作者：李建英, 谢寅凯, 谢帅

作者单位：哈尔滨理工大学机械动力工程学院，黑龙江 哈尔滨 150080

关键词：加载系统;动态力加载;快速性;加载误差;最小均方算法;自适应神经网络

摘要：

在电液负载模拟器进行动态力加载时，对其快速性指标有更高要求，此时加载误差比较大的问题就凸显出来，即输出加载力的数值随加载频率的增大而增大、相位也有超前，针对加载系统的这一性能特殊变化规律与实际现象，提出基于最小均方算法的自适应神经网络控制策略。该方法对神经网络的配置权值进行实时在线调整，调整效率和算法的收敛性都有明显提高，可以有效减小控制系统工作时的循环调整时间。当输入信号经过加权运算后作用于加载控制系统时，加载误差明显减小，加载系统的快速性与跟踪精度进一步提高。仿真和实验结果表明，采用文中方法后，加载力输出的幅值增大和相位超前量得以抑制，系统动态力加载时的综合性能有所提高，所提出算法调整规则有效。

Research of control on load system of electro-hydraulic load simulator based on least mean square
LI Jianying, XIE Yinkai, XIE Shuai
School of Mechanical and Power Engineering, Harbin University of Science and Technology, Harbin 150080, China
Abstract: Under the condition of dynamic force loading of electro-hydraulic load simulator, the request of speediness index of the system will be higher, and at this time, the problem of large loading error is highlighted. Namely, the value of the output loading force increases with the increase of the system loading frequency, at the same time, the phase will be ahead of the input signal. In view of this special change rule and practical problem of load system, in this paper, an adaptive neural network control strategy based on least mean square is proposed. The advantage of this way is that the weights of the neural network can be adjusted online in real time, and the designed adjustment rule can improve the adjustment efficiency and the convergence of the algorithm obviously, which can reduce the cycle adjustment time of the control system. When the input signal is applied to the load control system after weight operation, the loading error can be effectively reduced, thus further improving the rapidity and tracking accuracy of the load system. The simulation and experimental results show that, after adopting the method in this paper, the amplitude increase and phase advance of the load force output are effectively suppressed, and the comprehensive performance of the system dynamic force loading is improved, and the effectiveness of the proposed algorithm rule is verified.
Keywords: load system;dynamic force load;rapidness performance;load error;least mean square;adaptive neural network
2021, 47(3):133-138  收稿日期: 2020-10-31;收到修改稿日期: 2020-12-28
基金项目:
作者简介: 李建英(1980-),男,陕西宝鸡市人,教授,博士,研究方向为电液伺服系统、流体传动与控制
参考文献
[1] 甄玉杰, 康金. 舵机电液负载模拟器中CMAC神经网络的应用[J]. 舰船科学技术, 2020, 42(10): 88-90
[2] 刘晓琳, 李卓, 陈立东. 电液负载模拟器神经网络辨识器及控制器设计[J]. 科学技术与工程, 2020, 20(2): 834-840
[3] 林壮彬, 薛家祥. 自组织模糊控制能馈电子负载模拟策略[J]. 中国测试, 2020, 46(3): 115-122
[4] 刘伟良, 杨瑞峰, 郭晨霞, 等. 刚度和惯量对电动负载模拟器的影响分析[J]. 机械科学与技术, 2019, 38(3): 452-456
[5] 刘晓琳, 殷健敏, 袁昆. 飞机舵机电液负载模拟器控制方法研究[J]. 系统仿真学报, 2018, 30(9): 3472-3479
[6] 慕香永, 裴润, 刘志林, 等. 用于船舶舵机的电液负载模拟器之控制系统[J]. 控制理论与应用, 2008, 25(3): 564-568
[7] 王经甫, 叶正茂, 张辉, 等. 基于LMI的H∞控制在电液负载模拟器中的应用[J]. 机床与液压, 2007, 35(5): 169-171
[8] 张彪, 赵克定, 孙丰迎. 电液负载模拟器的神经网络参数辨识[J]. 航空学报, 2009, 30(2): 374-379
[9] YUAN Z H, WU J D, LIU W G. Hybrid control of load simulator for unmanned aerial vehicle based on wavelet networks[C]//Proceedings of the Second International Conference on Machine Learning and Cybernetics. 2003: 715-719.
[10] 李阁强, 赵克定, 袁锐波, 等. μ理论在电液负载模拟器中的应用[J]. 航空学报, 2007, 28(1): 228-233
[11] YOONSU N. QFT force Loop design for the aerodynamic load simulator[J]. IEEE Transaction on Aerospace and Electronic Systems, 2001, 37(4): 1384-1392
[12] KIM J G, BAE W G, CHOI W S. Adaptive step size control of LMS-based interference cancellation for ICS repeater in wibro environment[J]. Journal of Measurement Science and Instrumentation, 2010, 1(1): 101-105
[13] ZHANG W, MU X, YUAN Z H, et al. Adaptive back stepping design for electro-hydraulic loading simulator[C]//Proceedings of the 25th Chinese Control Conference. Harbin, 2006: 253-255.
[14] 姚建均. 电液伺服振动台加速度谐波抑制研究[D]. 哈尔滨: 哈尔滨工业大学, 2007.
[15] 杨宇, 郑淑涛, 韩俊伟. 基于动力学的Stewart平台振动控制策略研究[J]. 农业机械学报, 2010, 41(6): 20-24