Inversion Analysis of GEO Plasma Environmental Parameters Based on BP Neural Network
-
摘要: 空间等离子体环境诱发的表面充电效应会对航天器运行产生干扰，严重时将导致太阳电池等部件失效。通过神经网络反演方法，以GEO环境中介质表面充电电位曲线作为输入，在双峰麦克斯韦分布假设下，可以逆向得到高能峰的等离子体参数。分析了GEO等离子体环境参数对表面充电电位曲线的影响，表明高能峰在充电过程中起决定性作用；其次通过MATLAB搭建BP神经网络，采用 COMSOL计算得到多组充电曲线进行网络训练和反演计算，得到等离子体密度反演的平均相对误差为0.42%，温度反演的平均相对误差为0.03%，整体误差在0.1%~5.6%。结果表明，采用神经网络对等离子体环境进行反演具有可行性，该方法可以作为空间等离子体环境探测结果的对比参考和航天器非探测点表面电位计算的输入条件。Abstract: The surface charging effect induced by the space plasma environment may interfere with the operation of spacecraft, which can lead to the failure of solar cells and other components. It is shown that the high-energy plasma environmental parameters can also be obtained by taking the electrical potential curve of the dielectric surface in the GEO environment as input under the assumption of double Maxwell distribution. Firstly, the influence of GEO plasma environment parameters on the surface charging potential curve is analyzed, indicating that the high energy peak plays a decisive role in the charging process. Then, BP neural network is established by MATLAB, and the network training data is obtained according to multiple groups of charging curves calculated by COMSOL. Finally, the average relative error of plasma density and temperature inversion is 0.42% and 0.03%, and respectively, and the overall relative error is within 0.1%~5.6%. Results showed that it was feasible to use a neural network to invert the plasma environment, and the method can be used as a reference for comparison of the detection results of the space plasma environment. The inversion results can be used as the input conditions for surface potential calculation of non-detection points of spacecraft.
-
表 1 GEO最恶劣等离子体环境参数
Table 1. GEO worst-case environmental parameters
等离子体分布 ne,1/cm–3 Te,1/eV ne,2/cm–3 Te,2/eV ni,1/cm–3 Ti,1/eV ni,2/cm–3 Ti,2/eV SCATHA-Mullen1 0.20 400 2.30 24800 1.60 300 1.30 28200 SCATHA-Mullen2 0.90 600 1.60 25600 1.10 400 1.70 24700 ECSS-E-ST-10–04 C
(SCATHA 1979)0.20 400 1.20 27500 0.60 200 1.30 28000 NASA Worst Case 1.12 12000 － － 0.236 29500 － － ATS-6 2.36 29500 － － 0.236 29500 － － MIL-STD-1809 2.36 3100 0.625 25100 0.60 200 1.20 28000 Galaxy 15 4.58 55600 － － 0.10 75000 － － 
下载: 导出CSV
表 2 部分等离子体环境参数
Table 2. Partial plasma environmental parameters
等离子体分布 ne,1/cm–3 Te,1/eV ne,2/cm–3 Te,2/eV ni,1/cm–3 Ti,1/eV ni,2/cm–3 Ti,2/eV 环境1 0.20 400 2.3 24800 1.60 300 1.30 28200 环境2 0.25 450 2.2 25000 1.50 310 1.32 28000 环境3 0.30 500 2.1 25200 1.40 320 1.34 27800 环境4 0.35 550 2.0 25400 1.30 330 1.36 27600 环境5 0.40 600 1.9 25600 1.20 340 1.38 27400
下载: 导出CSV
表 3 部分电位数据
Table 3. Partial potential data
时间 /s 0.5 1 1.5 2 2.5 3 3.5 ··· 50 电位1 –7.24 –14.48 –21.72 –28.96 –36.21 –43.44 –50.66 ··· –702.24 电位2 –6.62 –13.23 –19.85 –26.47 –33.08 –39.69 –46.29 ··· –642.96 电位3 –5.96 –11.92 –17.89 –23.84 –29.79 –35.75 –41.69 ··· –580.86 电位4 –5.28 –10.56 –15.85 –21.12 –26.40 –31.68 –36.95 ··· –515.85 电位5 –4.58 –9.15 –13.72 –18.30 –22.87 –27.44 –32.01 ··· –447.92
下载: 导出CSV
表 4 BP神经网络反演结果
Table 4. Inversion results by BP nueral network
等离子体分布 ne,2/cm–3 Te,2/eV ni,2/cm–3 Ti,2/eV 反演环境1 2.30 24800 1.30 28200 反演结果1 2.22 25025 1.315 28090 相对误差/(%) 3.48 –0.91 –1.15 0.39 反演环境2 1.50 26400 1.46 26600 反演结果2 1.43 26068 1.3579 27103 相对误差/(%) 4.47 1.26 6.99 –1.89 反演环境3 1.00 27500 1.70 24700 反演结果3 0.997 27459 1.5844 25378 相对误差/(%) 0.30 0.15 6.80 –2.74 反演环境4 0.99 27500 1.70 24700 反演结果4 0.967 27272 1.6663 24972 相对误差/(%) 2.30 0.83 1.98 –1.10 反演环境5 0.73 25600 1.46 26800 反演结果5 0.729 25629 1.4603 26738 相对误差/(%) 0.14 –0.11 –0.02 0.23
下载: 导出CSV
表 5 GA-BP神经网络反演结果
Table 5. Inversion results by GA-BP neural network
等离子体分布 ne,2/cm–3 Te,2/eV ni,2/cm–3 Ti,2/eV 反演环境1 2.30 24800 1.30 28200 BP值 2.274 24910 1.327 28088 GA-BP值 2.259 25041 1.319 28071 反演环境2 1.50 26400 1.46 26600 BP值 1.488 26394 1.351 27587 GA-BP值 1.499 26392 1.433 26908 反演环境3 1.00 27500 1.70 24700 BP值 1.018 27467 1.631 25101 GA-BP值 0.998 27511 1.688 24755 反演环境4 0.99 27500 1.70 24700 BP值 1.016 27451 1.658 24920 GA-BP值 0.981 27404 1.705 24792 反演环境5 0.73 25600 1.46 26800 BP值 0.746 25691 1.457 26774 GA-BP值 0.741 25650 1.459 26753
下载: 导出CSV
表 6 1979年SCATHA卫星环境参数反演结果
Table 6. Inversion results of environmental parameters of SCATHA satellite in 1979
等离子体分布 ne,1/cm–3 Te,1/eV ne,2/cm–3 Te,2/eV ni,1/cm–3 Ti,1/eV ni,2/cm–3 Ti,2/eV 初始环境 0.60 100 0.60 26000 0.62 300 1.30 28000 反演结果 － － 1.03 24885 － － 1.34 27452 峰值环境 0.20 400 1.20 27500 0.60 200 1.30 28000 反演结果 － － 1.78 24335 － － 1.62 25127
下载: 导出CSV
表 7 所取电位数据
Table 7. Selected potential data
取点 1 2 3 4 5 6 7 8 9 10 初始电位/V –4.07 –4.41 –4.89 –5.21 –5.57 –6.01 –6.33 –6.70 –6.95 –7.21 峰值电位/V –295.90 –299.31 –306.83 –314.27 –322.63 –327.01 –330.44 –335.73 –338.10 –341.52
下载: 导出CSV
-
[1] 刘尚合, 胡小锋, 原青云, 等. 航天器充放电效应与防护研究进展[J]. 高电压技术, 2019, 45(7): 2108-2118 doi: 10.13336/j.1003-6520.hve.20190628007LIU Shanghe, HU Xiaofeng, YUAN Qingyun, et al. Research progress in charging-discharging effects and protection of spacecraft[J]. High Voltage Engineering, 2019, 45(7): 2108-2118 doi: 10.13336/j.1003-6520.hve.20190628007 [2] LAI S T, MARTINEZ-SANCHEZ M, CAHOY K, et al. Does spacecraft potential depend on the ambient electron density[J]. IEEE Transactions on Plasma Science, 2017, 45(10): 2875-2884 doi: 10.1109/TPS.2017.2751002 [3] NAKAMURA M, NAKAMURA S, KAWACHI R, et al. Assessment of worst GEO plasma environmental models for spacecraft surface charging by SPIS[J]. Transactions of the Japan Society for Aeronautical and Space Sciences, Aerospace Technology Japan, 2018, 16(6): 556-560 doi: 10.2322/tastj.16.556 [4] 方庆园, 周江波, 季启政, 等. 基于等离子体能量密度的航天器表面电位估计[J]. 微波学报, 2020, 36(5): 36-40 doi: 10.14183/j.cnki.1005-6122.202005007FANG Qingyuan, ZHOU Jiangbo, JI Qizheng, et al. Evaluation of spacecraft surface potential based on plasma’s energy density[J]. Journal of Microwaves, 2020, 36(5): 36-40 doi: 10.14183/j.cnki.1005-6122.202005007 [5] 方庆园, 王通, 季启政, 等. 基于平均二次电子发射系数的航天器表面起电特征分析[J]. 强激光与粒子束, 2021, 33(2): 023007 doi: 10.11884/HPLPB202133.200149FANG Qingyuan, WANG Tong, JI Qizheng, et al. Analysis of spacecraft charging onset using secondary electron yield[J]. High Power Laser and Particle Beams, 2021, 33(2): 023007 doi: 10.11884/HPLPB202133.200149 [6] 崔阳, 杨勇, 张晓峰, 等. SMILE卫星表面电位特性分析[J]. 真空科学与技术学报, 2021, 41(1): 22-28 doi: 10.13922/j.cnki.cjvst.202005017CUI Yang, YANG Yong, ZHANG Xiaofeng, et al. Analysis of surface potential characteristics for SMILE satellite[J]. Chinese Journal of Vacuum Science and Technology, 2021, 41(1): 22-28 doi: 10.13922/j.cnki.cjvst.202005017 [7] 关燚炳, 王世金, 梁金宝, 等. 航天器交汇对接电位控制研究与航天器电位监测[J]. 科技导报, 2011, 29(29): 27-31 doi: 10.3981/j.issn.1000-7857.2011.29.002GUAN Yibing, WANG Shijin, LIANG Jinbao, et al. Spacecraft potential control and potential monitor during spacecraft docking[J]. Science and Technology Review, 2011, 29(29): 27-31 doi: 10.3981/j.issn.1000-7857.2011.29.002 [8] 田立成, 石红, 李娟, 等. 航天器表面充电仿真计算和电位主动控制技术[J]. 航天器环境工程, 2012, 29(2): 144-149 doi: 10.3969/j.issn.1673-1379.2012.02.006TIAN Licheng, SHI Hong, LI Juan, et al. Simulation of spacecraft surface charging and active control of spacecraft surface potential[J]. Spacecraft Environment Engineering, 2012, 29(2): 144-149 doi: 10.3969/j.issn.1673-1379.2012.02.006 [9] 左应红, 王建国, 罗旭东, 等. 双麦克斯韦分布等离子体对航天器表面的充电效应[J]. 强激光与粒子束, 2015, 27(11): 114003 doi: 10.11884/HPLPB201527.114003ZUO Yinghong, WANG Jianguo, LUO Xudong, et al. Spacecraft surface charging effect of plasma in bi-Maxwellian distribution[J]. High Power Laser and Particle Beams, 2015, 27(11): 114003 doi: 10.11884/HPLPB201527.114003 [10] 武明志. 空间等离子体诱发太阳能电池表面充放电效应的仿真分析[D]. 南京: 南京航空航天大学, 2018WU Mingzhi. Analysis and Simulation of Surface Charging and Discharging on Solar Array[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2018 [11] ISO. ISO 19923: 2017 Space environment (natural and artificial)—plasma environments for generation of worst case electrical potential differences for spacecraft[S]. ISO, London, 2017 [12] 马晓敏, 王新. 基于遗传算法的BP神经网络改进[J]. 云南大学学报(自然科学版), 2013, 35(S2): 34-38MA Xiaomin, WANG Xin. An improved BP neural network algorithm based on genetic algorithm[J]. Journal of Yunnan University, 2013, 35(S2): 34-38 [13] 刘建林. BP神经网络模型在电磁环境预测中的应用[J]. 电力科技与环保, 2017, 33(4): 5-9 doi: 10.3969/j.issn.1674-8069.2017.04.002LIU Jianlin. Application of BP neural network prediction model to the electromagnetic fields[J]. Electric Power Environmental Protection, 2017, 33(4): 5-9 doi: 10.3969/j.issn.1674-8069.2017.04.002 [14] 刘天, 姚梦雷, 黄继贵, 等. BP神经网络在传染病时间序列预测中的应用及其MATLAB实现[J]. 预防医学情报杂志, 2019, 35(8): 812-816,821LIU Tian, YAO Menglei, HUANG Jigui, et al. Application of back propagation neural network in prediction of infectious disease time series and its MATLAB implementation[J]. Journal of Preventive Medicine Information, 2019, 35(8): 812-816,821 [15] 任谢楠. 基于遗传算法的BP神经网络的优化研究及MATLAB仿真[D]. 天津: 天津师范大学, 2014REN Xienan. Study on Optimization of BP Neural Network Based on Genetic Algorithm and MATLAB Simulation[D]. Tianjin: Tianjin Normal University, 2014 [16] GUSSENHOVEN M S, MULLEN E G. Geosynchronous environment for severe spacecraft charging[J]. Journal of Spacecraft and Rockets, 1983, 20(1): 26-34 doi: 10.2514/3.28353 -
下载:
点击查看大图
图(7) / 表(7)
计量
- 文章访问数: 30
- HTML全文浏览量: 8
- PDF下载量: 2
- 被引次数: 0
