Estimating the Axial Load of In-Service Continuously Welded Rail Under the Influences of Rail Wear and Temperature

Abstract

Linear guided wave based methods have been proposed to measure the axial load of continuously welded rail (CWR) in service. The underlying principle is that the propagation velocities of excited guided waves are sensitive to the axial load. However, the in-service CWR inevitably faces changes in rail wear and temperature, which also affects the propagating guided waves and results in severe degradation of existing methods. In this paper, we proposed the IGA-IWLS algorithm to estimate the axial load of in-service CWR using multiple guided wave modes. This novel load estimation method takes rail profile and phase velocities of a small set of wave modes as input, then uses an improved genetic algorithm to roughly search the candidate solutions of axial load and Young's modulus, and finally employs weighted least squares algorithm to iteratively converge to the estimated value of axial load. The paper presents the estimation theory in detail, including selection of the optimal set of guided wave modes and the IGA-IWLS algorithm. Numerical experiments show that the proposed method is able to estimate the axial load of CWR with an accuracy less than 2 MPa and is robust to measurement error and model error.