Institute models the utility of the future


Dario Gil, IBM Research’s director of energy and natural resources
Editor’s note: This article is by Dario Gil, IBM Research’s director of energy and natural resources.

Lightning hits telephone poles. Wind knocks down power lines. Mother Nature and power outages just seem to go hand-in-hand. IBM Research wants to help energy and utility companies with technology that predicts the future state of their assets – not just reacts to when they need repairing or replacing.

That help is coming in the form the Smarter Energy Research Institute. IBM, along with Canada’s Hydro-Québec; the Netherlands’ Alliander; as well as DTE Energy in the U.S. are researching and developing techniques that improve the balance between energy supply and demand using predictive analytics, optimization, visualization and advanced computation.

Deep Thunder

Deep Thunder is the weather forecasting element of the Insti-tute’s array of analytics and op-timization capabilities. Typical weather forecasts are on a scale that is too broad for a utility to use as a prediction of what they (and their customers) may face. It gives utilities a modeling of the weather at a service territory level, improving accuracy and specificity of the forecast – including the impact.
Sensors already tell us about the current or historical state of an asset, and can indicate what changes to make based on their environment. The Institute wants to further automate new predictive capabilities of a utility’s assets by layering algorithms that connect new data, such as weather information, on top of the data already available.

Improving reliability through automated prediction

Today’s distribution networks must incorporate all different kinds of energy sources, from renewables such as wind and solar, to standard coal, gas, and nuclear energy. But electric distribution networks typically have a low-level of instrumentation (a utilities manager might describe the situation as being “blind but happy”). And while there is a trend to increasingly instrument these networks (smart meters being a salient example), analytics and simulation will play a key role in achieving the desired objective of increasing the visibility of the distribution grid.

Let's look at the power of micro-forecasting as an example.

Not your local weather report, IBM’s Deep Thunder forecasting technology makes hyper-local accurate weather predictions, and helps measure the potential impact that weather may have on utilities. A utility company could use these outage predictions to assess the actions to take before a storm hits; the likely damage it may cause; and the restoration resources required.

Smart meters are great for telling the grid – and you, the consumer – about your home’s electrical load every 15 minutes. It’s even better when the smart meter can automatically adjust its load based on grid conditions – and even know to notify repair crews of weaknesses in the system, before a storm hits.

The Institute will create these models for utilities by introducing everything about and around the network: the location of trees in relation to power lines and utility pole locations, to how the entire distribution network is laid out; to even soil moisture (this makes a difference in gauging tree and root strength).

Smarter Energy Research Lab
Smarter Energy Research Lab
Predicting the uncertainty of renewable energy

Imagine a distribution company. They own the wires that carry the power to consumers. Their responsibility is to make sure that all the producers of power can push their electricity out to customers, and everyone is billed correctly.

Now introduce renewables into the mix. One house in the neighborhood with solar panels? No problem. A few electric vehicles charging over night? No problem.

What about 20 percent of an entire continent depending on wind, water, solar, and other renewable sources?

The European Union wants 20 percent of their member countries’ electricity to come from renewable sources by 2020. Fluctuations and instabilities will happen without an effective way to predict what will cause a change in supply or demand; and without a way to automate the actions to take based on those predictions.

Smart predictive meters and other responsive devices could also shift loads to places of need, say to a neighborhood charging hundreds of electric vehicles – and use distributed energy resources such as wind and solar.

The Institute’s projects will even sprinkle some behavioral psychology into its models to better predict usage patterns and what incentives might motivate a consumer.



The Smarter Energy Research Institute will operate as a collaborative research venture across the world. Hydro-Québec is one of the world’s largest hydroelectric power producers and the only North American electric utilities operating its own research center IREQ. Alliander is a major Dutch energy distributor specializing in renewable energy, serving three million customers in the Netherlands. DTE Energy is an investor-owned diversified energy company involved in the development and management of energy-related businesses and services across the United States.

Labels: , , , ,