流动流体黏度的声发射测量

Measurement of Flowing Fluid Viscosity Based on AE Signals

基于声发射（AE）测量技术,结合小波分析和R/S分析,获得了代表流体撞击壁面运动的特征信号频段（d1~d12频段）。根据声波特征信号频段能量随流体黏度的规律性变化,以流体撞击壁面产生信号特征区域的声能量值为特征参数,建立了流体黏度的预测模型。以搅拌釜中水-蔗糖体系为例,在搅拌转速1.67 r/s条件下,由预测模型得到的检测结果与真实值的相关系数（r）为0.970,平均相对偏差（ARD）为11.863%,交叉验证误差均方根（RMSECV）为1.228,表明预测模型对于流体黏度的检测具有较高准确性。由此获得了一种快速、准确和非侵入式的流体黏度检测技术,有利于工业生产流程的优化和控制。

A characteristic frequency band of acoustic emission（AE） signals,which represented the impact of the fluid on the wall（d1-d12）,was obtained by using wavelet transform and Hurst analysis.Then it was observed that the energy of AE signals at different frequencies changed regularly with the increase of fluid viscosity.A new method for building prediction model of fluid viscosity was presented based on this regular behavior.The prediction model of fluid viscosity was verified by the aqueous sucrose solutions system in a stirred vessel,showing that the correlation coefficient（r）,average relative deviation（ARD） and root mean square error of cross validation（RMSECV） between actual viscosity and predictive values were 0.970,11.863% and 1.228,respectively.It is concluded that multi-scale analysis of AE is feasible for predicting fluid viscosity and has the merit of sensitive,non-intrusive and accurate,being favor to the optimization and control of production process.