Cognitive Systems to manage the food-water-energy nexus
In 1759 Arthur Guinness signed a 9,000 year lease for the brewery in Dublin, Ireland that would become famous around the world. He agreed to pay £45 each year for this lease. Today, we would be less than 3 percent into the term of that lease which looks like a pretty bold and visionary move. Moreover, the rent is an absolute steal.
However, an often overlooked detail is that the same lease also granted Arthur Guinness use of the city’s water supply. It’s an example of someone, more than 300 years ago, realizing the importance of securing a reliable water source, even in a country where water seems ubiquitous. It’s also an example of the interconnected nature of physical resources. For example, city water is shared by many users, and in this case also used for multiple purposes: to manufacture a product and to transport it (by ship) to consumers around the world.
Today, we face similar challenges on a global scale, not just for water, but also for energy, food, and other constrained resources. Fundamentally, these are needed for human survival, but their relative abundance in developed countries also forms the basis of our modern lifestyles. Many of us can hardly imagine not being able to turn on a light by simply flipping a switch or not having clean running water available in every household. However, even in developed nations, managing a reliable supply of these constrained resources is becoming more costly and more complex.
The food, water, and energy nexus
These resources have several key properties in common, but most importantly, they are all connected and interdependent, which is why we refer to them as the food-water-energy nexus. For example, water from a single river can be used for energy generation, agriculture and drinking water. Anything that happens up-stream affects users downstream.
But the connection runs even deeper: while water is used to generate electricity, electricity is also used to pump water. Agriculture uses large amounts of water for irrigation, and bio-fuels are a source of energy for heating and electricity generation. See the infographic below for more information.
Managing these resources jointly enables much greater efficiency but comes at the cost of greater complexity.
The potential of cognitive computing
Cognitive computing can be used to alleviate this complexity by ingesting (often messy) data from multiple sources and presenting users with predictions and recommended actions based on this data. This in turn, enables automation and autonomous systems that help decision makers focus on the most important choices. In addition, food water and energy have several properties that make food-water-energy nexus a particularly promising field for the application of cognitive methods.
What these resources have in common is that they are often not produced where they are consumed, they are costly to transport, and they are hard to store efficiently in large quantities. This is where cognitive computing comes in: if you cannot store a resource you must have a good forecast of supply and demand. Cognitive algorithms excel at finding the non-linear relationships between driving factors and variables of interest.
Cognitive methods are most effective in cases where large amounts of data are available from multiple sources. Indeed, the management of food, water, and energy supplies fit that bill; input data consist of a mix of time-series, images, network graphs, and public policy constraints. Adoption of Internet of Thing technologies will further add to the data available for cognitive systems to work on.
IBM and the food water energy nexus
Scientists at IBM Research – Ireland are already working on managing the complexity of the food-water-energy nexus. We have developed state-of-the-art systems to forecast water usage, energy consumption, energy generation, and crop growth. These systems benefit from increasingly large amounts of data that are available about the use of constrained resources, and the factors that drive their use. This data is collected from many sources, including IoT devices like smart meters, and is often unstructured, like images from satellites.
For example, researchers in Dublin help better manage water rights in developing nations. “Water rights management can be challenging in developing countries with limited data, financial resources and increasing demand,” said Seshu Tirupathi, an IBM research scientist., “My work, in collaboration with IBM Research-Africa in Kenya, has been to develop a system that can improve the quality of the information and interpretations available to decision makers and staff. This helps the license provider to analyze the impact of a new well on neighboring wells, and helps the water resource board to constantly model and monitor the groundwater levels in any given region.”
In IBM Research – Ireland, we also developed algorithms that optimize the interaction between water infrastructure and energy use. “Dynamic pricing is becoming a more common form of electricity tariff,” said Bradley Eck, IBM research scientist. “The price of electricity varies in real time based on the realized electricity supply and demand. Hence, optimizing industrial operations to benefit from periods with low electricity prices is vital to maximizing the benefits of dynamic pricing.”
In my own work, I’ve helped IBM develop a system to manage energy in a collaborative project built around the pioneering vision of the Vermont Electric Company (VELCO) and researchers from Dublin, China, and Thomas J Watson Research Labs. Specifically, using huge volumes of (often messy) data from multiple sources, including high resolution weather models and power system telemetry, we built forecasting models for renewable generation and energy demand.
In a recent announcement from VELCO, the energy company explains how this tool “Sees precisely into the near future, and enables utilities to better protect communities, meet customer needs, and garner renewable energy’s full value.”
This project was recently nominated for the US-Ireland Research Innovation Awards by the Royal Irish Academy and American Chamber of Commerce in Ireland.
As resource management systems become more interconnected, cognitive systems can reduce some of the complexity by helping decision makers make sense of large amounts of messy, unstructured data. Cognitive systems can recommend actions, for example, to schedule power plants or irrigation, or to bid in markets. Three centuries ago, Arthur Guinness used clever management of the food-water-energy nexus to build a global business empire. Today, IBM is using cognitive systems to help its clients all over the world do the same.
More information about cognitive computing and the food-water-energy nexus can be found in the upcoming book Cognitive Computing: Theory and Applications.