基于多传感器信号特征融合的孔系钻削质量一致性评估

发布时间：2018-04-17 08:46

本文选题：孔系钻削 + 突变点检测　；参考：《湘潭大学》2015年硕士论文

【摘要】：钻削质量一致性是衡量与保证孔系类零部件加工质量和工作性能的关键因素。本文以实现孔系钻削质量一致性检测为目的,在孔系钻削实验平台搭建、监测信号突变点检测与提取、钻削过程信号特征提取与融合、模式识别等方面进行相关研究。(1)孔系钻削实验平台搭建和信号采集。根据钻削加工的特性,搭建孔系钻削实验平台。采用霍尔电流传感器监测主轴电流负载的变化情况、声发射传感器监测钻削工件材料断裂与变形和三向加速度振动传感器监测工件在钻削过程中的振动情况,并通过数据采集仪采集各孔钻削过程多传感器监测信号。(2)孔系钻削过程监测信号预处理与突变点提取。首先采用谱减法对多传感器监测信号进行去噪;然后,根据钻头刃尖与工件接触和钻头刃尖切出工件时刻钻削力以及主轴功率信号在突变点处的变化特点,对主轴功率信号进行RMS分析和信号变化率计算,提取出各孔钻削过程中钻入钻出时刻突变点;最后根据信号突变点将与钻削质量相关的钻削阶段信号分割出来,为后续的孔系钻削质量分析提供数据基础。(3)监测信号特征提取与分析。针对钻削阶段监测信号的非线性、非平稳、非高斯特性,对各孔监测信号进行希尔伯特黄变换和高阶谱分析,提取各孔监测信号的边际谱和双谱特征,分析监测信号特征变化与钻孔质量之间的有机联系。(4)基于主成分分析的监测信号特征融合与聚类。采用主成分分析对钻削阶段监测信号边际谱与双谱特征构建的特征矩阵进行特征降维处理。根据各孔监测信号第一主成分和第二主成分对钻孔数据进行k均值聚类,获得孔系钻削质量分布状态。将聚类结果与人工检测结果进行比对,评判聚类分析结果的有效性。最后,通过改变钻削工艺参数进行钻削试验验证本文所述方法的可行性。计算分析和试验验证结果表明,采用本文所述研究方法可有效实现孔系钻削质量进行一致性评估,快速分析和识别质量异常钻孔。该方法有效克服了孔系类零部件钻削质量检测中人工检测存在的漏检以及在线检测实现难度较大等问题,为实现孔系钻削质量一致性检测与控制提供了理论依据。
[Abstract]:The consistency of drilling quality is the key factor to evaluate and guarantee the machining quality and working performance of hole system parts.In order to realize the consistency inspection of drilling quality of drilling system, the paper builds the experimental platform of drilling in the hole system, detects and extracts the abrupt point of monitoring signal, and extracts and fuses the feature of the signal during drilling process.Related research on pattern recognition. 1) drilling experiment platform and signal acquisition.According to the characteristics of drilling, a drilling experiment platform is built.Hall current sensor is used to monitor the change of spindle current load, acoustic emission sensor is used to monitor the fracture and deformation of drilling workpiece material and the vibration sensor of three-direction acceleration is used to monitor the vibration of workpiece during drilling.The multi-sensor monitoring signal of drilling process of each hole is collected by the data acquisition instrument, and the monitoring signal of drilling process of the hole system is preprocessed and the abrupt point is extracted.Firstly, the multi-sensor monitoring signal is de-noised by spectral subtraction method, then, according to the contact between the bit tip and the workpiece and the cutting force of the workpiece and the variation of the spindle power signal at the abrupt point, the cutting force of the workpiece and the power signal of the spindle are obtained.The power signal of spindle is analyzed by RMS and the rate of change of signal is calculated, and the abrupt point of drilling time is extracted in the process of drilling. Finally, the signal of drilling stage related to drilling quality is segmented according to the abrupt point of signal.This paper provides a data basis for the subsequent analysis of drilling quality of hole system. The feature of monitoring signal is extracted and analyzed.Aiming at the nonlinear, non-stationary and non- characteristics of the monitoring signals in drilling stage, Hilbert Huang transform and high-order spectrum analysis are carried out to extract the marginal and bispectral characteristics of the monitoring signals.Analysis of the organic relationship between the variation of monitoring signal features and the quality of borehole. (4) the feature fusion and clustering of monitoring signals based on principal component analysis (PCA).The feature matrix constructed by the marginal spectrum and bispectral feature of monitoring signal in drilling stage is processed by principal component analysis (PCA).According to the first principal component and the second principal component of the hole monitoring signal, the drilling data were clustered by k-means clustering, and the drilling quality distribution state of the hole system was obtained.The validity of clustering analysis was evaluated by comparing the clustering results with the results of manual detection.Finally, the feasibility of this method is verified by changing drilling parameters.The results of calculation analysis and experimental verification show that the method presented in this paper can effectively realize the consistency evaluation of drilling quality of hole system and quickly analyze and identify the abnormal drilling quality.This method effectively overcomes the problems of manual inspection of drilling quality of hole system parts and the difficulty of on-line testing. It provides a theoretical basis for the consistency detection and control of drilling quality of hole system.
【学位授予单位】：湘潭大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TG52

【相似文献】