The Research On Indentification Of Inertia For Load Commutated Inverter Fed Synchronous Motor

The electrically excited synchronous motor, as an important equipment of high power AC drive field such as pumped storage power stations and synchronous compensator of grid, is attracting more attention. Load commutated inverter(LCI) is suitable for starting control of high power synchronous motor because of its advantage in driving capacity, cost performance and reliability. In the past, parameters of double closed-loop speed adjustment system controllers are adjusted according to experience or by trial and error during the regulating process for LCI-fed synchronous motor, since there lack of understanding about mathematic model parameters of the controlled object, which leads to failure of control performance optimization. As an essential parameter of the controlled object mathematic model, inertia is of great importance for the optimization of control performance. Through obtaining the system’s inertia, an accurate mathematic model is able to be established and online parameter auto-tuning for double closed-loop speed adjustment system controllers is able to be realized further. Thus the research on identification of inertia for a LCI-fed synchronous motor is of great importance for the engineering area.To begin with, this paper analyzes fundamental principles and recurrence formula of recursive least-squares method(RLS), which is applied for identification of inertia for a LCI-fed synchronous motor. An identification strategy is proposed based on the combination of RLS and mechanical equations of LCI-fed synchronous motors. At the same time, the essential input parameters are emphasized. In order to prove the viability of this strategy, an inertia identification model based on the soft start of a LCI-fed synchronous motor is built with the help of Matlab/Simulink.Secondly, as for the real-time detection problem of motor speed during the identification of inertia, a rotor position detection method based on pulse signals of field current is introduced and the corresponding analysis of theory and feasibility is provided. Meanwhile, the corresponding model is built in Matlab/Simulink to verify the strategy. On this basis, a concrete scheme and circuit design is provided for rotor position detection, which realizes the real-time detection of rotor position and speed without sensors. This method is independent of speed, simple and reliable and has high precision at low speed compared with other sensorless methods.Finally, the experimental platform for a 15 k W/380 V LCI-fed synchronous motor, which could realize various topology structures, is established to verify the strategies of inertia identification and sensorless rotor position detection. The control system based on TMS320F2018 is designed and a detailed description of software procedure is presented for inertia identification detection and sensorless rotor position. The corresponding circuits design include band-pass filter circuit for pulse signals, envelope extraction circuit with high precision and so on. In multiple conditions, the algorithm for identification of inertia and strategy for rotor position detection are tested. The experimental results verify the availability and conformity of the proposed methods and realize the accurate inertia identification.