Abstract

In this work, the strain reconstruction with interpolation methods was investigated based on measurements of distributed fibre optical sensors. The strain measurements under different sensor densities (5,10,15,⋯, 30mm) were adopted in this investigation. The measurements under 5mm distance were adopted as true strain distribution (high-fidelity data). The other measurements were interpolated to 5mm by barycentric Lagrange and cubic spline methods. These interpolated results were adopted as the low-fidelity data and to compare with the true strain distribution. The results show that the accuracy of strain reconstruction decreases with the increase of sensors distance. Specifically, the local patterns of strain distribution are lost after the reconstruction when the sensor distance of measurement data is much larger (≥6 times) than the sensor distance of reconstruction data. However, the main trends of the reconstructed strain distribution are consistent with the true strain distribution. In addition, the barycentric Lagrange interpolation shows better stability than the cubic spline interpolation, especially for measurements with large sensor distance.