Renewable energy in the United States accounted for 13.44% of domestically produced electricity in 2015. This is a number that is only going to increase as governments look to cut down on the amount of pollution emitted by traditional fossil fuels. The International Energy Agency (IEA) said Friday in its annual market report that renewable energy will represent the largest single source of electricity growth over the next five years, driven by falling costs and aggressive expansion in emerging economies.
This does, however, present a number of challenges, challenges that analytics is necessary to solve.
International Energy Agency’s Chief Economist, Laszlo Varro, has argued that renewables combined with big data could see conventional fuel sources such as natural gas increasingly ’in competition with Silicon Valley.’ The key problem it can solve is forecasting. The primary function of grid operators is to constantly balance the supply and demand of electricity, whether the energy source is natural gas, coal, wind, or solar. They must schedule power generation 24 hours ahead of time, using an hour-by-hour timetable of when to run and supply electricity. Failure to balance supply and demand is likely to result in blackouts or power surges.
The problem with wind and solar is that they are so unpredictably variable in comparison with other energy sources. The sun does not always shine nor the wind blow, making forecasting far harder than with fossil fuels. Grid operators must subsequently identify and use fossil fuels, predominantly natural gas, as back-ups for when supply is too low. The more back-up plants required to provide this, the higher the cost.
However, by using data around the weather, historical patterns of supply and demand, and information about technical features from the plants, grid operators are able to more accurately forecast generated power, and ramp down or even shut down plants accordingly. This means backup energy is needed less, both lowering costs and use of fossil fuels.
According to a recent article in Nature, power generation could increase by up to 16% as a result of analytics. This means a higher concentration of renewables in the energy mix regardless of how much electricity can actually be stored. A 10% improvement in the accuracy of wind forecasting on the western US grid alone would lead to cost savings of tens, potentially hundred of millions of dollars per year. One early adopter of wind forecasting models, Colorado-based Xcel Energy, has already managed $40 million in savings over four years, and this will increase as they collect more historical data and can further improve their forecasting models.
Vestas Wind Systems of Denmark is another renewable energy company to use analytics successfully. Vestas produces 20% of global wind capacity. Its real competitive advantage lies in what it calls the ‘biggest wind data asset in the world, which is the historical data it has collected over the past 30 years - a data set that is constantly growing as data is collected from its 50,000 turbines across the world. Applying analytics to this data enabled the company to create a performance indicator called the Lost Production Factor, which measures the amount of electrical production lost from turbine assets as a result of non-optimization. Using this, Vestas improved turbines’ orientation to changing wind strength and direction. This has seen its Lost Production Factor fall from 4.4% in 2008 to 1.5%, significantly below the industry average of 3.6%.
The world needs renewable energy, and given the power of oil and gas interest and the propaganda spread around the failure of wind farms, it is imperative that it is working to an optimal level.