Determine Q–V Characteristics of Grid-Connected Wind Farms for Voltage Control Using a Data-Driven Analytics Approach

Abstract

Due to the varying and intermittent nature of wind resource, grid-connected wind farms pose significant technical challenges to power grid on power quality and voltage stability. Wind farm Q-V characteristic curve at the point of interconnection (POI) can offer valuable information for voltage control actions and provide essential indication about voltage stability. Data-driven analytics is a promising approach to determine characteristics of a large complex system, a physical model of which is difficult to obtain. In this paper, data-driven analytics is used to determine Q-V curve of grid-connected wind farms based on measurement data recorded at the POI. Different curve-fitting models, such as polynomial, Gaussian, and rational, are evaluated, and the best fit is determined based on different graphical and numerical evaluation metrics. A case study is conducted using field measurement data at two grid-connected wind farms currently in operation in Newfoundland and Labrador, Canada. It is found that the Gaussian (degree 2) model describes the Q-V relationship most accurately for the two wind farms. The obtained functions and processed data can be used in the voltage controller design. The plotted QV curve can also be used to determine the reactive margin at the POI for voltage stability evaluation. As a generic method, the proposed approach can be employed to determine Q-V characteristic curve of any grid-connected large wind farms.