Impact of Load Modelling on Reactive Power Margin Measured in Probabilistic Voltage Stability Analysis

Voltage stability is a vital area of focus in power system studies, and load modelling, along with its parameters, significantly influences it, particularly in renewable-rich power networks. This paper determines the influence of load modelling and its attributes (employing various load model types and various estimations of their parameters) on reactive power margin (RPM) (utilising the QV-curve method) for assessing probabilistic voltage stability. In DIgSILENT PowerFactory, the IEEE 39-bus system is used to accomplish the QV-curve(s) analysis for measuring the RPM and its boundary under uncertain conditions. The Monte Carlo (MC) method is utilised in MATLAB software to generate the uncertain system datasets, and the input parameters in the MC method consist of the indeterminacies inherent in system load and the sporadic nature of wind power for a period of one year. According to the QV-curve-based probabilistic voltage stability analysis, constant power load model indicates better stability performance as they improve the RPM. Also, an increase in the dynamic load, reducing the system capacity, and increasing the values of the active and reactive power coefficients in the exponential load model results in an expansion of the RPM, rendering it less susceptible to system uncertainties.