Seasons: A Case Study of Using AI in a Water Treatment Plant

This case study is really exciting! This one is really exciting omg wow! This case study looks at a specific water treatment facility in Drayton Valley, Canada. This water treatment plant wants to both reduce their environmental impact and save some money. This one is going to be a bit of a harder read for me as it is less structured than most of the other articles (neurodiversity!) but let me do my best please - thanks.

One huge gap that I can see in water technology is technology in water treatment facilities. 

I am not sure just how much this is being done in the UK at the moment but as many of my previous blogs say, and as all the articles I am reading at the moment seem to suggest, AI in water treatment is one of the biggest opportunities we have for the technology. The main things that stand out are obviously leaks and leak prevention, maintenance of infrastructure, optimising water flow, balancing supply and demand, but then water treatment really does stand out as a front runner to me. 

There are so many aspects to it, ones that I can’t even begin to remember. Air quality is one that just blows my mind. But improving energy efficiency is a huge one too. And then monitoring water quality is a huge one too. 

Remember seasons. And weather. And changes in pollutants. 

And the introduction of new pollutants into the water sphere. And to be honest there is so much to revisit in my precious water articles that is now only just beginning to make sense. But let me continue with the other articles for now. Including this new one. Before I try and revisit it all collectively…

Continuing with this article 

One of the responsibilities that water treatment facilities face is that of providing clean drinking water to people.

This can include filtration and filtering out from the water system all of the things that we don’t want in it to drink - bacteria, pharmaceuticals.

The article then goes on to describe the contents of the water filtration system in extremely advanced technical detail. Basically, it says that if the wrong dose of a water filtration chemical is used, then this can damage the hardware and basically the whole machinery that they are using for the water. Oh sorry it’s more filters that can get clogged - by too much use of coagulants - but these can be expensive to clean and to replace. As “raw water conditions” are “constantly changing”, calculating this dosage can be quite hard for humans to get right. Okay I wonder if we are heading here in the direction which I think we are heading in…

Basically in the place in Canada they are piloting whether AI can be used in calculating how much dosage of coagulants to use. Hmmm, why do I find this so exciting.

This process aims to produce cleaner and more cost-effective drinking water

Right there is a lot going on here. Sorry. A bit overwhelmed. I am just trying to process this article and I am doing the best I can. It is a different format to the ones that I am used to reading. Anyway… yes.

Basically this pilot in Alberta test where AI can help to make the plant’s ultrafiltration system better. So we use optical sensors. Okay. What do these optical sensors do please? Thanks. These optical sensors…

• Send water quality data to the AI system
• And then the learning algorithm 
• Adjusts different things about the water filtration process

What does it adjust please? Thanks.

It adjusts:

• Coagulant dosage
• Pumping cycles
• Cleaning cycles

The goal is to:

Minimise the usage of:

• Water
• Energy 
• And chemicals

Interesting that one. I am pretty much obsessed with the reduction of energy consumption in the water filtration process.

I am also fascinated by the fact that in itself, what gets used a lot in the energy consumption process, is, well… water. As the case study is initially in a micro area of the plant - the goal is eventually to apply this to the rest of the body plant as well. It requires a combination of water system techniques and knowledge as well as AI expertise as well. Anyway so how does the pilot system work? In real time.

AI in real time 

If there was one thing I could say about AI and its role in water and renewable energy and all utilities - in real time. That’s what I would say. In real time. AI allows us to make decisions in real time about how we manage and run our water systems. How we manage and run our water treatment plants. How we manage and run our renewable energy. So in this pilot system. How does this apply? Well… in this pilot system 

In this pilot system AI helps the water filtration system to adjust in real time. What does it adjust to?

• It adjusts in real time the precise amount of coagulants that are needed for the water filtration system
• It optimises the frequency of the membrane cleaning
• It aims to use less water 
• And use less energy

It aims to provide clean drinking water that is more cost effective. And it hopes to increase the cleanliness of the water as well. Omg and it gets more exciting from here as well!!!! Wow! I have just read something really cool and I need to calm down.

AI and 24-hour monitoring data 

The AI has access to 24 hour data from the ultrafiltration system.

What does this mean. 

It means that the algorithms have access to data that can help it to predict seasonal changes in raw water quality. I repeat: the algorithm has access to data that can help it to predict seasonal changes in raw water quality. What does this mean? Sorry I am struggling a bit here not to lose it and get too excited. I absolutely love seasonality in water quality and water control. The raw water quality that the treatment plant gets from the North Saskatchewan River will change according to the seasonal changes.

Examples of seasonal changes in the North Saskatchewan River

• In the early spring
• Water is fresh and clear and blue
• Like water just straight out of the Rockies
• Later in spring however 
• More organic content can come with the meltwater
• And this can turn the river brown

The biggest challenge for the project is dealing with ever-changing water conditions.

Wildfires

Wildfires are of an increasing concern in the area of Canada where this research takes place. The debris can fall into watersheds and thus contaminate the area’s drinking water. AI can help to adapt to the changes in water quality for water treatment in real time.

Access to clean drinking water

Access to fresh drinking water can be a critical for some municipalities in Canada. But the savings costs of AIs contributions in water technology could be a game changer. For example, reducing energy costs in water treatment processes could save a lot of money. And you know energy as well. This means that the future will be more sustainable. And more resilient. We will have more access to clean, safe drinking water and this can only be an extremely good thing.

Closing Thoughts: Summary

This article is mega complicated. I could see myself needing to revisit this article or at least this blog post. I also see myself needing to revisit my last water tech blog post as I haven’t revisited it yet. But that one was quite similar to the very first one I did. Only really new aspect was water and water risks.

But this one feels completely new. And fresh.

Maybe I should prioritise revisiting this one first please - and then just focus on the risk factor in article 2 if I need and want to. Thanks. Thank you. Read on.

Source:

• This article