Deflection is of great significance for evaluating the safety performance of long-span bridges, while it is difficult or expensive to be directly measured in actual projects despite numerous displacement sensors have been developed. This paper proposes a displacement reconstruction framework to obtain the dynamic deflection of beam structures using strain data. The framework defines the judgment criteria of curvature symbols and standard side first, and then stochastic subspace identification algorithm is used to calculate the strain mode shapes involved in the vibration, which solves the unclear calculation process and the need for considerable measurement points of the traditional mode superposition method. Furthermore, the displacement mode shapes and corresponding modal coordinates are obtained, and then the dynamic deflections are derived. Numerical simulations of cantilever and simply supported beams illustrate the effectiveness of the proposed framework, and the results show that the error is only 0.69% with only four strain measurement points. The two corresponding model experiments have also been carried out, and the results demonstrate that the deflection estimation error is only 0.14% with five measurement points, which further proves that the developed displacement reconstruction framework can accurately reconstruct the dynamic deflection of the beam structures.

挠度对于评价大跨度桥梁的安全性能具有重要意义，而在实际工程中，尽管已经开发了许多位移传感器，但直接测量却困难或昂贵。本文提出了一种位移重建框架，以利用应变数据获得梁结构的动态挠度。该框架首先定义了曲率符号和标准边的判断标准，然后采用随机子空间识别算法计算振动中涉及的应变模态，解决了传统模态计算过程不清晰和需要大量测量点的问题。叠加法。进而得到位移模态振型和相应的模态坐标，进而推导出动态挠度。悬臂梁和简支梁的数值模拟说明了所提出框架的有效性，结果表明，只有四个应变测量点，误差仅为0.69%。还进行了两个相应的模型实验，结果表明，五个测量点的挠度估计误差仅为0.14%，进一步证明所开发的位移重建框架能够准确地重建梁结构的动态挠度。