| The Induction Motor (IM) based Variable Frequency Drives (VFDs) have played an important part in electrical systems of industrial production. However, compared with foreign enterprises, the high-end market share of domestic enterprises in the VFD industry applications is very low and the core technology lags behind relatively. Hence, the research of VFD control systems with high performance has great significance for the development of corresponding industry applications in our country. In addition, although the VFD control theory has been developed maturely, there are still some technical problems unsolved in traditional applications. With the progress of power electronic technology and the use of advanced digital control chips, it is necessary to further study appropriate control algorithms to explore final solutions of corresponding problems. Furthermore, the VFD control systems in new applications, such as Electric Vehicles (EV), also encounter some challenges. It is rather meaningful to timely discover and solve the new coming difficulties for the future development of VFDs. This dissertation aims to achieve high performance control of IM based VFDs and studies in-depth the control algorithms of speed-sensorless vector control system based on the combined voltage and current mode flux observer. The authors try to resolve several key problems of modern IM control, including:Firstly, as the basis of high-performance VFD control system; the IM Offline Parameter Identification (OPI) becomes the primary object of study in this dissertation. As the hardware configuration and corresponding procedures are not uniform dependent on different OPI schemes, this dissertation integrates and optimizes a set of OPI methods based on a deep research of this subject for the purpose of industrial applicability. In addition, it is difficult to ensure the traditional rotating and no-load test conditions for OPI in most applications. With reference to the existing OPI methods which maintain the rotor at standstill, the inverter nonlinearity error has a significant impact on the accuracy of identification results. To solve this problem, this dissertation presents a novel OPI method based on the DC biased excitations, which effectively avoid the influence of the inverter nonlinearity errors on the final results. Accurate OPI results of IM are obtained and verified by experiments. Secondly, the Magnetizing Curve Identification (MCI) of IM has been studied. In order to ensure the accuracy of steady-state performance, MCI and corresponding online updating methods play an important role in the variable flux control, which is just the core technology of optimal efficiency algorithm of high-performance VFD systems. Traditional MCI methods are based on specific analytical function assumptions with corresponding curve fitting strategies, the drawback of which is that final identification results are different dependent on different assumptive curve functions. This dissertation presents a direct calculation method to obtain the magnetizing curve with a slow-ramped excitation-thereby avoids the dependence on specific assumptive functions in the curve fitting algorithms. The proposed method is simple and convenient for implementation, and its principle has been verified by simulation. Further a detailed comparison between the proposed method and the traditional curve fitting methods in the optimal efficiency experiments are given. Experimental results show that the proposed scheme can be well used in practice.Thirdly, the regenerating mode instability and its improvements of combined voltage and current mode flux observer (also named as UI model) has been mainly studied. Compared with other observers, the UI model has some advantages, such as the simple structure, low parameter sensitivities etc., but the unstable regions have been observed when it works in the regenerating mode, and there is no effective solution for the moment. In fact, elevators and other applications which need load elevating usually require the VFDs to work in the regenerating mode, and also require the high reliability of the system. Hence, the research of regenerating instability improvements cannot be ignored, especially for the high-performance VFD control systems. This dissertation constructs the small signal model of the UI mode flux observer, and derives the characteristic matrix to realize the stability analysis, based on which, the stable range with corresponding conditions has been calculated explicitly. Furthermore, a cross-coupling feedback mechanism is proposed to solve the instability problem when the VFD runs in the low-speed regenerating or plugging-in mode. The improvement strategy is well verified by both the simulation and corresponding experimental results.Finally, a novel rotor time constant (Tr) online identification method has been proposed. The rotor time constant can produce large changes due to the thermal effect, which will lead to large speed estimation errors. Hence, effective Tr online identification is an important means to meet the steady-state precisions of high-performance VFD systems. The difficulty of Tr online identification lies in how to precisely obtain the AC signals used for calculations. With reference to the existing online identification methods, the precise separation and filtering of DC and AC signals is usually needed and the AC signals used for final identification cannot be obtained directly but need further complex calculations. However, in the proposed method of this dissertation, corresponding AC signals can be extracted from the UI model directly without any signal separation or filtering operations, which not only simplifies the identification process but also improves the final accuracy. Furthermore, both amplitudes and phases of corresponding AC signals have great impacts on the identification accuracy in existing methods whereas the newly proposed method requires only the amplitudes of AC signals without considering the phase drifts, which means the dependence on the AC signals for Tr online identification has been reduced from two dimensions (amplitude and phase) to only one dimension (amplitude). Corresponding simulation and experimental results effectively verify the correctness and practicality of the proposed method. |
PDF Full Text Request