Eoin Lane, IBM Smarter Water Architect

Eoin Lane, IBM Smarter Water Architect

People often say that water is the new oil, but really, it’s not. Oil is a fossil fuel that takes millions of years and a lot of pressure to create.

When we burn oil – for example, by driving our cars – it is gone forever (or at least for a few more millions of years before it can be created again!). Water, on the other hand, cannot be created or destroyed (this is not strictly true, but bear with me). The same amount of water is around today that was around when the Earth was formed.

The truth is there is a lot of water on Earth – just not a lot of drinking water. Here are some facts about just how little drinkable water is available:

* 97.5 percent of all water on Earth is salt water, leaving only 2.5 percent as fresh water 

* Of that fresh water, nearly 70 percent is locked in ice 

* Most of the rest of that freshwater is in aquifers which we are draining much more quickly than the natural recharge rate 

* Two-thirds of our freshwater is used to grow food 

* With 83 million more people on Earth each year, water demand will keep going up unless we change how we use it.[1]

Margaret Catley-Carlson of the Global Water Partnership has said, ”We cannot create water, but we can manage it better, much better.” [2] Take, for example, the longest water tunnel supplying NYC: it is 85 miles long, and it leaks 35 million gallons of water every day. We need to become much smarter about how we manage this precious resource and about how we collect, analyze and use water data.

Water cycle. via Wikipedia

Water cycle. via Wikipedia

There are three ways we can become smarter about water management:  Instrumentation, Big Data analytics, and cooperation.

Instrumentation involves smart meters and sensors that take digital readings (pressure, flow etc.) and stores them in a database. The combination of smart meter data and GPS location data allows for rich visualization of the information. On top of this we can also display water pipe information, asset information (such as manholes, pumps, and work orders), as well as customer information. We can be smart about this and organize all of the information using a semantic model. For now, think of a semantic model as a flexible model that allows us to connect information from different sources, for example, one that allows the water utility operator to determine what pipes are associated with which customers.

Big Data analytics can then be applied to the data collected through instrumentation. We can classify analytics into three broad categories:

Let’s look at concrete examples of these, as applied to an instrumented water system.

To understand the power of descriptive analytics, watch the movie Moneyball. Moneyball tells the story of a poor baseball team that takes a sophisticated descriptive analytics approach and wins 20 consecutive games. The same kind of descriptive analytics can be applied to water assets such as pipes, pumps, etc. Descriptive analytics could allow a water utility to pinpoint leaky water pipes down to a particular type of pipe that’s manufactured by a local company.

For prescriptive or optimization analytics, let’s travel to Europe. I come from a town in southern Ireland called Cork, and a couple of years ago I was walking home when I met some water utility workers making holes in the road. After chatting with them, I learned that the holes in the road were for acoustic sensors, which use sound to locate leaks. I also found out that the pipe networks were very old, and nobody knew where the pipes really were. But they did know they were leaking very badly. This story is echoed across many European cities and towns, and indeed, around the world.

One of the major contributing factors to these leaks is the water pressure in the pipes, which is kept high to ensure good pressure on the tap side. However, analytics can help here. We can run optimization analytics on the pressure reading data across the network to optimize the pressure in the network. This optimization will still maintain good pressure on the tap side, but bring down the overall pressure in the water network. Long term, this will reduce wear and tear on the pipes, and as a consequence reduce the leaks.

Lastly, predictive analytics can provide water utilities with information about when leaks or failures (pumps, pipes, sensors, meters, etc.) are likely to occur.

The theme of World Water Day 2013 is the International Year of Water Cooperation. Instrumentation and Big Data analytics are truly the foundation for cooperation about water management. Getting back to our Moneyball analogy, analytics enabled the Oakland A’s to build a better baseball team, for less money, but team players still had to cooperate and play as a team in order to win. Citizens and corporations, alike, must also cooperate when it comes to smarter water management.

In some communities such as Dubuque, Iowa, citizens use online water conservation portals to gain a better understanding of their water usage. They can also compare their water use with other similar local households. Here we are seeing cooperation between the citizens and the water utility company. Citizens can also collaborate by becoming the collective “eyes” of the water utility. This allows for authorities to get up-to-the-minute information on the state of the water infrastructure and prioritize fixing these problems. Check out the WaterWatcher initiative in South Africa.

I believe we can manage our water better by making our water networks smarter and by encouraging cooperation.

I will leave you with a few more sobering facts about how we currently use water:

Americans use about 100 gallons of water at home every day * Millions of the world’s poorest subsist on fewer than five gallons per day * 46 percent of the people on Earth do not have water piped to their homes * Women in developing countries walk an average of 3.7 miles daily to get water * In 15 years 1.8 billon people (a third of the world’s current population) will live in regions of severe water scarcity [1]