The relationships between the fatigue crack growth rate $\left(da/dN\right)$ and stress intensity factor range $\left(\Delta K\right)$ are not always linear even in the Paris region. The stress ratio effects on fatigue crack growth rate are diverse in different materials. However, most existing fatigue crack growth models cannot handle these nonlinearities appropriately. The machine learning method provides a flexible approach to the modeling of fatigue crack growth because of its excellent nonlinear approximation and multivariable learning ability. In this paper, a fatigue crack growth calculation method is proposed based on three different machine learning algorithms (MLAs): extreme learning machine (ELM), radial basis function network (RBFN) and genetic algorithms optimized back propagation network (GABP). The MLA based method is validated using testing data of different materials. The three MLAs are compared with each other as well as the classical two-parameter model ( $K^{*}$ approach). The results show that the predictions of MLAs are superior to those of $K^{*}$ approach in accuracy and effectiveness, and the ELM based algorithms show overall the best agreement with the experimental data out of the three MLAs, for its global optimization and extrapolation ability.

Thermal barrier coating (TBC) is a critical material in the aerospace domain to increase the lifetime of gas turbine components subjected to thermal load. The properties of TBC are strongly related to the growth of thermally grown oxide (TGO) whose main constituent is Al₂O₃. However, the oxidation of Cr and Co can affect the growth of TGO, which is not studied sufficiently. In this paper, high-temperature exposure at 1000 °C was performed to investigate the effect of Cr and Co oxides on TGO growth. The morphology and composition analysis of the interface between the ceramic top coat and the bond coat (TC/BC) were investigated by using scanning electron microscopy (SEM) and the energy dispersion spectrum (EDS). The thermodynamics and kinetics of oxidation were analyzed. The results indicated that the oxidation kinetics basically followed the sub-parabolic law with exposure time. Additionally, the major factor affecting the formation of oxides was the diffusion rate at the initial stage of exposure, then oxides depended on thermodynamics, and the oxidation was influenced by both of them in the last stage. The major elements to be oxidized were different at different stages. Moreover, the replacement reaction of Cr₂O₃ and the phase conversion of Al₂O₃ resulted in thickness variations of the TGO and Al-depleted zone during high-temperature exposure.

A Fiber Bragg Grating (FBG) interrogation system with a self-adaption threshold peak detection algorithm is proposed and experimentally demonstrated in this study. This system is composed of a field programmable gate array (FPGA) and advanced RISC machine (ARM) platform, tunable Fabry–Perot (F–P) filter and optical switch. To improve system resolution, the F–P filter was employed. As this filter is non-linear, this causes the shifting of central wavelengths with the deviation compensated by the parts of the circuit. Time-division multiplexing (TDM) of FBG sensors is achieved by an optical switch, with the system able to realize the combination of 256 FBG sensors. The wavelength scanning speed of 800 Hz can be achieved by a FPGA+ARM platform. In addition, a peak detection algorithm based on a self-adaption threshold is designed and the peak recognition rate is 100%. Experiments with different temperatures were conducted to demonstrate the effectiveness of the system. Four FBG sensors were examined in the thermal chamber without stress. When the temperature changed from 0 °C to 100 °C, the degree of linearity between central wavelengths and temperature was about 0.999 with the temperature sensitivity being 10 pm/°C. The static interrogation precision was able to reach 0.5 pm. Through the comparison of different peak detection algorithms and interrogation approaches, the system was verified to have an optimum comprehensive performance in terms of precision, capacity and speed.

Brittle fracture of large forging equipment usually leads to catastrophic consequences. To avoid this kind of accident, the brittle fracture behaviors of a large die holder were studied by simulating the practical application. The die holder is used on the large die forging press, and it is made of 55NiCrMoV7 hot-work tool steel. Detailed investigations including mechanical properties analysis, metallographic observation, fractography, transmission electron microscope (TEM) analysis and selected area electron diffraction (SAED) were conducted. The results reveal that the material generated a large quantity of large size polyhedral M₂₃C₆ (M: Fe and Cr mainly) and elongated M₃C (M: Fe mainly) carbides along the martensitic lath boundaries when the die holder was recurrently tempered and water-cooled at 250 °C during the service. The large size carbides lead to the material embrittlement and impact toughness degradation, and further resulted in the brittle fracture of the die holder. Therefore, the operation specification must be emphasized to avoid the die holder being cooled by using water, which is aimed at accelerating the cooling.

Titanium-alloy weldments have been extensively used in engineering structures, and accurate estimation of the fatigue life is beneficial for avoiding catastrophic failures in titanium-alloy structures. The weld pore is a common defect in these weldments that significantly influences the fatigue process. In this study, the weld-pore size and depth and its effect on the fatigue life of Ti-6Al-2Zr-1Mo-1V (TA15) alloy weldments are investigated by adopting fatigue tests and fracture observations. The results show that fatigue crack initiations occurred at the weld pores for all specimens. The diameter and depth of all weld pores on fracture surfaces were then measured, and an indicator, P, was proposed, defined as a combination of pore diameter and depth. It was found that all the fatigue cracks initiated from the pore have the smallest P indicator, which suggests that P can be used to judge the location of crack initiation in an individual sample. Moreover, a model was developed based on P to estimate the fatigue life of weldments, considering the effects of weld-pore size and depth. Finally, analogous fatigue tests were carried out for model verification, and results show that the proposed model has a higher accuracy compared with several typical models. The findings of this study can be helpful in estimating the fatigue life and fatigue design of titanium-alloy weldments.