轴向磁场盘式开关磁阻电机参数计算及其驱动控制系统研究

发布时间：2018-06-03 11:56

本文选题：轴向磁场 + 盘式开关磁阻电机　；参考：《山东大学》2014年博士论文

【摘要】：轴向磁场盘式开关磁阻电机具有开关磁阻电机和轴向磁场盘式电机的综合优势,因而具有功率密度高、转矩大、结构紧凑等优点。本文从电机结构、磁路计算、控制算法优化及控制系统设计等多个方面对轴向磁场盘式开关磁阻电机进行了分析与研究。磁路的解析分析是各类电机电磁设计、性能分析不可或缺的重要手段,对径向磁场开关磁阻电机,磁路的解析分析以几个关键转子位置处磁化曲线的解析计算为基础,目前已得到了很好的研究和应用。对轴向磁场盘式开关磁阻电机,这方面的研究还是空白,影响到该种电机的深入研究和应用开发。本文针对轴向磁场盘式开关磁阻电机,解析计算了定子盘和转子盘齿中心线对齐位置、齿槽中心线对齐位置和临界对齐位置三个关键位置处的磁化曲线。首先根据电磁场的有限元计算结果,确定了各关键位置处的磁路结构；然后给定绕组线圈电流,并将磁场磁力线等效为圆弧及直线,忽略铁心磁阻,解析计算相应产生的磁链,并由此得到磁链与电流之间的关系,即磁化曲线。对基于解析计算所得到的轴向磁场盘式开关磁阻电机关键位置处的磁化曲线,首先对其模化处理,然后借用常规径向磁场开关磁阻电机的设计方法进行该种电机的电磁设计研究。最后,设计制造了一台12/8极单定子盘、单转子盘轴向磁场盘式开关磁阻电机,并对该电机进行了关键转子位置处磁化曲线的三维有限元计算和实际测量,数值计算结果、实测结果与解析计算结果基本相符,证明了前述解析计算的正确性和有效性。传统的径向磁场开关磁阻电机为双凸极结构,而轴向磁场盘式开关磁阻电机为双平面凸极结构,这种结构的不同将导致两类电机的数学模型存在差异。而开关磁阻电机本身具有非线性的电磁特性,难以建立精确的数学模型,这对该种电机采用传统驱动控制方法带来很大的困难。本文提出并研究了基于神经网络的轴向磁场盘式开关磁阻电机的最优控制策略,首先,通过对电机样机的离散实验,研究了开关磁阻电机的开通角和关断角对输出转矩的重要影响,并由此定义了开关磁阻电机最优开关角的概念；其次,从控制的角度,确立了开关磁阻电机多输入、多输出的复杂非线性关系,从而引入了神经网络在开关磁阻电机驱动控制中的应用研究；然后,采用三层BP神经网络,设计了电流最优的开关磁阻电机非线性多变量静态神经网络控制器模型,其输出为目标电流、开通角及关断角,输入为目标转矩及电机当前转速,将这一神经网络控制器与传统的PID控制器相结合,可构成反馈控制系统,从而使系统具有一定的动态特性。在神经网络驱动控制的实施过程中,为了获得训练数据,设计了神经网络在线训练方法,这一方法利用基于最小二乘法的变步长拟合寻优方法,可以快速选择在线训练的数据；最后,初步实现了开关磁阻电机的神经网络驱动控制系统,并进行了样机试验,试验结果证明了前述分析的正确性及神经网络在开关磁阻电机驱动控制中的有效性。在轴向磁场盘式开关磁阻电机控制系统方面,本文从主电路结构、MOSFET驱动优化等方面进行了深入的分析研究。在主电路结构方面,本文提出了一种基于同步整流技术的H桥结构开关磁阻电机驱动控制方式,用多个功率MOSFET并联的形式代替不对称半桥结构中的续流二极管,通过合理的控制,实现续流功能。理论分析与实验证明,本文提出的基于同步整流技术的H桥结构开关磁阻电机驱动控制方式,MOSFET的续流压降明显低于原有技术中二极管续流时的续流压降,降低了续流功耗,提高了主电路的功率转换效率。在MOSFET驱动优化方面,本文提出了一种基于动态电源的MOSFET优化驱动方法,该驱动方法在专用驱动芯片直接驱动的基础上添加了动态电源辅助系统,实现了功率MOSFET的理想驱动,降低了电磁辐射,增加了系统运行的可靠性。这一驱动方法的工作过程分为动态电源与驱动芯片共同驱动和驱动芯片单独驱动两个阶段。共同驱动阶段为双电源驱动模式,通过选择合适的驱动参数,使该驱动阶段恰好工作于MOSFET的开通延迟阶段,可有效增加驱动电流,减少开通延迟时间；在单独驱动阶段,驱动系统首先工作在MOSFET的电流上升阶段,驱动芯片的输出电流一部分给动态电源充电,另一部分用于驱动MOSFET,驱动电流有所降低,从而减缓了漏极电流的上升速度；然后,当栅极电压升高到密勒电压后,MOSFET进入电压下降阶段,栅极电压固定为密勒电压值,此时驱动芯片的驱动电流停止给动态电源供电,全部输入到MOSFET的栅极电容中,有效的缩短了密勒效应的持续时间,加快了MOSFET漏源电压的下降速度；最后,当密勒效应结束后,MOSFET的栅极电压开始升高,此时驱动芯片的输出电流又恢复到给动态电源和MOSFET栅极电容充电的状态,直到驱动过程结束。实验表明,本文提出的基于动态电源的MOSFET优化驱动方法,能够有效地优化MOSFET的开通过程。
[Abstract]:The axial magnetic disk type switched reluctance motor has the advantages of the switched reluctance motor and the axial magnetic disk motor, so it has the advantages of high power density, large torque and compact structure. In this paper, the axial magnetic disk type switched reluctance motor is carried out in many aspects, such as the structure of the motor, the calculation of the magnetic circuit, the optimization of the control algorithm and the design of the control system. Analysis and research.
The analytical analysis of magnetic circuit is an important means for all kinds of motor electromagnetic design and performance analysis. The analytical analysis of the radial magnetic field switched reluctance motor and the analytical analysis of the magnetic circuit are based on the analytical calculation of the magnetization curves at several key rotor positions. At present, the magnetic circuit has been well studied and applied. In this paper, the magnetization curves of the axis alignment position of the stator disc and the rotor disc tooth center line, the alignment position of the cogging center line and the critical position position at the three key position position are analyzed and calculated. First, the electromagnetic field is based on the electromagnetic field. The magnetic circuit structure at each key position is determined by the finite element calculation, then the coil current of the winding is given, and the magnetic field line is equivalent to the arc and the straight line, the magnetic resistance of the core is ignored and the corresponding magnetic chain is calculated, and the relationship between the magnetic chain and the current is obtained, that is, the magnetization curve.
On the basis of the magnetization curve at the critical position of the axial magnetic disk type switched reluctance motor based on the analytical calculation, the electromagnetic design of the motor is studied by using the conventional radial magnetic field switched reluctance motor (SRM). Finally, a 12/8 pole single stator disk is designed and manufactured, and the axial magnetic field of the single rotor disk is designed. The field disk type switched reluctance motor is used to calculate the magnetization curve at the key rotor position and to measure the magnetization curve at the key rotor position. The results of the numerical calculation agree basically with the analytical results, proving the correctness and validity of the analytical calculation mentioned above.
The traditional radial magnetic field switched reluctance motor is a double salient structure, and the axial magnetic disk switched reluctance motor is a double plane salient structure. The difference in this structure will lead to the difference in the mathematical model of the two types of motor. And the switched reluctance motor itself has the nonlinear electromagnetic characteristics, and it is difficult to establish a precise mathematical model. In this paper, the optimal control strategy of the axial magnetic disk switched reluctance motor based on neural network is proposed and studied. First, the important influence of the opening angle and the turn angle of the switched reluctance motor on the output torque is studied by the discrete experiment of the motor prototype, and the definition is also defined. The concept of optimal switching angle of switched reluctance motor is introduced. Secondly, from the angle of control, the complex nonlinear relation of multiple input and multi output of switched reluctance motor is established, and the application of neural network in the drive control of switched reluctance motor is introduced. Then, the optimal switched reluctance is designed by using three layers of BP neural network. The nonlinear multivariable static neural network controller model, whose output is the target current, the opening angle and the turn off angle, is input to the target torque and the current speed of the motor. This neural network controller is combined with the traditional PID controller to form a feedback control system, so that the system has a certain dynamic characteristics. In the neural network drive, the system has some dynamic characteristics. In the implementation of the dynamic control, a neural network on-line training method is designed to obtain the training data. This method uses the variable step length fitting method based on the least square method, and can quickly select the online training data. Finally, the neural network driven control system of the switched reluctance motor is preliminarily realized, and the prototype is carried out. The test results prove the validity of the above analysis and the effectiveness of neural network in the drive control of switched reluctance motor.
In the aspect of the axial magnetic disk switched reluctance motor control system, this paper makes an in-depth analysis of the main circuit structure, MOSFET drive optimization and so on. In the main circuit structure, this paper proposes a H bridge structure switched reluctance motor drive control method based on synchronous rectifier technology, which is used in the form of multiple power MOSFET parallel. Instead of the continuous flow diode in the asymmetrical half bridge structure, the continuous flow function is realized by reasonable control. The theoretical analysis and experiment prove that the H bridge structure switched reluctance motor drive control method based on the synchronous rectifier technology is proposed in this paper. The MOSFET continuous flow pressure drop is obviously lower than the continuous current pressure drop of the diode in the original technology. The power consumption of the continuous circuit improves the power conversion efficiency of the main circuit.
In the aspect of MOSFET drive optimization, this paper proposes a dynamic power based MOSFET optimization drive method. The driving method adds a dynamic power supply auxiliary system on the basis of the direct drive of the dedicated driver chip. It realizes the ideal drive of the power MOSFET, reduces the electromagnetic radiation and increases the reliability of the system operation. This driver is the driver. The working process of the method is divided into two stages: the dynamic power supply and the driver chip co drive and drive the chip to drive separately. The common drive stage is a dual power drive mode. By selecting the appropriate driving parameters, the driving stage is exactly working in the MOSFET opening delay stage, which can increase the driving current and reduce the opening delay time. In the separate driving stage, the drive system works first in the MOSFET current rising stage, the drive chip output current is partly charged to the dynamic power supply, the other is used to drive the MOSFET, and the drive current is reduced, thus slowing the rise of the leakage current; then, when the grid voltage is raised to the Miller voltage, the MOSFET enters the voltage. In the decline stage, the gate voltage is fixed to the Miller voltage value. At this time the driving current of the drive chip is stopped to the dynamic power supply, all input to the gate capacitance of the MOSFET, effectively shortens the duration of the Miller effect and accelerates the decline speed of the MOSFET leakage source voltage; at the end, when the Miller effect is over, the gate voltage of the MOSFET is open. At this time, the output current of the drive chip is restored to the state of charging the dynamic power supply and the MOSFET gate capacitance until the end of the driving process. The experiment shows that the proposed MOSFET optimization method based on the dynamic power supply can effectively optimize the operation of the MOSFET.
【学位授予单位】：山东大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TM352

【参考文献】