No abstract

No abstract

No abstract

In order to avoid the failure of bolted joints used on ultra-super critical steam turbine inner casing, their mechanical behavior should be strictly assessed, especially under the in-service conditions with steam temperature and pressure fluctuations. In this study, a 3D bolted inner casing is modeled using the commercial finite element software ABAQUS. The Norton-Bailey constitutive model and strain-life–based Manson-Coffin equation are applied to study the mechanical behavior of the bolted inner casing under in-service conditions. The results demonstrate that the inner casing could exert significant influence on the bolts by conduction, structure restriction and thermal expansion difference of the materials. The temperature and von Mises stress of the bolts experience continuous fluctuations during the steady state operation phase due to the constant variation in the steam temperature and pressure, and the alternative stresses give rise to fatigue damage of the material. The steam temperature and pressure fluctuations decline the fatigue lifetime of these high-temperature bolts in long-term operation, which should be taken into account in the design.

Data-driven neural network methods have been widely applied for the prediction of stress–strain behavior, but have proven ill-suited for the extrapolation of time-dependent creep behavior. To overcome this problem, we embedded a physics-based model into feedforward neural networks (FFNNs) to construct a model-guided neural network (MGNN). We proposed a new initialization method for the weights in the model, based on selecting the appropriate physics-based model and activation function, and the resulting MGNN was used for predicting the creep behavior of blade-grooves in a steam turbine rotor under in-service conditions. We compared the performance of the MGNN with baseline methods, namely MGNN0, a FFNN, and a nonlinear autoregressive network with exogenous inputs (network). The results showed that the physics-based model and the neural network in the MGNN complemented each other: the model provided physical relationships to guide the neural network, and the neural network provided stress-fluctuation-tracking for the model. This functionality enabled primary creep behavior to be used as training data for the MGNN, enabling it to predict secondary creep behavior.