Abstract:Rail transportation plays a significant role in contemporary societal production activities. A large number of rails in service are built by using fishplate and bolt connections, leading to frequent impact loading on rail joints, which easily induces crack damage around bolt holes. The early detection and quantitative measurement of bolt hole cracks are crucial for ensuring railway transportation safety and preventing serious accidents. Phased array ultrasonic imaging technology, characterized by high precision, strong environmental adaptability, and convenient implementation, have been widely applied in non-destructive testing of rails. However, reconstructed images on cracks through traditional phased array sector scanning methods is influenced by the tilt angle of cracks, relying on operators′ subjective judgment in industrial practice, making it difficult to achieve quantitative and standardized detection. Therefore, in this paper, a quantitative detection method towards rail bolt hole crack based on ultrasonic plane wave imaging is proposed. Compounded plane wave imaging method is utilized to achieve high-precision reconstruction of the region of interest. Reconstructed images from multiple positions are fused to generate final inspection image, ensuring cracks with different tilt angles are covered. Principle components analysis is deployed to quantitatively evaluate the tilt angle of the cracks. After that, directional maximum intensity projection is utilized to measure the length of the cracks. Experiments on a railway test block shows that the proposed method makes a 100% detection rate for cracks in the range of [-45°,45°]. The maximal measurement error of the proposed method on the location of cracks is 1. 47 mm, which is close to a wavelength. The maximal measurement error of the length is 1. 17 mm which is less than a wavelength. The maximal error of tilt angle of cracks is 5. 01°. The required equipment is simple and can be conveniently mounted on mobile rail inspection vehicles to achieve automatic inspection.