**1. Introduction**

Most steel products are manufactured by the hot rolling process because the hot rolling process is one of the most efficient plastic forming processes in metal forming industries. In this process, the initial large and thick material, which is called slab, bloom, and billet, is changed to the desired shape by passing through two counter-rotating rolls in the temperature range of 900–1200 ◦C. Many researchers have been devoted to measuring and predicting the temperature of steels during the hot rolling process to improve the quality of hot-rolled products because the thermal history of steel has a strong influence on the quality of final products [1–6]. Typically, the hot rolling process is categorized into two groups: flat rolling and shape rolling. Plate and strip are made via the flat rolling process, whereas wire, rod, and bar are manufactured via the shape rolling process. The deformation behavior and thermal history of a rod during the shape rolling process are more complex than those of a plate during the flat rolling process because the rod experiences three-dimensional (3D) deformation during the shape rolling process, as shown in Figure 1. In contrast, the plate is deformed under the plain strain condition, i.e., two-dimensional (2D) deformation. This different deformation behavior between the two processes may induce different thermal history between plate and rod, leading to the different microstructures and mechanical properties between the two processes.

During the last three decades, there have been considerable studies on the temperature of the hot rolling process. Most of the researchers have analyzed the rolling temperature of a workpiece on the basis of the numerical simulations such as finite difference method (FDM) and finite element method (FEM), and then the numerical results were validated by experimental measurement with optical pyrometers. Meanwhile, in the 1990s, several researchers investigated the temperature distribution of a plate during flat rolling using thermocouples in pilot rolling mills [7–13]. They investigated the effects of rolling speed, reduction ratio of height, oxide scale on the surface, and lubrication conditions on temperature distribution of the plate, and reported that the cooling rate of the plate along the height direction is different. It should be noted that most of the studies aforementioned were concentrated on the strip and plate during the flat rolling process.

In contrast, a little attention has been paid to the thermal behavior of shaped products during the hot rolling process because the prime concern of industrial shape rolling mills is to produce many products with an appropriate cross-sectional shape as fast as possible. In addition, considerable research on the thermal history of the rod was developed on the basis of the approach of plate rolling. Komori and Kato [14] analyzed the temperature distribution of a rod during the hot rolling process with roll shape using the combination of FEM and FDM. They showed that the roll velocity and friction coefficient greatly influences the temperature distribution at the roll–rod interface. They also reported that temperature distribution of the rod is dependent on the roll shape—square-diamond pass and oval-round pass have a different pattern of temperature distribution. Using the four-pass hot rod rolling experiments with low carbon steel, Said et al. [15] reported that 3D modeling is highly necessary to accurately predict the roll force, torque, and temperature of a rod during the shape rolling process. Yuan et al. [16] simulated the temperature and deformation of the rod during multi-pass rod rolling by 3D FEM, and the predicted temperature was compared with the measured temperature with a pyrometer. Xue and Liu [17] developed the 2D numerical model for calculating the temperature distribution of the rod during the multi-pass hot rolling process. Despite the 2D simulation, the calculation time was very short because the constitutive equation was not solved by iterative scheme. The results were validated by FEM simulation and industrial measurement. Serajzadeh et al. [18] developed a mathematical model based on FEM and phase transformation to predict temperature and microstructure distributions of a steel rod. They showed that the rolling speed affects the temperature of the rod. The increase in the rolling speed leads to a lower temperature drop at the surface area of the rod due to the reduced heat transfer between roll and rod stemming from the shorter contact time. Serajzadeh [19] also developed a mathematical 2D model that considered the various process parameters such as caliber shape and rolling speed. Kown et al. [1] studied the temperature distribution of the rod during hot rolling using 3D FEM analysis and experimental measurement. They insisted that the temperature in the surface area drastically decreases during rolling and recovers right after rolling due to the heat redistribution within the rod.

Meanwhile, steel wire, rod, and bar manufacturing industries have suffered from the several surface defect problems during hot rolling process: surface flaw, decarburization, and abnormal grain growth [20,21]. On the basis of the author's experiences, these surface defects are strongly related to the uneven temperature distribution of the rod during the shape rolling process in comparison with the plate during the flat rolling process [1,22]. Consequently, it is necessary to investigate the detailed temperature distribution of the rod during the hot rod rolling process. In addition, comparative study on the thermal behavior between rod and plate is necessary in order to completely understand the thermal behaviors and the formation of surface defects of the rod during the shape rolling process. On the basis of literature reviews, numerical simulation was mainly used to understand the thermal behaviors of the rod because it is difficult to measure the temperature of the rod during the hot rolling process. Additionally, most of the experimental studies were conducted using the radiation-type pyrometers and thermal cameras to measure the temperature of the rod. However, there are the following five limitations or errors in measuring the temperature of the rod during the hot rolling process:

(i) Pyrometer cannot measure the rod temperature at the rolling time due to the existence of facilities around rolls; thus, only equalized temperature along the radial direction of workpiece is measured, stemming from the rapid rebound of surface temperature. That is, the sharp temperature drop at the surface area by the roll contact cannot be detected.

(ii) During hot rod rolling, measurement errors are generated due to the severe rolling conditions such as the high rolling speed, the vibration of the rod, and the evaporation of coolant for the rolls.

(iii) The measured radiation energy of the rod by pyrometers tends to be lower than that of the actual value, owing to the shape of the round rod, indicating that pyrometers should be carefully validated to measure the shaped materials compared to the flat materials [23].

(iv) Radiation-type pyrometers measure the only surface temperature of a workpiece. Therefore, we cannot understand the temperature distribution of a workpiece along the radial direction.

(v) The radiation energy is highly dependent on the surface conditions of the rod. The rod has various oxide scales according to the process conditions and chemical compositions. In other words, measuring the surface temperature is not easy due to the thick oxide scales as shown in Figure 2.

Accordingly, measuring the accurate temperature of the rod with area is not easy. In addition, many researchers measured temperature of the rod at only one point, such as measuring the temperature of the plate. As shown in Figure 1b, the surface of the rod experiences different conditions of heat transfer with area, which is very different from the plate rolling. For instance, one location of the surface strongly makes contact with roll and other location of the surface softly makes contact with roll, which can produce different thermal behaviors of the rod along the circumferential direction, leading to different microstructures and mechanical properties along the circumferential direction of the rod. This different thermal history of the rod with area can induce the surface flaw during hot rod rolling [1]. Consequently, it is necessary to understand the thermal behavior of the rod during the hot rolling process to solve the quality problems in wire, rod, and bar industries and to produce high quality products. Overall, it is insufficient to measure the temperature by pyrometers and to measure the temperature at only the center and surface areas in order to understand the thermal behaviors of the rod during the shape rolling process due to the complex roll contact of the rod during the process. This is totally different from the plate during the flat rolling process, as shown in Figure 1. To the best of the author's knowledge, no study has been attempted to measure the rod temperature during shape rolling in detail, although several works exist that focus upon plate rolling.

Therefore, temperature of the rod was measured using thermocouples at several points via off-line rolling simulator in order to understand the thermal behavior of the rod during hot shape rolling process in more detail, which can help to improve the surface properties of the rod, for instance through surface flaw, decarburization, and abnormal grain growth, during the shape rolling process. For comparison purposes, temperature of the plate was also measured during the flat rolling process. Then, the measured temperature of the rod and plate with thermocouples was analyzed using a numerical simulation.

**Figure 1.** Schematic description of (**a**) plate and (**b**) rod during the rolling processes.

**Figure 2.** Thermal image of billet during hot rod rolling using an infrared thermo-camera.
