**Túlio M. Campos 1, Gilvandro Bueno 1,2 and Luís Marcelo Tavares 1,\***


Received: 26 June 2020; Accepted: 24 July 2020; Published: 27 July 2020

**Abstract:** High-pressure grinding rolls (HPGR) have gained great popularity in the mining industry in the last 25 years or so. One of the first successful applications of the technology has been in iron ore pressing prior to pelletization. Piston-and-die tests can provide good insights on the material response in an HPGR. This work analyzed confined bed breakage of four iron ore concentrates under different conditions. Saturation in breakage of particles contained in the top size in the tests was observed to occur at specific energies of about 2 kWh/t, whereas full saturation in breakage, with no additional increase in specific surface area of the material, occurred at energies above about 6 kWh/t. An expression was proposed to characterize the propensity of a material to break under confined bed conditions. The phenomenology involved in confined bed breakage of such materials was then analyzed in light of the results.

**Keywords:** bed breakage; iron ore; comminution; saturation; piston-and-die; compaction; compression; breakage

#### **1. Introduction**

Comminution machines that rely on confined bed breakage, namely high-pressure grinding rolls (HPGR) and the vertical roller mill (VRM), have gained great popularity in the cement industry [1–7] and, in the case of the HPGR, also in the minerals industry [8–13]. Their main advantages include good energy efficiency, high throughput, and ability to operate in size reduction operations from secondary crushing down to fine grinding. One particularly important characteristic of these technologies is that, as the applied specific energies are raised, the efficiency of energy usage drops, partially owing to the phenomenon of breakage saturation.

One important approach to gain insights into the operation of comminution machines that rely on confined breakage is to conduct tests using the piston-and-die apparatus [14,15]. Through these tests, it is possible to analyze how the applied energy is dissipated under controlled conditions and is used to create new surfaces. Schubert [16] highlighted that the total energy input in a particle bed can be divided into several microprocesses, namely friction losses, plastic deformation work, elastic recovery, energy dissipated during breakage, and compaction [16]. As such, in order to assess the particle breakage behavior resulting from the application of different applied vertical stresses, several authors [14,15,17–24] used the force–displacement curve and the size reduction results from piston-and-die tests for different materials. Kalala et al. [15] presented results indicating an elastic recovery of around 40% of the total input energy for pressing iron ores in a piston-and-die apparatus with compressive forces of 1700 kN and with particle sizes of about 12 mm.

Great attention has been dedicated to the study of the compaction behavior of particulate materials under confined conditions. Such compaction occurs during processes that include metal powder smelting, as well as tableting and pelletizing in the pharmaceutical, ceramic, and food industries. Indeed, compaction has been studied and modelled for inorganic powders [25], metal powders [26], ceramic powders [27], fine grinding processes [17,28,29], and organic powders in the manufacture of pharmaceutical tablets [30–33]. The particle bed compaction behavior can be influenced by different parameters including the deformation or stress–strain behavior [34,35] as well as particle size and shape [33,36]. Assessing limestone breakage response under confined conditions, Cabiscol et al. [29] observed that the packing density reached a maximum of around 90% of the specific gravity of the material with particles with mean initial size of 236 μm, being equivalent to the value found by Wünsch et al. [28] for different materials used in the production of pharmaceutics. On the other hand, recent works have attempted to assess particle bed behavior under confined conditions from computational simulations using the discrete element method (DEM) [37–39]. Besides DEM simulations, Garner et al. [37] also showed several experimental results highlighting the ability to reach relative densities around 0.95 for a vertical stress of 150 MPa. Indeed, the complexity of this process is the main reason for the poor predictability and the limited understanding yet available.

One particularly successful application of confined bed breakage is the HPGR in association with ball milling in either pregrinding or regrinding of iron ore concentrates for pellet feed production [8,13]. The present work investigated breakage of iron ore concentrates in confined conditions in single and multiple pressings in a piston-and-die apparatus with the aim of identifying the onset of saturation during confined breakage.

#### **2. Experimental Section**

#### *2.1. Materials*

Samples were collected of four iron ore concentrates that are fed into Vale's pelletizing plants at Complexo de Tubarão (Vitória, Brazil). Three of these concentrates (Itabira, Brucutu, and Timbopeba) are produced by flotation of ores from the Iron Quadrangle from the state of Minas Gerais (Brazil), having a top particle size of 1 mm. The fourth consists of the result of the preparation of ore from the Carajás mineral province (Pará, Brazil).

Specific gravity was measured by Helium pycnometry. Bulk (apparent) density was measured by placing each weighed sample in a beaker and measuring the apparent volume after vibration. Prior to testing, all samples were dried in air.

#### *2.2. Piston-and-Die Tests*

Particles contained in three size ranges were tested, namely 150–125 μm, 106–75 μm, and 53–45 μm, which were prepared by careful wet sieving. Each previously dried sample, containing 30 g, was placed in the die for testing. The piston had a 40 mm diameter and the resulting initial bed height was about 13 mm. As such, these tests were conducted under ideal confined particle bed conditions, as precluded by Schönert [14,40], that is, the height of the bed is larger than six times the top size of the samples and the bed diameter is larger than three times the bed height, so that wall effects were minimized.

Figure 1 illustrates the servo-hydraulic press (Shimadzu, Kyoto, Japan) used, showing the linear variable differential transformer (LVDT) employed to measure the particle bed displacement (a) and the piston-and-die apparatus in detail (b). The elastic deformation of the system was subtracted from the value measured using the LVDT.

The compressive forces were applied in the range of 50–1000 kN (40–800 MPa), and the deformation rate was 5 mm/min. The output of the test was the force–displacement curve. The specific input energy in each test was calculated from the numerical integration up to the maximum load. Upon completion of the test, the bed of particles was removed from the die, dispersed in water, and the size distribution measured. Size analyses were conducted directly of the material from the test by laser scattering in a Malvern Mastersizer 2000 (Malvern Instruments Inc., Malvern, UK) while the Blaine specific surface area (BSA) was measured using a PCBlaine-Star (Zünderwerke Ernst Brün GmbH, Marl, Germany). In addition to that, each pressed sample was also wet sieved using the bottom size of the interval in order to assess the proportion of material broken.

An additional set of tests consisted of conducting multiple pressings of material contained in the 106–75 μm size range. The initial bed was pressed up to 200 kN (159 MPa), the load removed, the material dispersed and analyzed, then it was reloaded to the die and the same procedure repeated. This specific energy was selected since it is in the range of values used in the pelletizing industry [12].

**Figure 1.** Press used in the piston-and-die tests highlighting the linear variable differential transformer (LVDT) used to measure the (**a**) particle bed displacement and (**b**) piston and die apparatus.
