基于扩张状态观测器的级联无刷双馈电机并网同步和发电鲁棒预测电流控制

Robust Predictive Current Control for Grid Synchronization and Power Generation of Cascaded Brushless Doubly-Fed Generators Based on Extended State Observer

级联无刷双馈电机无需电刷和集电环,具有电流谐波小、维护成本低等优点,在风力发电等领域具有广阔的应用前景。目前,级联无刷双馈电机常用的控制方法有矢量控制和无差拍控制,当控制器中使用的电机参数由于温度、饱和等因素与实际值不同时,控制效果均有不同程度的恶化。为了克服这一问题,该文提出一种基于扩张状态观测器的级联无刷双馈电机鲁棒预测电流控制方法,同时适用于并网同步和发电过程。该方法基于超局部模型使用扩张状态观测器来估计系统的动态部分和总扰动,并进一步结合无差拍控制,实现控制绕组电流的快速准确控制。与传统的矢量控制和无差拍控制进行了比较,实验结果表明,该方法不仅在电机参数准确时有良好的稳态和动态性能,在电机参数变化时同样能保持较好的控制效果,具有较强的参数鲁棒性。

The cascaded brushless doubly-fed generator(CBDFG)comprises two coaxial wound rotor induction motors,and its rotor windings are directly connected to eliminate the need for brushes and slip rings.Therefore,it has the advantages of long service life,high reliability,and low maintenance cost.CBDFG can be used for independent generation and grid-connected generation,which has broad prospects in wind power and hydroelectric power generation.At present,vector control and deadbeat control are commonly used to control the CBDFG.Because the above two methods use a large number of motor parameters,when the machine parameters in controllers are different from the actual values due to temperature,saturation,and other reasons,the control effects have varying degrees of deterioration.Therefore,this paper proposes an extended state observer(ESO)based robust predictive current control(RPCC)method for CBDFG.Firstly,the proposed RPCC can be applied to grid synchronization and power generation processes.The only difference between the two processes is the reference value calculation of the control winding current.In the process of grid synchronization,because the power side does not emit power,the virtual flux is needed to calculate the reference value of the control winding current.During the power generation process,the reference value of the control winding current is calculated directly according to the power reference value.Secondly,the proposed RPCC is based on an ultra-local model and uses an ESO to rapidly estimate the total disturbance caused by model uncertainty.The ESO can extend the total disturbance to a new disturbance and feed this total disturbance back to the controlled system.Based on the state space equation of ESO,the closed-loop transfer function of the ESO for CBDFG is obtained,and the two parameters of ESO are simplified to one.Then,the control block diagram of the proposed RPCC in the z domain is given,and the parameters of ESO are designed according to the characteristic equation of the proposed RPCC.In addition,the parameter stability of the proposed ESO is analyzed.Finally,according to the designed ESO parameters and the obtained disturbance values,one-step delay compensation is carried out for the current value of the control winding,and deadbeat control is further incorporated to achieve fast and accurate control of the control winding current.Compared with the traditional field-oriented control and deadbeat control,the experimental results show that this method has good steady-state and dynamic waveforms when the motor parameters are accurate and maintains a good control effect when the CBDFG parameters change.This method has good parameter robustness.The switching frequency is usually set at several kHz to reduce the switching loss in an actual wind power system.The simulation results of the proposed RPCC at low switching frequencies show that although the decrease in switching frequency increases the current THD,it is still far less than 5%,which meets the grid-connection requirements.