基于小波尺度-能量的钢绞线张拉力识别研究

Research on Tension Identification of Steel Strand Based on Wavelet Scale-Energy

目的研究预应力钢绞线长期服役后剩余应力水平,开展无损检测方法评估剩余应力.方法以工程中常用的钢绞线为研究对象,进行钢绞线逐级加载实验与超声导波实验,对端面传感器与侧面传感器采集到的导波信号进行小波分析与多尺度分析;定义不同拉力状况下能量特征向量Ej之间的差值为LWE,能量Shannon熵特征向量Sj之间的差值为LWES,并以该值作为钢绞线张拉力识别指标.结果随着张拉力的增大,实测导波信号幅值降低,其频谱图幅值也降低,能量分布发生了改变;不同张拉力下,小波能量在尺度上分布有相似的规律,其峰值随张拉力的增大而下降;识别指标LWE和LWES随张拉力变化显著,线性规律强,可靠性系数R2均在0.94以上,传感器布置策略对其影响较小.结论构建的识别指标均能识别出钢绞线张拉力的变化,与小波尺度-能量Shannon熵图相比,小波尺度-能量图更能清晰地观察到能量随张拉力的变化趋势,基于小波能量的识别指标LWE敏感系数K值更大,可靠性越高,识别效果更佳.

The purpose of this paper is to study the nondestructive testing method of estimating the remained stress in steel strands in the long term lifecycle.The steel strand that commonly used in engineering was taken as the research object,and the stepwise loading experiments and ultrasonic guided wave experiments were carried out conjointly.Through adopting the multi-dimensioned wavelet analysis to processing the waveguide signals collected at the end and side sensors.The difference between the energy eigenvectors Ej and energy Shannon entropy eigenvectors Sj under different tension conditions is defined as LWE and LWES,and the value is taken as the tension identification index of steel strand.The results illustrate that the amplitudes in time domain and frequency domain decline with the increase of tensile force,and the energy distribution changes as well;the energy distribution has similar law on scale in different tensile forces,and its peak value decreases with the increasing of tensile force;the identification indexes LWE and LWES change significantly in the variation of tensile force,which have a linear regularity and the reliability coefficient R2 are all above 0.94.The diverse sensors arrangement have a little influence on the results.In conclusion,the constructed identification indexes can well identify the variation of tensions in steel strands,and the scale-wavelet energy graph can more clearly observe the change trend of energy,of which the recognition index LWE has a higher sensitivity coefficient and better recognition effects.