NICTA to develop algorithms for optimising joint gas and electricity networks

NICTA to develop algorithms for optimising joint gas and electricity networks

Algorithms will underlay the predictive modelling and analytics used to optimise the networks

NICTA will develop a new algorithms to help multiple gas and electricity providers integrate and optimise their networks using predictive data modelling and optimisation.

The organisation is working with Los Alamos National Laboratory in the United States on this project, under a 12-month Cooperative Research and Development Agreement and co-funded by the US Department of Energy, and is hoping to have the algorithms tested at Australian utility companies.

“We are trying to sign an agreement with an Australian company that has both gas and electricity network,” said Professor Pascal Van Hentenryck, optimisation research group leader at NICTA.

“In the United States, the gas industry is creating around more than half a million jobs with it having a substantial impact on their GDP. We also have significant gas resources, so the research that we are doing here at NICTA can have a major impact on the Australian economy as well.”

Van Hentenryck said joining up the gas and electricity networks will allow companies to better maximise natural gas and exploit renewable energy.

The algorithms will underpin the predictive modelling and analytics used to optimise the networks, Van Hentenryck said.

“The underlying mathematics is much more complicated, the underlying computation is more complex and challenging. That’s what the project is all about – if we can actually get the computational complexity at a level where we can operate these two systems jointly,” he said.

“I think the key in the future is integrating predictive models based on data analytics and what we call prescriptive analytics, which is essentially optimising the system. So the future, from a computational standpoint, will be integrating the predictive and prescriptive models.”

As there's a lot of uncertainty when it comes to generating renewable energy – weather patterns can suddenly change, for example – optimisation is computationally harder to achieve. Predictive models determine what is likely to happen based on current and historical data. Prescriptive models not only suggest a course of action to take based on what is likely to happen but also offer an alternative if things were to change. Using these two models together helps significantly in hedging uncertainty.

Read more: AI software in freight logistics, gamification in universities feature in state iAwards

Van Hentenryck said the use of gas is growing in the United States because it is cheaper to produce than coal and it pollutes less.

“I think the US said they met all their targets for the Kyoto Protocol, although they haven’t signed it, just because they moved away from coal power plants into gas power plants. In New England in the United States, for example, [the region] moved from 5 per cent [coal] to 50 per cent [gas]. So the share of gas has moved tremendously.”

Another advantage to integrating the gas network into electricity is that it can store power in the pipes, which can then be used later on when there’s short supply or peak demand, Van Hentenryck said.

“You can address the peaks must faster, you can also address the fact that you lose some of your resources typically renewable energy.”

NICTA will also look at weather forecasting and predictive models to help utilities prepare for spikes in electricity when people increase use of their heaters and air conditioners.

“If you have a joint gas and electricity network, obviously you will integrate renewable in there too. So you will want to have predictions on clouds and wind, and things like this. So it’s trying to find out how the cloud will move, hour many hours of [sunlight] you will have; that is going to help tremendously," he said.

“The better predictive models we have, the more update data and real time information we have, the better we can operate the network.”

Read: Building smart utilities.

When it comes to addressing concerns around different utility companies sharing their information, Van Hentenryck said the algorithms will be developed and tested onsite.

“We need to test it we can embedded it inside their own teams. So it’s a good model because they don’t have to worry about confidential information, we never have access to it really.”

Follow Rebecca Merrett on Twitter: @Rebecca_Merrett

Follow CIO Australia on Twitter and Like us on Facebook… Twitter: @CIO_Australia, Facebook: CIO Australia, or take part in the CIO conversation on LinkedIn: CIO Australia

Read more: Technology falling short for HR workers

Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the CIO newsletter!

Error: Please check your email address.

Tags predictive analyticsinfrastructurenictadata modellingLos Alamos National Laboratory

More about FacebookNICTA

Show Comments

Market Place