Climate change is one of the most difficult challenges facing the world today. To prevent climate change, profound changes in the production, distribution, and consumption of energy are required. The increased emission of carbon dioxide due to various human activities is directly responsible for the increase in the Earth’s average temperature. Recently, there has been an immense discussion on environmental protection. To reduce carbon dioxide (CO
2) emissions and avoid the related environmental problems, scientists and engineers are always looking for methods with which the emission of exhaust gases can be mitigated. Firstly, the use of renewable energy has become increasingly important in meeting future energy demands and limiting the exposure of CO
2, such as solar power plants, wind mills, and geothermal. The key issue associated with greener plants is the amount of energy being extracted from renewable sources of energy, i.e., the energy efficiency of renewable power plants. For example, the maximum theoretical efficiency of a wind turbine is 40% [
1] and that of solar power plants is about 20% [
2], while geothermal energy is 12% [
3]. There is also an immense amount of ongoing research focused on finding solutions for increasing energy output from renewable energy power plants. Secondly, emission abatement through the improvement of the efficiency of thermal power plants is comparably cost-effective and therefore has great effects on fuel consumption and environmental impact. The development of novel materials for steam turbines and their related components have always been a major issue in the power sector. Steam turbines are exposed to ultra-high temperatures of above 700 °C. During the service period, material is subjected to severe conditions, static and dynamic loads, temperature sequences, weathering, and chemical influences. Nickel alloy 617 has proven to be the most promising due to its high metallurgical stability, oxidation resistance, and ease of fabrication [
4]. However, in using this alloy as a structural material for various components, new welding technology is required. Therefore, 12 Cr steel is also a suitable candidate due to its excellent corrosion resistance properties and reasonable cost. The key technology for the application of 12 Cr steel and Alloy 617 to low pressure and temperature stages is to develop and design dissimilar material welding (DMW) technology. Previously, it was suggested that explosion welding (EXW) could be used as a joining method for similar or dissimilar materials [
5]. It consists of a solid-state welding process with controlled explosive detonation on the surface of a metal. It was proven that the reflection and superposition of stress waves caused by explosive loading led to redistribution and remarkable reduction of residual welding stresses. ASTM A516 low carbon steel with A5086 aluminum alloy [
6,
7] and Al–Cu, Ti–Cu, and Cu–Ti [
8] explosion bonding have been extensively studied.
In operation, performance deterioration or failure of the critical components can result in huge economic loss or catastrophic consequences, so determining the life time of the key components has now become more important when considering reliability. This prediction is based on how materials behave under stress. As their life time tends to be several years, in this case, an accelerated life test (ALT) becomes a feasible way to accelerate the failure process and shorten the test time. Several ALT models are being used today [
9]. In accelerated failure time (AFT) models, it is assumed that failure time will follow the same type of distribution under different levels of stress, and time to failure would be shorter at higher levels of stress [
10]. The proportional hazards (PH) model assumes that the applied stresses act multiplicatively on the hazard rate [
11]. An extended hazard regression (EHR) model was proposed, which encompasses both the PH and AFT models [
12]. Other ALT models also being used are the extended linear hazard regression (ELHR) model [
13], proportional mean residual life (PMRL) model [
14], and proportional odds (PO) model [
15].
Fatigue life prediction can also be done using an artificial neural network, considering the tensile properties, volume fraction, and statistical parameters as the input, and receiving the number of fatigue life cycles as the output. The neural network is supposed to evaluate the degradation on components under mechanical stress in real time to predict when they will eventually fail.
In this work, optimum welding conditions were used to perform dissimilar material welding of Alloy 617 and 12 Cr. The fatigue and corrosion fatigue strength of dissimilar materials welded were found and compared. Fatigue life prediction was done using accelerated life tests and an artificial neural network method. Two training algorithms, Bayesian regularization (BR) and Levenberg–Marquardt (LM), were employed for training ANN. Finally, the effectiveness of the probabilistic prediction methods was investigated.