电液力伺服系统的BP神经网络辨识  

Application of BP neural network based on genetic optimization in electro-hydraulic force servo system

作　　者：王文丽 周挺[1] Wang Weni;Zhou Ting(National Key Laboratory of Strength and Structural Integrity,Aircraft Strength Research Institute of China,Xi'an,Shaanxi,710065,China)

机构地区：[1]中国飞机强度研究所强度与结构完整性全国重点实验室,陕西西安710065

出　　处：《飞行器强度研究》2023年第3期19-24,共6页Aircraft Strength Research

摘　　要：介绍电液力伺服系统的结构及原理,针对系统建模中参数时变和非线性问题,采用BP神经网络进行系统辨识建模;利用遗传算法优化BP神经网络,克服单纯BP算法容易局部收敛、训练速度慢的问题;借助MATLAB神经网络工具箱、全局优化工具箱编写系统辨识算法,建立系统的神经网络辨识模型。分析神经网络模型辩识结果,将其与ARMAX线性参数模型的辨识结果作对比,验证遗传优化BP神经网络系统辨辩识建模的高效性和适用性。The structure and principle of electro-hydraulic force servo system were introduced.BP neural network was used for system identification modeling aiming at parameter time-varying and nonlinear problems in system modeling.By genetic algorithm,BP neural network was optimized to overcome the problem of easy local convergence and slow training speed.With the help of MATLAB neural network toolbox and global optimization toolbox,system identification algorithm was written and neural network system identification model was established.The identification results of the neural network model were analyzed and compared with those of ARMAx linear parameter model to verify the efficiency and applicability of genetic opti-mized BPneural network system identificationmodeling.

关 键 词：电液力伺服系统 BP神经网络 遗传算法 系统辨识建模 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]