Responding to Larger and More Critical Patterns In Water Use with AI

In my last few blog posts, we have explored what "critical" and "larger" patterns means in water, and now I would like to look at how AI can solve them.

As a reminder - what are critical and larger patterns in AI again? I will discuss this a little bit more below again, thanks.

How AI Can Help

So how can AI help to solve these problems?

AI can help to solve these problems by being trained to consider weather, seasonality and other broad factors. 

We have looked into what weather and seasonal factors are in water tech, and we have looked at what "other broad factors mean." While I could do with revising these, the ones that always jump out at me are demographic patterns - population growth, social changes, industry, economic changes, agriculture. Also things like climate change consequences.

So let's review the whole concept in full - as one sentence sent me into over 5 blog posts' worth of research and I could have easily done many many more if I had not pulled myself in just a little bit. This was the quote that did it all for me: 

"When AI models are trained to also consider weather, seasonality and other broad factors, they can help utility and other government leaders identify and respond to larger and more critical patterns in water use" Source: SAND Technologies

So what is this quote saying? We have AI models.

We have AI models in water tech.

These AI models can be used to respond to certain types of issues:

• Emergencies
• Spikes in demand
• Drops in reservoir levels
• Sudden drops in water pressure
• Urgent water quality change alerts
• Rapid shifts in demand for water for industrial processes
• Common trends
• Long term trends in urban growth
• Such as population growth or decline
• Urbanisation
• Increases in industrial activity within urban areas
• The effects of climate change
• Increase drought
• Higher rainfall
• This can lead to bust pipes
• Groundwater depletion
• Agricultural use

These are the types of helpful situations that AI can be very helpful to responding to (when it comes to water network distribution optimisation). However, in order to help advise us to make the following decisions, AI models need to be trained on the following things:

• Seasons in water infrastructure
• Seasonal events:
• Dry spells
• Rainfalls
• Winter freezes 
• Ice melting
• Seasonal demand changes:
• Reduced or increased irrigation
• Need for cooling at home (showers, pools)
• Need for cooling in public spaces (sprinklers, misters, fountains, pools)
• Gardening
• Agriculture
• Weather in water infrastructure
• Temperature, high or low
• "Other broader factors"
• Population and demographics
• Holidays
• Environmental change
• Days of the week
• Climate change
• Water pricing
• Infrastructure and Supply
• Unusual events
• Industry demand

Once they know all about these things, then AI can help to advise on making these decisions around these larger and more critical issues.

I am still not 100% satisfied with my understanding of all of the above - most of my knowledge comes from articles but when I want to go really deep into something and there are no relevant articles I ask Chat GPT. I never learn as much though.

Hence I will try and stick to the articles.

I always learn more from them.