Obtaining a submerged arc welding flux of the MnO–SiO 2 –CaO–Al 2 O 3 – CaF2 system by fusion

This article was downloaded by: [University of Sydney] On: 13 August 2013, At: 05:14 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Welding International Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/twld20 Obtaining a submerged arc welding flux of the MnO–SiO2–CaO–Al2O3 – CaF2 system by fusion A.C. Crespo a , R.Q. Puchol a , L.P. Goncalez a , L.G. Sanchez b , C.R. Gomez Perez a , E.D. Cedre a , T.O. Mendez a & J.A. Pozol a a Marta Abreu de las Villas Central University, Welding Investigations Centre, Villa Clara, Cuba b Centre for Metallurgical Investigations, City of Havana, Cuba Published online: 05 Oct 2007. To cite this article: A.C. Crespo , R.Q. Puchol , L.P. Goncalez , L.G. Sanchez , C.R. Gomez Perez , E.D. Cedre , T.O. Mendez & J.A. Pozol (2007) Obtaining a submerged arc welding flux of the MnO–SiO2–CaO–Al2O3 – CaF2 system by fusion, Welding International, 21:7, 502-511, DOI: 10.1080/0950711070159746 To link to this article: http://dx.doi.org/10.1080/0950711070159746 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions 502 Crespo et alWelding International 2007 21 (7) 502–511 Selected from Soldagem & Inspecao 2006 11 (3) 147–155 Welding International ISSN 0950 7116 © 2007 Taylor & Francis http://www.informaworld.com/journals DOI: 10.1080/0950711070159746 Obtaining a submerged arc welding flux of the MnO–SiO2–CaO–Al2O3 – CaF2 system by fusion A.C. CRESPOa, R.Q. PUCHOLa, L.P. GONCALEZa, L.G. SANCHEZb, C.R. GOMEZ PEREZa, E.D. CEDREa, T.O. MENDEZa and J.A. POZOLa aMarta Abreu de las Villas Central University, Welding Investigations Centre, Villa Clara, Cuba bCentre for Metallurgical Investigations, City of Havana, Cuba 1. Introduction MnO – SiO2 system fluxes are generally located according to their percentage ratio on the right of the eutectic zone of the binary system, with percentage ratios of MnO / SiO2 ≤ 1.12.1–5 This location of the fluxes is justified within the system in the search for lower consumption manganese materials to directly reduce costs and also to reduce the amount of P used in the weld metal. However, the MnO / SiO2 ratio is limited in terms of its suitability, not only at higher temperatures, but also because it is not certain that this ratio will influence all the properties of the flux. It is precisely this point that has led the majority of authors to consider adding modifiers (MgO, TiO2, Na2O, K2O, BaO, ZrO2), in attempts to correct the properties, but these have not achieved the same effect as a better MnO / SiO2 ratio, above all in the activities of the SiO2 and MnO, which translate into lower levels of Mn and higher levels of Si, C and S in the deposits, using the same electrodes and welding conditions. In these cases, when manganese minerals have the right levels of phosphorous and no financial limits exist, it is useful to work in the eutectic region (1.128 ≤ MnO/ SiO2 ≤ 2.4).6 In addition, the minerals of manganese are characterized after a heating process, by a higher MnO / SiO2 ratio.6–9 Therefore, the synthesis of an MnO-SiO2 flux in generally implies a displacement in the binary state diagram to an area of lower MnO / SiO2 percentage ratio values, which may possibly be achieved by the disassociation of higher manganese oxides present in the minerals.6 There are only two possible ways to achieve this: reduce the contents of MnO by means of reducing the Mn using a reducer (carbon, silicon, aluminium) when the end result obtained will be the desired flux, as shown in a number of works10–14 or increase the contents of SiO2 (by the addition of quartz, silica sand or other minerals containing this oxide), which has been the practice of the majority of the authors.1,6,15–17 The first way, even without allowing for financial considerations, has a low feasibility from the processing point of view, due to the complexity of the reductive process and particularly due to the high temperatures required, which do not allow a granulated end result to be obtained with the physical characteristics for direct use as a flux for the weld, and therefore necessitating a process of agglomeration, as has been reflected in a number of works.10,11,13,14 In accordance with the characteristics of the magnesium ore from the “Margarita de Cambute” deposit [in eastern Cuba], it is feasible to obtain an MnO – SiO2 flux which meets the MnO/SiO2 ratio, known to be favourable.6 Despite this fact, it was considered opportune to add modifiers to the system that equate to the cited binary ratio, contributing to improved behaviour by the flux and a better quality of the welded joint. In addition, the basic objective that covers this work is aimed at the synthesis of an MnO – SiO2 flux from the pyrolucite ore of the “Margarita de Cambute” deposit, and with the addition of correctors to the system (CaO, Al2O3, and CaF2) on the basis of an experimental design of mixtures for a restricted area. The high consumption of flux in the Republic of Cuba is due to the manufacture of domestic gas bottles, calling attention to the fact that lack of this application would signify a limitation on the versatility of use of the flux in carbon steel welds and light alloys in general. 2. Design of the test to obtain the flux Using the known criteria regarding the peculiarities of high silicon and high manganese fluxes, and the specific properties of the ore from the “Margarita de Cambute” deposit and its behaviour during the thermal process, as well as the characteristics of other valuable minerals, each one of the specific variables of the design and its ranges were defined. The MnO (X1) acts in a decisive way, together with the silicon (SiO2) (X2), in the metallurgical processes (the processes of oxidation – reduction and refining) during the welding, conferring specific metallurgical and D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 503Submerged arc welding flux technological characteristics on fluxes with significant contents of these oxides. The contents of these may be varied across a wide range, depending on the chemical, physical, technological and even financial factors, with the first being the most important in that it defines the flux properties and its behaviour in the weld in which it is used.4,5 The levels of impurities formed from inclusion in the manganese ore have frequently been evaluated.18,19 The specialized literature, in terms of the authors mentioned above and others, such as 1, 15, and 20 -22, offers a full range of MnO and SiO2 levels within the limits of 22 ≤ X1 ≤ 53 % and 35 ≤ X2 ≤ 52 %, respectively, which have been taken in this work as the variation ranges for these variables. The manganese ore from the “Margarita de Cambute” deposit (X1*) to be used has SiO2 within its composition at level values that do not meet the required MnO / SiO2 ratios, so it is necessary to add SiO2 (X1) in this instance, in the form of silicon sand (X2*).6 Other oxides that may appear as a constituent material in the manganese ore are CaO, MgO, Al2O3 and others, acting as modifiers within the diagram and on the technological and metallurgical properties of the flux. The CaO (X3) content has an influence on a combination of properties. In the specialized literature, levels of CaO ≤ 10 % are frequently reported in high silicon and high manganese fluxes.22 In every case, an increase in CaO (an increase in alkalinity) encourages the process of manganese reduction and diminishes the reduction in silicon, favouring the processes of removing sulphur and phosphor from the weld pool, in addition to i ts technological properties. It was therefore decided to increase its content to 20 %, with a view to evaluating the consequent intensification of the oxidation processes – reduction and refining. The desired content of CaO (X3) (0 ≤ X3 ≤ 20 %) falls within the determined proportions for limestone (X3*), (see table 1), allowing for the possible input of CaO from the other components of the load. The Al2O3 (X4) that may be present in the manganese ore frequently appear in the fluxes as a consequence of this. Its presence may affect the technological behaviour of the flux and above all its behaviour in the synthesis as a result of its influence over the silicate framework, but on the other hand it favours certain physical-chemical properties that have a positive effect, particularly because of its tendency to reduce the level of carbon in the metal.12 To evaluate this influence, without affecting the other properties, the Al2O3 (X4) must be moved in to the 0 ≤ X4 ≤ 13 % range, in accordance with that established by an important number of authors.4,5,15,19–21 For the addition of aluminium, kaolin and feldspar will be evaluated (see table 1). This does not cover the whole range of combinations of the test design, to cause of a lower Al2O3 / SiO2 percentage ratio, for which kaolin is defined as X4*. The level of SiO2, which supplies the kaolin, is achieved by the addition of silicon sand (X2*). The CaF2 (X5) exercises an influence on the properties of the flux and its interaction with the weld metal. The most suitable range for the CaF2 (X5) is 0 ≤ X5 ≤ 9 %, according to the reports from a substantial number of authors.1,12,20–22 The desired level of this compound in the composition of the flux is achieved by the addition of fluoride (X5*) (see table 1). As the properties of the experimental mixes are defined for the constituents, if these do not move into the range stnenopmoC OCaC 3- niloaC rapsdleF etiroulF etisuloryP OiS 2 43.0 65.54 03.66 42.2 06.8 lA 2O3 32.0 89.63 08.02 34.0 60.2 eF 2O3 71.0 53.1 07.0 - 71.1 OnM 2 - - 50.0 - 68.67 OiT 2 - 91.0 70.0 - - OgM 86.0 70.0 74.0 01.0 23.0 OaC 02.55 01.0 16.1 22.0 00.2 aN 2O - 61.0 59.5 - - K2O - 11.0 00.4 - - H2O - 24.41 - - 22.5 OC 2 72.34 - - - - P2O5 - - - - 72.0 FaC 2 - - - 01.69 - Table 1 Chemical composition of the load minerals, wt.% D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 504 Crespo et al of composition, at a predefined interval, the experimental planning uses a close approximation represented by a McLean Anderson design for determined boundary conditions, in which the number of experiments (N) is established as N = q·2q–1, with q being the number of variables (5, in our case) and which must fulfil the condition of normality ΣXi = 1.23 The number of experiments to be carried out in the full matrix is N = 5·24 = 80, but the application of the condition of normality and the elimination of points of variables outside the range made it only feasible to carry out 24 experiments. In light of that set out above, it was decided to establish, as a restriction on the design, a composition region of the eutectic zone in the MnO – SiO2 diagram (1.128 ≤ MnO/SiO2 ≤ 2.4) with the end of producing a suitable MnO / SiO2 ratio, at lower processing temperatures in the synthesis and better behaviour of the fluxes in the weld, which would ensure the quality of the joints.6 With the application of this restriction, only 8 experimental points were applicable, which, with the central point sand duplicates gave a total of 11 experiments (see table 2). To establish a criterion of the synthesised process of the flux as a function of the entry variables of the test design, the following were defined as variables: the casting time (Y1, min) and the mass of the flux obtained (Y1, kg). These variables allow sufficiently solid evaluation of the influence of the entry variables (fundamental oxides in the system) on the mass of the flux obtained in each case (Y2) and the time required to obtain the same (Y2). Both responses, as we can see in epigraph 4, are very closely related to the productivity of the process and in this particular case the time of casting is very closely related to the fusibility of the system of oxides. 3. Obtaining fluxes from the experimental design The system to obtain the set percentage ratios - MnO: SiO2, CaO: Al2O2: CaF2 in this case – then had to be defined (see table 2), along with the particularities of the specific minerals that supply these oxides (fundamentally their chemical composition and their behaviour in the pyro- metallurgical process) (see table 1) [6]. A process of calculation and conformation of the loads was carried out, taking into account for the first the transformation and interactions of the chemical compounds of the minerals on the basis of their stoichiometric ratios. The concrete results of the calculation of each one of the mineral mixes (experimental points) are shown in table 3. For each one of the mixes, (see table 3) a load of 4 kg tnemirepxE .oN )%(,OnM OiS 2 )%(, )%(,OaC lA 2O3 )%(, FaC 2 )%(, 91 * )1( 25 53 0 31 0 22 * )2( 35 53 21 0 0 32 * )3( 54 53 02 0 0 42 * )4( 35 74 0 0 0 83 * )5( 34 53 0 31 9 93 * )6( 35 53 0 3 9 24 * )7( 35 53 3 0 9 84 * )8( 35 83 0 0 9 1C * )9( 6.05 9.63 4.4 6.3 5.4 2C * )01( 6.05 9.63 4.4 6.3 5.4 3C * )11( 6.05 9.63 4.4 6.3 5.4 Table 2 Experimental points to be carried out. egrahC etisuloryP aciliS OCaC 3 eniloaC etiroulF 91 )1( 23.46 99.11 0 96.32 0 22 )2( 55.36 90.32 63.31 0 0 32 )3( 26.25 14.32 79.32 0 0 42 )4( 08.56 02.43 0 0 0 83 )5( 66.86 41.61 0 32.13 73.9 93 )6( 31.76 18.22 0 26.2 44.7 24 )7( 31.76 71.42 41.1 0 64.7 84 )8( 52.66 14.62 0 0 43.7 )9(1C 35.46 23.42 3.3 1.4 57.3 )01(2C 35.46 23.42 3.3 1.4 57.3 )11(3C 35.46 23.42 3.3 1.4 57.3 Table 3 Percentage mineral composition of the loads of the experimental points D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 505Submerged arc welding flux was prepared with the following granular sizes: pyrolucite, limestone and fluorite 4 / 0.037 mm, silicon sand 2 / 0.1 mm and kaolin, 0.05 / 0 mm. The components of the load were weighed in a SARTORIUS technical balance, which is accurate to 0.1g, and submitted to a mixing process for a period of 30 minutes in a gyrating drum mixer, to achieve the required homogeneity.24 The components of the load were added to the drum to ensure the best mixing in ascending order of their densities; kaolin (2.6 g/cm3), silicon (2.65 g/cm3), limestone (2.7 g/cm3), fluorite (3.18 g/cm3) and pyrolucite (4.7–5.0 g/cm3).25 The fusion and pouring process is considered the most important stage in the synthesis, as this is the stage in which the chemical-physical transformation of the minerals that enable the growth of new compounds imprint the flux with the desired characteristics. This stage depends greatly on the efficiency of the process in which the highest energy consumption is concentrated.17,24 The loads, duly mixed as in the previous stage, were submitted to electro-thermal fusion in an electric arc oven with a carbon element, coupled to a direct power supply at a current of 1000 A. The parameters of the ovens operation during all the castings were 33 V at 500 A. The loads were added in a continuous way, and in a way that formed a cone around the electrode to diminish any energy losses. Once an entire load had been added, it was kept in the oven at a temperature of 1500º C (measured with an optical pyrometer), after which the arc was interrupted, and it was then left for two minutes to achieve homogenization of the temperature. The end product was discharged into a granulation pool of water to achieve, after cooling, the outcome of a granulated flux with a vitreous appearance, a dark caramel colour, and polyhedral grains. In table 4, the results obtained are shown for each one of the castings, giving the time (Y1) and mass of the flux (Y2). The discharge conditions were: static water mass equal to 3.5 kg, discharge height between 0.28 m and 0.33 m, constant flow of 10–3 kg / (m2 min), Tpouring = 1480 – 1500º C.17 After the results of the casting, the statistical processing was carried out using STATGRAPHICS software, establishing regression models for the Y1 and Y2 variables that define the dependency of these with the components of the loads produced (X1). These models are: After obtaining the granulated masses, they were dried in the air and calcinated for 2 hours. The calcination temperature was selected from a table as a function f the alkalinity,1 with a result for all the fluxes obtained of around of 400º C. As part of this stage, the fluxes were put through a granulation process of 0.25 mm to 2.5 mm by means of a process of measuring and crushing grains larger than 2.5 mm, using a rolling mill in a closed cycle as previously defined.22,24 The fluxes obtained were then analysed for their chemical composition. The SiO2 was determined by a gravimeter, phosphorus and fluoride by using a colorimeter, and the rest of the elements by atomic absorption. The results are set out in table 5. The values for alkalinity, B and the chemical activity related to flux AF, as well as the related chemical activity (AMnO and ASiO2), and thermo-dynamic activity (aMnO and aSiO2), of the MnO and SiO2, respectively, were determined by the procedures given by Potapov1 and shown in table 6. 4. Selection of the suitable composition zone for the flux From the results reflected in table 4, we can see that the best responses were related to experimental points 39 (6), 42 (7), 48 (8) and C1 (9), C2 (10) and C3 (11). However, the inevitable variability of the chemical composition of the prime materials, although within small ranges, must be taken into account, as must the chemical composition of the fluxes (see table 5), and therefore it can be considered that there is not a single valid point in our selection, but an area, marked out by the in the boundaries of the points with the best results. It was decided that the ranges in which the fundamental compounds of the flux will be found (see table 5) are the following: MnO SiO2 CaO Al2O3 CaF2 50.6-53 35-38 0-4.4 0-3.6 4.5-9 The experimental points corresponding to this area (the flux selected hereafter is known as FFMN 1) were characterised by excellent behaviour during casting, no presentation of difficulties during fusion, and having egrahC emiT )nim( Y1 egrahC thgiew Y)gk( 2 egahC emiT )nim( Y1 egrahC thgiew Y)gk( 2 egrahC emiT )nim( Y1 egrahC thgiew Y)gk( 2 91 * )1( 52.35 082.3 83 * )5( 0.93 643.3 1C * )9( 05.43 643.3 22 * )2( 00.34 181.3 93 * )6( 0.53 053.3 2C * )01( 00.53 443.3 32 * )3( 00.54 221.3 24 * )7( 0.43 253.3 3C * )11( 52.43 243.3 42 * )4( 05.74 063.3 84 * )8( 5.33 263.3 - - - Table 4. Results of the locations of the experimental tests. 2 1 1 2 3 4 5 ajY 0.3497X 0.5792X 0.3593X 0.9051X 0.878X R 99.62%= + + + − = [1] 2 2 1 2 3 4 5 ajY 0.03099X 0.03677X 0.02266X 0.030197X 0.03876X R 99.62%= + + + − = [2] D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 506 Crespo et al maintained a good flow of end product during the discharge, leaving the oven clean . Points 19 (1), 22 (2), 23 (3), 24 (4) and 38 (5) demonstrated difficulties in behaviour during the casting. In the case of point 19 (1), li t t le fusibility was observed, which provoked a considerable increase in the casting time (see table 4) and the fluidity of the flux was not as desired, making pouring difficult. Points 22 (2) and 23 (3) demonstrated fusibility difficulties, although their high fluidity temperatures were good, but they both required a longer period in the oven (see the casting time in table 4). For flux 24 (4), there were difficulties due to high viscosity and it was necessary to over-heat it, with a subsequent increase in casting time. For flux 38 (3), as with the others, although to a lesser extent, the casting time was prolonged and its fluidity was not as desired. As is seen in the results of tables 2 and 4, from that set out above, and from regression equation Y1, the major influence on the fusibility and fluidity of the flux is the action of the fluorite. This influence can be defined as having values close to 4 %, for higher values its influence on these properties is less appreciable. In addition, the difficulties in the synthesis of experimental points (2) and (3) are related, not due to high CaO, but due to the .pxE OnM OiS 2 OaC lA 2O3 FaC 2 OgM aN 2 K+O 2O OiT 2 OeF P2O5 ytinilaklA )1(91 28.94 35.33 49.1 97.21 00.0 51.0 13.0 61.0 22.1 780.0 94.1 )2(22 25.05 24.33 44.11 5.0 00.0 36.0 54.0 41.0 22.1 780.0 15.1 )3(32 39.24 94.33 80.91 7.0 00.0 58.0 74.0 51.0 21.1 570.0 82.1 )4(42 22.94 56.34 42.2 23.0 00.0 26.0 36.0 02.0 13.1 680.0 31.1 )5(83 41.14 87.33 16.1 57.21 16.8 73.0 23.0 41.0 32.1 270.0 22.1 )6(93 24.05 03.33 70.2 80.7 65.8 95.0 34.0 41.0 12.1 880.0 15.1 )7(24 46.05 44.33 78.2 72.0 06.8 45.0 54.0 41.0 33.1 980.0 15.1 )8(84 64.94 64.53 11.2 52.0 04.8 36.0 84.0 51.0 02.1 780.0 04.1 )9(1C 32.94 09.53 82.4 78.3 83.4 25.0 94.0 61.0 11.1 070.0 73.1 )01(2C 69.84 20.63 13.4 19.3 43.4 84.0 25.0 61.0 99.0 170.0 63.1 )11(3C 20.94 67.53 33.4 30.4 14.4 25.0 54.0 51.0 11.1 760.0 73.1 Table 5 Chemical composition of the fluxes obtained, wt.% tnemirepxE B AF A 2OiS A OnM a 2OiS a OnM )1(91 56.0 37.0 25.0 41.0 92.0 20.0 )2(22 30.1 55.0 33.0 22.0 91.0 30.0 )3(32 71.1 15.0 92.0 12.0 71.0 30.0 )4(42 55.0 39.0 08.0 11.0 44.0 20.0 )5(83 06.0 57.0 65.0 01.0 13.0 10.0 )6(93 08.0 16.0 24.0 71.0 42.0 30.0 )7(24 48.0 95.0 04.0 81.0 32.0 30.0 )8(84 47.0 56.0 94.0 61.0 72.0 30.0 )9(1C 67.0 76.0 84.0 61.0 72.0 30.0 )01(2C 57.0 66.0 84.0 51.0 72.0 30.0 )11(3C 67.0 56.0 74.0 61.0 62.0 30.0 Table 6 Values for the alkalinity and the activities of the fluxes and MnO and SiO2 oxides D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 507Submerged arc welding flux absence of CaF2 in their loads. The variation in the mass of the flux obtained, given in table 4 and expressed by regression equation Y2 was less appreciable, being very close for the fluxes falling within the compound area selected. In the evaluation of the technological properties, the results were good in all cases, in that the fluxes falling within the compound area selected (flux FFMN 1) exhibited good behaviour. 5. Evaluation of the process of obtaining FFMN1 with average composition To evaluate the flux selected (FFMN 1), falling within the compound ranges selected above, 12 loads of ore were configured with the average composition of the flux, consisting of 2.675 kg of pyrolucite, 0.977 kg of silicon sand, 0.029 kg of limestone, 0.092 kg of feldspar and 0.227 kg of fluorspar (the substitution of kaolin for feldspar, in regard to the fluxes of the test, did not cause significant alteration, being calculated on the basis of its pseudo-components – pure oxides – and that these only differ in the Al2O3 / SiO2 ratio) and a process of synthesis was carried out, using identical procedures, parameters and equipment for all the test fluxes. Four castings were carried out, with three loads each, the results of which are shown in table 7. Table 8 shows the average chemical composition of each one of the fluxes, determined in an equal way as in egrahC )nim(emiT )gk(thgieW 1 5.84 668.9 2 3.05 758.9 3 6.25 029.9 4 52.94 110.01 latoT 61.05 419.9 S 87.1 170.0 No OnM OiS 2 OaC lA 2O3 FaC 2 OgM aN 2 -+O K2O OiT 2 OeF P2O5 1 12.05 43.53 26.2 64.2 55.6 46.0 27.0 41.0 02.1 880.0 2 95.94 30.53 8.2 45.2 17.6 36.0 27.0 41.0 32.1 880.0 3 59.94 29.43 46.2 93.2 84.6 46.0 07.0 41.0 91.1 580.0 4 30.05 21.53 95.2 84.2 6.6 36.0 96.0 41.0 02.1 970.0 ??? 59.94 01.53 66.2 74.2 95.6 46.0 17.0 41.0 12.1 580.0 S 62.0 81.0 90.0 60.0 01.0 00.0 20.0 00.0 20.0 400.0 No B AF A OnM A 2OiS a OnM a 2OiS 1 77.0 46.0 61.0 64.0 30.0 62.0 2 87.0 36.0 61.0 54.0 30.0 52.0 3 87.0 36.0 61.0 54.0 30.0 52.0 4 77.0 36.0 61.0 64.0 30.0 52.0 latoT 77.0 36.0 61.0 54.0 30.0 52.0 Table 7. Results of the locations for the average composition of the FFMN 1 flux. Table 8 Average chemical composition of the FFMN 1 flux in different locations Table 9 Values for the alkalinity and the activities of the FFMN 1 fluxes and its oxides in different locations the case of the values in table 5. In table 9, the alkalinity values and chemical activity related to the flux are given, as well as the related chemical activity and thermo- dynamism of the SiO2 and the MnO, determined in a similar way to those set out in table 6. As can be seen, the values of casting time and flux mass are very close in all cases. In addition, the values for the chemical composition of the flux in each case and the consequence of it, and the alkalinity values and activities were all also very close. The proximity of the results obtained corroborates the fact that FFMN 1 flux is reproducible. D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 508 Crespo et al For the technological and metallurgical evaluation of the flux in the weld, a representative 2 kg sample was taken from the total homogenized mass of flux from the 4 castings. The technological behaviour of this flux was evaluated by carrying out welds (5 overlapping welds) with 2 mm EM12K wire, and with the following parameters: Weld current, IS = 280 A; arc voltage, Ua = 30 – 32 V; weld speed V s = 42 m/h, inverted polarity and at a peak height of 25 mm.17 These tests were carried out using the OK Flux 1040 flux as a reference, and the results of the same were equivalent (i.e., excellent) in all cases, in agreement with the methodology of evaluation employed, based on the criterion of experts.26 The metallurgical evaluation of the flux was carried out by spectrum analysis of the atomic emission of the welds, obtained in the evaluation of their technological behaviour. The results of this evaluation are set out in table 10. These results for the flux obtained (FFMN 1) are better in terms of the Mn, Si, C, P and S obtained than with the commercial OK Flux 1040 flux, which is evidence that this test flux satisfies the requirements for the welding of gas bottles, in combination with EM12K wire. 6. Assessment of the mass and energy used in the process of obtaining the flux The assessment of the mass was carried out by processing the mineral load of the average composition of the FFMN 1 flux, and comparing that with the chemical composition of the minerals of the load. The results of these assessments are shown in table 11, where it is observed that of the total of the mix fed into the oven, 83.80 % was converted into flux, and the rest (16.19 %) corresponded to the exit gases. This result is close to that actually obtained in table 7, when for a mass of mineral load of 12 kg in each casting fed into the oven, an average mass tnenopmoC %,C %,nM %,iS %,P %,S 1NMFF 480.0 134.1 692.0 610.0 210.0 0401xulFKO 880.0 232.1 363.0 120.0 610.0 Table 10 Average chemical composition of the weld metal obtained using the FFMN 1 flux and with the OK Flux 1040 flux, with an EM12K electrode wire. Carbon and sulphur by gas volume analysis and the rest by colorimeter tnenopmoC laitinI tcudorP noitpmusnoC dleiY OiS 2 27.92 27.92 lA 2O3 60.2 60.2 eF 2O3 21.1 21.1 OnM 2 04.15 04.15 OnM 59.14 59.14 OiT 2 21.0 21.0 OgM 64.0 64.0 OCaC 3 17.0 17.0 OaC 97.1 04.0 91.2 aN 2O 43.0 43.0 K2O 72.0 72.0 P 70.0 70.0 FaC 2 05.5 05.5 latobuS 08.38 H2O 34.6 34.6 OC 2 13.0 13.0 O2 54.9 54.9 sesaG 91.61 latoT 99.99 99.99 Table 11 Weight of the mass of the process of obtaining the flux, expressed as a percentage D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 509Submerged arc welding flux of flux of 9.914 kg was obtained, corresponding to 82.61 % of the stated initial charge, and representing 98.58 % of the theoretical maximum. Alongside the results of the assessment of the mass, an assessment of the energy required for the load was carried out, which permitted the evaluation of the energy requirements while processing the load to obtain the flux. To determine the total latent heat, a sum of the products of all the latent heat was made (λ) for the number of moles (n) (Qlatent = Σnλ) [25]. These results are shown in table 12. The sensible heat of each one of the components s determined as 2 1 , T T Qsensible mcpdT=∑ ∫ taking the values for Cp given in table 12.25 In the case of the decomposition reactions of MnO2 and CaCO3, present in the load, the heat involved in these processes is taken into consideration as follows: R1 prod. Reacc 1 2 H H H 92 ( 124.5) 32.50 kcal/mol → ∆ = ∆ − ∆ = − − − = 2 2MnO MnO + O � (3) R2 prod. ReaccH H H 151.9 ( 94.0518) ( 288.45) 42.49 kcal/mol → ∆ = ∆ − ∆ = − + − − − = 3 2CaCO CaO + CO � (4) Taking into consideration the efficiency of the reactions, the content of MnO2 or CaCO3 supplied for the given load and the formation heat is: ∆H°R1=(32.5·51.401/ 87)·1=19202 cal and ∆H°R2=(42.4982·0.714/100)·1=303 cal. On the basis of the balance of the mass given in table 11, taking into account that it generates 0.4 g of CaO, 0.314 g of CO2, 41.948 g of MnO and 9.453 g of O2 and with the values of Cp given in table 12, the sensible heat for the products of the reactions is determined, and are also given in table 12. The total heat of the previous reactions, consisting of the formation heat, the sensible heat and the latent heat: QR1=46482.16 cal and QR2=633.93 cal. Finally, the total heat is determined as a result of all the heat, being the sensible heat, the latent and the reaction heat: QTotal=73.82 kcal. The energy consumption will be 0.08578 kWh (73.82 kcal) per 100 g of load. In addition, processing 100 g of load produces 83.80 g of flux, according to table 11, and therefore the consumption of energy per kg of flux will be 0.08578 kWh / 0.08380 kg = 1.024 kWh / kgFLUX If the average results of the castings of table 7 and the oven are taken to have average values of 500 A and 33 V, we can determine with close approximation the energy consumption of the oven: P – (500·33·50.1625 / 60) / 9.91= 1.392 kWh / kgFLUX. As a result, the consumption of theoretical energy, determined by the assessment of the Table 12. Cp, QLATENT and QSENSIBLE values for the components of the mineral load dnuopmoC ,ygrenE lom/lac lac,tnetalQ lom/lac,pC lac,elbisnesQ OiS 2 0043 47.3861 T5500.0+59.01 1.42931 lA 2 3O 00062 21.625 T/005225-T179800.0+80.22 2 8.6401 eF 2O3 00211 91.87 T/004324-T40610.0+27.42 2 71.694 OiT 2 00411 28.61 T/00914-T45700.0+18.11 2 58.94 OgM 00581 38.112 T/007802-T791100.0+68.01 2 16.112 OaC 04221 76.874 T/000801-T48400.0+01 2 27.508 aN 2O 0014 24.22 T4500.0+7.51 68.091 K2O 0515 69.41 T4600.0+9.51 55.801 rP 516 74.1 55.5 63.91 FaC 2 0014 13.982 T8300.0+7.41 93.3112 H2O 7179 12.2743 )oidemorppC()K892-833(240.1 3.839 latotbuS 47.5976 17.40991 snoitcaeR OnM 2T26300000.0-T83010.0+34.7 23.71891 O2 2T/007781-T852000.0+32.8 23.3647 OaC 2T/000801-T48400.0+01 50.081 OC 2 2T/000801-T47200.0+43.01 44.051 D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 510 Crespo et al energy (1.024 kWh / kg flux) represents 86.49 % of the actual average energy supplied to the oven. 7. Certification tests for the flux obtained With the objective of validating the possible industrial application of the FFMN1 flux as a substitute for commercial fluxes, certification tests were carried out under the terms set out in AWS Standard AWS A5.17 (1998).27 For the test section, ASTM A 36 steel was employed with the geometry as stated in the standard cited. The test section weld was carried out with an EM12K electrode of 4 mm, which is within the range recommended by the standard, which in turn recommends the following welding regimen: I s – 500 + -25 A, V a = 28 + -1V (direct current, inverted polarity), V s = 6.8 + -0.4 mm/s (24.48 + - 1.44 m/h). Pre-heating was not carried out, and the temperature was maintained at 135–163 º C between passes, for which a digital contact thermometer was employed at the point of the section established in the standard. The first weld was carried out with 2 passes, while the following had 2 or 3 passes. The section weld was carried out in a flat position, with the plates constrained by means of a fixing tool to ensure that angular deformation of the welded section would be ≤5º. The test section had thermal post-welding treatment with the end of reliving stresses. The section was placed in a Mufla oven at a temperature of 316º C, and gradually increased at a rate of 222º C / h, up to a temperature of 621ºC +/- 14ºC, where it remained for a period of one hour. Later, it was cooled in the oven at a speed no greater than 177ºC / h. The section was given a radiographic check, as set out in the standard, in order to eliminate the support and reinforcement given to the weld by machining. The result obtained was acceptable, being free from cracks, incomplete areas of fusion and rounded indications that would not be acceptable. For the tensile tests and impact tests, test pieces were extracted as set out in the standard. These were tested in an IBERTES machine, giving the results as shown in table 13. The impact test was carried out on test pieces at-18ºC. These tests used a Charpi pendulum to obtain the results shown in table 14. From the previous results, as recommended in the standard,27 the highest and lowest values were discounted, with the average value of energy absorbed being 30.33 J. In accordance with the results of tensile and impact obtained, it can be concluded that the FFMN 1 flux, in combination with an EM12K electrode, conforms to the F7PO – EM12K classification of AWS. It should be highlighted that the commercial OK Flux 1040 flux used, as referred to above, has an identical AWS classification. In accordance with the result of the certification tests, the flux responded successfully to a wide variety of applications. With the objective of corroborating if the flux obtained by the FFMN 1 – EL12K metallurgical system complies with the standards and requirements of the production of combustible gas bottles in the production process of the CONFORMAT Company of the Ministry of Extractive - Mechanical Industry of Cuba, tests were carried out under the specific conditions established by this company. Welds were made around three 10 kg gas bottles, under the welding conditions set out for the procedure (I s = 280 A, U = 30–32 V, V s = 42 m / h, 2 mm EM12 K wire, inverted polarity). From the technological point of view, the results obtained are similar to those of the commercial fluxes that are used in the process (Ok Flux 1040 and F-103). The three bottles were then annealed and pneumatic and hydraulic tests were carried out, which gave satisfactory results. In addition, the evaluation of the properties of the joints, for which a bottle was submitted to an explosion test, produced the following results: Fracture pressure – 122 kg / cm2 (technical specifications ≥ 85 kg / cm2, fracture of the base metal). From the other bottles, 4 test pieces from diametrically opposing positions were extracted, and were submitted to tensile tests as set out by the company. The results of these tests are shown I table 15 (in all cases, the fracture occurred in the base metal). The results obtained with the FFMN1 flux as a substitute for the commercial OK Flux 1040 (ESAB) and F-103 (UTP) fluxes in welding bottles, combined with the results of the homologation tests satisfied the established requirements, which allows an assurance that the flux’s application from both the technological point of view and the quality of its joints is feasible. 8. Conclusions The experimental plan, based on the design of McLean and Anderson mixes, allowed the study of the behaviour in the synthesis process to be carried out, and established the equations that relate the composition of the flux to its fundamental components (MnO (X1), SiO2 (X2), CaO (X3), Al2O3 (X4) and CaF2 (X5)) with the casting time and the mass of the flux (Y1 and Y2 respectively). The region of the flux compound that responded with the most suitable results is characterised by an average ratio of MnO / SiO2 = 1.42, and its fundamental components ,htgnertselisneT )aPM( ssertsdleiY )aPM( ytilitcuD )mm( noitagnolE )%( 46.415 78.824 50.66 5.92 Table 13. Results of the tensile test nemicepS 1 2 3 4 5 J,ygreneetulosbA 62 13 82 93 23 Table 14. Results of the impact tests nemicepS 1 2 3 4 htgnertselisneT mm/fgk( 2) 8.46 8.46 7.56 9.36 Table 15. Tensile tests on the welded joints D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3 511Submerged arc welding flux fall into the following ranges: MnO SiO2 CaO Al2O3 CaF2 50.6 – 53 35 – 38 0 – 4.4 0 – 3.6 4.5 – 9 In accordance with the assessment of the masses, 83.79 % of the mass fed into the oven was transformed into flux, while for the actual results this value was 82.61 %. In terms of the return from the synthesis process of flux, this constitutes 98.53 % of the theoretical maximum. The average real energy consumption of the over represents 1.392 kWh / kg FLUX. The theoretical energy consumption for the processing of the load of the oven, determined from the balance of the energy (1.024 kWh / kgflux) represents 86.49 % of the real energy supplied to the oven. In accordance with the results of the certification tests that the flux obtained (FFMN 1 flux), they meet the AWS F7PO – EM12K classification. From the point of view of its behaviour in the weld and its application in industrial conditions, the flux responded satisfactorily as a substitute for commercial fluxes with an identical AWS classification. 9. References 1 Potapov, N.N., Welding material, Vol. 1, Shielding gases and fluxes. Mashinoestroenie, Moscow, 1989, 544 p. 2 Stukalo, V. et al., Avt Svarka 1 (1987) 40–42. 3 Nishio, V.K., Manufacture of fused flux, Patent No. JP 3027892A2, Kawasaki Steel Corp., 1991, (02.06) 4 Noguchi, Y., et al., Fused flux for submerged arc welding, Pat- ent No. JP 11019795, SUMIKIN YOSETSU KOGYO K.K., SUMIKIN DAIKEIKOUKAN, K.K., 1999, (01.26) 5 Okabe, Y.Y. and Sakaguchi, S. Fused flux for submerged arc welding, Patent No. JP 8187593, Kawasaki Steel Corp. 1996, (07.23). 6 Cruz, A., et al., Rev. Metal, Madrid 39, 2 (2003), 114–122. 7 Jimenez, G., et al., Technology for Fe Mn production, Techni- cal Information PPE, 15-029-05, CIME, 1984, 123 p. 8 Jimenez, G., et al.,, Rev. Metal, Madrid, 28, 2 (1992) 114 – 1 1 8 9 Gomez, P., et al., Rev. Metal, Madrid, 34, (1998) 391 – 394 10 Perdomo, L., Obtaining an agglomerated flux from refined products of Cuban prime resources, Doctoral Thesis, Higher Institute of Mineral Metallurgy, Ma 1994, 164 p. 11 Perdomo, L., et al., Use of prime resources to obtain agglom- erated fluxes, used in the automated submerged arc welding (SAW), Rev. Metal Madrid, 39, (2003) 268 – 278, 12 Perdomo, L., et al., Carbo-thermal reduction of metallic min- erals: one method for obtaining ferroalloys and scrap em- ployed in the development of welding consumables, Notes of MATANICA, 2003, Havana, July 2003. 13 Cruz, A., Obtaining high carbon ferro-manganese using carbo- thermal reduction of a pyrolucite mineral, Notes of MATANICA, 2003, Havana, July 2003. 14 Perdomo, L., et al., Processing automatic weld residuals to obtain Fe C R Mn alloys and matrices for the fabrication of fluxes using SAW, Notes of the II Material Engineering Con- gress, Havana, September 1998. 15 Suruhashi, S., Fused flux for submerged arc welding, Patent No. JP 9085488, Kobe Steel Ltd, 1997, (03, 31). 16 Volovel, S.D., Fused flux for welding deposition, Patent No. RU 2116529, TsNIISKM, 1998 (07, 20) 17 Cruz, A., Obtaining, by fusion, a fused flux of an MnO – SiO2 system for submerged arc welding, Doctoral Thesis, Higher Institute of Mineral Metallurgy, Moa 2002, 158 p. 18 Kondipka, V.P. and Paramonchuk, I.B., Study of the possibil- ity of using scrap of the production of ferro-alloys for the fusion of fluxes contained in manganese, Notes of the 5th “Manganese Metallurgy” Conference, Ukraine, Nikopol, 10– 12 Sep, 1991, pages 129–132. 19 Abe, T., Fused flux for submerged arc welding, Patent No. JP 7303990, Kobe Steel Ltd, 1995, (11, 21) 20 Suruhashi, S. Fused flux for submerged arc welding, Patent No. JP 8267279, Kobe Steel Ltd, 1996, (10.15) 21 Suruhashi, S., Fused flux for submerged arc welding, Patent No JP 9262692, Kobe Steel Ltd, 1997, (10.07) 22 Podgayetskii, V. and Kuzmenko, G., Theory of scrap, Naukova Dumka, Kiev, 1988, 255 p. 23 Aknazarova, S. et al., Experimental optimization in chemis- try and chemical engineering. Mir, Moscow, 1982, 471 p. 24 Gomez, C.R., Obtaining fused flux from S.A.W. using Cuban rocks, Doctoral Thesis, UCLV, Santa Clara 1995, 167 p. 25 Perry, J.H. and Chilton, C.H., Chemical Engineer Handbook, Ed. McGraw-Hill, New York 1999. 26 Marcelo, D., and others, Rev. Energetica, 22 (2001) 27 AWS A5.17, Specification for carbon steel electrodes and fluxes for submerged arc welding, (1998). D ow nl oa de d by [U niv ers ity of Sy dn ey ] a t 0 5:1 4 1 3 A ug us t 2 01 3