The morphology and EDS and XRD spectrograms of granules obtained by hydro-vacuum granulation of the most demanded slag melt GPIS presented in Fig. 3, where image (а) shows that the grain surface is smooth and glassy, without crystalline-ribbed roughness, grains do not contain open (through) pores and metallic inclusions. Relatively small granules consist smoother and more transparent glass formations. For a comparative analysis of surface morphology, Fig. 4 shows SEM images of the granules obtained by traditional and proposed wet granulation methods. The average height of the protrusions on the surface of conventional granules, as determined by optical profilometry, is on average 35 μm and the average depth of the grooves is 45 μm. The average height of the protrusions on HVG granules is only 5 μm, and the depth of the grooves is practically non-existent, which gives the latter a significant advantage in terms of moisture holding capacity.
The result of X-ray phase spectral analysis of granulated GPIS melt Fig. 3(b) (grain size 0.5–1 mm) shows that the obtained diffractogram does not reveal clearly distinguished diffraction peaks of the crystal component, the structure mainly consists of a glassy, hydraulically active amorphous phase, but the background presence of inclusions of mineral formations close in composition to gehlenite Ca2Al2SiO7, merwinite Ca3Mg(SiO4)2, shennonite (γ-2CaO∙SiO2), larnite (β-2CaO∙SiO2), goethite (α-FeO∙OH), and hematite (α-Fe2O3) is still identified. The number of inclusions of the noted mineral formation (MF) is reduced when the grain size of granulate (D, mm) decreases by dependence MF=0.8D+0.36, which is possible by increasing the degree of hydro-vacuum crushing and quenching by increasing the pressure of water supplied to the hydro-vacuum disperser (Fig. 1, positions 2 and 3). It is noteworthy that the slags granulated by traditional methods also contain crystalline inclusions of acermanite (Ca2Mg(Si2O7)), manganolite (Ca3Mg(SiO4)2), melilite (Ca2(Mg0.5Al0.5)·(Si1.5Al0.5O7)) and merwinite (Ca3Mg(SiO4)2), which makes them noticeably inferior in hydraulic activity to HVG granulate. In contrast to the conventional granulation technologies, in which the glassy phase content (HAGP) in the obtained granules is unstable and varies depending on their grain size, degree of loosening, swelling and volume overcooling (quenching), the HVG process can stably produce granules homogeneous in key characteristics and properties, which once again emphasizes its high efficiency.
The high degree of glazing of HVG granules is also evidenced by the results of differential thermal analysis, which are illustrated in Fig. 5. The derivatogram of HVG granules shows that at 1170 K an exothermic effect (release of latent heat) corresponding to the glass crystallization (reverse phase transformation from amorphous to crystalline state) is observed. It is preceded by endothermic effect (without mass loss) at 1130 K related to energy expenditures on bond rearrangements in the glassy phase before its crystallization. The endothermic effect at 1090 K may be caused by the polymorphic transformation of shennonite (γ→α). The broad endothermic effect at 1225 K corresponds to the beginning of glassy amorphous phase decomposition. For hematite (α-Fe2O3), a small endothermic effect at 990 K caused by its transition into γ-Fe2O3 maghemite is characteristic, accompanied by an abrupt change in properties.
Crystalline iron oxide monohydrates have three modifications: α (goethite), β, and γ (lepidocrocite). The differential curve of goethite (α-FeO∙OH) has an endothermic effect in the temperature range from 600 to 690 K caused by its dehydration with transition to anhydrous crystal form α-Fe2O3 (hematite). On the DTA curve, the endothermic effect also appears at 990 K, which is associated with the polymorphic transformation of α-Fe2O3 into γ-Fe2O3. A broad endothermic effect at 410 K is associated with removal of adsorbed moisture on the surface of the slag particle. Small endothermic effects at 470, 550, and 600 K are caused by the removal of hygroscopic moisture in β-modified iron oxide monohydrate (β-FeO∙OH), the content of which does not exceed 0.5 wt.%. Analysis of phase transformations with accompanying mass losses shows that the content of the crystalline phase in the granules does not exceed 1.5–1.6%. Thus, we can conclude that the GPIS granulate obtained by hydro-vacuum granulation on average by 98.5% consists of glassy, solidified in the amorphous state grains with micro inclusions of crystalline dicalcium silicate (shennonite, larnite), iron oxides (hematite, maghemite), and their hydrates (goethite, lepidocrocite). The phase composition of HVG granulate produced from FPIS melt practically did not differ from GPIS granulate. The only difference is the decrease in the ratio between of mineral compounds of Ca2Al2SiO7 to Ca3Mg(SiO4)2 by 18–20 % (graphically unrepresented). In turn, the phase composition of the LCCESS granulated by the hydro-vacuum method was slightly different from composition of GPIS. In particular, the samples were 95–96% composed of the glassy phase, with inclusions of merwinite (Ca3Mg(SiO4)2), magnesioferrite (MgFe2O4), larnite (Ca2SiO4), wustite (FeO), and goethite (FeO∙OH). In spite of the fact that HVG granulate LCCESS also consisted to a high degree of the glassy phase, its hydraulic activity was still significantly lower (by 25–20%) than that of HVG slag GPIS and FPIS, which can be explained by its relatively low AM (Table 1).
The results of the study of hydraulic activity of the most highly demanded GPIS granulate are given in Fig. 6, where the impact of granule size (D, mm) and the length of reaction with the lime (in days) per value of its absorption in milligram per gram is illustrated. The comparative deceleration of the lime absorption rate observed in granules of higher grain size can be explained by both a decrease in the specific area of the reacted surfaces and a possible rise in the presence of above noted microcrystalline mineral formations (MF=0.8D+0.36).
Figure 6 shows that the average СаО-absorption activity of granules produced in the HVG process makes 600–650 mg/g, which is almost two times higher than that of slag granulated by traditional technologies (320–360 mg/g [41]). This testifies to the obvious advantage of the technological conditions of high-speed degassing, globular-thin-walled (flake) dispersion, and high-gradient quenching, created during hydro-vacuum granulation of the slag melt.
Since the granulate activity depends to a considerable extent on its grain particle size (D, mm), an investigation for detecting the factors of its control was carried out. The main factors of technological were found to be the temperature (T, K) and the internal diameter of the melt feeding nozzle, i.e., the flow fed in the diffusor (d, mm).
The main results of this research are given in Fig. 7. From Fig. 7 (1), it can be seen that at hydro-vacuum granulation of the close in chemical composition GPIS and FPIS melts, at the 15 mm diameter of the feeding flow, the treatment temperature decrease from 1720 to 1600 K which leads to the coarsening of the produced at the output granules from 0.2 to 1.9–2 mm. At treatment, flow d=20 mm, and the average diameter of the granules grows from 0.26–0.3 mm to 2.4–2.5 mm. An increase in the flow diameter up to 25 mm gives an increase in the size of granules from 0.33–0.35 to 2.9–3 mm; during dispersion of the 30-mm flow, the granules coarsen from 0.4 mm to 3.5–3.7 mm. In turn, Fig 7 (2) shows that during treatment of the relatively viscous LCCESS melt, a reduction of its initial temperature from 1830 tо 1690 K, at dispersion of the 15 mm diameter flow, is accompanied with the coarsening of the average size of granules from 0.25 tо 2.2 mm. An increase in the diameter up to 30 mm and accordingly as relative doubling of the melt supply rate at the given temperature range led to the coarsening of the granules from 0.4 to 4.3 mm.
Figure 7 graphical dependences indicate on the existence of a stable essential correlation between the controlled parameter (grain size) and regulated technological melt treatment modes, which, in turn, can become an functional lever to operating control of the HVG process efficiency by choosing the rational values of these modes.
Analysis of the laboratory data showed that the fine, thin-walled semi-transparent particles of average size of 0.5–0.4 mm and lower are distinguished by the highest activity. Based on the above, the temperature meeting the condition D ≤ 0.5 mm could be taken as the temperature threshold of high efficiency of the HVG process. Based on the noted threshold criterion and data with given in Fig. 7 (1), the rational temperature for the GPIS and FPIS melts can be considered 1680 K. In its turn, the recommended temperature for the LCCESS melt is 1790 K.
Laboratory studies on the porosity and abrasion resistance of granules obtained by the HVG method, similarly to studies on hydraulic activity, showed a clear advantage of the HVG process over traditional granulation methods. On average, the porosity of HVG granules, depending on their fractionality (0–4 mm), was in the range of 5–20%, which is actually 2.5–3 times lower than the porosity of granules produced by traditional wet methods. In turn, the loss of mass during self-abrasion was 3–5%, which is 3.5–4 times lower than the index of abrasibility of granulated slag produced by traditional technologies.
The application of the similar method for assessing the HVG process efficiency by the operational characteristics of WHC and AMC showed that in contrast to the conventional hydromechanical INBA or VNIIMT technologies, where the WHC of the obtained from GPIS и FPIS granulate, depending on the grain particle size, can vary within 45–50%, the AMC 24–20%, during hydro-vacuum granulation, the WHC reduced to 25–13%, and the AMC to 4–6%. The noted quality indicators also significantly improve in the LCCESS granulate. Dependence of the WHC and AMC of granulate obtained by the HVG technique on its grain size is illustrated in Fig. 8.
The effect of perforation reduction degree and improvement of the WHC indicator is also clearly depicted in SEM images (Fig. 9), illustrating the surfaces of GPIS melt granules produced by the conventional hydro-channel (Fig. 9(1)) and the proposed hydro-vacuum technology under different temperature conditions of treatment (Fig. 9 (2–4)). The SEM images of surfaces of the granules produced by the HVG technique indicate on both an essential reduction of the chadded perforation and a reduction of the scale of volumetric swelling. At that, with a decrease in the granulation temperature from 1720 to 1680 K, i.e., with a limiting increase in viscosity and surface tension of the processable melted slag, the structure and morphology of the granules are improved and stabilized. This effect is another proof of essential advantage and necessity of selecting an optimum temperature range of processing the melted slags in the process of their hydro-vacuum granulation. So, unlike new HVG process, in conventional technological processes of wet granulation, at technological losses of heat, decrease in temperature, and fluidity of processed slag melt, structure and morphology of granules with an increase in their grain size do not improve; on the contrary, their degree of swelling increases, degree of quenching decreases, and as a result, content of hydraulically active glassy amorphous phase can fall below the limiting value of 66.6%, fixed by the European standard BS EN 15167-1: 2006. The self-abrasion rate also deteriorates, reaching 30% or more.
The HVG process also showed its high efficiency in terms of separation and extraction of metallic spherical particles and flake graphite included in the slag. Figure 10 shows the effect of separation of metal inclusions from slag melt (fractionality 50–200 μm). The degree of extraction depending on the temperature and viscosity of the slag varies in the range of 90–95%, a slightly lower index of graphite extraction—75–80%. Studies on determination of rational approach and regimes of separation of granulated slag and metal and graphite inclusions present in it are still in progress.
A system analysis of advantages and features of the semi-industrial pilot HVG-plant showed that its operation for granulation of slag melts of ferrous metallurgy in enterprises equipped with modern arc electro-thermic ore-smelting furnaces is feasible according to the schematic diagram presented in Fig. 11, where 1 is arc ore-thermal furnace, 2—chute for intake and primary processing (liquid-phase separation) of slag-metal melt mixture [42], 3—conveyer for casting liquid metal (LM) cleaned from associated slag, 4—container for receipt of solidified metal pigs (MP), 5—plant for hydro-vacuum suction and granulation of molten slag, 6—high-pressure water pump, 7—technological unit of continuous pulp (P) processing: separation and precipitation of metal inclusions, dehydration of slag granulate (SG), extraction of graphite (G), and return water supply, 8—granulated slag intake and final dehydration tanks, and 9—collector of extracted metallic beads (MB).
The pilot scale research of the proposed technological approach and plant for hydro-vacuum granulation of metallurgical slag melts showed that its scaling under the real industrial conditions will not be associated with outages and essential reconstruction of melting furnaces or the adjacent casting platform.
It was also revealed that in the case of significant overheating of the slag and the impossibility of regulating the temperature during hydro-vacuum treatment, the need to a proportional increase the pressure of the high-speed water flow used for granulation becomes paramount. A magnification in the feed rate and water pressure, together with an increase in the rate of suction (vacuuming capacity) and granulation of liquid slag, intensifies the heat removal process (accelerates the process of cooling and quenching the forming granules), which sufficiently compensates the deviation from the desired temperature regimes of processing. Due to the need to increase the water pressure, the power consumption increases only slightly, but with an increase in the productivity of the HVG-plant, this consumption is also fully compensated. Accordingly, the quality, safety, and economic performance of the HVG process do not deteriorate.