[MUSIC] In order to optimize the space and electrical infrastructure, wind or marine turbines are generally packed together in small or large farm. It is then needed to consider the interaction between the turbines. When a turbine operates in the wake of another, it is expected to have a reduced power production due to the lower incident flow speed. Hence, a large group of turbines will not deliver the same power than the sum of the power output of isolated device. This image is famous and amazing. Due to that mismatch condition, the turbine wakes become visible and reveals the multiple interaction between the wind turbines. When we consider such a large number of turbine displayed in a farm with regular spacing, we could wonder, what is the optimal spacing distance? The standard spacing distances, which are given by the turbine supplier, are around 6-10 diameters in the downstream distance. And 3-5 diameters for the lateral distance. We have seen that the typical length and width of individual turbine wake are not constant values, but depends on the turbulence of the upstream flow. Nevertheless, if we consider an average value, big turbines will have larger and longer wake. Hence, we could ask whether for a given area, the power extraction will be optimized with a large number of small turbines or a small number of larger turbines. And if we consider a single turbine, the mean extent of the wake fixes the minimal area occupied by the turbine in the farm. This area, AT, depends on the turbine diameter and skates as D squared. Hence, the total number of turbines that can be deployed on the given area A, will scale as A divided by D squared. And the total power produced by the wind farm will then be proportional to the number of turbines times the swept area of the turbines, D squared. And the product of these two terms correspond exactly to a number which is proportional to the total area, A. Hence, these very simple calculations shows that the total power available will depends only on the total area A and the upstream velocity, V. But, not on the individual turbine diameter. However, for our wind turbine, and also for a tidal or a river turbine, the amplitude of the upstream velocity depends on the hub height. And turbine with larger diameter will be at a higher altitude than smaller ones. Hence the vertical velocity shear will favor larger diameter turbines, which are able to harvest stronger winds or currents. If we increase this, the turbine diameter, we will install a smaller number of device, but the available power of the whole farm will increase. Once we fix the size and the number of turbine in the given area, once the velocity deficit in the wake is accurately modelled, we are in front of another challenging problem. Optimize the turbine positions. This is called the farm layout optimization problem. Let's consider for instance, a periodic layout with eastward wind. The first hole of turbines gets the upstream and unperturbed wind, which will therefore maximize the power extraction. Now, if the second hole is laterally shifted, we could expect to diminish the wake interaction and therefore reduce the downstream distance between these two holes. However, if now the wind turns by a few tenths of degree, we could obtain a strong shadowing of the second turbine hole by the first one. A concrete example of this negative alignment was measured at Horns Rev 1 offshore wind farm. For real wind data, there is always statistical distribution of the wind directions, which are quantified with the wind hose. For the Horns Rev 1, the wind mainly blows from the south, southwest. According to the wind farm layout, when the wind blows from the southwest, the downstream distance between two consecutive turbines is ten diameters. If now we consider a wind blowing from the northeast, so downstream distance between two consecutive turbines will goes up to 11 diameters. And for this optimal wind angle the power output by each turbine slightly decays with the downstream distance. The mean output power of the individual turbines located in the first hole is weakly affected, while the ones located in the last holes could be reduced by 30 to 35%. However, when the wind turns just by 10 to 20 degree, and is coming from the east, the distance between two consecutive turbines would be reduced down to seven diameters. With a drastic impact on the power production. For this specific angle, the power of almost all the turbines behind the first hole will be reduced by 40%. To sum up, the wake effect may have a drastic impact on the production of a turbine farm. The length and the width of a single turbine depend mainly on the turbine diameter and the turbulent intensity of the upstream flow. Some empirical models could give a relatively correct description of the wake. But to get a more accurate and precise estimation of the velocity deficit in the wake, laboratory experiments or numerical models are still needed. When several turbines are packed together in a large farm, the layout optimization becomes crucial for the power production. And a precise description of the flow viability, mainly the turbulence intensity and its direction are needed to solve this complex optimization problem. Thank you.