Probabilistic Fatigue Life Prediction of Dissimilar Material Weld Using Accelerated Life Method and Neural Network Approach

Abstract

Welding alloy 617 with other metals and alloys has been receiving significant attention in the last few years. It is considered to be the benchmark for the development of economical hybrid structures to be used in different engineering applications. The differences in the physical and metallurgical properties of dissimilar materials to be welded usually result in weaker structures. Fatigue failure is one of the most common failure modes of dissimilar material welded structures. In this study, fatigue life prediction of dissimilar material weld was evaluated by the accelerated life method and artificial neural network approach (ANN). The accelerated life testing approach was evaluated for different distributions. Weibull distribution was the most appropriate distribution that fits the fatigue data very well. Acceleration of fatigue life test data was attained with 95% reliability for Weibull distribution. The probability plot verified that accelerating variables at each level were appropriate. Experimental test data and predicted fatigue life were in good agreement with each other. Two training algorithms, Bayesian regularization (BR) and Levenberg–Marquardt (LM), were employed for training ANN. The Bayesian regularization training algorithm exhibited a better performance than the Levenberg–Marquardt algorithm. The results confirmed that the assessment methods are effective for lifetime prediction of dissimilar material welded joints.

Soft computing methods for fatigue life estimation: A review of the current state and future trends

“…Ahmad et al 82 proposed an accelerated life testing method and ANN with Bayesian and Levenberg-Marquardt (LM) training algorithms that confirmed the effective lifetime prediction of different materials welded joints.…”

“…Al Assaf and El Kadi, 12 Bezazi et al, 13 Rohman et al, 18 Kong et al, 19 Han, 20 Sohn and Bae, 24 Vassilopoulos et al, 28 Cai et al, 30 Kumar et al, 33 Zhaohua, 35 Vadood et al, 43 Liu et al, 46 Barbosa et al, 48 Lotfi and Beiss, 49 Razzaq et al, 50 Vassilopoulos et al, 55 Mathew et al, 57 Mohanty et al, 58 Mohanty et al, 60 Karakas, 63 Karakas and Tomasella, 66 Jin et al, 67 Abdalla and Hawileh, 75 Gao et al, 76 El Kadi and Al-Assaf, 77 Tian, 79 Song et al, 81 Ahmad et al, 82 Zhang et al 93…”

“…Rohman et al, 18 Kong et al, 19 Pleune and Chopra, 23 Vassilopoulos et al, 28 Mohanty et al, 40 Mishra et al, 44 Mohanty, 45 Liu et al, 46 Lotfi and Beiss, 49 Mohanty et al, 58 Xiao et al, 59 Moghaddam et al, 69 Durodola et al, 70 Yan et al, 72 Ahmad et al, 82 Zhang et al, 93 Deveci and Artem, 95 Vassilopoulos et al, 100 Azarhoosh et al, 102 Sai et al 131…”

Fatigue modeling using neural networks: A comprehensive review

“…Compared with the nonlinear support vector machine, the linear kernel function and multilayer perceptron with a linear transfer function obtained better results in terms of R 2 , RMSE, MAE, and MAE. Ahmad et al 93 modeled the fatigue life using a CNN. Solano‐Alvarez et al 94 studied the effect of heat treatments and chemical compositions on steels' rolling contact fatigue life.…”

Effect of microstructural heterogeneity on fatigue strength predicted by reinforcement machine learning

“…The neural network was extremely good at discriminating between low‐life and run‐out outcomes. One main outcome was that the Bayesian regularization training method outperformed the Levenberg–Marquardt approach in terms of performance 23 . The influence of corrosion was studied, but there was no correlation between fatigue strength and microstructural parameters.…”