Parameter Identification Methods Of Thermal Power Unit Model Based On The Real-time Data

The thesis studies the parameter identification of. The thesis is composed of four parts, i.e. the basic principles of identification and methods of parameter identification, Thermal Power Unit model and analysis of identifiable parameters, data preprocessing and parameters identification. First, study the basic least squares method, methods of time-varying parameter identification and the time delay identification. The simulation study on the basis of MATLAB proves the effectiveness of the algorithm. This thesis also discusses identifiable issues of the system and come to conditions of identifiable system. Second, analyze and simplify identifiable parameters of Thermal Power Unit model. Third, eliminate and correct the unreasonable data; remove the noise from the corrupted data with wavelet transform; And Smooth data again by means of central interpolation algorithm. Finally, identify the parameters of the model via auxiliary variable algorithm. The examinations show the validity of the identification model.