Multi-hole parts made of difficult-to-cut materials like discs, blisks and casings are common and require high surface quality in aero engines. In workshops, a large number of holes in these parts are drilled successively in one process to ensure their positional accuracy. Due to the fast time-varying drill wear, the surface roughness of the holes is unstable and difficult to be satisfied. To this end, this paper presents a varying-parameter drilling (VPD) method to improve machining efficiency and hole surface roughness for multi-hole parts made of Ni-based superalloy. This method uses varying cutting parameters for each hole to adapt to the varying drill wear. The main issue of this method lies in an optimization problem in which the optimal sequence of cutting parameters need to be found, with the objective of the processing time and the constraint of the hole surface roughness. As the cutting parameter sequence has a high dimension and the surface roughness of all the holes must be guaranteed, the challenge of this optimization problem is the strict constraint with a complicated non-linear boundary of the feasible zone. To address the convergence difficulty of the searching algorithm, a soft computing method based on particle swarm optimization (PSO) algorithm with a self-adaptive penalty method (SAPM) is applied. The hole surface roughness is predicted with a radial basis function (RBF) neural network. Different types of drill wear comprising flank wear, crater wear, chisel wear and outer corner wear are considered, and the grey relational analysis (GRA) is employed to select the input drill wear parameters to the network. The PSO algorithm coupled with the SAPM is used to search the global optimal solution of the optimization problem. It is found that the satisfied solutions can be searched in all the three trials with the proposed algorithm, even though the proportion of feasible solutions is severely fluctuant during the searching process. The drilling experiment confirm that, when compared with the fixed-parameter drilling, the proposed VPD and the soft computing method for solving the optimization problem can effectively improve machining efficiency and surface quality for drilling Ni-based superalloy multi-hole parts.

由难以切割的材料制成的多孔零件（如圆盘，叶盘和外壳）很常见，并且在航空发动机中要求具有较高的表面质量。在车间中，这些零件中的大量孔是通过一个过程连续钻出的，以确保其位置精度。由于随时间变化的钻头磨损快，孔的表面粗糙度不稳定并且难以满足。为此，本文提出了一种可变参数钻削（VPD）方法，以提高镍基高温合金制成的多孔零件的加工效率和孔表面粗糙度。该方法对每个孔使用不同的切削参数，以适应变化的钻头磨损。该方法的主要问题在于优化问题，其中需要找到切削参数的最佳顺序，以加工时间和限制孔表面粗糙度为目标。由于切削参数序列的尺寸较大，并且必须确保所有孔的表面粗糙度，因此该优化问题的挑战在于严格限制了可行区域的复杂非线性边界。为了解决搜索算法的收敛难度，提出了一种基于粒子群算法（PSO）和自适应罚分法（SAPM）的软计算方法。用径向基函数（RBF）神经网络预测孔的表面粗糙度。考虑了不同类型的钻头磨损，包括后刀面磨损，月牙洼磨损，凿子磨损和外角磨损，并且采用了灰色关联分析（GRA）选择输入到网络的钻头磨损参数。结合SAPM的PSO算法用于搜索优化问题的全局最优解。结果发现，即使在搜索过程中可行解的比例发生很大的波动，使用该算法也可以在所有三个试验中搜索到满意的解。钻孔实验证实，与定参数钻孔相比，本文提出的VPD和软计算方法能够解决优化问题，可以有效提高Ni基高温合金零件的加工效率和表面质量。即使在搜索过程中可行解的比例严重波动。钻孔实验证实，与定参数钻孔相比，本文提出的VPD和软计算方法能够解决优化问题，可以有效提高Ni基高温合金零件的加工效率和表面质量。即使在搜索过程中可行解的比例严重波动。钻孔实验证实，与定参数钻孔相比，所提出的VPD和求解优化问题的软计算方法可以有效地提高镍基高温合金多孔零件的加工效率和表面质量。