Optimized Identification Strategy for Composite Load Model Parameters Based on Sensitivity and Correlation Analysis

Li Peiqiang,Li Hui,Li Xinran
DOI: https://doi.org/10.3969/j.issn.1000-6753.2016.16.021
2016-01-01
Abstract:Load modeling and parameter identification are important and difficult in power system analysis. To study the correlations between parameters and the effects of parameters on model response, an optimized identification strategy based on parameter sensitivity and correlation analysis is presented. Firstly, parameter sensitivities are analyzed based on analytic sensitivity and eigenvalues of Hessian matrix. Secondly, through the Pearson correlation coefficient, the correlations among 1st order sensitivity are identified to select the parameters that have the similar effects on model responses. Then, fix the parameters with small sensitivities and simplify the identification of the linearly dependent parameters proportionately, using the hybrid algorithm combined genetic algorithm with levenberg-marquardt algorithm. Finally, efficiency of the optimized strategy is verified by comparing the measure data with the simulations of the integrated load model.
What problem does this paper attempt to address?