How far can you go with wind power?

During the last decades, wind power has known an immense growth over the world, and particularly in Europe. The traditionally mentioned advantages are the infinite availability of the ‘fuel’ (wind), and the fact that wind turbines produce electricity without emitting carbon dioxide or other toxic gases.But can wind power be installed without limits, in a country like Belgium? Except for the lack of space, there is the difficulty that wind power is never as reliable as conventional power plants, because wind can never be predicted with unlimited accuracy, let alone controlled.

However, accurate prediction and control is necessary. The European electricity grid is divided in a number of control areas. The grid operator in each control area (e.g. Elia in the control area Belgium – Luxembourg) makes sure that the instantaneous power production in its area is always equalto the total consumption. Otherwise, the grid becomes instable, with severe disturbances resulting from it. Because the grid operator can never know in advance the exact amount of total power consumption in its control area, an amount of ‘regulatingpower’ is needed: this consists of power plants that can control their output power or that can be put into service within a very short timeframe, to immediately respond to load fluctuations.

With an increasing number of wind turbines connected to the grid, more regulating power is required. While the variation of power consumption in a control area is well predictable, because it shows the same pattern every day, the power production by wind turbines is much less predictable. When the wind speed is high and all wind turbines are operating at full power, the risk that the wind speed suddenly drops must always be taken into account. In that case, the system depends on the regulating power. Regulating power consists of e.g. turbojets, well controllable but consuming a lot of fossil fuel. With high penetration of wind power in a control area, the uncertainty of power production becomes an order of magnitude higher than was previously the case. The turbojets will operate more, as a consequence the total fuel consumption and carbon dioxide emissions of all power plants together will possibly be higher than for a scenario with less wind power.

What if the wind turbines are geographically well spread over the control area, rather than all turbines installed close to each other in large wind farms? Is there then a high probability that always somewhere in the control area there will be much wind, and somewhere else not, making sure that the fluctuations of total power production by the aggregated wind park are limited? The answer is no. The correlation between wind speeds in Belgium is very high, or, when the wind speed changes in the coastal region, the probability is high that de wind speed changes to the same extent at the same hour in Brussels or elsewhere. The advantage of spatially distributing wind turbines over the control area is only relevant when large areas can be considered, e.g. Western Europe. Unfortunately, it is at this moment technically not feasible for a large region such as Western Europe to operate permanently as a single control area. This would result in large and unpredictable power flows between the actual control areas, requiring high investment costs for reinforcement of the many congested high voltage lines connecting the different control areas.

Then how far can you go with wind power in Belgium? The conclusion of the thesis is that in Belgium there is space – from the electrotechnical point of view – for 500 to 700 MW of wind power, or 250 to 350 medium-sized turbines. From then on, the negative effect of the extra required regulating power becomes relevant. Considering that nearly 200 MW is installed in Belgium by now, and 216 MW will possibly be constructed offshore in a near future, this limit is already relatively near. Meanwhile, it is worth the effort to look for alternatives for responding to the increased demands of regulating power. ‘Demand side management’ (actively controlling the power consumption, as a function of the available power) is a very promising option for this.