摘要
针对ROV变量液压推进器伺服控制系统的控制模型,提出了采用改进遗传算法进行控制模型参数辨识。为解决遗传算法易早熟,难以找到精确解等问题,采用一种基于均匀设计的种群初始化方法和一种改进变异方式的深度捕食策略,有效提高了ROV变量液压推进器伺服控制模型辨识算法的全局收敛性和搜索效率。同时,利用液压试验平台搭建液压回路模拟液压推进器伺服控制系统,采集试验数据辨识得到了ROV变量液压推进器的精确控制模型,仿真和实验结果证明了辨识算法的可行性和模型的正确性。
An improved genetic algorithm was designed to identify the parameters of variable displacement of an ROV （ remote operated vehicle） hydraulic thruster. To avoid premature convergence and determine an accurate solu-tion ,a new method to create an initial population based on a uniform design and new mutation method combined with a predatory search strategy were proposed. This approach improved the ability of the algorithm to perform local searches and solved the problem of global convergence and low search efficiency effectively. An identification exper-iment was performed on a hydraulic test platform to simulate the hydraulic thruster servo control system. Based on the results of the experiment,an accurate control model of the variable displacement of an ROV hydraulic thruster was identified by the improved genetic algorithm. Comparison of the experimental and simulation results confirmed the validity of the improved algorithm and accuracy of the final control model.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2017年第2期168-174,共7页
Journal of Harbin Engineering University
基金
国家高技术研究发展计划(2012AA092101)
关键词
ROV推进器模型
变量马达
遗传算法
捕食策略
种群初始化
ROV（remote operated vehicle） thruster model
variable displacement motor
genetic algorithm
preda-tory search strategy
initialization of population
作者简介
丁汉卿（1992-）,男,硕士研究生
王旭阳（1977-）,男,助理研究员.通信作者：王旭阳,E-mail：wangxuyang@sjtu.edu.cn.
