Wrong weather forecasts are a disaster for the climate crisis


Predict the weather It may be a frustratingly imprecise science. The weather app on your phone is very good at predicting whether it will rain at some time of the day, but if you want to know whether it will rain heavily in central London at 3pm this Sunday, then it won’t help much . If you absolutely must keep it dry, it is best not to bring an umbrella or stay inside.

For most people, they don’t know what inconvenience the weather will cause in the next hour. However, when it comes to power networks, not knowing that the next weather will be not just an annoying thing, but an important source of carbon emissions. If we can better predict when and where the weather will change, then we can stop large amounts of carbon dioxide emissions, simply because we don’t know what the cloud will do next.

This is the problem. On a sunny spring day in the UK, solar energy can account for about 30% of all electricity produced on the island. The exact numbers vary greatly, but under ideal conditions (solar panels perform best in cool but sunny conditions), they may consume 9 gigawatts (GW) of energy, which is 30 GW average energy demand a big part. So far so good. However, if a large cloud covers the southwestern part of the UK where many solar panels are located, a large part of renewable energy will suddenly disappear from the grid, which is equivalent to the immediate offline of the entire gas station. In this way, hundreds of megawatts of energy disappeared.

It is obviously not ideal to eliminate the power of the entire power station within a few minutes. Therefore, in order to compensate for this, the power network arranges some backup energy production to intervene and solve any turbulence caused by changes in solar energy production. In the UK, the responsibility for balancing and distributing this energy lies with the National Grid Electricity System Operator (ESO), which requires fossil fuel power plants (usually burning natural gas) to provide additional energy to prevent unexpected drops in solar production.

Fossil fuel plants are slow-moving beasts. “We really hope that a power plant can accelerate in five minutes or half an hour, because that is the speed at which wind and solar power generation may change,” said Jan Kleisel, a professor of renewable energy and environmental mobility at the university. Jan Kleissl) said. University of California, San Diego. But fossil fuel plants cannot work like this. They take a long time to turn on and are most efficient when running at full power. This restriction further encourages the grid to overproduce energy in case the power from solar or wind power drops.

One way to solve this problem is to better predict the weather. If we know exactly how much solar energy the UK may generate at any time, then ESO can dial back its reserves, thereby reducing the total carbon footprint of the energy grid. In other words, if we know exactly how much solar energy will flow to the grid every five minutes, then we can ensure that every kilowatt of energy is used instead of using the excess electricity generated by fossil fuel power plants to hedge our bets.

Jack Kelly thinks he knows a way to greatly improve these predictions.Former researcher of Alphabet’s former artificial intelligence research company DeepMind co-founded in 2019 Open climate repair, A non-profit organization dedicated to reducing greenhouse gas emissions through machine learning.Kelly said: “I am a machine learning researcher. He is afraid of climate change and is keen to do his best to solve it.” He estimated that better solar forecasts in the UK might Save 100,000 tons The total amount of carbon dioxide emitted each year, which is essential for the State Grid ESO to achieve its 2025 goal, that is, as long as there is sufficient emissions to achieve zero emissions Renewable Energy.


Source link