基于协同进化多目标遗传算法的复式液压摆动缸结构设计

Structural design of compound hydraulic swing cylinder based on co-evolution multi-objective genetic algorithm

针对舰船操舵过程中舵叶与水动力之间存在强力位耦合的问题,本文设计出一种新型复式液压摆动缸转舵机构,该机构由内外两层转子组成,内层转子控制舵叶转角,外层转子进行力矩解耦。为减小复式液压摆动缸的质量和体积,本文提出一种基于协同进化的多目标遗传算法,算法包含两类种群,一类用于进化决策解,并评价罚因子,另一类用于调整罚因子,并指导第一类种群进化,两类种群通过信息交换,来指导自身的搜索,从而使整个系统协同进化,以此求解复式液压摆动缸各几何参数的数值,直至获得满意解。最后,将该算法得到的参数数据,在三维软件中完成建模,并利用有限元软件进行应力应变分析。结果表明,相较于传统遗传算法,协同进化多目标遗传算法得到的复式液压摆动缸的质量减少了约25.5%,其得到的几何参数数据更为合理。

To address the problem of strong potential coupling between the rudder blade and hydrodynamic forces during the steering of a ship,a new compound hydraulic oscillating cylinder rudder mechanism was de⁃signed,which consisted of two rotor layers inside and outside,with the inner rotor controlling the rudder blade rotation angle and the outer rotor performing torque decoupling.To reduce the mass and volume of the compound hydraulic oscillating cylinder,in this paper,a multi-objective genetic algorithm based on co-evo⁃lution was proposed.The algorithm consisted of two kinds of population.One was used to evolve the decision solution and evaluate the penalty factor,and the other was used to adjust the penalty factor and guide the evo⁃lution of the first kind of population.The two kinds of population guided their own search through information exchange,so that the whole system could co-evolve.By this method,the geometric parameters of the com⁃pound hydraulic swing cylinder could be solved until a satisfactory solution was obtained.Finally,the parame⁃ter data obtained by this algorithm were modeled in 3D software and the stress-strain analysis was performed using finite element software.The results show that the mass of the compound hydraulic oscillating cylinder obtained by the co-evolutionary multi-objective genetic algorithm is reduced by about 25.5%compared with the traditional genetic algorithm,and the geometric parameter data obtained by it are more reasonable.