Ranking of Power System Loads Based on Their Influence on Power System Stability

Abstract

Load models play an important role in power system stability analysis and accurate load models are essential for and will benefit power system stability assessment. It is not necessary and practical though to model all loads very accurately, as load modelling requires significant financial and human resources. This thesis thus develops a methodology for identifying critical loads and load model parameters based on their influence on power system stability. Therefore, only those important loads and load model parameters would need to be modelled accurately, saving significant investments. The main outcome of this research is being able to identify the critical loads and load model parameters in the power system for different types of power system stability, and to determine the required accuracy level of critical load model parameters so that the accuracy of power system stability analysis is not affected. This research mainly contributes to the following areas. First, an automatic load modelling tool (ALMT) which was originally developed in the past, as part of the final year project, has been substantially improved and fine tuned by adjusting software parameters for higher accuracy and implementing composite load model. The tool can be used to automatically build load model from measured power system data without human intervention. Second, a framework for ranking power system load model parameters for four types of power system stability has been developed. The framework is based on Morris screening method. Load model parameters are ranked according to their impact on power system stability indices. The ranking has been performed for four types of stability studies, three types of load models, and four different loading conditions. Third, a methodology for determining the accuracy levels of critical load model parameters has been developed. The methodology is based on Monte Carlo simulation. Different uncertainty levels of load model parameters will result in various confidence levels of stability indices. The required accuracy level of load model parameters can be chosen according to the desired confidence level of stability indices. Fourth, the influence of stochastic dependence of load model parameters on the parameter ranking is investigated. Load model parameters have been obtained from actual field measurements. Based on them, the Gaussian Copula is used to generate correlated load model parameters, which is then used for parameter ranking. Finally, the critical load locations in the power system are identified for different types of power system stability, and load buses are ranked irrespective of the load models used to represent load of these buses.

Date of Award	1 Aug 2019
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Jovica Milanovic (Supervisor) & Robin Preece (Supervisor)