MPP awareness in Automatic Generation Control (AGC) with wind farm participation

Abstract

The AGC is a system that automatically controls the generation of power plants participating in secondary frequency regulation [1]. Currently common AGC systems are designed to control conventional power plants such as thermal power and hydropower power plants. However, due to the ongoing expansion of renewable energy sources (RES) all over the world, system operators are facing increasing regulation problems due to power deviations resulting from renewable energy source within their portfolio and reduction of regulation reserves as synchronous generation is displaced by RES generation [2]. A possible approach to tackle this challenge could lie in the inclusion of RES into the AGC system. In past Cigre contributions it has been shown, that by actively incorporating wind power into AGC regulation, wind deviation could be compensated whilst maintaining regulation performance [3]. Nevertheless, the results also indicated, that with increasing amount of wind power participation, dynamic upper wind farm power limit estimation and communication is becoming more and more important. In recent studies such dynamic MPP awareness has been assumed for AGC regulators in order to allow more appropriate power set point assignment to wind farms in regulation [4]-[6]. Nonetheless, this assumption implicates that, despite operating beneath MPP, the theoretical wind farm’s MPP is calculated based on the prevailing wind conditions and subsequently communicated with the AGC regulator. In this context the present study analyses the influence of such MPP awareness (from perfect knowledge to reasonable accuracy), considering both general MPP calculation inaccuracy (noise) and communication delays between AGC regulator and wind farm controller. Furthermore, to increase the level of influence within AGC regulation, the amount of wind power generation participating in regulation has been increased to 150 MW installed power (6 wind farms), based on real wind farm operation data and the balancing area size has been reduced. Hereupon, the observed AGC and balancing area behavior is evaluated in terms of wind power spillage compared to standard MPP operation and AGC performance according to the Spanish dynamic regulation evaluation criteria. The obtained results show, that the AGC regulator clearly benefits from MPP awareness, compared to assuming upper wind power generation limit at nominal power. Additionally, it is shown, that even communication delays of up to 24 seconds (6 AGC cycles) as well as reasonable inaccurate MPP forecast barley affect AGC performance, not least because of the assistance of conventional generation, compensating small regulation inaccuracies of wind farms. In summary, the obtained results illustrate that despite rather imprecise MPP awareness proper AGC regulation can be ensured, if some conventional generation is in place to mitigate still prevailing wind farm MPP estimation inaccuracies.