基于声发射信号的激光选区熔化增材制造过程飞溅缺陷检测方法

The spatter defect is easy to cause secondary defects such as slag defects, porosity and crack defects even cause the process failure in the Selective Laser Melting (SLM) processing. There are no quantitative characterization methods to solve spatter defect until now, so it is difficult to achieve the in-situ control technology in processing. We has studied and verified the feasibility of acoustic emission technology in SLM processing. This project intends to use acoustic emission signal to characterize spatter defects. Firstly, the release mechanism and its propagation characteristics of elastic wave which generated by recoil pressure of continuously moving molten metal will be studied. Then, a method of noise reduction based on multi-scale wavelet packet decomposition is established to explore the sensitive characteristics of spatter defects. Afterwards, according to the weight of influence in the processing parameters on the acoustic emission characteristic parameters, the relation among the weighted density of input-energy, acoustic emission characteristic parameters and spatter defect will be studied, the function model of spatter defect will be established. Combined with the optimization of intelligent algorithm, the optimized and controlled methods and techniques of the in-situ monitoring and processing parameters about spatter defects in the SLM technology will be built. This project provides a new method for the in-situ monitoring and control for spatter defect in SLM technology, there are important scientific and engineering application significance for improving kinds of performance such as the forming efficiency, stability of processing and the quality of parts.

激光选区熔化（SLM）增材制造过程中飞溅缺陷容易导致加工零件产生夹渣、孔洞和裂纹等次生缺陷甚至加工失败，目前尚无飞溅缺陷定量表征方法，进而较难实现制造过程的实时调整控制。项目组前期研究并验证了声发射信号作为SLM增材制造过程检测的可行性。本项目拟利用声发射信号实现飞溅缺陷的检测，首先研究熔融金属气化产生的连续、移动反冲压力在加工基体中的弹性波释放机制及传播特性；建立基于多尺度小波包分解的降噪方法，获取反映飞溅行为的敏感特征量；根据加工工艺参数对声发射特征参数的影响权重，研究加权能量输入密度、声发射特征参数和飞溅缺陷之间的映射关系，构建飞溅缺陷的函数模型，并结合优化智能算法，建立SLM增材制造过程飞溅缺陷的在线检测和加工工艺参数的优化与控制的方法和技术。该项目为SLM增材制造飞溅缺陷在线检测和工艺控制提供了新方法，对提高SLM增材制造零件成形效率和质量等方面具有重要的科学意义和工程应用价值。

针对激光选区熔化（SLM）增材制造过程中飞溅现象具有的连续性、随机性和微弱性，以及飞溅现象声发射信号呈现出多声发射源共存和特征信号微弱等特点，深入研究了基于声发射信号的激光选区熔化增材制造飞溅缺陷检测方法。①在飞溅缺陷形成的动力学行为及声发射响应机理方面。通过数值模拟还原了离散金属粉末熔融过程复杂演变过程。分析了影响熔道成形与导致熔滴飞溅的潜在因素，分析了熔体内部流动与飞溅行为产生的关联性，建立了飞溅行为形成的熔池动力学模型。②在复杂噪声背景下飞溅现象声发射信号特征信息获取方面。搭建SLM增材制造过程飞溅现象声发射检测实验平台。并针对背景噪声和不同加工条件下的飞溅现象声发射信号进行幅频时域特性分析。通过频谱分析得到三个固有背景噪声声发射信号频段与代表飞溅现象声发射信号特征频段。③飞溅缺陷声发射信号多源信号分离方面。提出一种基于遗传变异粒子群-变分模态分解（GMPSO-VMD）算法的声发射多源信号分离方法，通过对声发射多源信号进行GMPSO-VMD分离可验证该方法的有效性，同时对比EMD算法与GMPSO-VMD，准确性和速度方面，后者明显由于前者。利用GMPSO-VMD 算法对不同加工条件下的飞溅现象声发射多源信号进行分离，实现了代表飞溅现象声发射信号特征频段的准确快速分离。④ 基于声发射信号特征的飞溅缺陷表征及加工工艺优化方面。研究了不同工艺参数下熔池表面形貌及尺寸的变化规律。根据不同加工条件下SLM成形零件质量分析，当激光功率固定为250W时，飞溅现象声发射信号频域幅值范围处在550.77-584.69内SLM增材制造加工零件质量较好。当扫描速度固定为200mm/s时，飞溅现象声发射信号频域幅值范围处在507.17-550.77内SLM增材制造加工零件质量较好，初步构建了利用特征频段幅值范围选择最优加工工艺参数的方法。. 所取得的成果为机械系统故障诊断与健康预测、SLM增材制造缺陷检测与加工工艺优化提供了理论基础，也拓展了机械动力学与增材制造交叉领域的纵深研究，具有较强的理论意义和应用价值。