留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

RBF神经网络在静电悬浮位置控制中的应用

陆潇晓 刘晓珂 李虎 郑福 孙志斌 于强

陆潇晓, 刘晓珂, 李虎, 郑福, 孙志斌, 于强. RBF神经网络在静电悬浮位置控制中的应用[J]. 空间科学学报, 2022, 42(5): 952-960. doi: 10.11728/cjss2022.05.210927103
引用本文: 陆潇晓, 刘晓珂, 李虎, 郑福, 孙志斌, 于强. RBF神经网络在静电悬浮位置控制中的应用[J]. 空间科学学报, 2022, 42(5): 952-960. doi: 10.11728/cjss2022.05.210927103
LU Xiaoxiao, LIU Xiaoke, LI Hu, ZHENG Fu, SUN Zhibin, YU Qiang. RBF Neural Network in Electrostatic Levitation Position Control (in Chinese). Chinese Journal of Space Science, 2022, 42(5): 952-960 doi: 10.11728/cjss2022.05.210927103
Citation: LU Xiaoxiao, LIU Xiaoke, LI Hu, ZHENG Fu, SUN Zhibin, YU Qiang. RBF Neural Network in Electrostatic Levitation Position Control (in Chinese). Chinese Journal of Space Science, 2022, 42(5): 952-960 doi: 10.11728/cjss2022.05.210927103

RBF神经网络在静电悬浮位置控制中的应用

doi: 10.11728/cjss2022.05.210927103
基金项目: 中国科学院科研仪器设备研制项目(YJKYYQ20190008)和空间科学战略先导专项(XDA15013600)共同资助
详细信息
    作者简介:

    于强:E-mail:yuqiang@nssc.ac.cn

  • 中图分类号: V524,TP273.2

RBF Neural Network in Electrostatic Levitation Position Control

  • 摘要: 静电悬浮位置控制系统具有非线性、时变性的特点,传统的控制方法不能有效抑制扰动的影响。针对该问题提出了一种神经网络与PID相结合的RBF-PID控制策略。以球形样品为例分析其受力情况,推导出静电悬浮位置控制系统的机理模型。搭建基于RBF-PID控制器的静电悬浮位置控制系统,并根据仿真结果实时在线调整控制参数。仿真结果表明,当样品带电量从10–9 C突变至3×10–9 C时,RBF-PID控制器只需0.12 s即可使样品达到稳定状态。实验结果表明,当样品处于加热状态时,实时调整参数后系统的平均绝对误差为0.0416 mm,控制效果比传统PID控制策略提高了70%。所提出的控制方法辨识精度高,具有比传统PID方法更强的鲁棒性和稳定性。

     

  • 图  1  静电悬浮位置控制系统工作原理

    Figure  1.  Position control system of electrostatic levitation

    图  2  样品位置控制回路

    Figure  2.  Control loop of sample position

    图  3  位置平衡处样品位置与控制电压的关系

    Figure  3.  Relationship between the sample position and the controlling voltage when sample position is in equilibrium

    图  4  RBF神经网络结构

    Figure  4.  Structure diagram of RBF neural network

    图  5  RBF-PID控制器结构回路

    Figure  5.  Structure loop diagram of RBF-PID controller

    图  6  静电悬浮位置控制系统非线性仿真结果

    Figure  6.  Nonlinear simulation result of position control system of electrostatic levitation

    图  7  各控制器的初始响应曲线

    Figure  7.  Initial response curve of different controllers

    图  8  电荷量突变时各控制器的响应曲线

    Figure  8.  Response curve of different controllers when the charge changes suddenly

    图  9  静电悬浮实验装置

    Figure  9.  Experimental facility of electrostatic levitation

    图  10  样品锆悬浮实验

    Figure  10.  Diagram of sample zirconium in levitation state

    图  11  室温下各控制器的样品位置对比

    Figure  11.  Sample position comparison of different controllers at room temperature

    图  12  加热状态下各控制器的样品位置对比

    Figure  12.  Sample position comparison of different controllers at heating state

    表  1  系统仿真参数

    Table  1.   Parameters of system simulation

    系统参数非线性系统
    参数值
    线性系统
    参数值
    样品质量/mg2626
    上下电极间距/mm1010
    样品带电量/C10–910–9
    学习速率 η3070
    动量因子 α0.040.09
    参数P学习速率 ηp500700
    参数I学习速率 ηi850950
    参数D学习速率 ηd350490
    下载: 导出CSV

    表  2  仿真结果对比

    Table  2.   Comparison of simulation results

    样品电荷量突变值/C
    0.5×10–92×10–93×10–9
    最大
    误差
    /mm
    稳定
    时间
    /s
    振荡
    情况
    最大
    误差
    /mm
    稳定
    时间
    /s
    振荡
    情况
    最大
    误差
    /mm
    稳定
    时间
    /s
    振荡
    情况
    传统PID0.2478.6剧烈0.1260.242较大0.1700.249较大
    最优整定PID0.1760.2260.0830.138较小0.1120.173较小
    RBF-PID0.1740.1180.0790.1100.1250.122
    下载: 导出CSV

    表  3  实验结果对比

    Table  3.   Comparison of experimental results

    控制器室温加热
    误差范围
    /mm
    平均绝对
    误差/mm
    误差范围
    /mm
    平均绝对
    误差/mm
    传统PID–0.130~0.1100.0384–0.430~0.6900.1879
    最优整定PID–0.080~0.1100.0269–0.143~0.2960.0685
    RBF-PID–0.095~0.0800.0172–0.140~0.1850.0416
    下载: 导出CSV
  • [1] 孙一宁, 王飞龙, 于强, 等. 静电悬浮条件下的材料典型热物理性质测量[J]. 材料导报, 2016, 30(S2): 253-258

    SUN Yining, WANG Feilong, YU Qiang, et al. Thermophysical property measurements by electrostatic levitation in material science[J]. Materials Review, 2016, 30(S2): 253-258
    [2] WANG F L, DAI B, LIU X F, et al. Containerless heating process of a deeply undercooled metal droplet by electrostatic levitation[J]. Chinese Physics Letters, 2015, 32(11): 114101 doi: 10.1088/0256-307X/32/11/114101
    [3] MOHR M, WUNDERLICH R K, KOCH S, et al. Surface tension and viscosity of Cu50Zr50 measured by the oscillating drop technique on board the international space station[J]. Microgravity Science and Technology, 2019, 31(2): 177-184 doi: 10.1007/s12217-019-9678-1
    [4] China Manned Space. High quality domestic products, strong as the sky![EB/OL]. (2021-07-21)[2021-09-25]. http://www.cmse.gov.cn/kjkx/kjkxyjyyy/202107/t20210726_48473.html
    [5] RULISON A J, WATKINS J L, ZAMBRANO B. Electrostatic containerless processing system[J]. Review of Scientific Instruments, 1997, 68(7): 2856-2863 doi: 10.1063/1.1148208
    [6] ISHIKAWA T, KOYAMA C, PARADIS P F, et al. Densities of liquid Re, Os, and Ir, and their temperature dependence measured by an electrostatic levitator[J]. International Journal of Refractory Metals and Hard Materials, 2020, 92: 105305 doi: 10.1016/j.ijrmhm.2020.105305
    [7] HERLACH D. Crystal nucleation and dendrite growth of metastable phases in undercooled melts[J]. Journal of Alloys and Compounds, 2011, 509(S1): S13-S17
    [8] LEE G W, JEON S, PARK C, et al. Crystal-liquid interfacial free energy and thermophysical properties of pure liquid Ti using electrostatic levitation: Hypercooling limit, specific heat, total hemispherical emissivity, density, and interfacial free energy[J]. The Journal of Chemical Thermodynamics, 2013, 63: 1-6 doi: 10.1016/j.jct.2013.03.012
    [9] HU L, WANG H P, XIE W J, et al. Electrostatic levitation under the single-axis feedback control condition[J]. Science China Physics, Mechanics and Astronomy, 2010, 53(8): 1438-1444 doi: 10.1007/s11433-010-4068-0
    [10] ZOU Z Z, LUO X H, YU Q. Droplet image super resolution based on sparse representation and kernel regression[J]. Microgravity Science and Technology, 2018, 30(3): 321-329 doi: 10.1007/s12217-018-9597-6
    [11] 陈东阳, 郭清远, 董文博, 等. 基于高速视觉的静电悬浮控制系统[J]. 光学 精密工程, 2019, 27(11): 2343-2353 doi: 10.3788/OPE.20192711.2343

    CHEN Dongyang, GUO Qingyuan, DONG Wenbo, et al. Control system of electrostatic levitation based on high-speed vision[J]. Optics and Precision Engineering, 2019, 27(11): 2343-2353 doi: 10.3788/OPE.20192711.2343
    [12] NAKAMURA T, AWA Y, SHIMOJI H, et al. Control system of electrostatic levitation furnace[J]. Acta Astronautica, 2002, 50(10): 609-614 doi: 10.1016/S0094-5765(01)00219-3
    [13] MEISTER T, WERNER H, LOHOEFER G, et al. Gain-scheduled control of an electrostatic levitator[J]. Control Engineering Practice, 2003, 11(2): 117-128 doi: 10.1016/S0967-0661(02)00102-8
    [14] NUELLA I, CHENG C, CHIU M S. Adaptive PID controller design for nonlinear systems[J]. Industrial & Engineering Chemistry Research, 2009, 48(10): 4877-4883
    [15] QIAO J F, LI F, YANG C L, et al. A self-organizing RBF neural network based on distance concentration immune algorithm[J]. IEEE/CAA Journal of Automatica Sinica, 2020, 7(1): 276-291 doi: 10.1109/JAS.2019.1911852
    [16] 郭益深, 陈力. 漂浮基姿态受控空间机械臂关节运动的自适应神经网络控制[J]. 空间科学学报, 2008, 28(2): 173-179 doi: 10.11728/cjss2008.02.173

    GUO Yishen, CHEN Li. Adaptive neural network control of free-floating space manipulator with an attitude controlled base[J]. Chinese Journal of Space Science, 2008, 28(2): 173-179 doi: 10.11728/cjss2008.02.173
    [17] RAHIMZADEH H, SADEGHI M, GHASEMI-VARNAMKHASTI M, et al. On the feasibility of metal oxide gas sensor based electronic nose software modification to characterize rice ageing during storage[J]. Journal of Food Engineering, 2019, 245: 1-10 doi: 10.1016/j.jfoodeng.2018.10.001
    [18] LI Y, LIU M Y, ZHANG X J, et al. Global approximation based adaptive RBF neural network control for supercavitating vehicles[J]. Journal of Systems Engineering and Electronics, 2018, 29(4): 797-804 doi: 10.21629/JSEE.2018.04.14
    [19] MUNRO K, MILLER T H, MARTINS C P B, et al. Artificial neural network modelling of pharmaceutical residue retention times in wastewater extracts using gradient liquid chromatography-high resolution mass spectrometry data[J]. Journal of Chromatography A, 2015, 1396: 34-44 doi: 10.1016/j.chroma.2015.03.063
    [20] 周扬扬, 韦韧, 王超, 等. 静电悬浮位置控制系统设计与实现[J]. 空间科学学报, 2011, 31(5): 675-681 doi: 10.11728/cjss2011.05.675

    ZHOU Yangyang, WEI Ren, WANG Chao, et al. Design and implementation of electrostatic levitation position control system[J]. Chinese Journal of Space Science, 2011, 31(5): 675-681 doi: 10.11728/cjss2011.05.675
    [21] LIU N J, CAI Z H, ZHAO J, et al. Predictor-based model reference adaptive roll and yaw control of a quad-tiltrotor UAV[J]. Chinese Journal of Aeronautics, 2020, 33(1): 282-295 doi: 10.1016/j.cja.2019.08.001
    [22] 李明, 封航, 张延顺. 基于UMAC的RBF神经网络PID控制[J]. 北京航空航天大学学报, 2018, 44(10): 2063-2070 doi: 10.13700/j.bh.1001-5965.2017.0777

    LI Ming, FENG Hang, ZHANG Yanshun. RBF neural network tuning PID control based on UMAC[J]. Journal of Beijing University of Aeronautics and Astronautics, 2018, 44(10): 2063-2070 doi: 10.13700/j.bh.1001-5965.2017.0777
    [23] ATTARAN S M, YUSOF R, SELAMAT H. A novel optimization algorithm based on epsilon constraint-RBF neural network for tuning PID controller in decoupled HVAC system[J]. Applied Thermal Engineering, 2016, 99: 613-624 doi: 10.1016/j.applthermaleng.2016.01.025
    [24] SHI K J, LI B, WANG F M, et al. Research on the RBF-PID control method for the motor actuator used in a UHV GIS disconnector[J]. The Journal of Engineering, 2019, 2019(16): 2013-2017 doi: 10.1049/joe.2018.8728
  • 加载中
图(12) / 表(3)
计量
  • 文章访问数:  157
  • HTML全文浏览量:  35
  • PDF下载量:  20
  • 被引次数: 0
出版历程
  • 收稿日期:  2021-09-27
  • 录用日期:  2022-03-15
  • 修回日期:  2022-05-11
  • 网络出版日期:  2022-09-19

目录

    /

    返回文章
    返回