【作者】 张云；

【导师】 徐衍亮；

【作者基本信息】 山东大学 ， 电机与电器， 2014， 博士

【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。更多还原

【Abstract】 Axial field disk type switched reluctance motor integrates the advantages of switched reluctance motor and disk type motor, so this kind motor has high power density, large torque, compact structure, etc. This paper has studied the motor structure, calculation of magnetic circuit, optimization of control algorithm and design of control system.Magnetic circuit analysis is an important means for the electromagnetic design and performance analysis of all kinds of motor. At present, the magnetic circuit analysis method widely used in switched reluctance motor is based on analytical calculation of magnetization curve of several key rotor positions, and this method has not been applied to the study of axial field disk type switched reluctance motor, which made this kind of motor has not been widely used. For the axial field disk type switched reluctance motor, this paper has calculated the magnetization curve of the center alignment position of the stator tooth and the rotor tooth, the center alignment position of the stator tooth and the rotor groove and the critical position of the stator tooth and the rotor tooth. First of all, the key positions of the magnetic circuit structure had been determined according to the results of finite element calculation. Then, the winding coil’s flux, which had been produced by a given current, was analytical calculated. In the calculating process, the magnetic field lines were equivalent to circular arc or straight line, and the core magnetic resistance was ignored. So we could get the relationship between the flux and current, namely magnetization curve.First of all, the key position’s magnetization curve of the axial field disk type switched reluctance motor was modeled. Then, the design method of radial magnetic field switched reluctance motor was used to study the motor. Finally, a12/8single stator and single rotor axial field disk type switched reluctance motor has been manufactured, and the comparison of numerical calculation result, magnetic circuit analytical calculation result and testing result has proved the correctness and effectiveness of the analytical calculation.The axial field disk type switched reluctance motor has double convex pole structure, which is different from the radial magnetic field switched reluctance motor, and the differences will lead to different mathematical models of the two types of motor. It is difficult to establish more accurate mathematical model because of the nonlinear electromagnetic characteristics of switched reluctance motor, and it is not feasible using traditional control method to drive this motor. This paper has proposed the optimal axial field disk type switched reluctance motor control strategy based on neural network. First of all, through the discrete experiments on the prototype motor, it has been found that the switched reluctance motor’s turn-on and turn-off angle has important influence on the output torque, and the optimal switching angle of the motor has been defined. Secondly, the complicated nonlinear relationship between the inputs and outputs of switched reluctance motor was studied, and the neural network has been introduced to the motor driving. Then, current optimal switched reluctance motor nonlinear multivariable static three layer BP neural network controller model has been designed, which output variables were target current, turn-on angle and turn-off angle, the input variables were target torque and current speed of motor, and this neural network controller can be combinated with the traditional PID controller to constitute the speed feedback control system so that the system could have some dynamic characters. The methods of online training neural network have been designed, and variable step size fitting optimization method based on least squares has been designed in order to get online training data quickly. Finally, the prototype experiments have proved the correctness of the analysis and the validity of neural network control in switched reluctance motor. In the axial field disk type switched reluctance motor control system, this paper has studied the main circuit structure and MOSFET driving optimization etc. In the main circuit structure, this paper has proposed an H bridge structure based on synchronous rectification to control switched reluctance motor, in this main circuit structure, the plurality parallel power MOSFETs were used to instead the freewheeling diodes used in half bridge structure, and rational control method should be designed to realize the freewheeling function. Theoretical analysis and experiments have proved that the proposed control method could reduce the wheeling power significantly and improve the efficiency of power conversion.This paper has also proposed a MOSFET optimization method based on dynamic driving power, and the dynamic power assistant system designed in this paper has realized the power MOSFET ideal driving, which could reduce electromagnetic radiation and increase the reliability of the system. This driving method has two working stage:double power driving stage and single power driving stage. In the double power driving stage, appropriate driving parameters should be selected to make the driving circuit working in the opening delay stage, which can increase the driving current effectively and reduce opening delay time of MOSFET. In single power driving stage, the drive system works in the current rising stage of MOSFET, one part of the driving output current is charging the dynamic power, and the other part is driving the MOSFET, which can reduce the driving current and slow down rising rate of the drain current. Then, after the gate voltage increased to Miller voltage, MOSFET would get into voltage drop stage, the gate voltage of MOSFET will be fixed to Miller voltage value, and all output driving current is charged to gate capacitance of MOSFET, which will shorten the duration of the Miller effect and accelerate the MOSFET drain source voltage drop speed. Finally, the gate voltage will begin to increase after the Miller effect is finished, and all of the output driving current come back to the state of charging the dynamic power and MOSFET gate capacitors. Experiments have shown that the optimized driving method based on dynamic power could effectively optimize the opening process of MOSFET.更多还原