基于伴随方法的舰船推进器优化设计

Propeller optimization design based on the adjoint method

  • 摘要:
      目的  为探索高效的螺旋桨优化设计方法,基于面元法开展伴随优化方法的研究。
      方法  通过桨叶表面法向速度为零条件和等压库塔条件建立伴随方程,得到敏感导数求解式。以DTMB 4381螺旋桨为对象,分别运用伴随方法和传统的求解控制方程方法计算螺旋桨性能与参数之间的敏感导数;基于伴随方法对某螺旋桨进行敏感导数分析,再根据敏感导数分析结果进行几何参数优化,并将结果与ISIGHT优化平台中的粒子群算法(PSO)得到的结果进行对比。
      结果  结果表明,采用伴随方法与传统的求解控制方程方法计算得到的结果具有较好的一致性,但伴随方法的计算效率更高,优化结果也优于PSO算法,且优化所用时间也少。
      结论  研究表明,伴随方法在多参数螺旋桨优化设计中的计算效率优于智能算法。

     

    Abstract:
      Objectives  In order to develop a highly efficient method for propeller design optimization, the adjoint method is studied based on surface panel method.
      Methods  An adjoint equation is established under the conditions of zero normal velocity of blade and equal-pressure Kutta to obtain a formulae for solving sensitive derivative problem. A DTMB 4381 propeller is used as the research objective to calculate the sensitive derivatives of propeller performance to design parameters using the adjoint method and traditional method respectively for solving governing equation. Next, an analysis of senstive derivatives on a ship propeller design is carried out based on the adjoint method.The sensitive derivatives are then obtained and applied to optimize the geometric parameters of the propeller, achieving the optimal solutions which are compared with that by particle swarm optimization (PSO) algorithm of ISIGHT.
      Results  The results indicate that the sensitive derivatives caluculated via the adjoint method are not only in good agreement with that by traditional method, but also offers much higher computation efficiency, generating optimal solutions of the propeller design superior to that by PSO algorithm with less time-consumption.
      Conclusions  The research shows that the computation efficiency of the adjoint method is superior to traditional intelligent algorithms in multi-parameters optimization design of ship propellers.

     

/

返回文章
返回