基于HSMM和EEMD的熔融沉积成型3D打印过程故障诊断研究

发布时间：2018-04-11 05:32

本文选题：熔融沉积成型3D打印 + 故障诊断　；参考：《浙江大学》2017年硕士论文

【摘要】：论文课题来源于国家基金项目"熔融沉积成型3D打印的声发射监控理论与方法研究"(编号:51675481)。本文针对熔融沉积成型3D打印过程,以声发射技术作为信号检测手段,结合 EEMD(Ensemble Empirical Mode Decomposition)信号处理方法和 HSMM(Hidden Semi Markov Model)故障状态识别方法,围绕产品缺陷检测以及同步带齿故障状态诊断展开研究,旨在提出一种切实可行的熔融沉积成型3D打印过程故障诊断方法。论文的主要工作内容包括:(1)分析了熔融丝料的粘结机理,在对熔融沉积成型分层破坏力学模型进行分析的基础上,结合丝料的粘结机理研究了产品缺陷状态的故障演化过程;对同步带的受力情况进行分析,得到其应力模型,结合熔融沉积成型3D打印机的工作特点探究了同步带齿健康状态的故障演化过程;对熔融沉积成型3D打印过程中的声发射源进行分析,以实际采集到的声发射信号说明熔融沉积成型3D打印过程中声发射技术的可用性。(2)阐述了 EMD的基本原理和算法,分析了 EMD方法的缺陷,在对EMD的改进算法EEMD进行研究的基础上,提出了基于EEMD的熔融沉积成型3D打印过程故障声发射信号处理方法,基于试验采集到的故障声发射信号验证了 EEMD方法的抗混叠性,使用相关系数方法对EEMD分解得到的有效IMF分量进行筛选,分析有效IMF分量的能量特征,证明了 EEMD方法分解熔融沉积成型3D打印过程故障声发射信号的有效性。(3)研究了 HSMM的基本原理和算法,为了提高HSMM的鲁棒性和稳定性,对多组观测序列训练HSMM的问题进行研究。结合HSMM和EEMD提出基于HSMM和EEMD的熔融沉积成型3D打印过程故障诊断方法,并使用Java语言对HSMM程序进行了实现。(4)设计多组对比试验,对论文中提出的故障诊断方法进行研究,试验结果表明,基于HSMM和EEMD的熔融沉积成型3D打印过程故障诊断方法具有很高的诊断准确度,非常适合于熔融沉积成型3D打印过程故障诊断。
[Abstract]:The thesis is based on the National Foundation project, "study on the Theory and method of Acoustic Emission Monitoring for 3D Printing of Melt deposition Molding" (No.: 51675481).In this paper, for the 3D printing process of melt deposition molding, acoustic emission technology is used as signal detection method, combined with EEMD(Ensemble Empirical Mode signal processing method and HSMM(Hidden Semi Markov Model fault state identification method.Based on the research of product defect detection and fault state diagnosis of synchronous belt teeth, a feasible fault diagnosis method for 3D printing process of molten deposition molding is proposed.The main work of this paper includes: (1) analyzing the bonding mechanism of the fused filament. Based on the analysis of the delamination failure model of the melt deposition forming, the fault evolution process of the defect state of the product is studied in combination with the bonding mechanism of the filament.The stress model of the synchronous belt is analyzed and the fault evolution process of the healthy state of the synchronous belt tooth is discussed according to the working characteristics of the melt deposition forming 3D printer.In this paper, the acoustic emission sources in the process of 3D printing of melt deposition are analyzed, and the availability of acoustic emission technology in the process of 3D printing of melt deposition is illustrated with the collected acoustic emission signals. (2) the basic principle and algorithm of EMD are expounded.The defects of EMD method are analyzed. Based on the study of the improved algorithm EEMD of EMD, a method of processing fault acoustic emission signal in 3D printing process of melt deposition molding based on EEMD is proposed.The anti-aliasing property of EEMD method is verified based on the fault acoustic emission signals collected from the experiment. The effective IMF components obtained from EEMD decomposition are screened by correlation coefficient method, and the energy characteristics of the effective IMF components are analyzed.It is proved that the EEMD method is effective to decompose the fault acoustic emission signals in the 3D printing process of melt deposition molding. The basic principle and algorithm of HSMM are studied. In order to improve the robustness and stability of HSMM,The problem of training HSMM with multiple observation sequences was studied.Combined with HSMM and EEMD, a fault diagnosis method based on HSMM and EEMD for 3D printing process of molten deposition molding is put forward, and a multi-group comparative test is designed by using Java language to implement HSMM program. The fault diagnosis method proposed in this paper is studied.The experimental results show that the fault diagnosis method based on HSMM and EEMD for 3D printing process of molten deposition molding has high diagnostic accuracy and is very suitable for fault diagnosis in 3D printing process of melt deposition molding.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP334.8

【参考文献】