Optimisation of the reactive power management of a wind power plant

The wind farm Les Hauts-Pays in France has a rated power of 78 MW and comprises 39 wind turbines. The transmission system operator requires that the plant participates in primary voltage control by means of reactive power. The reactive power is produced or consumed by both the wind turbines and static units such as capacitor banks and shunt reactors. It is suspected that the reactive power dispatch between the two types of sources could be suboptimal. Indeed, the reactive power produced or consumed by the wind turbines must be transferred across the collector system of the plant, which leads to additional energy losses and a loss of revenue. On the other hand, the static units are installed at the substation and their reactive power is almost directly exchanged with the transmission network. Currently, the wind turbines receive equal reactive power setpoints and are used as much as possible to perform the voltage control, while the static units are only used as a complement. For this reason, the static units correspond to available but inadequately used equipment. The reactive power capacity of the plant appears to be oversized and the static units constitute a potential of energy savings that is only partially exploited. In addition, it could be considered to allow for unequal reactive power setpoints among the wind turbines. Following these observations, the main objective of this work was to determine to what extent the energy losses across the collector system could be reduced by improving the reactive power dispatch. A data analysis was performed and confirmed that the static units were rarely used. A model of the wind power plant was implemented and validated by comparison with measurement data. Two axes of improvement were investigated. The first axis of improvement concerned the reactive power dispatch between dynamic and static sources. The model of the wind power plant was used to optimise the control of the static units. Simulations showed that a gain equal to 0.16% of the yearly energy input could be reached. The risks related to the ensuing increase of the number of switching cycles were examined. It was concluded that the risk of degrading equipment integrity or power quality was low. The second axis of improvement concerned the reactive power distribution among the wind turbines. Optimal power flows demonstrated that allowing for unequal reactive power setpoints could lead to a gain of 0.04% of the yearly energy input when the tap of the substation transformer was not included in the optimisation. A gain of 0.22% could be reached when the tap was part of the optimisation. Finally, the sizing of the static units was analysed. Although the capacitive capability appeared to be somewhat oversized, it was concluded that the sizing of the external equipment is coherent with the reactive power requirements at Les Hauts-PaysMaster [120] : ingénieur civil électromécanicien, Université catholique de Louvain, 201