基于可变激励条件下多维磁信号的机械零部件状态检测方法

The performance and life of equipment is closely related to the state of ferromagnetic mechanical components. A fast detecting technology to ensure the production of machinery manufacturing is very important but not well developed. In different exciting conditions, magnetic signals on the ferromagnetic material surface contents different residual stress, defect and other information; On the other hand, the magnetic nondestructive testing method can be well used in the rapid detection and assessment of the stress and micro defect state of ferromagnetic material. A new detecting method under multidimensional magnetic excitation field will be carried out to solve the problems of magnetic memory testing method. The characteristics of this study will be not significantly increase the number of sensors, but also getting abundant information of part state by changing the exciting conditions. Further work will be done to study the behaviors of ferromagnetic samples based on the former research, a magnetic detecting device with variable excitation will be designed and developed, and different magnetic signal detecting experiments under different exciting conditions will be conducted; Then, signals will be analyzed with multi-fusion and decoupling algorithm technology. finally,the relationship between the three-dimensional signals and residual stress and micro defect states of specimens will be established. On the other hand, the optimal magnetizing conditions in material and evaluation of residual stress will be summarized under different direct or changing magnetic field and geomagnetic field. All of these will be helpful to explore a new detecting method in magnetic non-distructive testing field.

机械零部件状态和设备性能、寿命密切相关。研究能够应用于生产和使用现场的零件状态快速检测技术，是机械制造领域十分重要而又未能很好解决的问题。各种激励条件下的铁磁材料表面磁信号，含有残余应力、缺陷等丰富的被测零件状态信息；且磁无损检测方法适于快速检测和评定铁磁性材料应力和微缺陷状态。针对磁记忆等检测方法存在的问题，拟开展改变外激励磁场条件下的多维磁信号检测技术研究，其特点是不显著增加传感器数量，通过改变激励条件，即可获得零件状态的丰富信息。工作内容包括设计研制可变激励的多维磁信号检测装置；进行不同激励条件下磁信号的检测实验；进行多信号融合和解耦算法研究；研究在地磁场和不同强度的交、直流磁化条件下的磁感应输出信号特征，探求建立残余应力、微缺陷等状态量和磁感应多维信号的定量关系。通过系统性试验，确定检测不同材料的最佳磁化条件组合；开展材料学微观实验和铁磁学机理研究，努力探索磁检测领域新方法。

机械零件的状态直接影响设备的性能和寿命。研究一种能够在现场使用的快速的无损检测技术，是机械制造领域的挑战之一。本研究针对磁记忆技术的存在的问题，提出了弱磁场激励下的强化磁记忆技术。从铁磁材料应力-磁导率关系出发，讨论了弱磁激励提高磁记忆技术应力检测灵敏度提高的原理，并通过实验验证了这一设想。通过分析磁记忆信号的特点，深入讨论了磁记忆检测技术在应力定量化检测方面的缺陷。选取不同牌号的铁磁材料，对比研究静动载荷作用下，塑性区的产生，微裂纹的萌生和扩展等不同磁记忆信号和弱磁激励下强化磁记忆的检测能力；开发了多维磁信号检测装置和交、直流变磁激励装置，进行多维磁信号-应力关系的研究；从磁偶极子模型出发，利用主成分分析对裂纹信号进行降维和解耦，利用GA-BP神经网络和PSO-LSSVM算法建立了裂纹信号特征与裂纹几何特征之间的映射关系，实现了裂纹定量化识别；分析了裂纹长度、宽度和深度的识别原理，研究了一种自然裂纹三维轮廓重构技术。