DESCRIPTION

Techni cal field 5 The present invention relates to a method for monitoring and controlling of smeltmetallurgical,. endothermic and exothermic processes, particularly pyrometallurgical processes by means of optical spect roscopy .

10 The object of the present invention is to obtain a possibi lity to control different smeltmetallurgical, endothermic and exo¬ thermic processes, particularly pyrometallurgical processes by spect rophotomet ry in a simple and rational way thus making it possible to improve the yield of the processes in a qualitative

15 and quantitative way.

Background of the invention

In pyrometa I lurgy, such as copper production, lead production, and iron and steel production fire and flames are generally en-

20. countered. The visual observation of these flames has for a long time been used to judge if and when such a process is to be in¬ terrupted or .the running mode to be changed in any other way. In doing so one has then more or less on a firm ground supposed, that the process has reached a certain stage, phase or tempera- 5 ture as the flames have changed their colour, or seemingly have changed thei r colour.

In iron production certain quantitative methods of analysis are in use, whereas in /the copper production on-line methods are not 0 in use t o —t-h-e s.ame. extent. Particularly the copper converting orocess uses purely visual ethoαs for process control. In par¬ ticular during the slag blowing phase and the production of blister copper it is important that the orocess is interruoted at the correct moment. An incorrect blowing time Leads to an im- 5 paired yield and difficulties during the continued process. An incorrect temperature leads to disturbances of the Drocess pos¬ sibly leading to a b ^' reak down. One reαui rement is thus that the

end point of e.g. the slag blowing stage has to be determined to within better than one minute.

Depending on the composition of the melt the flame present above the converter will look different, and thus it can not easily be determined visually, if, and when, the process is to be inter¬ rupted .

Thus there exists a considerably technical problem to be solved within the smeltmeta llurgy, at endothermic and exothermic pro¬ cesses, and in that connection particularly within pyrometa Ilur- gy but also in plasma smelting and electric arc smelting.

Description of the present invention It has now surprisingly been shown possible to solve the above mentioned technical problems by means of the present invention which is characterized in that one first determines, for each endothermic and exothermic smeltmetallurgical process ^' and/or process step, characteristic optical emissions or absorptions and identifies the atomic and/or molecular origin of the emis¬ sions/absorptions, that one during an ongoing process run re¬ gisters changes in the characteristics of the emiss ons/absorp¬ tions and relates these changes to changes of the status of the process, and that one controls the process with reference he- reto.

One calculates preferably the ratio between intensities of at least two emitting products and/or the intensities at at least two different wave lengths of one emitting product.

Further character stics are evident from the accompanying clai ms .

The invention wi ll be described more in detai l in the following with reference to the example below, which example is directed to a copper converting process, however, without restricting the invention thereto.

Examp le

Spect rophotomet ri c measurements of a copper converter flame were carried out using a modern optical system capable of measuring either a complete spectrum or part of a spectrum with a high re¬ solution in a very short time period, typically 10 ms. A Jarrel- Ash model 1233 spectrometer containing three interchangeable gratings was used to disperse the incoming light. The gratings covered the spectral range 32008, 8008, and 2008 (and had 150, 600, and 2400 grooves per mm). The measurements were carried out in the wave length range 200 to 800 nm, and more particularly in the range 400 to 650 nm. The resulting spectra were recorded on a PARC model 1421 detector with 1024 diodes placed in a row in the exit plane of the spectrometer. The detector included an im- age intensifier. A PARC model 1460 OMA (Optical Multichannel An¬ alyser) console was used to display the data, and spectra were stored on floppy discs for further processing and evaluation. The system was used both for emission measurements, when the light from the converter flame was focused on the entrance slit of the spectrometer, and for absorption measurements when light from a Xe-lamp after having passed the flame was fed through an optical fibre to the entrance slit.

It turned out that emission spectra from the slag and copper making stages were quite different. On the other hand spectral characteristics were reproducible from one process cycle to an¬ other. Fig. 1 shows parts of the spectra emitted during slag and copper making stages. The spectra consist of a continous back¬ ground caused by Planck radiation from particles in the flame and contributions from gaseous elements. It turned out that the dominating gaseous emitter during the slag phase was PbS whi le during copper making phase PbO dominated the emission spectrum. Close to the end of the slag phase one could observe that the emission spectrum started to change from a pure PbS spectrum to a mixture of a PbS and PbO spectra. A series of dete rm i nat i c~ s were made in which the ratios between the intensity of certain PbO and PbS bands were recorded during the last minutes of the

slag phase. Fig. 2 shows the end-point value of the ratio deter¬ mined versus the copper content in the white metal (concen rated matte). When the ratio increases the copper content reaches an asymptotic value of 77.5 %. The Cu content of the white metal was determined using XRF (X-ray fIuorescense) determinations on a sample taken after the interruption of the process. By means of the present measurements it could be determined that when PbOCg) : PbS (g) approached 1 at the actual choice of wave lengths the slag blowing should be interrupted, however, preferably at a value of 0.2 to 0.5, whereat less amounts of metallic copper is formed .

Due to the high load of particles during slag phase it was pos¬ sible to carry out absorption measurements only during the cop- per making phase. Preliminary observations showed a strong ab¬ sorption at wave lengths shorter than 350 nm, which was in good agreement with laboratory measurements of SO, at 1200 C. When the end point is reached the concentrat on of S0-, decreases and the transmission of light down to 350 nm can gradually be obser- ved. A calculation of the ratio of the intensities at 320 nm and 350 nm was hereby used to determine the end point of the copper making phase. Fig. 5 shows the intensity of the S0-, absorption versus the wave length at some different instant's before the end point .

The theoretical composition of the gas phase in equilibrium with the matte and slag was calculated for comparison and as a guide in the analysis of the emission spectra recorded. In these cal¬ culations equi librium conditions were assumed to be present dur- ing both slag and copper making phases. The assumption of equi¬ librium is not necessarily true for all elements but is likely to describe the ma or changes during the converting cycle. Ne¬ cessary thermodynami c data were taken from a previous calcula¬ tion of the copper converting process and assessment of slag da- ta completed by data from thermodynami c tables. The calculation of the composition of the matte, white metal and slag phase in the converter process were carried out using the free energy mi-

nimiziπg programme SOLGASMIX.

The metal-containing gaseous constituents which were considered in the calculations were PbS, Pb, PbO, Bi , Bi_, BiS, Zn, ZnS, As, As ₂ , As,, AsS, AsO, As.S,, As.O,, Sb, Sb , Sb,, SbO, and SbS.

Fig. 3 shows the calculated partial pressure exceeding 10 bar during the slag production phase versus the amount of air in¬ jected per ton matte and the copper content of the matte. The calculated variation of the partial pressures of PbO ⁽ g) and PbS(g) is in good agreement with spectrometric measurements, and show a greater increase in the ratio p (PbO,g) :p(PbS,g) at matte concentrations exceeding 75% of Cu.

Fig. 4 shows the calculated composition of the gas phase in the copper making stage versus the amount of air being injected per ton matte during the in tial slag blowing stage. The white metal was supposed to contain 77.1% of Cu and 2% of the slag formed during the slag forming stage were supposed to be suspended in the white metal. From about 60 Nm air per ton of initial matte metallic copper starts to form and from that point the p(PbO,g) :p(PbS,g) natio remains unchanged during the whole pro¬ cess with a value of about 2.6 in a quantitative agreement with the optical observation.

The calculated distribution in the gas phase is in good agree¬ ment with the values given in Table 1, below, with exceptions for Pb and Zn, for which the calculated values are lower than n practical operation. Consequently the partial pressures of lead and zinc compounds in the gas phase have to be higher than the values given in Fig. 3.

The determination of the end point of the slag forming stage and copper making stage in the converting process is today mainly dependent on ski lful operators. Computer programmes for the cal¬ culation of-blowing time, addition of sand, and heat balances

exist. However, the reliability of the calculated blowing time is dependent on the amount of Cu left in the converter from a previous cycle, the amount of slag in the matte added, the com¬ position of scrap added and recirculated slag from different steps of the process, as well as the oxygen efficiency. All the¬ se parameters thus influence the process to a greater or lesser extent and thus have to be estimated by the operator. The calcu¬ lated blowing time can thus be used as a coarse guide only to determine the end point and the final decision is today in the hands of the operator. However, as has been shown above the light emitted from the process can be used, after spectrometri c determinat ons and analysis, to exactly determine the end point of the process steps, independent of the ski I of the operators.

Table 1

The values given are in % by weight for the composition of mat¬ te, slag and white metal in the actual plant.

By means of this new method the process can also be closed to a greater extent which leads to less emissions of sulphur dioxide through the ventilation system of the plant.

The invention is thus based on a technique where, in this case, the end point of the slag producing phase of a copper process can be determined by measurements of the intensity ratios bet¬ ween the emitted light from PbO(g) and PbS(g) whereby variations in the background due to disturbances are restricted, and dis- turbances from variations in the total Lead content are restric¬ ted as long as spectra from lead compounds can be identified. The lead content in the white metal formed in the slag phase va-

ried in the actual plant between 0.5 to 2 % by weight and even at lower concentrations the emitted light was strong enough to be detected. The variations of the silicon content of the slag influences the oxygen partial pressure and thus also the PbOrPbS ratio in the gas phase. These differences dependent on varia¬ tions of the silicon content are small compared with the rapid increase of the PbO:PbS ratio at the end of the slag producing phase .

The utilization of the present method for an exact monitoring and control of the slag producing phase at a copper process leads to potentially great advantages as the number of over blows and to early interrupted slag blows are reduced to a mi¬ nimum. Normally the copper content of the white metal is 76 to 77 % after the final blow. If this step is interrupted too early the blow has to be restarted after an analysis, which takes about 20 minutes to carry through. If on the other hand one dri¬ ves the copper making stage with too low Cu content there exists a considerable risk for of the formation of magnetite and an in- creased risk of slag foaming with a great time loss as a conse¬ quence .

The exact monitoring and control of the copper content is also essential to optimize the removal of impurities to the slag pha- se during the slag stage. Several impurities such as Ni and Sb show an improved distribution to the slag phase at higher copper concentrations but if the blow continues above 78% of copper in the white metal metallic copper starts to form as indicated above, and this counteracts the elimination of impurities to the slag as the distribution of impurities between slag and metallic copper is more unfavourable than between slag and white metal.

It is not only the- Cu concentrations wh ch .are of interest to control, but it is also possible to determine and control the release of halogenides which are known to form compounds with precious metals. The precious metals are a valuable constituent in the form of an impurity in the copper metal and these pre-

cious metals should not be evaporated in the form of gaseous molecules with a halogen.

In the same way the temperature can easi ly be followed in a pro- cess by recording and comparing the intensities in different emiss on/absorption bands from one and the same molecule and/or the width of an em ssion/absorpt on band from one and the same molecule. For example one can record and compare the intensities of the PbO bands in a copper process, and in an iron steel pro- cess different atomic Fe lines can be compared with each other. Fig. 9 shows the changes in relative intensities of two emis¬ sions of one product of an iron/steel process, where the inten¬ sities are recorded from 1500°C (top), and to 1650°C (bottom). Thus there is a gap of approximately 35 C between each record- ing. As is evident from these recordings the temperature can be easi ly determined by re-cording and calculating the ratios be¬ tween the two intensities.

The use of the present invention means that open converter pro- cesses can easily be closed and thereby it is possible to radi¬ cally improve the environment in and around a smelter. Further by closing the process gas collecting systems do not need to be dimensioned for t«1ιe large gas volumes necessary to handle when using an open converter to remove hazardous gases leaking to the environment but can be adopted to the amount of gas which is re¬ ally produced in the process.

In producing iron and steel one can, in the way described above, record and compare the intensities of different carbon compounds (CO, CO , CN ) or work with different atomic states in for exam¬ ple Fe, Mn, K, Na, etc. in order to determine the condition of the process. The measurement is carried out in the interval 260 to 600 nm, more preferably in the interval 300 to 400 nm. FIG. 6 shows the emitted spectrum from a converter flame in a small re- solution in the interval 200 to 700 nm, and FIG. 7 shows a spec¬ trum of the hot spot in the wave length interval 200 to 300 nm, in which spectrum the majority of the discrete structures come

from atomic iron, but also atomic manganese, sodium, and potas¬ sium have been identified. Strong iron and manganese lines have been marked in Fig. 7. The recordings show that the ratio Fe/Mn changes during the blow in a converter, as well as the K/Fe ra- tio varies considerably.

During the production of alloy steel, stainless steel and other qualities, Ni, Cr, Mn, Mo, Al, MgO, CaO, are present in the smelt, and, in particular, FeSi is present during the reduction phase. Mn and Cr are present as compounds and can be easi ly de¬ ter ined.

In the same way as described above the temperature in a smelter can be determined accurately, and the elementary composition of the melt can be determined in order to reach a correct alloy point. It is important in a steel making process to reach the correct concentration of carbon at the correct temperature of the melt. This can easi ly be achieved by means of the present invention. One can also use the method for measuring the compo- sition of the alloying metals of the iron. The elementary com¬ position can be used to control the addition of alloying metals which has become more important because of the increasing use of scrap metal, the total composition of which is very uncertain.

It is also very important to be able to monitor the status of the lining of a melting furnace so that the lining is not worn out and a penetration of the converter /furna ce takes place, which can cause very high costs. The quality of the lining layer can easily be determined by monitoring through the oxygen lance the K, Na, and Al presence in the hot spot. Particularly moni¬ toring of the em ssion/absorption of Na makes it possible to check the lining. The deter ination hereby involves a study of the intensity and shape of the absorption and emission lines of Na and K, particularly at 600, and 400 nm, respectively. Compare Fig. 8 in which graph a) shows the sodium lines in absorption only, and graph b) the lines as self-reversed emission lines.

At the production of lead the ratio PbO:PbS can of course be used as well as bands from imnpurities like Hg .

As evident from above impurities, but also a product to be pro- duced but present in a characteristic intermediary form can be utilized for these recordings and comparisons for process con¬ trol.

As indicated the method can use different measuring methods such as d rect optical measurement using conventional optics, but al¬ so modern fibre optics whereby process control based on the con¬ ditions underneath a protecting and cooling slag layer is tech¬ nically possible. A protecting and cooling slag layer may some¬ times change the status of the atom or molecule on which a de- termination is based and thus it may be of utmost interest to follow the conditions at the actual reaction site. This situa¬ tion is encountered for example, in a steel melt during oxygen b lowing .