Blowing in the wind

The case for renewable energies such as wind and solar is clear. But while tapping into these sources of power is becoming more popular around the world, there is a difficulty in that the source is not constant; the wind does not always blow, the sun does not always shine. And, which can be just as bad, solar and wind farms can be generating energy when the demand is low.

This is not a new problem; the use of renewable energy has been increasing for the past twenty years. However, while not new, it is a growing problem; a decade ago wind farms were usually small, maybe ten or 15 turbines, but now the power from them is being aggregated. The grids across Europe are interconnected and companies can own multiple farms and sell the output to the grid as one block.
The second difficulty is that companies have invested in these farms and the price of electricity has fallen. This is putting pressure on the operators to make sure the farms are running optimally.
“We are seeing a lot of work with analytics and big data,” said Serge Younes, Director of Sustainability Services at Accenture. “They are developing more advanced methods for weather prediction. They are creating more advanced models and applying them to big wind and solar farms. This lets them sell the energy on the exchanges; they can predict how much energy they will have available in the next 30 minutes and sell it. Before, they would decide how much they were going to create in a week and sell that quantity over the week. This has advanced so much they can now say with high accuracy how much will be produced in the next five to 30 minutes.”
Such analytics can use weather analysis to predict output plant by plant, wherever they are. Having this information means they can increase their profit margins as they know exactly how much they can sell and when. Operators can work hand-in-hand with the transmission companies to make sure the right amount of capacity is on the grid when it is needed.

Controlling demand
In the UK, National Grid uses aggregators, such as Open Energi, to help balance the supply versus demand equation. The aggregator will work with large consumers such as industrial and commercial premises to get them to match their consumption to the supply.
“Renewable energy brings an element of unpredictability,” said David Hall, Director of Business Development at Open Energi, “so we use automatic controls to change the demand. For example, if the temperature in a building needs to be kept between limits, there can be flexibility there as to when the heating and cooling are on or off.”
Traditional power stations with big steam turbines have an inertia, which effectively means energy is being stored in the rotating equipment and that helps keep the system steady. “With solar, there is not that inertia,” said Tony Coleman, Open Energi’s Director of Innovations, “so to keep the system balanced you need to access energy storage more than you used to. If a power station trips and goes off line, you had time to deal with it. You don’t have that with solar. It is much quicker to turn off lots of devices than to ramp up an old style power station.”
Industrial companies have control systems made by the likes of Mitsubishi and Siemens that can link with Open Energi’s Honeywell controller. The overall demand for electricity in the country is measured by measuring the frequency of the grid. This is meant to be 50Hz, but fluctuates slightly depending on supply and demand; if it is above 50Hz, the amount of energy coming into the grid is more than demand, and below 50Hz indicated demand is above supply.
“National Grid has an understanding of how much capacity it needs at any time, but if there is an unexpected surge from renewables National Grid can deal with it, but there is an economic problem,” said Hall. “It wants to buy in just the right amount to keep the system in balance. If it buys too much, it is a waste. So, we help by controlling the demand.”

Visibility
Another problem is that utilities often do not have the visibility to know what is being generated by renewables. A recent pilot in Hawaii tried to solve this by integrating solar inverters with a smart energy platform, the two-year project is just coming to an end and there is hope that it can now be commercialised.
“The solar inverter was connected as a grid asset in the same way as a smart meter,” said Matt Smith, Senior Director of Silver Spring Networks, which ran the project with the University of Hawaii at Manoa’s Hawaii Natural Energy Institute. “This way the utility can have the information, and with the information some planning can be done. This type of real-time control can be important. It was useful for the consumer as well to see what they were generating versus what they were using.”
One of the difficulties in balancing a grid even with the information is getting the different bodies to work together. “It is all fragmented,” said Smith. “You have the solar energy people, you have the energy storage people and you have the smart meter people. You need to bring all this together so it works in harmony. There are a lot of views on how to do that.”
He believes that the distribution system operator should run a data network grid in parallel with everything plugged into it. “A communications network infrastructure is a requirement to coordinate all these things,” he said. “You need to be making decisions in real time at the grid edge. It is a lot easier to balance a microgrid in such a way because decisions can be made quickly, in real time, rather than data having to make a round trip to a central location.”

Households
The other side of the market, which is also growing, is individual household solar panels. While the surplus can be sold back to the grid, it is more economical to balance generation and consumption within the household. The spread of smart meters will help with this as users will be able to get a more accurate view of consumption. But what will also help is the spread of electric vehicles. In Germany and the Netherlands, people are starting to rent out their driveways to neighbours so they can plug their car in while the householder is at work. This can earn the householder credits in a car-sharing scheme.
“There is a formal process for this in Germany driven by BMW,” said Younes. “We are also seeing it in North America with Tesla, where the utility company will install batteries that can be charged up by the solar power and used when needed.”
UK Power Network had a different problem during the time when the British government was heavily subsidising householders to install solar panels. To install pipes capable of handling this extra energy coming onto the system was both expensive and time consuming.
In response, it ran a project in Cambridgeshire with Silver Spring Networks to allow local control of the network; only those people that could be serviced by the existing infrastructure were connected at any one time.
“You turn people on and off the network as you need,” said Rachel Eyres, UK Sales Lead for Silver Spring Networks. “Now the trial is over, it is running as a live generating service. They are now looking to add more zones early next year.”

Predictive maintenance
Another plus from using analytics is the ability to carry out predictive maintenance. Take a wind turbine for example; because the owner knows to a high degree of accuracy how all its parameters should be performing and what its electrical signal should look like, it is possible to see quickly if something is starting to go wrong. This would be obvious if it changed output by, say, 20 per cent, but not by 0.01 per cent, but that can be the first indication of a fault and over time even a small loss such as that can add up to a big loss.
As well as the electrical signal, the hydraulic liquid can be analysed from which it is possible to tell when the metal in the turbine is wearing thin and when it is likely to fail. Other parameters can also be measured.
Now, if a turbine fails during high winds, the operator will lose money, especially as doing maintenance in strong winds can be impossible. However, knowing when it is heading towards failure means maintenance can be scheduled when the wind is low, so the loss of the turbine being offline is reduced. Similar techniques can be applied to solar farms.
“You can check if something is going wrong by analysing the electrical output and see which panels are underperforming,” said Younes. “You can thus work out when to replace them. If one panel in a string of 20 goes down, then all 20 go down and that can be costly and time consuming to find and replace the panel. If you know when something is going to fail, you can schedule the repair to fit the weather.”
The growth of renewable energy has thankfully happened at a time when the industry is beginning to get to grips with big data and the analytics that go with that. This is letting the operators gain a greater insight into electricity consumption plus control over that consumption. These analytics also let them predict with far more accuracy the amount of electricity that will be generated by renewable sources as well as helping them know the best time to take resources offline for maintenance purposes. The sun is shining on renewable energy and the answers to the problems this causes are no longer blowing in the wind.