No abstract

Fatigue cracks often initiate at the weld toes of welded steel connections. Usually, these cracks cannot be identified by the naked eyes. Existing identification methods like dye-penetration test and alternating current potential drop (ACPD) may be useful for detecting fatigue cracks at the weld toes. To apply these non-destructive evaluation (NDE) techniques, the potential sites have to be accessible during inspection. Therefore, there is a need to explore other detection and monitoring techniques for fatigue cracks especially when their locations are inaccessible or cost of access is uneconomical.Electro-mechanical Impedance (EMI) and Lamb wave techniques are two fast growing techniques in the Structural Health Monitoring (SHM) community. These techniques use piezoelectric ceramics (PZT) for actuation and sensing. Since the monitoring site is only needed to be accessed once for the instrumentation of the transducers, remote monitoring is made possible. The permanent locations of these transducers also translate to having consistent measurement for monitoring.The main focus of this study is to conduct a comparative investigation on the effectiveness and efficiency of the EMI technique and the Lamb wave technique for successful fatigue crack identification and monitoring of welded steel connections using piezoelectric transducers. A laboratory-sized non-load carrying fillet weld specimen is used in this study. The specimen is subjected to cyclic tensile load and data for both techniques are acquired at stipulated intervals. It can be concluded that the EMI technique is sensitive to the crack initiation phase while the Lamb wave technique correlates well with the crack propagation phase.

2.1 BACKGROUND 9 '.

In cylindrical structures such as pipelines and pressure vessels, cracks are most likely to occur along the longitudinal (axial) direction and they can be fatal to the serviceability of the structures. Unfortunately, the conventional ultrasonic crack detection techniques, which usually use longitudinal wave, are not very sensitive to this type of cracks. This paper focuses on the detection and monitoring of axial cracks in cylindrical structures using torsional wave generated by piezoelectric macro-fiber composite (MFC). The first order torsional wave is a kind of non-dispersive pure shear wave which propagates at a fixed wave speed. Torsional wave is utilized in this work because, intuitively, it is more sensitive to axial cracks than the family of longitudinal waves. Numerical simulation has been performed using ANSYS to show the effectiveness of torsional wave in detecting and monitoring axial cracks. The time of flight (TOF) of the waves is used to determine the crack position, while the crack propagation is monitored by measuring the variation in the crack induced disturbances. Experiments have also been conducted to investigate the feasibility of the proposed method. MFC transducers oriented at 45˚ against the axis of the specimen are used to generate and receive torsional waves. The experimental results demonstrated that the crack position can be indentified and its growth can be well monitored with the presented approach using torsional wave.