曲面基体绝缘涂镀层涡流测厚高精度标定方法

为了提高金属零件表面绝缘涂镀层厚度测量精度，以某型电涡流位移流传感器为例，开展3种不同曲率的曲面试件标定试验；对比分析9次多项式、7项傅里叶级数、多峰高斯函数和径向基函数神经网络这4种标定方法的误差平方和、均方根、决定系数以及单点运算时间；综合评价4种标定方法在测量精度、计算速率方面的优劣，为涡流位移传感器在不同应用场合的标定方法选择提供参考. 针对燃料贮箱外表面聚氨酯泡沫层厚度测量要求，选用基于径向基函数神经网络的曲面基体涡流测厚法进行验证试验. 结果表明，通过优化标定方法，测量误差可以控制在?0.15~ 0.15 mm.

High-precision calibration methods of thickness measurement for insulation coation on curved surface based on eddy current

A certain type of eddy current displacement sensor was taken as an example, and three calibration testing experiments were conducted, in order to increase the accuracy of thickness measurement for insulating coating on curved surface. The 9th-degree polynomial, 7th-order Fourier series, Gaussian multi-peak function, and radial basis function neural network were used for fitting. The squared sum of errors, root mean square, coefficient of determination and single-point operation time were compared and analyzed in terms of measurement accuracy and operation rate, for reference on how to choose the calibration method in different applications. Aiming at the requirements of thickness measurement of polyurethane foam on the outer surface of the fuel tank, chose the method based on radial basis function neural network for verification test. Results show that the measurement error can be controlled between ?0.15 mm to 0.15 mm by optimizing the calibration method.