Identifiability Analysis of Load Model Parameter Identification with Likelihood Profile Method

Abstract

Load model parameter identification from practical measured data has become an essential method to build load models for power system simulation, analysis and control. With different power system practical measurement and operation conditions, which may include disturbance magnitudes, measurement errors and data lengths, the difficulty to identify load model parameters is also different, which would lead to the problem of practical identifiability. In this paper, a likelihood profile based parameter practical identifiability analysis method for load model identification is proposed. The load model structure and parameters used for identification and the method to identify parameters based on ambient signal are introduced first. The definition of identifiability together with the likelihood profile analysis method are then proposed, after which the procedures of load model parameter identifiability are given. Simulation is conducted in WSCC 9 bus system to show the effectiveness of the proposed method. Impact factors of load model parameter identifiability are also analyzed and simulated.