This paper presents a modified fatigue life model of the Basquin equation using the stress parameter of the magnetic flux leakage signal. Most pipeline steels experience cyclic loading during service and the influence of the load history makes assessing fatigue behaviour more difficult. The magnetic flux leakage signal’s response to a uniaxial cyclic test of API X65 steel was measured with eight levels of ultimate tensile stress loads. The influence of dH(y)/dx on fatigue failure was the main concern in this study, the aim being to represent localised stress parameters in the modified Basquin equation. Both fatigue lives, experimental and predicted from the modified Basquin equation, were validated through reliability analysis, producing a 60% value when approaching 1.8 × 10⁵ cycles. The fatigue data from the experiment produced a higher mean-cycle-to-failure value than the prediction data, with slightly different values of 3.37 × 10⁵ and 3.28 × 10⁵. Additionally, the modified Basquin equation’s predicted and the experimental fatigue lives were found to have a high R² correlation value of 0.9022. The Pearson correlation also showed a good relationship between the fatigue lives, with an r value of 0.9801. Finally, the modified Basquin equation based on dH(y)/dx signals provided an accurate and alternative method for durability assessment.

The behaviour of the steel bar in concrete under cyclic loading has been evaluated using magnetic flux leakage associated with acoustic emission monitoring technique. Visual observation was used to observe the deformation of the beam under cyclic loading. The sensors of metal magnetic memory were scanned in the middle of the beam at a distance of 320 mm at the bottom part. Twenty-two cyclic ranges were performed for cyclic loading of 100 or 200 cycles for each range, with a frequency of 1 Hz and a sinusoidal wave mode. The magnetic flux leakage signal, acoustic emission characteristics and crack width were measured and analysed to evaluate the behaviour of the steel bar in the concrete beam. The magnetic flux leakage signal and acoustic emission energy results were well matched with the occurrence of cracks at the centre of the beam. It was found that the relationship between the magnetic leakage flux signal and crack opening showed a strong correlation with R² of 0.969. A high acoustic emission energy of 1300 nVs is observed at the centre of the beam. Based on the results, the behaviour of the steel in the concrete beam can be determined by the integrity assessment of a structure.

This paper proposes a novel semi-supervised framework, time-frequency Contrastive Learning (CL), to address the challenge of accurate rolling bearing fault diagnosis under industrial small-sample conditions. Raw vibration signals are transformed into discriminative time-frequency images using short-time Fourier transform (STFT). A CL network with a ResNet18 model is pre-trained on a lot of unlabeled samples to learn generalized feature, and the ResNet18 model is fine-tuned using small labeled samples for fault classification. Experimental validation on bearing fault datasets demonstrates that the proposed STFT-CL method achieves above 99% diagnosis accuracy with only 50 labeled samples per fault type, outperforming conventional semi-supervised methods by 6-12%. The proposed method provides a potential solution to the "small sample dilemma" in industrial applications through the synergistic effect of physically driven signal processing and self supervised representation learning.

This work focuses on the investigation to evaluate a material parameter, namely a contraction value ψ, at failure in estimating the fatigue life of ferromagnetic steel based on magnetic metal memory method. It begins with specimen preparation made from mild carbon steel and it was machined as specified in the ASTM E467-01 standard. These cyclic tests were performed using the 25-kN servo-hydraulic machine with the applied axial loading levels were at 60% UTS and 70% UTS. The magnetic flux leakage was then measured at every hour of the load applied, and the nature of change of magnetic leakage intensity H_p over the surface of the specimen. A further analysis on the collected signals or pattern was carried out based on the H_p distribution and its gradient at dH_p/dx, and both parameters were later being used to estimate total loading time before specimen failure. A further analysis on the collected signals or pattern was carried out based on the H_p distribution and its gradient at dH_p/dx, and both parameters were later being used to estimate total loading time before specimen failure. At every hour of loading cycles, the fatigue life of the specimen is estimated using the specific calculation of this MMM approach.