI has proven useful in helping to reduce the cost of operations and human error. AI technology can be applied to source-water quality determination, coagulation/flocculation, disinfection, membrane filtration, desalination, modeling wastewater treatment plants, prediction of membrane fouling, removal of heavy metals, and monitoring of biological oxygen demand and chemical oxygen demand levels.2

Being able to proactively perform maintenance to avoid costly leaks or manage membrane fouling improves worker efficiency. AI can also optimize chemical dosing. Instead of manual testing and adjustments, AI continually monitors certain parameters and can adjust dosages in real time.3

AI can also increase the reuse of wastewater, helping to conserve resources and reduce environmental impact. Accurately predicting the quality of treated wastewater allows for wider reuse in irrigation, industrial processes, and drinking water.3

Digital Twins
To better prepare for unexpected scenarios, many treatment facilities have adopted the use of digital twins. Digital twins offer safe, controlled testing—essentially, they are simulations. Water-treatment facilities that utilize AI can accurately simulate environments and scenarios without the potential harm or drawbacks of physical experimentation. Digital facility replicas can accurately simulate input conditions and make predictions based on that information. They sometimes utilize live data streams (like supervisory control and data acquisition, Internet of things, and automated meter reading/advanced metering infrastructure) in their predictive models alongside designated variables, such as power outages, pipeline breaks, and other emergency conditions, to test the system’s responses.

The American Water Works Association (AWWA) has been publishing technical resources for the use of digital twins since 2021 and continues to provide more materials as AI advances. The most recent text provided by AWWA is about using digital twins to improve pumping and distribution system operations.4

AI Against PFAS and Other Contaminants
For years, AI has been used in the fight against contamination by per- and polyfluoroalkyl substances (PFAS). There have been many efforts to lower the barriers of cost and complexity surrounding the testing of these forever chemicals, and AI’s potential for producing a solution may be on the horizon.5 AI could scrub thousands of molecules in a sampling to locate PFAS molecules that could then be targeted by sorbents for removal.

Research has shown that real-time detection and responses to PFAS contamination is feasible with AI technology. In a recent study by the Centre for Technology in Water and Wastewater at the University of Technology Sydney, Australia, scientists successfully used AI technology to respond to PFAS materials in wastewater systems.

According to the study report, the “modeling provided accurate adsorption capacity, equilibrium, and removal estimates with the ability to predict the adsorption mechanisms.”6 Based on the information presented, the model predicted that pH measurement was the most effective parameter to predict PFAS removal. Although this was not the first study developed to determine PFAS removal methods, it was noted as the first to have PFAS and environmental conditions considered through its machine-learning programs. The findings suggest that this method of using AI could be used to create a similar framework to address other micropollutants in the wastewater stream.

Engineering Solutions with AI
A subfield of AI called generative design has been embraced by engineers for its abilities to help produce solutions to design problems. This design method allows engineers to set up chosen parameters (such as cost, performance, and/or restrictions) to create a solution, and the program then produces several designs as options for the operators to use in their projects.

Chosen designs can undergo further rounds of iteration until the specified conditions are met. This process allows for rapid prototyping before any physical options are manufactured, as many of the expected constraints or setbacks are already accounted for.

AI for Customers
Customers routinely engage with AI tools, sometimes without realizing it. Chatbots are frequently used as customer service tools to engage with customers who want to learn more about a company’s products or services.

Tech giants like IBM and Microsoft offer AI programs that integrate into customer support systems on websites. Some have even developed advanced voice features to better interact with customers over the phone. These tools can provide off-hours support, create support tickets if problems are not easily rectified, and even generate sales by offering products based on customer needs that have been identified in conversations. Since many customers are weary of nonhuman interaction, integrating a sales team into the final steps of engagement will help to maintain the expected levels of human interaction.

Job Security in a World with AI
In a business landscape with more and more AI applications, some may fear that increased use of AI may lead to job displacements in the water-treatment industry. With the advent of new and groundbreaking technology, there is naturally some pushback from those who may not fully understand the technology or its capabilities, or from those who fear AI could become entirely autonomous.

For now, though, human operators or specialists are involved throughout the process when AI is used. Digital twins are simulations run by water-treatment operators, generative design is utilized by engineers to streamline solutions testing, and consumers will need assistance installing water-treatment solutions, even if some product sales are generated by conversational AI tools.

The human element in these AI applications is currently required for the water industry; there is no AI tool that can be operated without an operator. There is no AI tool that can replace a lifetime of experience in a water-treatment plant, the physical knowledge of an installation technician, or the savvy of a water-business manager to adapt the tools to the company’s or municipality’s specific needs.

There are ethical questions that have arisen regarding the use of AI for artistic creation (such as to generate works of art and now even realistic video), and conversations exploring the regulation and obligations of using AI must be had. But for the water industry, you don’t need an algorithm to see that AI is already here. The water-treatment field has been using this technology for some time. And as AI advances, its use as another tool to bring clean and safe drinking water to people everywhere will increase.

References

  • WSP. “Five Ways Artificial Intelligence Is Going to Shape the Future of Water and Resilient Infrastructure,” published March 21, 2023. https://www.wsp.com/en-us/insights/2023-artificial-intelligence-shaping-future-of-water
  • Soma Safeer, et al. “A Review of Artificial Intelligence in Water Purification and Wastewater Treatment: Recent Advancements,” Journal of Water Process Engineering 49 (October 2022). https://doi.org/10.1016/j.jwpe.2022.102974
  • David Cain. “Water Management Enhanced by AI,” LinkedIn, published August 5, 2023. https://www.linkedin.com/pulse/making-splash-how-ai-diving-water
    -management-david-cain/
  • American Water Works Association. “Digital Twins.” https://www.awwa.org/Resources-Tools/Resource-Topics/Digital-Twins
  • Jesper Bruun Petersen. “Artificial Intelligence to Help Remove PFAS,” AU Engineering, Aarhus University, published November 29, 2023. https://ingenioer.au.dk/en
    /current/news/view/artikel/artificial
    -intelligence-to-help-remove-pfas
  • Elika Karbassiyazdi, et al. “XGBoost Model as an Efficient Machine Learning Approach for PFAS Removal: Effects of Material Characteristics and Operation Conditions,” Environmental Research 215, no. 1 (December 2022). https://doi.org/10.1016/j.envres.2022.114286

About the author Keller O’Leary is managing editor at Water Conditioning & Purification International magazine.

Share.

Comments are closed.