Renewable energy is at the heart of the transition to a less carbon intensive world. Wind and sun surely seem like endless resources. Water quite less. If using these resources to produce electricity is now mastered, having them available when needed is not so simple.
The share of renewable energy in electricity production is increasing every year. As per IEA, it reached 29% in 2020, highest figure so far. It’s good but still a far cry from the 60% needed to reach by 2030 to come close to the Net Zero Scenario’s objective.
The challenge of operating these hydro, solar and wind power plants is the availability of their respective energy source: water, sun or wind, all relying on weather conditions. The evolution of the climate since the middle of last century is already adding to their intrinsic variability, extending it to longer periods (years, decades). As a consequence, the way the climate is changing impacts the electricity spectrum.
As several countries are facing droughts, limiting the quantity of water available, hydro power plants face limits to the amount of electricity produced. In Germany or in California, the large installed capacity of solar PV (with respectively 58.5 and 35 GW) can disrupt the grid with high levels of production mid-day and limited availability during cloudy days or at night.
As the slightest change in wind, rain, sunlight will impact the output level of tens of thousands of generation assets, their operating conditions could change and fluctuate within minutes. Even within a site, weather conditions could vary, one part could be under the rain when the opposite side benefits from the sun.
Weather is influencing not only the quantity of electricity produced by renewables but also the consumption. Temperature sensitivity is high in systems relying on electricity for heating or cooling. In France, it is estimated that a drop in temperature of 1°C leads to a rise of demand of 2 100 MW while an increase of 1°C during the summer months will increase consumption by 500 MW.
Extreme weather events could have an even stronger effect. On top of endangering people, they could take down transmission or distribution lines, damage the structure of power plants, crack PV cells or harm wind turbines. Weather conditions impact the operations but also the economics of a renewable power plant. Site selection relies on irradiation, wind, water flows and how much electricity can be generated. Once in operation, a reduced output, not available for trading could result in lower revenues.
Operating conditions and the volatility of the electricity wholesale prices make it even more important than before to be able to sell electricity on the wholesale market at its optimum price. Accurate forecasts are crucial in achieving this. And with renewable energy, high definition weather forecasts have a key role to play.
The main weather measurements are pressure, humidity and temperature, collected by weather stations or satellites. The latest provides more extensive but less accurate data than the ground stations.
For precipitations, two types of measures can be used: the first, a direct measurements with a rain gauge or a disdrometer for example. The second type is indirect and measures a disturbance caused by the rain. This is the case of weather radars that emit an electromagnetic wave before analyzing the disturbed signal which is returned. More cleverly, a company like HD Rain uses sensors analyzing electromagnetic waves already present, the ones used for TV reception.
To predict the weather, hours, a day or even several days in advance, a mathematical model is used to anticipate the future rain, wind, temperature, pressure based on the past and on current conditions. The more we want to anticipate and the more it is necessary to look further away from the place for which we want to obtain a prediction and we are losing data accuracy.
HD Rain provides shorter term predictions - between 1 and 2 hours in advance. The precise pooling of its sensors delivers a high spatial (500 meters) and high temporal (1 minute) resolution. Furthermore each station provides a measurement along several kilometers instead of a small area. 1 000 of such stations have already been installed around the Mediterranean arc. Once collected by each station, the aggregation of data via advanced Deep Learning techniques enables the production of rainfall maps for a detailed forecast.
Such a solution could be of great help to better assess the upcoming weather conditions and therefore better integrate intermittent sources of energy in the electricity generation mix.
Harvesting natural sources of energy such as wind, sun and water is increasingly crucial to reduce the impact of climate change. However, systems need to become smarter, decision making faster to optimize their use.
(Sources: IEA, Irena, SEIA)