Abstract:

For improving the surface integrity and processing quality in the precision cylindrical grinding of cemented carbides, and investigating the prediction technique of the surface quality, a prediction model of the surface roughness in the precision cylindrical grinding of YG3 cemented carbide based on the adaptive network-based fuzzy inference system (ANFIS) was proposed. Then, it was improved by the hybrid Taguchi genetic algorithm (HTGA). At last, the grinding parameters and corresponding surface roughness in the experiments was used as the training and testing samples. Trough the comparative analysis of the prediction results in the BP neural network model, traditional ANFIS model and improved ANFIS model, the effectiveness and prediction accuracy of three models were verified. The results show that the prediction accuracy of the improved ANFIS model was better than that of an other two in the distribution of the relative error Er and the size of mean square error EMSRE, and it is an effective prediction method of the surface quality.

Key words: cemented carbide, surface quality prediction, adaptive network-based fuzzy inference system (ANFIS), hybrid Taguchi genetic algorithm (HTGA)

摘要：

为提高硬质合金材料精密外圆磨削的表面完整性和加工质量,研究其表面质量的预测技术,建立了基于自适应模糊推理系统(ANFIS)的YG3硬质合金精密外圆磨削表面粗糙度预测模型,并引入混合田口遗传算法(HTGA)对预测模型进行了改进。采用工艺试验中所用的磨削参数及相应条件下测得的表面粗糙度数据作为训练样本和测试样本,通过对BP神经网络模型、传统ANFIS预测模型及改进ANFIS预测模型的预测结果进行对比分析,对三种模型的有效性和预测精度进行了验证。结果表明,所提出的改进ANFIS预测模型从预测值相对误差Er的分布及均方根相对误差EMSRE的大小来看,均优于其他两种预测模型,预测精度较高,是一种有效的表面质量预测方法。


关键词: 硬质合金, 表面质量预测, 自适应模糊推理系统, 混合田口遗传算法