基于磁致伸缩的电磁超声钢板机械性能检测方法

钢板广泛应用于航天、汽车、石油管道等国民经济行业,因此需要寻找合适的方法对其机械性能进行检测,包括屈服强度、抗拉强度和延伸率,否则会留下安全隐患；钢板机械性能的检测方式目前大多依赖于破坏性检测,且钢板的微观结构可通过电磁参数反映,对此,提出了基于磁致伸缩的电磁超声机械性能检测方法；对钢板进行实验,得到了相关电磁超声信号；对相关信号进行特征提取,分析了特征参数与机械性能之间的相关性；分别采用了逐步回归和径向基神经网络函数,建立特征参数与机械性能之间的关系；两个模型均具有较高的预测精度,代表了所提出方法的可行性。

Testing Method of Mechanical Properties of Steel Based on Magnetostriction and EMAT

Steel plate is widely used in aerospace, automobile, oil pipeline and other national economy industries, so it is necessary to find an appropriate method to test its mechanical properties, including yield strength, tensile strength and elongation, otherwise it will leave a safety hazard; At present, most detection methods of steel plate mechanical properties rely on destructive detection, and there is a certain relationship between steel plate microstructure and electromagnetic characteristic parameters. Therefore, a mechanical properties detection method based on Electromagnetic Acoustic Transducer(EMAT) and magnetostriction was proposed. The EMAT signals of steel plate were obtained by experiments. Feature extraction was carried out to analyze the correlation between characteristic parameters and mechanical properties. Stepwise regression and radial basis neural network functions were used respectively to establish the relationship between characteristic parameters and mechanical properties. Both models have high prediction accuracy, which represents the feasibility of the proposed method.