Estimation of cyclic behavior of unalloyed, low‐alloy and high‐alloy steels based on relevant monotonic properties using artificial neural networks

Abstract

Within the framework of this research, a new methodology is proposed for estimation of cyclic Ramberg–Osgood parameters i. e. cyclic stress–strain behavior of steels based on their monotonic properties using artificial neural networks. A large number of experimental data for steels were collected from relevant literature and divided into unalloyed, low‐alloy and high‐alloy steels, since previous research confirmed that statistically significant differences exist among cyclic parameters of these subgroups of steels. Detailed statistical analysis is performed by means of forward selection and monotonic properties relevant for estimation of each cyclic parameter of each subgroup of steels are determined. Based on results of performed statistical analyses, artificial neural networks are developed, separately for each parameter and each steel subgroup, using only monotonic properties that proved to be relevant for estimation of particular parameter. Neural networks are evaluated on an independent set of data, in comparison with experimental values and values obtained by existing empirical estimation methods. The new approach is more successful than empirical methods for estimation of most of cyclic stress–strain parameters and behavior of different steel subgroups.

Estimation of Cyclic Stress–Strain Curves of Steels Based on Monotonic Properties Using Artificial Neural Networks

“…First, most of the existing studies developed models on the datasets consisting of the variety of steel without considering differences in individual subgroups of steel. To counter this, in the presented study, the estimations were performed separately for steels grouped according to alloying element content since in [9,21] it was shown that the cyclic stress-strain behavior of these groups is statistically significantly different.…”

“…They reported better accuracy of their approach when compared to the chosen empirical method. More recently, Marohnić [21] developed a new approach to the estimation of cyclic and fatigue parameters of unalloyed, low-alloyed, and high-alloyed steels that includes the identification of monotonic properties relevant for the estimation of each particular cyclic and fatigue parameter and each steel subgroup and incorporating that knowledge to modeling of artificial neural networks for the considered problem. Separate ANNs were developed for each cyclic Ramberg-Osgood parameter as well as strain-life Basquin-Coffin-Manson parameter.…”

“…The lowest values of MAPE and RMSE are underlined in Table 5 for each steel group, indicating models with the best generalization capability. For unalloyed steels, the lowest values of both MAPE and RMSE were observed for the ANN-CSSC-2 approach, closely followed by ANN-param [ 22 ]. For low-alloy steels, the lowest MAPE value is observed for the method by Li et al [ 5 ] (Li 1), closely followed by Lopez 1 [ 4 ] and ANN-CSSC-2.…”

“… Percentage of Δ σ /2 values estimated by selected empirical methods and various approaches to ANNs that deviate up to 5, 10, 15, and 20% from values calculated on experiment-based values of cyclic stress–strain parameters for high-alloy steels [ 2 , 4 , 5 , 22 ]. …”

“… Experimental and points on cyclic stress–strain curves for DIN Ck 45, AISI 4142, and AISI 304 steel datasets using ANN-param [ 22 ], ANN-CSSC-2, and ANN-CSSC-5 approaches. …”