| Permanent magnet synchronous motor direct torque control (PMSM-DTC) isfinding expanded in industrial and agricultural production and in aerospace fieldbecause of its excellent dynamic performance. In the dissertation, two kinds of highprecision and high dynamic response special servo applications are the mainbackground: the dynamic load simulator for the steering gear, and the high-frequency-response missile borne electric servo system. On the premise of the superior dynamicperformance of PMSM-DTC, some related issues are studied, and the maincontributions of this dissertation are as follows:1. On the background of the missile borne electric servo system, the traditionalcontrol strategies of PMSM-DTC are studied. Currently, DTC has not as widely usedas vector control (VC) in PMSMs. Control strategy of PMSM-DTC is generally theconstant stator flux magnitude (CSFM), and there are some literatures proposed themaximum torque per ampere (MTPA). As the missile borne system having the strictrequirements on volume and weight, in this dissertation, a control strategy with unitypower factor (UPF) operation of PMSM-DTC is proposed. Under this proposedstrategy, the inverter capacity requirement is minimized, that reduces the volume andweight of the inverter. At the same time, taking into account DTC has no current loopand the current cannot be limited directly, the relationship between the stator currentvector and the stator flux vector is analyzed, a detailed calculation method toindirectly limit the current is proposed. The correctness of the proposed controlstrategy was validated in Matlab/Simulink environment, and the simulation resultsshow that the proposed control scheme achieves UPF without deterioration of thedynamic properties of the traditional DTC.2. Due to the cogging torque, the ripple torque and the errors generated from thecurrent-measurement, periodic speed ripple is presented in the PMSM-DTC system,which has seriously affected the speed servo performance of the high-frequency-response missile borne electric servo system. To minimize the speed ripple ofPMSM-DTC system, the traditional speed ripple suppression methods are discussed,the main components of speed ripple in PMSM-DTC system are analyzed and itsperiodic characteristic is described. Base on this, a new algorithms is proposed tominimize the periodic speed ripple, iterative learning control (ILC) is used to compensate the torque signal.This algorithm needs a small amount of softwarechanges to the original system, the structure and algorithm are both simple. Thisalgorithm is the beneficial attempt to improve the traditional direct torque control ofPMSMs; the purpose is to keep the excellent dynamic performance of DTC, tosuppress the periodic speed ripple and to expand the ability of PMSM-DTC for highperformance speed or position servo application.3. With the dynamic load simulator for the steering gear as the researchbackground, the loading motor often falls from synchronism at high speed, especiallywhere a large load angle is necessary to produce a high torque. The traditional DTChas no natural mechanism to prevent it. To solve this problem, a model predictivedirect torque control (MPDTC) with load angle limitation for PMSM is proposed inthis dissertation. An exact discrete-time state-space model of SPMSM is presented,which improves the state prediction accuracy comparing to simple Eulerapproximation. A finite control set type MPDTC (FCS-MPDTC) is used to select theoptimum voltage vectors applying to the voltage source inverter (VSI). It makes fulluse of the inherent discrete nature of VSI, and according to the predefined costfunction it chooses the optimal solution from the possible switching states. With theload angle limitation in the cost function, the proposed scheme can prevent thePMSMs falling from synchronism. And in order to improve the performance, the timedelay introduced by sampling and computation time is compensated during thepredictive process. The simulation and experimental results of the proposed controlscheme show that with the proposed scheme SPMSM drives show adequate dynamictorque performance and considerable torque ripple reduction as compared totraditional DTC. The proposed method provides strong support for applying DTC tomore PMSM drives.4. The errors generated from the current-measurement path are inevitable, andthey can be divided into two categories: offset errors and scaling errors. Current dataincluding these errors cause the periodic rotor speed ripples, which are one and twotimes the fundamental stator current frequency. Since these undesirable ripples canharm the motor drive system, a compensation algorithm must be included in the motorcontrol drive. In this dissertation, a combination algorithm of the online iterativelearning compensation and FCS-MPDTC is proposed. Firstly, the design process ofcontinuous control set type MPDTC (CCS-MPDTC) is analyzed, and optimizationproblems is converted from CCS type to FCS type using the vector quantizer. Then, ILC is used to compensate the current-measurement-errors, and the convergence isproved. The MATLAB simulation results verify the usefulness of the proposed currentcompensating algorithm. In this method, model predictive control (MPC) focuses onthe torque and flux trajectory tracking, and ILC is used to suppress the periodic torqueripple. With this proposed control, PMSM can keep the excellent dynamiccharacteristic, and the steady-state torque accuracy is improved with the proposed ILCcompensation. Whether for the torque servo systems like the dynamic load simulatorfor the steering gear, or for the position servo systems like the high-frequency-response missile borne electric servo system, the proposed control has essentialapplied value. |