基于残余应力的激光冲击强化参数多目标优化 *

Multi-Objective Optimizationon Laser Shock Peening Parameters Basedon Residual Stress

关键词:  GH4169高温合金  激光冲击强化  多目标优化  RBF模型

Key words:  GH4169 superalloy  Laser shock peening  Multi-objective optimization  RBF model

摘要:

为了获取镍基高温合金 GH4169激光冲击强化( Laser Shock Peening，LSP)过程中最佳工艺参数，利用 ABAQUS/EXPLICIT软件，对搭接率、光斑尺寸、峰值压力和脉冲宽度 4个关键工艺参数对残余应力的影响规律进行了研究。以最大残余拉应力、表面平均残余压应力及残余压应力层深度为优化目标，采用径向基神经网络代理模型( Radial Basis Function，RBF)和遗传算法( GA)(包括快速非支配排序遗传算法( NSGA-II)、相邻培养式遗传算法( NCGA)、自适应变异遗传算法( AMGA))相结合的方法对这 4个工艺参数进行了多目标优化研究，得到了最优的工艺参数。结果表明，采用 AMGA所得的表面平均残余压应力绝对值比 NSGA-II所得高 22.9MPa，比 NCGA所得高 59.4MPa。优化后最大残余拉应力降低了 6.99%，表面平均残余压应力提高了 9.78%，表明了优化方法的有效性。

Abstract:

In order to obtain the optimal processing parameters combination of GH4169 superalloy duringthe laser shock peening(LSP)process，ABAQUS/EXPLICIT package was used to examine the effects of four keyprocessing parameters including overlap rate，spot size，peak pressure and pulse width on the residual stress based on a plate specimen. Then，maximum residual tensile stress，average residual compressive stress on sur.face and the depth of residual compressive stress were taken as optimization objectives. With the help of radial ba.sis function(RBF)surrogate model and genetic algorithm(GA)( including non-dominated sorting GA(NSGA-II)，neighborhood cultivation GA(NCGA)，adaptive multi-population GA(AMGA))，the multi-objective opti. mization of four processing parameters was carried out. Finally，the optimal processing parameters of LSP wereobtained. Numerical results show that the absolute value of average residual compressive stress on surface ob.tained by AMGA is 22.9MPa higher than that of NSGA-II and 59.4MPa higher than that of NCGA. The maximumresidual tensile stress is decreased by 6.99%，and average residual compressive stress on surface increased by 9.78%，demonstrating the effectiveness of the optimization analysis.