基于PID神经网络控制的阻力干扰效应补偿  

The Hydrodynamic Resistance Compensation by Using PIDNN Control

作　　者：朱战霞[1] 李玥[1] 马岩[1] 魏奇章 

机构地区：[1]西北工业大学航天学院,陕西西安710072

出　　处：《控制工程期刊（中英文版）》2014年第1期16-22,共7页Scientific Journal of Control Engineering

摘　　要：摘要：航天器中性浮力实验中的水动阻力是导致实验逼真度降低的主要因素之一，如何消除阻力对实验的干扰影响是进行地面实验天地一致性研究必须解决的关键问题。为此，提出了基于预测值的干扰效应多推进器补偿方案，通过推力的精确分配和控制，实现阻力效应的抵消。在此基础上，设计了PID神经网络控制算法，利用神经网络的自学习功能训练权系数，实现PID参数的在线调整，该算法可适应实验体系统的非线性和参数变化。最后，基于MATLAB／simulink软件搭建了仿真系统，并对航天器在中性浮力环境中的六自由度运动实验的阻力干扰效应进行补偿仿真，结果表明基于本文所提补偿方案和算法，可实现阻力干扰效应的有效补偿，为提高中性浮力实验逼真度提供了思路和有效解决方案。During neutral buoyancy experiment, the hydrodynamic resistance is a main factor leading to test results poor similarity to real spacecraft＇s. In order to achieve more credible experiment result, the effect caused by hydrodynamic resistance should be compensated. In this paper, we proposed a compensate scheme in which the hydrodynamic resistance is counterbalanced using multi-propulsion based on thrust refined distribution and control. Then we designed the Proportional-Integral-Derivative Neural Network （PIDNN） control law, in which weight coefficients are trained by using self-study of Neural Network and then to adjust and determine the parameters of PID for nonlinear system and parameter change. We also built the simulation system based on MATLAB/simulink and accomplished the hydrodynamic resistance compensation simulation for the six degrees freedom motion of the Neutral buoyancy＇s experimental body （NBEB）. The results show that the performance of the controller based on the PIDNN is good at fast response and small overshoot. And it also proved that the method proposed in this paper is effective for the hydrodynamic resistance compensation

关 键 词：中性浮力 阻力干扰 PID神经网络控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]