Building smarter grids

sFor over 80 years, homes in the UK other industrialised countries have been powered using a national grid system. This means that power stations all around the country can share the load and if one of them breaks down, the others take up the slack. And because our patterns of energy use are quite easy to predict, it's been a relatively easy task to turn power stations on and off in anticipation of when we will need them.

But, as we introduce new ways to make electricity, including sources like solar and wind,  the grid has a problem, because these alternatives don't produce power at a constant rate like a coal-fired station. When the wind drops or the sun goes behind a cloud, the power supply dwindles. So this week we're exploring how new technology - and even maths - can solve the problem...

One way to deal with this is to invent a new "smart" national grid system that controls how we use power according to how much is available.

Guenter Conzelmann is working on this at Argonne National Laboratory in Chicago...

Guenter - We've done this experiment for 80 years. We can fairly well forecast what the demand will be the next hour, 5 hours from now or tomorrow. However, there's new technologies that come into play that change that now. If you now put a solar panel on your rooftop, what happens on a day when we have intermittent clouds moving through the area? In essence, your net consumption changes from almost minute to minute to second to second. And so, those fluctuations are very different than what we've observed for the last 70, 80 years. So, we have to be able to handle that and manage that.

Chris - Is it the dream that the grid will become switchable? It would direct energy into different places at different times in order to make it reactive.

Guenter - Very much so. But we're trying essentially to convert the grid from a flip phone to a Smartphone. And so, with that smarter grid, we can react to unanticipated changes. We can accommodate those changes much more rapidly.

Chris - So, the obvious question is, what's stopping you?

Guenter - Some of the technologies are available. It's just a matter of implementing those technologies and advancing some of the technologies.  4% of our electricity comes from wind power. If we're going to go 20, 30, 40% wind power that fluctuates all the time then we need to do better than what we do now.

Chris - At the moment, a lot of these smart metering technologies in people's own homes, it's telling them where they're using the power and encouraging them led by price to change their behaviour. But to do what you're saying would involve the power station or the grid itself sending a signal to my house to say, "Right, the grid is under heavy load at the moment. Have you got some systems running that actually, you probably could turn off for 10 minutes or so to give the grid a bit of a rest?" That's a bit different though isn't it?

Guenter - That technology exists. I have one of those meters. I've had it for the last 5 years and I get messages every day when the grid gets strained. Meaning, that translates to a higher price. My phone tells me, now will be a good time to turn something off. Now, that still depends on me then making a decision and actually, actively doing something. So, we want to move away from that.  so for example, if your refrigerator gets turned off for 2 minutes, you won't notice the difference whatsoever. The food will still be cold. So, what we want to do is we want to get to a situation where it's fully automated and your appliances do this all in the background without you, even have to worry about it.

Chris - It is not a problem with that though that if the grid goes under load, lots of things turn themselves off all at once and altogether in synchrony and then you get a huge surge and everything goes, "Ooh! There's a big surfeit of power!" Everything turns itself back on again and the whole grid goes to these horrible cycles of boom and bust which ultimately puts enormous load on the technology.

Guenter - Good point. The Germans have realised this with the solar PV panels where all the solar PVs, small scale intallations, all on people's rooftops had a controlled signal built-in that said, if the grid is strained to a certain point, at that point, turn it off. And if you had a million of those and if that were to happen, a million of them turning off at the same time, the grid would collapse. It couldn't handle that because this would happen within a fraction of a second. So, one research area clearly is to develop control theories and control algorithms and underlying technology to move to something where we have not just one controller but we have millions of them. They talk to each other. They're sensitive to what the others are doing. Then we can avoid these problems.

Chris - Have we got a grid to practice on?

Guenter - You don't want to do these experiments on a large scale. So, we build these, we test them in the laboratory and then we try to validate them on a computer too. Eventually, we'll move them out to small scale real grids. And so, right now, we don't do this but you can envision that we work with the Argonne on-site grid. You usually don't experiment with a real system that has millions of customers on it until you're really, really sure the thing works the way it's designed to.

Chris - And how far do you think we are from that dream?

Guenter - We sort of have a 5 to 10-year outlook so that we develop the tools, the control tools, the modelling tools in the first 5 and then in the second 5-year period, that we actually would deploy them and roll them out and test them.

Chris - Do you envisage any problems with retrofitting this into people's homes? Is that going to be a huge cost of implementation?

Guenter - One of the criterias that we always include when we make a plan is, clean, reliable, resilient, but also, affordable because that is the key thing. It's key thing to you as a consumer - you and your household as well as to industry.