METHOD FOR GRINDING A HYDRAULIC BINDER

Abstract

Disclosed is a method for grinding a hydraulic binder, including: a) introducing:—a hydraulic binder, and—a composition B including at least one grinding aid B into the first chamber of a horizontal grinder including several chambers, including a first chamber and a last chamber, each chamber being separated from the adjacent chamber by a diaphragm, whereby a composition β including the hydraulic binder and composition B is obtained in the first chamber; and b) grinding composition β in the horizontal grinder, whereby composition β moves from the first chamber to the last chamber and a ground composition C is obtained at the outlet of the last chamber. At the grinding step, the method includes introducing into the last chamber a composition A including at least one grinding aid A including an aminoalcohol. Also disclosed is a corresponding grinding unit.

Claims

1. A method for grinding a hydraulic binder, comprising: a) introducing: a hydraulic binder, and a composition B comprising at least one grinding aid B, into the first chamber of a horizontal grinder comprising several chambers, including a first chamber and a last chamber, each chamber being separated from the adjacent chamber by a diaphragm, whereby a composition β comprising the hydraulic binder and composition B is obtained in the first chamber, b) grinding composition β in the horizontal grinder, whereby composition β moves from the first chamber to the last chamber and a ground composition C is obtained at the outlet of the last chamber, wherein the grinding step comprises introducing into the last chamber a composition A comprising at least one grinding aid A comprising an aminoalcohol, composition A differing from composition β.

2. The method for grinding a hydraulic binder according to claim 1, comprising after step b): c) separating, by a separator, composition C ground into fines and separator rejects, where the mean size of the particles of the separator rejects is greater than that of the particles of the fines; d) recovering the fines; e) returning the separator rejects to the first chamber of the horizontal grinder.

3. The method for grinding a hydraulic binder according to claim 1, wherein the horizontal grinder only has two chambers.

4. The method for grinding a hydraulic binder according to claim 1, wherein the hydraulic binder is cement.

5. The method for grinding a hydraulic binder according to claim 1, wherein the grinding aid B comprises a polyol.

6. The method for grinding a hydraulic binder according to claim 1, wherein grinding aid B comprises: an aminoalcohol or one of the salts thereof, and a carboxylic acid or salt thereof.

7. The method for grinding a hydraulic binder according to claim 1, wherein the aminoalcohol of grinding aid A comprises: from 2 to 8 carbon atoms, and/or 1, 2 or 3 alcohol functions.

8. The method for grinding a hydraulic binder according to claim 1, wherein composition A is injected into the last chamber; either at the diaphragm separating the last chamber from the adjacent chamber, or into the enclosure of the last chamber, or at the outlet of the last chamber.

9. A grinding unit intended to implement the method according to claim 1, comprising: a source of hydraulic binder, a source of composition B comprising at least one grinding aid B, a source of composition A comprising at least one grinding aid A comprising an aminoalcohol, a horizontal grinder comprising several chambers, including a first chamber equipped with at least one inlet and a last chamber equipped with an outlet, each chamber being separated from the adjacent chamber by a diaphragm, wherein the last chamber is equipped with an inlet connected to the source of composition A.

10. The grinding unit according to claim 9, wherein the outlet of the last chamber is connected to the inlet of a separator able to separate particles according to their particle size and equipped with two outlets, one of the outlets being connected to an inlet of the first chamber of the horizontal grinder.

11. The method according to claim 5, wherein the polyol of grinding aid B is selected from among: a diol, a triol, and a tetraol, or a mixture thereof.

12. The method according to claim 11, wherein the diol is an alkylene glycol.

13. The method according to claim 12, wherein the alkylene glycol has 1 to 20 carbon atoms.

14. The method according to claim 11, wherein the grinding aid B comprises an alkylene glycol having 1 to 20 carbon atoms, or a mixture thereof.

15. The method according to claim 11, wherein the polyol of grinding aid B is selected from among: a diol selected from among 2-methyl-1,3-propanediol, monoethyleneglycol, diethyleneglycol, triethyleneglycol, tetraethyleneglycol, propylene glycol and a mixture thereof, a triol, which is glycerol, a tetraol, which is erythritol.

16. The method for grinding a hydraulic binder according to claim 6, wherein the aminoalcohol of the grinding aid B comprises: from 2 to 8 carbon atoms, and/or 1, 2 or 3 alcohol functions.

17. The method for grinding a hydraulic binder according to claim 16, wherein the aminoalcohol of the grinding aid B is N-methyldiethanolamine (MDEA), diisopropanolamine (DIPA), triisopropanolamine (TIPA), triethanolamine (TEA), ethanol-diisopropanolamine (EDIPA), diethanolisopropanolamine (DEIPA) or a mixture thereof.

18. The method for grinding a hydraulic binder according to claim 1, wherein the hydraulic binder is cement comprising mineral additions.

19. The method for grinding a hydraulic binder according to claim 1, wherein grinding aid B comprises an aminoalcohol or one of the salts thereof.

20. The method for grinding a hydraulic binder according to claim 2, wherein the horizontal grinder only has two chambers.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0101] FIG. 1: Schematic of a grinding unit according to the second alternative of the invention.

[0102] FIG. 2: Schematic of a prior art grinding unit and such as used in Example 1, indicating the sampling points 1, 2, 3, 4 5 and 6 for Example 1.

[0103] FIG. 3: Example of a Tromp curve. Percentage of separator rejects returned to the grinder as a function of particle size in μm.

[0104] FIG. 4: Efficiency C of the separator as a function of the quantity of grinding aid entering the separator for each grinding aid and each dosage of grinding aid measured in m.sup.2 of cement. The values indicated in the graphs correspond to the initial dry dosages in ppm of grinding aid relative to cement weight. The crosses correspond to grinding aid A comprising an aminoalcohol and the circles to grinding aid B comprising an alkylene glycol.

[0105] FIG. 5: Slope β of the separator fish-hook as a function of the quantity of grinding aid entering the separator for each grinding aid and each dosage of grinding aid measured in m.sup.2 of cement. The crosses correspond to grinding aid A comprising an aminoalcohol and the circles to grinding aid B comprising an alkylene glycol. The values indicated in the graphs correspond to the initial dry dosages in ppm of grinding aid relative to cement weight.

[0106] FIG. 6: Dosage of grinding aid (in dry g) measured per m.sup.2 of cement in the chambers of the grinder and at different points of the unit where samples were taken. The crosses correspond to grinding aid A comprising an aminoalcohol and the circles to grinding aid B comprising an alkylene glycol.

[0107] FIG. 7: Relationship between the quantity of grinding aid (in dry g) measured per tonne of cement on the samples and the specific surface area of the cement in cm.sup.2/g. The crosses correspond to grinding aid A comprising an aminoalcohol and the circles to grinding aid B comprising an alkylene glycol.

[0108] FIG. 8: Schematic of a grinding unit according to the second alternative and positions of sampling points in Example 2.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0109] FIG. 1 illustrates a grinding unit according to the second alternative of the invention, in the case in which the grinding unit comprises a filter. A horizontal grinder 11 comprising two chambers: [0110] a first chamber 12 equipped with a first inlet 13; and [0111] a second chamber (last chamber) 14 equipped with an inlet 15 and outlet 16, the first chamber 12 being separated from the second chamber 14 by a diaphragm 17. The inlet 13 of the first chamber 12 is connected to a hydraulic binder source 18, to a source 19 of composition B, and to an outlet 24 of the separator 23. The inlet 15 of the second chamber 14 is connected to the source 20 of composition A. The grinder 1 is configured so that air is able to circulate from the inlet 13 of the first chamber towards the outlet 16 of the second chamber 14. The outlet 16 of the second chamber 14 is connected to: [0112] a filter 21 configured to filter air and to return the filtered particles to the inlet 22 of a separator 23; and [0113] the inlet 22 of the separator 23 able to separate the particles according to particle size and equipped with two outlets: an outlet 24 connected to the inlet 13 of the first chamber 12 and an outlet 25.

[0114] The points 1, 2, 3, 4, 5 and 6 do not correspond to elements of the unit but indicate the different points where samples are taken, with reference to the following examples.

EXAMPLES

[0115] In the following examples, the grinding aids A and B have been fed in varying methods, either into the first chamber 12 via inlet 13, or at the discharge grate equipping the outlet 16 of the second chamber, or at both these points. The experiments conducted evidence the distribution of grinding aid A comprising an aminoalcohol or of grinding aid B comprising an alkylene glycol, and the impact thereof on particle size and cement flow, as well as the flow rates of material in the method, and show the advantage of injecting grinding aid A into the second chamber 14 of the horizontal grinder 11.

Example 1

Materials

[0116] The tested cement was of CEM I 42.5R type (94% clinker; 5.5% gypsum; 5.5% limestone).

[0117] The grinding aids were specifically formulated for the study. Their compositions are given in Table 1 below:

TABLE-US-00001 TABLE 1 Composition of the grinding aids A and B used in Example 1 Active material Compounds (%) Grinding aid A Water comprising an Triethanolamine 8.33 aminoalcohol Triisopropanolamine 33.33 (diluted) Grinding aid B Waater comprising an Diethylene glycol 22.70 alkylene glycol Glycerol (diluted) 12.09

[0118] The grinding unit used is such as illustrated in FIG. 2.

[0119] The horizontal grinder 11 comprised two chambers separated by a diaphragm 17.

[0120] For a cement without grinding aid, and for cements each containing grinding aids at different concentrations, samples were taken at different points on the circuit shown in FIG. 2, once the state of equilibrium had been reached on the grinding line.

[0121] Also, for cements containing grinding aids, samples were taken every 1.2 m in the first chamber 12 of the horizontal grinder 11 and at every metre in the second chamber.

[0122] In this first example, the grinding aid in liquid form was injected dropwise into the feed hopper of the hydraulic binder. The mixture of hydraulic binder and grinding aid was fed into the inlet 13 of the horizontal grinder 1.

[0123] Table 2 below summarises the different analysed samples

TABLE-US-00002 TABLE 2 Samples taken along the grinding line-initial dosages Samples Initial Dry * Reference taken in Samples taken dosage dosage given to the along the circuit Test Grinding aid (ppm) (ppm) dosage grinder (Points 2, 3, 4, 5, 6) T3 (ref) None 0 0 X X T4 Grinding aid 250 104 D1 X T5 A comprising 350 146 D2 X T6 an aminoalcohol 450 187 D3 X X T7 Grinding aid 300 104 D1 X T8 B comprising 400 139 D2 X T9 an alkylene 500 174 D3 X X glycol * Dosage of active material without taking water into account.

[0124] Test T3, without grinding aid, is the reference for the study. The different initial dosages are called D1, D2 and D3 hereafter

Methods

Mixing

[0125] A Kenwood Chef Elite mixer KVC5305S was used to mix the cement and ultrapure water with the desired water/cement ratio («W/C» hereafter). 400 g of cement were added to ultrapure water prepared in the bowl of the mixer, following the sequence indicated in Table 3:

TABLE-US-00003 TABLE 3 Mixing protocol Speed Time (rpm) Action 0′-30″ 43 Pouring of powder 30″-1′ 96 Mixing 1′-1′30 0 Scraping edges 1′30-2′30 96 Mixing
If there were many coarse particles in the cement, mixing was performed manually. The cement was added to the ultrapure water for 30 s after which the paste was mixed with a spatula for 2 min.

Assay of Grinding Aids

[0126] Assay of grinding aids was obtained by washing the cement and measuring the carbon concentration in the cement grout.

[0127] The cement was mixed with ultrapure water following the above-described mixing protocol, then left to stand for 30 min. After manual homogenisation, the grout was filtered through a Büchner, and the filtrate collected in a haemolysis tube after 0.2 μm filtration, and acidified to overcome any carbonatation. These solutions were passed through a Total Organic Carbon analyzer («TOC» hereafter) to determine the carbon concentrations.

[0128] The amount of carbon in the cement without grinding aid was deducted from measurements taken on the cements containing a grinding aid.

[0129] To verify that the grinding aid had no adsorption isotherm on the solid phase in water, measurements were taken with two water/cement W/C ratios: 0.4 and 0.6, on samples of fine particles taken at outlet 25 of the W/C separator (tests T3, T6, T9, point (6) on the grinding circuit). In the absence of an adsorption isotherm, the ratio of carbon in solution/cement is not dependent on the initial W/C ratio, and the entirety of the grinding aid is assayed.

[0130] Measurement of Total Organic Carbon (TOC) was performed on the acidified filtrates using a SHIMADZU TOC-VCPN analyzer. TOC was calculated by the difference between the quantity of total carbon (obtained by carbonisation of the solution and measurement of the quantity of CO.sub.2 released under infrared) and the quantity of inorganic carbon (obtained by acidification of the solution to pH<1 and release of dissolved CO.sub.2 by bubbling with synthetic air). A calibration curve for each of the grinding aids allowed determination of the concentration thereof in the cement grouts. It is expressed in g/L.

Results

[0131] The quantities of grinding aid are expressed in ppm (g dry weight of grinding aid per tonne of cement) or in g/m.sup.2 (g dry weight of grinding aid per square metre of cement surface area).

TOC Calibration

[0132] The calibration curves obtained for grinding aid A comprising an aminoalcohol and for grinding aid B comprising an alkylene glycol showed a coefficient of correlation of 1, and were therefore able to be used to calculate the quantity of grinding agent from the quantity of TOC measured in the samples.
The concentration of grinding aid A (C.sub.A) comprising an aminoalcohol was therefore calculated from the TOC value according to:

C.sub.A=0.0019*TOC+0.0115

The concentration of grinding aid B (C.sub.B) comprising an alkylene glycol was therefore calculated from the TOC value according to:

C.sub.B=0.0026*TOC+0.0111

Verification of Cement Washing

[0133] To ensure the absence of adsorption isotherm of the grinding aids on the cement, TOC measurements were taken with two W/C ratios: 0.4 and 0.6, for tests T6 and T9, at points (6) and the results were as follows:

TABLE-US-00004 TABLE 4 Efficacy of cement washing - TOC value as a function of W/C ratio TOC value at TOC value at Test W/C 0.4 (ppm) W/C 0.6 (ppm) T6 (6) 149 146 T9 (6) 132 137

[0134] The results show that the TOC values are not substantially dependent on the W/C ratio. The grinding aids are not adsorbed on the surface of the cement particles, and washing is therefore efficacious for determining the quantity of grinding aid in the different samples.

For the experiments described below, the cements were analysed with a W/C ratio of 0.4, to obtain a stronger concentration of carbon in solution.

TOC on Samples not Containing a Grinding Aid—Test T3

[0135] TOC was measured at points (2), (3), (4), (5) and (6) on the grinding line and inside the horizontal grinder 11, on samples ground without grinding aid (reference). The TOC values related to cement weight are given in Tables 5 to 7:

TABLE-US-00005 TABLE 5 TOC values on the different samples taken outside the grinder TOC Point on grinding line (ppm of cement) (2) Grinder surplus 2.8 (3) Filter 4.6 (4) Separator feed 3 (5) Separator rejects returned to the grinder 2.4 (6) Fines leaving the separator 3.5

TABLE-US-00006 TABLE 6 TOC inside the grinder, 1.sup.st chamber Distance from TOC inlet (m) (ppm of cement) 0 1.3 1.2 4.2 2.4 2.1 3.6 1.9

TABLE-US-00007 TABLE 7 TOC inside the grinder, 2.sup.nd chamber Distance from TOC diaphragm 17 (m) (ppm de ciment) 0 2.2 1 3.1 2 2.5 3 2.8 4 5.0 5 2.5 5.9 3.1

[0136] The cement of the samples in the tests conducted without grinding aid contained less than 5 ppm TOC along the grinding line.

[0137] In the corresponding samples containing grinding aid in the experiments described below, this quantity was deducted so as only to take into account the grinding aid.

TOC on Samples Ground with Grinding Aids

[0138] The results are expressed in dry dosage of grinding aid per tonne of cement or ppm.

Influence of Initial Dosage

[0139] Each grinding aid was fed at three initial dosages D1, D2 or D3 (Table 2). The quantities of grinding aid measured in the different cement samples are given in Table 8 below.

TABLE-US-00008 TABLE 8 Active material by weight of cement for each of the samples Initial dosage of grinding aid A Initial dosage of grinding aid B comprising an aminoalcohol comprising an alkylene glycol D1 D2 D3 D1 D2 D3 Active (2) Grinder 65 80 97 61 64 84 material surplus by weight (3) Filter 134 108 150 97 92 120 of (4) Separator 72 86 107 52 63 85 cement feed (g/t, (5) Separator 33 35 44 32 37 40 BWOC) rejects returned to grinder (6) Fines 92 108 142 93 99 123 leaving separator
It was observed that: [0140] The quantity of grinding aid in the cement at different points along the grinding line increases with initial dosage, with the exception of the filter 21 in which the dosage measured for dosage D1 is higher than that measured for dosage D2, irrespective of grinding aid. [0141] In the filter 21, for grinding aid A comprising an aminoalcohol at dosage D1, the quantity of grinding aid measured in the cement is higher than the initial dosage. The inventors assume that this excess could be due to the quantity of grinding aid on the surface of the particles of separator rejects reinjected into the grinding line, or to adsorption of grinding aid in suspension in the circuit. [0142] The dosage of grinding agent on large particles (those of separator rejects returned to the horizontal grinder 11) is lower than on the particles of small size (Fines leaving the separator). More specifically, the inventors have observed that, for both grinding aids, the quantity of active material of the grinding aids measured in the different samples is correlated with the specific surface area of the cement particles, as per the following equation:

quantity of active material (in g/t)=−0.0329*(specific surface area of the cement (in cm.sup.2/g))−1.637

[0143] with a coefficient of correlation R.sup.2 of 0.9528.

The measured quantities of grinding aid were therefore related to the specific surface area of the particles and are given in Table 9 below.

TABLE-US-00009 TABLE 9 Active material per m.sup.2 of cement for each of the samples Initial dosage of grinding aid A Initial dosage of grinding aid B comprising an aminoalcohol comprising an alkylene glycol D1 D2 D3 D1 D2 D3 Active (2) Grinder 2.0 2.5 3.2 2.1 2.5 3.0 material surplus (g) per (3) Filter 2.5 2.4 3.1 2.4 2.3 2.9 m.sup.2 of (4) Separator 2.1 2.5 3.3 1.7 2.0 2.9 cement feed (5) Separator 1.9 2.3 3.0 1.8 2.2 2.7 rejects returned to grinder (6) Fines 2.3 2.7 3.6 2.4 2.6 3.3 leaving the separator

[0144] Expressed in g per m.sup.2 of cement, the difference between the quantities of grinding aid at the different points of the circuit decreases for one same initial dosage: the specific surface area of the cement particles and hence their particle size governs the interactions between the grinding aids and the cement.

Separator Efficiency—Fish-Hook (β) and Bypass Complement (C)

[0145] The Tromp curve describes the efficacy of a separator. It is calculated for each particle size class as the ratio between the flow of separator rejects (returned to the horizontal grinder 11) and the flow of separator feed. In the case of perfect separator efficacy, the percentage of rejects would be zero up until the maximum acceptable particle size is reached, and then 100%. In actual cases, the Tromp curve of the separator has the shape given in FIG. 3. [0146] Bypass evidences that there exists, for every class of particle size, a fraction of particles that is always reinjected into the horizontal grinder 11. Separator efficiency is described by:

C=1-bypass [0147] the separator is all the more efficient the higher the value of C [0148] The «Fish Hook» is that part of the curve at which particle size is smaller than that corresponding to the bypass, and evidences escape of fines towards the horizontal grinder 11. The gentler the slope (β) the lesser the quantity of fines returned to the horizontal grinder 11 and the better the efficiency of the separator.

[0149] The results given in FIG. 4 and FIG. 5 show that separator efficiency is dependent on the grinding aid used: [0150] C is always higher and β lower with grinding aid A comprising an aminoalcohol; [0151] β exhibits non-monotone variation with the initial dosage of grinding aid B comprising an alkylene glycol, and decreases monotone fashion when the initial dosage of grinding aid A comprising an aminoalcohol increases. [0152] limit separator efficiency is reached on and after the intermediate dosage of grinding aid A comprising an aminoalcohol (C no longer varies, β stabilises)

Inside the Grinder—Tests T6 and T9

[0153] Samples were taken every 1.2 m in the first chamber 12, and every metre in the second. The dosages of grinding aid per m.sup.2 of cement are given in FIG. 6.

[0154] For grinding aid B comprising an alkylene glycol, the quantity decreases slightly in the first chamber 12 and then drops and stabilises in the second chamber 14. For grinding aid A comprising an aminoalcohol, this quantity increases in the first chamber 12 and stabilises in the second chamber 14.

[0155] In the first chamber 12 of the horizontal grinder 11, the quantity of grinding aid by weight of cement does not progress linearly with the specific surface area of the cement. This relationship becomes true as from the second chamber, as illustrated in FIG. 7.

[0156] Inside the horizontal grinder 11, the two grinding aids differ in the first chamber 12, grinding aid B comprising an alkylene glycol being more abundant per unit surface area than grinding aid A comprising aminoalcohol. The quantity of grinding aid per unit surface area of cement stabilises in the second chamber of the horizontal grinder 11, and the difference between the two grinding aids decreases.

Example 2

Materials

[0157] The studied cement was of CEM I 42.5R type (94% clinker; 5.5% gypsum; 5.5% limestone).

[0158] The grinding aids were specifically formulated for Example 2. Their compositions are given in Table 10 below. Grinding aid B1 is of same type as grinding aid A. Grinding aid B2 differs.

TABLE-US-00010 TABLE 10 Composition of grinding aids A, B1 and B2 used in Example 2. Active material Compounds (%) Grinding aid A Water comprising an Triethanolamine 8.33 aminoalcohol Triisopropanolamine 33.33 (diluted) Grinding aid B1 Water comprising an Triethanolamine 8.33 aminoalcohol Triisopropanolamine 33.33 (diluted) Grinding aid B2 Water comprising an Diethylene glycol 22.70 alkylene glycol Glycerol (diluted) 12.09
The grinding unit used was such as illustrated in FIG. 8. FIG. 8 is identical to FIG. 1 except that: [0159] the sampling points 2 to 6 are shown in FIG. 8; [0160] the inlet 15 of the second chamber 14, that is connected to the source 20 of composition A, is positioned at the discharge grate equipping the outlet 16 of the second chamber. The inlet 15 and outlet 16 are therefore at the same position.

[0161] The horizontal grinder 11 used comprised two chambers separated by a diaphragm 17.

[0162] In this second example, the advantage is shown of injecting grinding aids into each of the two chambers.

[0163] Different cases were tested: [0164] either no grinding aid was injected (reference—test T1); [0165] or grinding aid B1 was injected into inlet 13 of the first chamber and no grinding aid was injected into the second chamber (comparative—tests T2, T3 and T4); [0166] or no grinding aid was injected into the first chamber and grinding aid A was injected at the discharge grate equipping the outlet 16 of the second chamber (comparative—tests T5 and T6); [0167] or grinding aid B2 was injected via inlet 13 of the first chamber and no grinding aid was injected into the second chamber (comparative—test T8); [0168] or grinding aid B was injected via inlet 13 of the first chamber and grinding aid A was injected at the discharge grate equipping the outlet 16 of the second chamber, via an injection tube on the axis of the discharge grate (invention—tests T7, T9 and T10), with two distinct cases: [0169] either grinding aid B is B1 (same as grinding aid A) and the grinding aids injected into the first and second chamber are therefore of same type (test T7), [0170] or grinding aid B is B2 (differing from grinding aid A) and the grinding aids injected into the first and second chambers are of different type (tests T9 and T10).

[0171] Samples were taken at different points 2, 3, 4, 5 and 6 of the circuit illustrated in FIG. 8, once the state of equilibrium was reached on the grinding line.

[0172] Table 11 summarises all the tests conducted, initial dosages and points of injection.

TABLE-US-00011 TABLE 11 1st chamber 2.sup.nd chamber Dosage of Dosage of Dosage of Samples taken Samples taken at Test aid B1 (ppm) aid B2 (ppm) aid A (ppm) in the grinder points 2, 3, 4, 5, 6 T1 (ref) 0 0 0 X X T2 (comp) 250 0 0 X T3 (comp) 350 0 0 X T4 (comp) 450 0 0 X X T5 (comp) 0 0 262 X T6 (comp) 0 0 350 X T7 (inv) 250 0 138 X T8 (comp) 0 300 0 X T9 (inv) 0 300 113 X T10 (inv) 0 300 199 X X

[0173] The different flows were recorded at the sampling points of the circuit and percentage circulating load was calculated as the ratio between the flow of fresh material entering the circuit (here 48 tonnes/h in all tests) and the flow of material returned by the separator from outlet 24. This percentage measures the manner in which the process becomes saturated with material i.e. its relative congestion, and directly impacts efficiency. It is therefore sought to reduce this percentage so that it is possible to increase the flow rate of the method.

[0174] Particle size analyses on samples also allowed measurement of the fineness of the cement obtained at the outlet via rejects on 45 μm screen. The smaller the rejects the finer the cement.

[0175] The bypass values of the separator were also measured during the tests. The higher the value of parameter C, the better the quality of filtration at the separator.

[0176] Flow rate at the filter was also measured. The filter 21 can easily be saturated since it becomes charged with fine particles. It is therefore preferred to have a low flow rate at the filter 21.

[0177] All measurements are grouped together in Table 12.

TABLE-US-00012 TABLE 12 Flow at filter 21 Test Percentage circulating load C = 1-Bypass Rejects at 45 μm (tonnes/h) T1 (ref) 44.56% 94% 5.0 19.73 T2 (comp) 86.67% 90% 4.2 8.28 T3 (comp) 136.96% 83% 2.8 13.53 T4 (comp) 244.75% 73% 1.0 7.25 T5 (comp) 97.02% 89% 3.2 13.38 T6 (comp) 114.58% 87% 2.0 13.74 T7 (inv) 98.83% 89% 2.0 3.53 T8 (comp) 68.04% 91% 3.4 16.86 T9 (inv) 77.96% 90% 4.0 8.61 T10 (inv) 71.79% 91% 2.8 11.00

[0178] The injection of grinding aid B1 into the first chamber without any injection into the second chamber (tests T2 to T4) allows a fine cement to be obtained (smaller rejects than for reference T1) but leads to a strong increase in circulating load and to strong degradation of filtering quality at the separator as shown by parameter C.

[0179] Injection of grinding aid A into the second chamber: [0180] without an injection into the first chamber (tests T5 and T6), [0181] or by injecting grinding aid B1 into the first chamber (test T7),
allows easing of the method with lower circulating loads and better separation factors C than those obtained in tests T2 to T4.

[0182] Injection of grinding aid B2 into the first chamber (test T8) allows a lower circulating load and good filtering quality at the separator. Nonetheless, the flow rate at the filter 21 is the highest. Test T8 places a heavier load on this filter than tests T9 and T10, meaning that the latter offer a better compromise.

[0183] The most efficient combination is injection of grinding aid B2 into the first chamber and of grinding aid A into the second chamber (tests T9 to T10), for which the lowest circulating loads, highest separation factors C and cements of acceptable fineness are obtained. The circulating loads and flow rates at the grinder outlet filter measured in tests T9 and T10 allow an increase in feed flow rate hence in the general productivity of the method without risk of saturating the method.

CONCLUSION

[0184] These results show that: [0185] the quantity of grinding aid found on the cement particles is essentially governed by the specific surface area of the cement particles as soon as they leave the first chamber 12 of the horizontal grinder 11; [0186] The more the initial dosage of grinding aid is increased, the more the quantity of grinding aid increases along the grinding line. However, for grinding aid A comprising an aminoalcohol, at an intermediate dosage D2, there is a greater quantity of grinding aid than initially added. This difference could be due to the fact that the filter 21 selects the particles of smallest size for which the concentration of grinding aid is the highest further to the developed area effect. [0187] The loss of grinding aid along the grinding line increases with initial dosage, which could result from gas-solid adsorption at dynamic equilibrium in the horizontal grinder 11. To limit this loss, lower dosages should be used but the grinding aid would then not allow optimization of the parameters of the horizontal grinder 1 to reach the desired fineness. [0188] Inside the horizontal grinder 1, the two grinding aids differ in the concentrations thereof in the first chamber 12, grinding aid B comprising an alkylene glycol being the most concentrated. [0189] At the separator, grinding aid A comprising an aminoalcohol has a more favourable effect on the fish hook slope (β) and on bypass complement (C) than grinding aid B comprising an alkylene glycol. The intermediate dosage D2 of grinding aid A comprising and aminoalcohol already allows an efficiency plateau to be reached in the separator, contrary to grinding aid B comprising an alkylene glycol. [0190] Injection of grinding aid A into the second chamber and of grinding aid B2 into the first chamber allowed obtaining of the best compromise between separator efficacy, circulating load, cement fineness and flow rate at the filter, compared with: [0191] injection of grinding aid B1 into the first chamber without injecting any grinding aid into the second chamber; [0192] injection of grinding aid B1 into the first chamber and of grinding aid B2 into the second chamber.
These results show the advantage of injecting the two grinding aids at different points along the grinding line: grinding aid B at the inlet 13 of the horizontal grinder 11 to allow a sufficiently long residence time of the cement in the horizontal grinder 11 without removing fines too rapidly, and grinding aid A comprising an aminoalcohol in the second chamber 14 of the horizontal grinder 11 for better fluidification of the powder in the separator line.