**1. Introduction**

As a type of precipitation-reinforced nickel-based high-temperature alloy, the GH4698 alloy has excellent strength, toughness, fatigue resistance, and corrosion resistance at up to 750 ◦C. Thus, this alloy has been widely used to manufacture machine parts working at high temperatures, such as airplane engine compressor disks, guide vanes, and gas turbine disks. However, this material is extremely sensitive to thermal processing parameters, and cracking could easily occur in the billet opening of GH4698 large forgings, owing to the addition of aluminum and titanium [1]. One solution could be to place the billet in a sleeve and forge as a whole, so that the material is under a three-dimensional compressive stress state, but this would increase the cost. A more economical method is to deform under optimized parameters, but thus far, hot processing maps of GH4698 have not been established, hindering the hot working parameters optimization in practical production. Therefore, there is an urgent need for systematic research on the flow behaviors and hot working maps of GH4698.

Various flow stress models have been proposed to describe the flow behaviors of alloys at high temperatures. The phenomenological Johnson-Cook model was successfully used to describe the exponential stress-strain relationships of GH4133B byWang et al. [2]. However, the Johnson-Cook model was inadequate for materials with nonexponential type stress-strain curves, and the Zerilli-Armstrong model was established for an NiTi alloy by Shamsolhodaei et al. [3]. To achieve a higher prediction accuracy, the Arrhenius model was modified to incorporate the influence of strain, and based on the modified Arrhenius model, flow stress models were established for various nickel-based alloys, e.g., the GH4169 alloy by Chen et al. [4], N08028 alloy by Wang et al. [5], and 80A alloy by Gu et al. [6]. The results by Lin et al. [7] and Wang et al. [8] indicated that the accuracy of the flow stress models could be further improved by a neural network, but the applications were limited, owing to the difficulty in finite elemental integration. Moreover, physical-based models were proposed to investigate the underlying mechanisms of the influences of creep, dislocation motion, and grain size on flow stresses by Lin et al. [9], Haan et al. [10], and Zhou et al. [11]. By comparing the above-mentioned models, the Arrhenius model showed an advantage in applicability and accuracy, and thus it has been widely used in the flow stress modeling of nickel-based alloys [4–6].

The processing maps of nickel-based alloys have also been investigated intensively in recent years. Hot working maps of a nickel-based alloy for power plant applications were established by Wu et al., and it was revealed that the different recrystallization mechanisms could be reflected by the hot working maps [12]. The microstructures on the different domains of the processing maps of the IN028 alloy were inspected by Wang et al. [13], and the study showed that the deformation mechanism maps agreed well with the processing maps. The processing maps of the N08028 alloy [14], the 617B alloy [15], and the GH4169 alloy [16] showed that the efficiency peaks of the processing maps were associated with dynamic recrystallization nucleation and dramatic grain growth of the N08028 alloy, whereas incomplete recrystallization, twinning, and adiabatic shear bands occurred in the instability domain. By comparing the different instability criterions from Gegel et al. [17–20], the different shapes of the deformation instability domains of GH79 alloy were compared by Ge et al. [21], and it was noted that the deformation instability domains from Prasad's criterion could effectively predict the deformation instability of GH79. It was shown by the result of Chen et al. [4] that the optimal hot working parameters of GH4169 were located at areas whose dissipative efficiencies were 30–35%. Specifically, for the GH4698 alloy, the flow behaviors were investigated by Zhang et al. [22], and a flow stress model was established. Nevertheless, hot working maps of GH4698 have not been established, forming a barrier for hot working parameter optimization in large forging production. Thus, systematic research is required on the flow behaviors and hot workability of the GH4698 alloy at high temperatures.

Therefore, in this study, the hot deformation behaviors of the GH4698 alloy were studied via hot compressions. An Arrhenius model was established to calculate the flow stresses. Processing maps at various thermal processing conditions were constructed, and an optimal hot working parameter range for GH4698 was recommended. This study provides a reference for hot working parameter optimization of GH4698 large forgings during the forging process.
