声发射信号的时序分析法在磨削烧伤预报中的应用研究

Study on Application of Time Series of Acoustic Emission Signals in Grinding Burn Prediction

高温合金、表面硬化钢、钛合金等难加工材料，由于具有热稳定性好、热强度高、耐腐蚀、抗磨损性能好等特点，因而被广泛应用于航空航天等工业部门中。但该类材料在磨削加工中极易出现磨削烧伤现象，降低了工件表面完整性，对零件的使用性能产生不利的影响，制约生产率的提高。本文在对高温合金（ＤＺ４）磨削过程中产生的声发射信号特性进行深入分析的基础上，应用时序分析法，建立起关于磨削过程中声发射信号的自回归时序模型，探讨磨削烧伤的在线监测和预报。研究结果表明，声发射信号功率谱结构变化能真正反映磨削烧伤发生与否，自日归时序模型参数及残差方差对工件表面状态变化敏感，均可作为特征参量进行磨削烧伤在线预报。

The difficult--to--machine materials such as superalloys, casehardened steels and titanium alloys have wide applications in the field of aeronautics, astronautics and other industries,for their super physical and mechanical properties such as heat stability, strength, corrosionresistance and wear--resistance. But grinding burn and thermal microcracks usually occur in theprocss of grinding these materials, which will deteriorate the integrity of workpiece surface layer seriously,and have undesirable influences on the practical performance of parts. From an enhancing production efficiency point of view, it is imperative to devise an intelligent sensing system to detect and predict grinding burn on line. This paper proposes an extensive researchwork on characteristics of acoustic emission (AE)signals when grinding superalloy DZ4, and uses a time series analytical technique to characterize the AE signals with an autoregressive model. The result suggests that the power spectra of AE amplitude signal show a strong sensitivityto the grinding burn. The autoregressive model parameters and the residual square error showa strong sensitivity to the changes of ground surface, which can be used as characterized parameters for predicting the grinding burn on line.