Research On The Full-Order Flux Observer And The Speed Estimation Method Of The Induction Motor

The theory and the implementation method of the flux observer and the speed estimation are the necessary issues of the speed sensorless control of the induction motor and received a lot of attention in the past research works. Basic on the review of the past research works, this paper focus on the study of the full-order adaptive flux observer which estimates the rotor speed from the current error.The past research works of the speed sensorless technology are reviewed, especially the foil-order adaptive flux observer. The poles locus of the induction motor are derived and analyzed. The foil-order adaptive flux observer is established basic on the observer theory. The speed estimation using the foil-order adaptive flux observer is described.The discretation of the foil-order flux observer is an important technology for the impletation of it in practical. The precision and the stability are crucial for the discrete foil-order flux observer. The discrate models of the observer using the simple forward Euler discretization and the trapezia discretization are presented. The discrete errors are analyzed by means of the norm 2. A new discretation strategy is proposed, which apply different discretation according to the rotor speed. The stability of the discrete foil-order flux observer is also analyzed and the limitations for the placement of the observer's poles are presented. As to the unstability in field-weakened region, a new feedback matrix is proposed to improve the stability of the discrete obsever.The foil-order flux observer is sensitive to the motor parameters. The parameters sensitivity of three kinds of flux observer are studied, which includes the voltage-model-based observer, the current-model-based observer and the foil-order observer. The comparative study of three observers is done by means of a ratio fuction of the estimated flux and the practical flux. A clear comparative result is presented in a table.The global stability of the speed estimation system is studied in the paper. The PI adaptation law is derived using the Popov's hyperstability criterion and the Lyapunov's stability theorem. The hypothesis of the PI adaptation law is pointed out. The unstability at low speed is analyzed and the design guidelines of the feedback matrix are proposed.The performance of the speed estmation system based on the foil-order flux observer is decided by the selection of the feedback matrix and the parameters of the PI adaptation law. Therefore, the relationship between the performance of the speed estmation system and the parameters of the PI adaptation law is studied. The principle of the selection of the parameters of the PI adaptation law is presented based on the analysis in the frequency domain. The self-optimization of the PI parameters is also studied in the paper. A new off-line optimization method based on the presetting Genetic Algorithm is proposed.The simulation system is established with the Saber software to verify the study results in the paper. An indirect field oriented control system of the induction motor is also established using the DSP TMS320F2407. A lot of experiments are carried out to verify the full-order flux observer and the speed estimation system, which are valuable to the further study of the speed sensorless control based on the full-order flux observer.