| 题名 | 4m级望远镜主轴交流伺服控制系统研究 |
| 作者 | 邓永停 |
| 学位类别 | 博士 |
| 答辩日期 | 2015-05 |
| 授予单位 | 中国科学院大学 |
| 导师 | 王建立 |
| 关键词 | 大型望远镜 永磁同步力矩电机 空间矢量控制 |
| 其他题名 | Research on AC Control System for the 4m Scale Telescope Main Axis |
| 学位专业 | 机械电子工程 |
| 中文摘要 | 望远镜是集光、机、电于一体的综合远程观测系统。为实现对深空目标的精密跟踪，要求主轴伺服系统达到角秒级的跟踪精度。随着望远镜口径和负载的增大，要求直驱电机提供的力矩值超过105N·m，具有该驱动能力的有刷直流力矩电机在功率、体积和低速性能方面已经无法满足望远镜跟踪性能的要求。交流永磁同步力矩电机以其较高的功率密度、较小的体积和良好的低速性能可以满足跟踪精度要求，故在国外望远镜控制系统中得到了较多的应用。目前，永磁同步力矩电机在国内望远镜控制系统中的应用还处于初级阶段，因此，探索一套适用于望远镜交流伺服系统的设计方法成为亟需解决的问题。本文主要解决的问题如下：(1)如何设计专用的高精度驱动控制器；(2)如何实现望远镜转台的平稳启动和控制模型的辨识；(3)如何设计基于控制模型的控制策略；(4)如何提高望远镜的低速跟踪精度。 相比于永磁同步电机的直接转矩控制方法，矢量控制方法具有较小的转矩波动，更适合于望远镜的低速平稳运行。因此，本文采用空间矢量的控制方法建立了基于永磁同步力矩电机驱动的望远镜转台仿真模型，对永磁同步力矩电机的电流解耦控制进行了仿真，从而验证了空间矢量控制方法原理，确定了望远镜主轴控制系统对电机驱动器的最大电流和驱动电压指标要求，以指导控制系统硬件的选型和设计。 在上述仿真的基础上，设计了永磁同步力矩电机的驱动器和控制器。驱动器采用三菱公司的智能功率模块(IPM)，该驱动器具有过压、欠压、过流、高温、和电机温度检测功能；控制器采用高性能的DSP和FPGA协同处理的模式， DSP负责控制空间矢量的调制和控制器算法的实现，FPGA作为协控制器负责电机相电流的检测、编码器的数据采集和故障保护。为了进一步提高电流环的运算速度和系统的实时性，采用FPGA并行电路实现电流环运算。实验结果表明，该硬件系统不仅具有高精度电流检测能力，而且满足望远镜控制系统的实时性要求。 在硬件系统设计的基础上，对望远镜交流伺服控制系统的关键技术—转台的平稳启动和控制模型的辨识进行了研究。首先，设计了磁极搜索算法，通过给电机施加电流矢量，根据电机转向与施加的电流矢量角之间的关系，不断改变电流矢量角，从而缩小磁极定位范围，定位过程转子转动微小，解决了永磁同步力矩电机驱动的望远镜转台的平稳启动问题。其次，采用白噪声信号作为系统的激励信号，然后记录系统的激励响应序列，对输入输出序列进行离散傅立叶变换获得系统的频率特性；为了获得系统的控制模型，采用特征辨识算法(ERA)对系统的频率特性曲线进行了辨识，实验结果表明，控制模型的辨识结果与有限元分析结果相吻合。控制模型的辨识为控制策略和控制系统带宽的设计提供了模型依据。 在控制模型辨识的基础上，为了减小结构模态引起的谐振，根据结构的锁定转子频率和谐振频率设计了结构滤波器；采用传统的控制策略(超前滞后补偿)对望远镜主轴控制系统进行了仿真和实验。为了克服电机参数变化对系统性能的影响，进一步提高望远镜控制系统的低速跟踪精度，针对速度回路设计了自适应滑模控制器(ASMC)，系统的参数摄动由自适应律估计得到；为了保证望远镜转台的无静差跟踪，设计了含有积分作用的滑模面，采用了边界层的设计方法以抑制滑模控制引起的抖振，实验结果表明，在最大速度和最大加速度条件下的正弦引导误差RMS值达到0.489″，相比于传统的控制策略提高了40%。 针对望远镜低速跟踪精度的指标要求，对影响望远镜主轴低速性能的因素进行了建模和补偿；针对轴系的摩擦力矩扰动问题，通过实验辨识出了转台的LuGre摩擦模型，采用上述辨识的模型对望远镜转台的低速性能进行了仿真分析。为了减小摩擦力矩扰动对低速性能的影响，设计了扩张状态观测器对上述扰动进行了补偿，仿真和实验结果验证了扩张状态观测器补偿方法的有效性。针对电机齿槽力矩扰动的影响，建立了电机的齿槽谐波力矩小信号模型；然后，采用最小二乘法对谐波力矩模型进行了辨识；最后，基于谐波力矩模型，通过重复实验的方法对齿槽力矩波动进行了补偿，仿真结果验证了上述补偿方法的有效性。针对低速时编码器存在测速量化误差大的问题，采用卡尔曼滤波器对望远镜转台的低速进行了估计，实验结果表明卡尔曼滤波器具有较好的低速测速性能，提高了望远镜转台的低速跟踪精度。最后，对上述低速补偿方法的控制效果进行了对比分析。 |
| 英文摘要 | The Telescope is an observer system, which involves a mix of optical, mechanical and electronic system. In order to track the task accurately, the precision of servo control system is required to arcsecond. With the increasing of telescope diameter and resolution, it is required that the motor could provid torque more than 105N·m. In this situation, the brushed motor can’t meet the requirement of the telescopes’ tracking precision on power, volume and low speed. The permanent magnet synchronous motor (PMSM) is used widely in the international telescope, for its higher power density, less volume and better low speed performance. At present, the application of PMSM in the home telescope is at primary stage. So the problem of telescope servo control system design based on PMSM, is urgent to solve. In this paper, we will put mian force on the following four problems: (1) how to design the specified driver; (2) how to start the motor smoothly and identify the control model; (3) how to design the control strategy based on the control model; (4) how to increase the telescope low speed tracking accuracy. Compared with the direct torque control method, the space vector control method has less torque ripple, which is good for the telescope low speed tracking. Firstly, a PMSM simulation model, based on the space vector control, is built for the decoupling control. Using the result of the simulation, we can verify the theory of space vector control, and guide the hardware control system design on the capacity of max drive current and voltage. Based on the simulation result, the PMSM diver and controller are designed. The driver uses the intelligent power module made by Mitsubishi, which has the protect functions of over voltage, under voltage, over current and over temperature. The controller contains DSP and FPGA, as the main controller, DSP is responsible for the control algorithm. FPGA is used as assist controller, for the acquisition of motor phase current, encoder data and fault handling. In order to increase real time of the current loop, parallel hardware circuit is adopt to realize the current loop. The result of experiment verifies that the hardware system has higher detection accuracy and meets the requirement of the telescope control system. Based on the hardware design, the key technologies of telescope servo control system are studied, which are the smooth start and control model identification. Firstly, the pole position detection method is designed. With injection of current vector, changing the position angle of the current vector based on the motor direction and the angle of current vector. In the process of positioning, the rotor has small movement and meets the requirement of non-impact machine. Secondly, the white noise signal is used as the incentive signal, and the response of the system is recorded. Then the frequency feature is obtained through the DFT. In order to identify the control model, the eigensystem realization algorithm is used to fit the frequency feature curve. The result of experiment indicates that the result of control model identification fits the theory analysis. The identification of control model provides basis for the controller design. On the basis of control model identification, In order to reduce the modal resonance, a structure filter is designed. Then the lead-lag controller is designed for the telescope main axis control system. An adaptive sliding mode controller (ASMC) is designed for increasing the low speed tracking performance, the varying system parameters are evaluated by the adaptive law. The integral action contained in the sliding surface can ensure the steady state error of tracking velocity zero. And the boundary layer is employed to suppress the chattering of sliding mode control. The result of experiment indicates that the tracking accuracy reach to 0.489″ after using ASMC, compared with the traditional controller, the accuracy increases by 40%. In this chapter, the factors, which affect the telescope tracking performance, are modeled and compensated. Firstly, the friction of the telescope axis is modeled through LuGre model. Then the identification model is used for the low speed performance simulation. In order to reduce the influence of friction, the extended stated observer is adopted to estimate the disturbance and compensate the friction. Secondly, the harmonic model is built. Then the least squares method is used to identify that model. At last, that model is used to iteratively improve the cancellation. The result of simulation verifies the validity of the cogging compensation method. Thirdly, Kalman filter is used to increase the speed detection accuracy, and the filter could estimate the low speed more accurately than numerical difference. Lastly, the above low speed compensation methods are analyzed and evaluated. |
| 公开日期 | 2015-12-24 |
| 内容类型 | 学位论文 |
| 源URL | [http://ir.ciomp.ac.cn/handle/181722/48833]  |
| 专题 | 长春光学精密机械与物理研究所_中科院长春光机所知识产出 |
| 推荐引用方式 GB/T 7714 | 邓永停. 4m级望远镜主轴交流伺服控制系统研究[D]. 中国科学院大学. 2015. |
| 个性服务 |
| 查看访问统计 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
修改评论