基于Elman的荧光油膜厚度近效测量

为解决荧光油膜灰度与厚度现标定方法（微小高度装置法）存在采集工况复杂、标定周期较长等缺陷，提出了一种新的近效测量方法，该方法只需极少量标定数据，便可达到与现标定方法相近的效果。近效方法利用了Elman动态神经网络对小样本数据进行量程扩展并引入一维插值算法使扩展后的数据项平滑，对解算出的三维荧光油膜厚度数据采用二维插值算法进行二次平滑以求得完整的厚度分布图。经模拟试验结果显示，通过该近效方法进行厚度测量最终可以清晰、准确、定量地显示荧光油膜厚度分布，与目前广泛使用的微小高度装置法相比效果相近，在荧光油膜汇集处（较厚区域）误差最大不超过±2.5μm，在平滑适中及较薄区域误差不超过±2μm，达到荧光油膜厚度工程测量标准，为飞行器全局摩阻测量提供了一种新标定思路，具有一定的实际工程应用意义。

Fluorescent oil film thickness approximate efficiency measurement based on Elman

To address the shortcomings of the current calibration method for fluorescent oil film grey scale and thickness (tiny height device method) with complex acquisition conditions and a long calibration period, an approximate measurement method was proposed, by which only a very small amount of calibration data were required to achieve similar results to the current calibration method. The Elman dynamic neural network was used to extend the range of the small sample data and a one-dimensional interpolation algorithm was introduced to smooth the extended data items. The solved 3D fluorescent oil film thickness data were smoothed twice using a two-dimensional interpolation algorithm to obtain a complete thickness distribution. The simulation results showed that the thickness measurement by this near-efficient method can finally display the fluorescent oil film thickness distribution clearly, accurately and quantitatively, and the effect was similar to that of the currently widely used tiny height device method, with the maximum error not exceeding ±2.5 μm at the pool of fluorescent oil film (thicker area), and not exceeding ±2 μm in the smooth and moderate and thinner area; hence the engineering measurement standard of fluorescent oil film thickness could be met. It provides a calibration idea for the global friction measurement of aircraft, and has certain practical engineering application significance.