珩磨缸套表面粗糙度预测及多目标优化研究

为了对粗珩阶段缸套内孔表面粗糙度Rk粗糙度集中的Rk、Rpk和Rvk进行预测，进而对粗珩加工参数进行优化，以珩磨压力(P)、珩磨头旋转速度(VR)和往复速度(VRe)为决定因素，Rk粗糙度集为目标响应，进行多目标优化. 建立基于广义回归神经网络(Generalized regression neural network, GRNN)与响应曲面法(Response surface methodology, RSM)的粗糙度预测模型，并采用三因素三水平的全因子珩磨试验进行验证，结果表明所建立模型的预测结果与试验结果具有很好的一致性. GRNN预测模型决定系数R2的均值为0.959，RSM多元回归预测模型决定系数R2的均值为0.963，与RSM所建立的多元回归预测模型相比，GRNN预测模型在预测Rk和Rpk时，预测精度更高，预测误差更小，R2分别提高了0.025和0.020，在预测Rvk时RSM多元回归模型更优，R2提高了0.057. 进一步结合响应曲面法分析了3个决定因素对粗糙度的影响显著性并进行了排序，对于Rk：VRe>P>VR；对于Rpk：P>VRe>VR；对于Rvk：P>VRe>VR. 结合多元回归模型与NSGA-Ⅱ (Non-dominated Sorting Genetic Algorithm Ⅱ)优化算法进行多目标优化，获得Pareto最优解的Pareto前沿.

To predict the Rk, Rpk, and Rvk parameters within the Rk roughness family of the inner hole surface roughness of cylinder liners during the rough honing stage, and optimize the rough honing processing parameters, a multi-objective optimization was carried out. Honing pressure (P), honing head rotational speed (VR), and reciprocating speed (VRe) were taken as the determining factors, and the Rk roughness family was set as the target response.

Two roughness prediction models were established based on the Generalized Regression Neural Network (GRNN) and Response Surface Methodology (RSM), respectively, and validated using a three-factor, three-level full factorial honing experiment. The results demonstrated that the predicted outcomes of the established models exhibited excellent consistency with the experimental results. The mean coefficient of determination (R²) of the GRNN prediction model was 0.959, while that of the RSM multiple regression prediction model was 0.963. Compared with the RSM multiple regression prediction model, the GRNN prediction model achieved higher prediction accuracy and smaller prediction errors when predicting Rk and Rpk, with R² values increased by 0.025 and 0.020 respectively. For Rvk prediction, the RSM multiple regression model was superior, with R² increased by 0.057.

Furthermore, combined with Response Surface Methodology, the significance of the influence of the three determining factors on the surface roughness was analyzed and ranked: for Rk: VRe > P > VR; for Rpk: P > VRe > VR; for Rvk: P > VRe > VR. Finally, combining the multiple regression model and the Non-dominated Sorting Genetic Algorithm II (NSGA-II) optimization algorithm, multi-objective optimization was conducted to obtain the Pareto front of Pareto optimal solutions.

搜集汇总

背景与挑战

背景概述

该数据集聚焦于珩磨缸套表面粗糙度的预测与多目标优化研究，通过结合广义回归神经网络和响应曲面法建立预测模型，并利用三因素三水平试验验证模型准确性。研究分析了加工参数对粗糙度的影响，并采用NSGA-Ⅱ算法进行优化，旨在提升粗珩加工效率和质量，适用于摩擦学及机械加工领域。