基于电流信号提取技术的SRD故障检测方法的研究

发布时间：2018-05-02 20:38

本文选题：开关磁阻电机 + 故障检测　；参考：《河北工业大学》2014年博士论文

【摘要】：开关磁阻电机（SRM）由于其结构简单及独立的相间控制，具备较高可靠性，与其它电机相比具有一定的容错能力。但其固有的容错能力也是有限的，电机长期故障运行，必将对整个系统造成严重危害。因此，对开关磁阻电机调速系统（SRD）故障检测技术的研究具有十分重要的意义，以便采取适当的容错策略。本论文以SRD故障检测技术的研究为主题，基于电流信号提取技术分别对电机定子绕组匝间短路故障、功率变换器故障、气隙偏心这三类故障提出了新的检测方法。首先，本文基于GA-BP算法对SRM进行建模。建立以电机绕组相电流和转子位置角为输入、磁链和转矩为输出的神经网络模型，利用GA-BP混合算法对网络进行训练。与传统BP神经网络所建模型相比较，该法建模具有较高准确度和可靠性。并基于Matlab/Simulink环境，提出一种独立功能模块化的8/6极SRM建模方法。分别对电机本体、功率变换器、速度控制、电流控制、转子位置角选择、转矩计算及参数计算等各部分进行模块化建模，再将各模块有机组合在一起，构建出SRD整体仿真模型。模型的速度控制器由PI调节器构成，电流控制模块将采用角度位置控制（APC）与电流斩波控制（CCC）相结合的控制方式，有效地实现了电机实时控制。其次，基于电磁分析软件Maxwell与电路分析软件Simplorer建立SRD仿真系统，深入分析了绕组匝间短路故障原理，分别从磁路和场路的角度对电机一极绕组完全短路和匝间短路两种情况进行仿真研究。仿真实验结果证明绕组短路故障破坏电机磁极间对称性，对电机输出性能造成影响。在此基础上，基于相电流频谱分析，提取基波分量重构SRM四相的对称电流，利用对称分量法计算重构电流的正序分量及负序分量，以正序分量与负序分量的比值作为故障特征量。不同工况下的仿真与实验结果证明了该方法可有效地实现SRM定子绕组匝间短路故障检测。此外，基于电机正常运行与故障运行时直流母线电流、相电流的不同表现提取故障特征量，并结合功率元件通断状态进行综合分析，提出一种能够及时准确地检测出故障发生、诊断出故障相、判断出故障类型、定位出故障元件的综合检测方案。仿真与实验结果证明该检测方案在电机高速、低速运行时的准确性和可靠性。同时，基于SRM电路方程的推导，得到以气隙长度为变量的非励磁相电流的函数解析式。利用有限元分析法计算电机处于正常状态及不同偏心等级情况下非励磁相电流，通过对比，提出一种以非励磁相电流差值为特征量的气隙偏心故障检测方法。该法能够实现偏心故障发生、故障相定位及偏心类型辨别的综合检测。最后，通过对SRM基本电路方程的推导，得出电机同相两定子极绕组的电流差值与气隙长度的函数关系式，利用有限元分析法计算出电机正常状态及不同偏心等级情况下绕组电流差值，并将不同励磁模式下绕组电流差值转化为矩阵表达式。通过分析矩阵表达式各元素的对称性，针对8/6极SRM提出一种以同相两绕组电流差值为特征量的气隙偏心故障检测方法。该法能够实现偏心故障发生诊断、故障相定位、偏心方向判定及偏心类型辨别的综合检测。仿真与实验结果表明，该法具有较高准确性、可靠性及鲁棒性，，并总结了上述两种偏心故障检测方案特点的对比。
[Abstract]:Switched reluctance motor (SRM) has high reliability because of its simple structure and independent interphase control. It has a certain fault tolerance ability compared with other motors. But its inherent fault tolerance is also limited. Long term failure of the motor will cause serious harm to the whole system. Therefore, the switch reluctance motor speed control system (SRD) failure The research of detection technology is of great significance in order to adopt appropriate fault tolerance strategy. This paper takes the research of SRD fault detection technology as the subject. Based on the current signal extraction technology, a new detection method is proposed for the three types of faults such as interturn short circuit fault, power converter fault and air gap eccentricity of motor stator winding.
First, this paper builds a model of SRM based on GA-BP algorithm. A neural network model is built with the input of phase current and rotor position angle as input, magnetic chain and torque output. The GA-BP hybrid algorithm is used to train the network. Compared with the traditional BP neural network model, the model has high accuracy and reliability. And it is based on M. In atlab/Simulink environment, an independent function modularized 8/6 polar SRM modeling method is proposed. The modules of motor body, power converter, speed control, current control, rotor position angle selection, torque calculation and parameter calculation are modular modeling, and each module is organically combined to build a SRD overall simulation model. The speed controller of the type is composed of PI regulator, and the current control module will use the control mode of angle position control (APC) and current chopper control (CCC), which effectively realizes the real time control of the motor.
Secondly, based on the electromagnetic analysis software Maxwell and the circuit analysis software Simplorer, the SRD simulation system is established, and the principle of the winding interturn short circuit fault is deeply analyzed. The simulation of two cases of the complete short circuit and the interturn short circuit of the one pole winding of the motor is carried out from the angle of the magnetic circuit and the field path. The simulation experiment results show that the short circuit fault of the winding is damaged by the electricity. On the basis of phase current spectrum analysis, the symmetric current of the SRM four phase is reconstructed based on the phase current spectrum analysis. The positive sequence and negative sequence components of the reconstructed current are calculated by the symmetric component method, and the ratio of the positive sequence component to the negative sequence component is used as the fault characteristic quantity. Simulation and experimental results show that the proposed method can effectively detect the inter turn short circuit fault of SRM stator windings.
In addition, based on the different performance of the DC bus current and phase current in the normal operation and fault operation of the motor, the fault features are extracted from the different performance of the phase current. And combined with the comprehensive analysis of the power element's off state, a kind of timely and accurate detection of the fault, the diagnosis of the fault phase, the fault type, and the integrated detection of the fault components are proposed. The simulation and experimental results prove the accuracy and reliability of the detection scheme when the motor is running at high speed and low speed.
At the same time, based on the derivation of the SRM circuit equation, the function analytical formula of the non excitation phase current is obtained with the air gap length as the variable. By using the finite element method, the non excitation phase current is calculated under the condition of normal state and different eccentricity grade. By contrast, a kind of gas gap eccentricity fault detection with the characteristic amount of the non excitation phase current difference is proposed. This method can realize the comprehensive detection of eccentricity failure, fault location and eccentricity type discrimination.
Finally, through the derivation of the basic circuit equation of SRM, a function relation between the current difference and the air gap length of the motor phase two stator winding is obtained. The current difference between the normal state of the motor and the different eccentricity grades is calculated by the finite element method, and the current difference of the winding current in different excitation modes is transformed into a matrix expression. By analyzing the symmetry of each element in matrix expression, a method of detecting air gap eccentricity fault detection is proposed for 8/6 pole SRM with the current difference of two winding current as the characteristic quantity. This method can realize the diagnosis of eccentricity fault occurrence, fault location, eccentricity direction determination and eccentricity type discrimination. The method has high accuracy, reliability and robustness, and summarizes the comparison of the above two eccentric fault detection schemes.

【学位授予单位】：河北工业大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TM352

【参考文献】

中国期刊全文数据库前10条

1 纪良文,蒋静坪,何峰;基于径向基函数神经网络的开关磁阻电机建模[J];电工技术学报;2001年04期

2 张兰红,胡育文,黄文新;三相变频驱动系统中逆变器的故障诊断与容错技术[J];电工技术学报;2004年12期

3 崔博文;任章;;基于傅里叶变换和神经网络的逆变器故障检测与诊断[J];电工技术学报;2006年07期

4 修杰;夏长亮;王世宇;;开关磁阻电机的Pi-sigma模糊神经网络建模[J];电工技术学报;2009年08期

5 丁文;梁得亮;;基于RBFN-AFS的开关磁阻电机非线性模型与动态仿真[J];电工技术学报;2009年09期

6 卢胜利;陈昊;昝小舒;;开关磁阻电机功率变换器的故障诊断与容错策略[J];电工技术学报;2009年11期

7 宋建成;郑建斌;曲兵妮;张宏达;;开关磁阻电机的最小二乘支持向量机建模与仿真[J];电机与控制学报;2010年05期

8 陈小元;邓智泉;连广坤;范娜;许培林;;高容错性模块化定子开关磁阻电机[J];电机与控制学报;2010年06期

9 许爱德;樊印海;李自强;;基于GA-ANFIS的开关磁阻电机建模[J];电机与控制学报;2011年07期

10 蔡永红;齐瑞云;蔡骏;邓智泉;;基于RBF神经网络的开关磁阻电机在线建模及其实验验证[J];航空学报;2012年04期