A two-level online parameter identification approach

In this paper, a novel online parameter identification approach with two-level architecture is proposed. In this approach, identifiability analysis, parameter identification and parameter validation is developed and integrated in a two-level manner. Firstly, an identifiability index based on trajectory sensitivity is defined to select measurements scans that satisfy the condition of identifiability. Parameter identification is then done with the selected measurements scans, which has the ability to compress measurements error. These two functions are deployed at substations or plants to guarantee efficiency. A hybrid simulation based parameter validation method is also developed which is deployed in control center. With this two-level online parameter identification approach, parameters of models can be identified and verified automatically. Numerical tests from real application show that the proposed approach has good performance.