利用多岛遗传算法的侧式进/出水口体型优化研究

水工建筑物体型优化一般通过物理模型试验或数值模拟对设定体型进行水力特性研究,针对不利的水力指标调整体型再进行研究,直至获得满意的较优体型,这是传统的做法。双向水流条件下的侧式进/出水口体型优化可视为多目标优化问题。本文将加权后的水头损失系数定为目标函数,以孔口断面流速不均匀系数和孔口间流量不均匀系数为约束条件,将多目标优化转化为单目标优化,利用多岛遗传算法进行进/出水口体型优化,可以获得最优体型。此外,参数化建模方法和基于CFD的响应面模型的应用,加快了建模与数值模拟速度,提高了优化效率。以某下水库侧式进/出水口为例,按照本文方法得到的推荐体型与原体型相比,进/出流的总水头损失系数降低3.35%,出流时流速不均匀系数减小14.50%,进/出流的流量不均匀系数均小于20%。该研究方法为水工建筑物体型优化提供了便捷的途径。

Optimization of the shape of lateral intake/outlet using multi-island genetic algorithm

The traditional method for optimizing hydraulic structures shape is generally carried out by model test or numerical simulation to study the hydraulic characteristics,and then adjust the shape of the inlet/outlet for the unfavorable hydraulic characteristics until obtaining satisfactory shape. The shape optimization of lateral inlet/outlet under two-direction flow conditions can be regarded as multi-objective optimization problem. The multi-objective optimization is transformed into a single objective optimization problem by using weighted method, and the multi-island genetic algorithm (MIGA) is used to search the global optimal solution. The weighted head loss coefficient is defined as the objective function, and velocity uneven distribution coefficient and discharge uneven distribution coefficient of the four orifices are taken as constraints. Parametric modeling method and the approximation of response surface model based on CFD calculation, which speed up the modeling and numerical simulation in the scheme selection,improve the optimization efficiency. Compared with the original shape, total head loss coefficient is reduced by 3.35%, velocity uneven distribution coefficient is reduced by 14.50% and discharge uneven distribution coefficient is reduced by 20%. The optimization methods provide a convenient way to optimize the shape of hydraulic structures.