An intelligent 3-D reconstruction method for arbitrarily oriented subsurface cracks based on the multi-frequency RACFM technique

Non-ferromagnetic metallic materials, such as aluminium alloys and stainless steels, are extensively utilised in aerospace, electrical power, and other industries. Due to harsh operating environments and complex stress states, arbitrarily oriented subsurface cracks with strong concealment but significant harm are prone to occur, posing a substantial threat to structural safety. Therefore, the detection and sizing of arbitrarily oriented subsurface cracks are crucial for maintaining the safe operation of equipment. In this paper, a novel intelligent 3-D reconstruction method for arbitrarily oriented subsurface cracks is proposed. A simulation model for a multi-frequency Rotating Alternating Current Field Measurement (RACFM) is established, and the relationship between subsurface cracks of different angles, lengths and depths and multi-frequency magnetic field images is analysed. A database is constructed through both simulation and experimentation. Guided by the physical information contained in the magnetic field images, an intelligent 3-D reconstruction network for arbitrarily oriented subsurface cracks is established. Experimental results indicate that the method can achieve 3-D reconstruction and sizing of arbitrarily oriented subsurface cracks, with an average length error of 0.79 mm and an average depth error of 0.35 mm.