x<vxv<xv
Steel is a pivotal material for modern societies. It is used in every industrial branch as well as a construction material and accounts for 95% of the world wide metal production [1]. However, the global steel industry has a share of approximately 8% of the global energy demand and therefore, contributes to 7%–10% of energy related CO2 emissions [2], [3], [4]. The hot rolling process accounts for approximately 20% of these emissions [5], [6]. The reheating furnaces, used to heat up the steel to 900 to 1280 °C to increase the deformability and subsequently enhance the product quality, are natural gas fired [7]. In order to significantly decrease the greenhouse gas emissions from these reheating furnaces, oxy–fuel combustion and hydrogen enrichment are promising solutions, as they are for many high temperature applications [8], [9], [10], [11]. Literature also shows that such measures can also be implemented in an economically sensible way [12]. Via oxy–fuel combustion the amount of exhaust gas losses, which can be as high as 43% in natural gas–air fired furnaces, can drastically be reduced [13], [14]. However, the influences of the differences in the furnace atmosphere, especially the increased moisture and the absence of CO2, may influence the amount of scale formed as well as the steel quality [15], [16], [17], [18], [19], [20]. The oxygen content of the furnace atmosphere also has a major effect on the scale formation kinetics [21], [22]. The influence of the residual oxygen content, however, is not caused by oxy–fuel or hydrogen combustion, but by changing the fuel–air equivalence ratio ϕ independent of the fuel and oxidizer in use. While hydrogen combustion reduces the CO2 emission to zero, the influence of hydrogen on the reheated steel has yet to be fully investigated [23]. As shown in literature, the scale formation can be significantly reduced by lower exposure times, lower residual oxygen and keeping the product temperature at the lower limit for the following steps [24], [25]. While optimization of these product parameters can also lead to significant energy savings and different scale formation, the presented work focuses on the comparison of different fuel and oxidizer combinations. Therefore, all these parameters were kept constant for the performed investigations, as the main focus of the presented work lies in the investigation of fuel and oxidizer combinations under industry-like conditions. To ensure these conditions, the experiments were carried out in a scaled down industrial furnace equipped with an industry-ready multi-fuel multi-oxidizer burner, instead of a smaller thermogravimetric furnace, as oft used in literature [26], [27], [28].
Fig. 1. Phase diagrams for iron and alloyed steel calculated using Thermo-Calc software [29].
2. Theoretical basics of scale formation in steel reheating
When heating up steel under the presence of oxidizing agents like oxygen or water vapor several chemical reactions appear, leading to a layer of different oxides. This process is referred to as scale formation, the surface layer is subsequently called scale. These conditions are unavoidable in reheat furnaces. The resulting scale layer needs to be removed before further processing of the intermediated products, to ensure consistent quality of the final products [30], [31]. As the reheating processed is crucial for the further processing, this means an unavoidable loss of produced steel. As mentioned in Section 4, an increased scale formation also leads to higher surface quality, due to the reduction of surface defects. For pure iron this scale layer consists of different iron oxides, namely wustite (FeO), magnetite (Fe3O4) and haematite (Fe2O3) [32]. For alloyed and especially high-alloy steels, this processes become more complex, due to the alloying elements also interfering with the scale formation reactions. Basically, the scale formation process can be described as Eq. (1)(1)M(s)+12O2(g)=MO(s)After a start phase of the oxidation a layer of solid metal oxide (MO(s) is formed on the metallic core, separating the metallic core (M(s)) from the oxidizer (O(g)) [33]. Therefore, the degradation of the material can be divided into internal and external oxidation, as well as phase transformations in the metallic core of the sample. As the insulating scale layer thickens over time, the diffusion length increases and the internal oxidation is slowed down over time [34]. This leads to a shift from a linear weight gain over time, to a parabolic rate law [35]. The overall rate of the scale formation reactions is dependent on the temperature, the availability and type of oxidizing species and the examined steel grade [36], [37]. In addition to the scale formation so called decarburization can occur, lowering the carbon content and therefore influencing the physical properties of the reheated steel. This decarburization is a combination of a surface reaction, the interstitial diffusion of carbon atoms and the dissolution of carbides within the steel [38]. The rate of the oxidation processes in reheating processes is defined by temperature, atmosphere and time. In Fig. 1, the theoretical phase diagrams for pure iron (Fig. 1(a)) and for the 1.2379 grade (Fig. 1(b)) can be seen. These phase diagrams show the equilibrium concentrations of different microstructures, at different temperatures. For the sample temperature used for the presented experiments, this temperature was 1250 °C, as indicated by the dashed red line. While this already leads to several different components when heating up pure iron or carbon steels, such as transformed austenite, wustite and haematite as well as solid solutions of the before mentioned oxides, the system gets much more complex for alloyed and high-alloy steels. As the 1.2379 phase diagram shows, not only many more different solid mixtures need to be considered, it also shows gaseous CO as well as influences from the alloying elements, such as molybdenum. Furthermore, the formation of oxides of all included species is contingent upon the availability of a suitable oxidizing agent [24]. While these very complex chemical processes, cannot be investigated in depth with the presented experiments, the influences of different furnace atmospheres on the final product quality can be determined. Nevertheless, the diagrams in Fig. 1 illustrate that the oxidation processes occurring in reheating furnaces unfold in phases of mainly austenite, intermingled with varying phases depending on the alloying elements.
3. Experimental material and procedure
3.1. Test furnace and measurement equipment
In order to ensure direct applicability of the obtained results, the steel samples were processed in a semi-industrial scale furnace. The internal dimensions of the combustion chamber are 2300 × 540 × 600 mm (L × W × H). For all presented experiments an industry ready multi-fuel multi-oxidizer burner that can be operated in the full range from 21 to 100% of oxygen as well as with hydrogen enrichments in natural gas between 0 and 100% [8], [9]. The fuel and oxidizer composition is controlled via a separate gas control line. Via an exhaust gas flap, the furnace pressure can be kept in slight overpressure, to ensure a tight furnace. This results in a constant, false air free furnace atmosphere throughout the treatment, which is critical for comparability of the obtained results. The furnace used allows for temperatures of up to 1430 °C and burner power up to 180 kW. It is equipped with extensive measurement equipment, including 13 thermocouples to measure the temperature distribution along the furnace ceiling, a differential pressure transmitter to monitor the overpressure as well as an ABB gas analyzing unit recording the dry flue gas concentration in the furnace atmosphere. In the presented experiments, no cooling lances were used. Instead, the cooling lance openings were used to insert up to 4 sample holders in cooling lance positions 2 to 5. The unused cooling lance openings were sealed with ceramic fiber insulation. The sample weight before and after treatment was measured on a lab scale with a repeatability of 0.1 grams. The furnace and the burner in use, as well as the temperature distribution in the given furnace for different fuels and oxidizers were already thoroughly investigated by the authors, both experimentally and numerically. An image of the test furnace in use can be found in Fig. 2. More detailed information about the furnace and the auxiliary equipment can also be found in these publications [8], [9], [39], [40]. The experiences gained via investigations combined with extensive pre-testing allowed for the definition of a measurement protocol for the heat treatment trials, described in Section 3 (see Fig. 2).
