The invention relates to a method to predict and control slag foam incidents in a smelting process in a metallurgical vessel.
Slag foaming is a well-known aspect in steelmaking in electric arc furnaces and basic oxygen furnaces. In an oxygen converter unstable slag foaming is known to result in slopping, which is the blow-up of the material contained in the converter.
Slag foaming is also known to occur in smelting reduction vessels for instance in the HIsarna smelting reduction vessel.
The HIsarna process is carried out in a smelting apparatus that includes (a) a smelting vessel provided with solids injection lances and oxygen-containing gas injection lances and is adapted to contain a bath of molten metal and (b) a smelt cyclone for partly reducing and smelting a metalliferous feed material which is positioned above and is in communication with the smelting vessel. The HIsarna process and the apparatus for the process are described in WO 00/022176.
In the HIsarna pilot plant a number of foaming events have occurred in the campaigns so far. Two types of foaming event were observed, a first type of slag foaming with a sudden and major evolution of foam with no advanced warning from the plant signals and a second type of slag foaming with a slower build up to foaming conditions. The first type of slag foaming, the sudden evolution of foam, was most likely triggered by the fall of large oxide accretions from the off-gas duct into the smelting reduction vessel which brought a large and uncontrolled oxygen input into the slag-metal system leading to a rapid increase in gas evolution. The second type of slag foaming, the slower type, appears to be related to a low process temperature with a partially solid slag and a high effective viscosity leading to high gas hold-up in the slag.
The second type of slag foaming event is a gradual event in which slag on top of the hot metal bath builds up, aided by bad slag composition and/or excessive gas formation, to a point that it starts to capture more and more gas and grows rapidly in volume. Once it passes a certain volume threshold in the smelting reduction vessel, being the level of the oxygen injection lances, the oxygen from these lances will suddenly pump up the volume, similar to blowing bubbles in soap water, and slag can reach or even pass the pressure relieve valve of the smelting reduction vessel. When the slag passes the pressure relieve valve the damage by the solidifying slag and hot metal not only concerns the smelting reduction vessel but also the outside and direct vicinity of the vessel.
The aim is to provide a solution to predict and prevent slag foaming events of the second type.
It is an objective of the present invention to provide a method to predict slag foaming events.
It is another objective of the present invention to provide a method to prevent slag foaming events.
It is another objective of the present invention to provide a method to predict and prevent slag foaming events that can be applied in a cost effective manner.
It is another objective of the present invention to provide a method to predict and prevent slag foaming events that can be applied in a simple manner.
The invention relates to a method as defined in claims 1-17. One or more of the objectives of the invention are realised by providing a method to control slag foaming in a smelting process in a vessel for smelting an iron-containing feed material comprising the steps of:
It was found in practice that it is possible to predict slag foaming events beforehand by using accelerometer data collected real time and control the input rates to the vessel and cyclone with sufficient time left to stabilize the process. The threshold value is determined on basis of historical accelerometer data, wherein the values derived from the historical accelerometer data comprise values which indicate the onset of a slag foaming event. By comparing the values derived from real time accelerometer data with the threshold value, it can be established whether or not a slag foaming event is about to occur.
The threshold value is a value representing a stage before a foaming event wherein the actual foaming event is not yet taking place and which is sufficiently in advance of the predicted foaming event to occur to be able to prevent the foaming event by adjusting the process. It was found that with the method a possible foaming event could be predicted in about 20 minutes in advance giving ample time to control the smelting process and therewith to prevent that the foaming event could occur.
A number of measures can be taken if it is established that a foaming event is going to occur without adjusting the process.
A typical adjusting step is adjusting the amount of oxygen injected in the smelting process. If the slag foam reaches the oxygen injectors in the smelting reduction vessel, oxygen injection will result in an increase in slag foaming and in most if not all circumstances will make the foaming event no longer controllable. The adjustment means decreasing the amount of injected oxygen and could include the total oxygen shut off, certainly if the foam is already rising in the vessel. A foaming event will also result in a pressure increase in the vessel, but at the time that the pressure has increased sufficiently to be able to measure the pressure increase the foaming event is already taking place and it will no longer be possible to control the foaming event.
Another adjusting step is adjusting the amount of coal injected in the smelting process. The gasification of coal contributes to slag foaming and should be controlled carefully and adjusted when a foaming event is predicted on basis of the values derived from the vibration measurement data. Adjusting of the amount of coal injected in the process means decreasing the amount of injected coal and could include that the coal injection is shut off completely.
Another adjusting step is adjusting the amount of iron-containing feed material injected in the smelting process. Adjusting means decreasing the amount of injected iron-containing feed material and could include that the iron-containing feed material injection is shut off completely. The reduction of iron ore has a direct effect on the amount of slag accumulating in the vessel and the associated gas formation. Reducing slag build up will reduce the likeliness of.
Another adjusting step is adjusting the amount of lime injected in the smelting process. The lime injected is one of the factors that determine the basicity and viscosity of the slag and therewith the extent of the slag foaming. Adjusting means decreasing the amount of injected lime material and could include that the lime material injection is shut off completely.
Typically it is provided that the basicity of the slag is monitored and the amount of lime injected in the smelting process is adjusted to keep the basicity of the slag in a predefined range or restore the basicity of the slag to within the predefined range.
Adjusting the amounts of gaseous and solid components is reducing the amounts of gaseous and solid components when the value derived from the accelerometer data corresponds to the threshold value. The term “corresponds” means that the value is equal to the threshold value or has passed the threshold value or is within a predefined range around the threshold value. Reducing includes decreasing as well as shutting off. Adjusting of the amounts of gaseous and solid components can be done successively or two or more components can be adjusted at the same time.
According to a further aspect it is provided that the method further comprises draining slag from the vessel. With this measure the amount of slag present in the smelting vessel is reduced and is only started when it is safe to start slag tapping. Opening the slag tap hole typically means drilling an open connection with the outside. When a slag foaming event is already progressing the slag might be forced out of the open connection. Therefore slag tapping will only be started either when the build-up is detected early on or after the process and therewith the forming of slag is back in control. Moreover, starting when the foaming event is already progressing will be too late because of the speed with which the foaming event will fill the smelting vessel.
According to a still further aspect when adjusting the amount of injected oxygen, the amount of oxygen injected in the smelting process is adjusted to a predefined excess of CO gas. The reason for this is that only a limited amount of CO is allowed in the off gas in order to prevent that an explosive mixture in the off gas is formed. Therefore adjusting the amount of injected oxygen is only possible up to a certain CO level in the off gas. The limit of the CO level is as stipulated in the safety regulations.
Typically the adjustment of the gaseous and solid components injected in the process is started with the adjustment of the amount of oxygen injected in the smelting process followed by the adjustment of solid components in the following order: coal, iron-containing feed material and lime. Although typically adjustment is started with the adjustment of the components that contribute significantly to a possible foaming event, it is of course also possible to start with the adjustment of all components simultaneously. However, the idea is to start with adjustment as soon as a possible foaming event is detected and to control the process with adjustments as minor as possible to be able to continue the process after preventing a foaming event without too much disturbance of the smelting reduction process.
The method further comprise that the adjustment of the amounts of the gaseous and the solid components injected in the smelting process when the value derived from the accelerometer data comes at the right side of the threshold value is increasing the amounts of the gaseous and the solid components injected in the smelting process.
According to a further aspect it is provided that the vibration of the metallurgical vessel is measured with one or more accelerometers for predefined periods of time at predefined time intervals. With the accelerometers the vibration of the installation or of the smelting reduction vessel can be measured at different positions and in different directions. By using multiple accelerometers it can be determined easily at which location and in which direction the most relevant and/or reliable information can be obtained. Measuring the vibration in a horizontal direction provided very useful measurements.
Accelerometers provide data over a wide frequency range but in case of large installations such as a smelting reduction installation the frequency range in which the relevant vibrations will occur will be far narrower than the total frequency range over which an accelerometer is capable of measuring vibrations. For that reason it is provided that a relevant frequency range is determined wherein the vibrations of the smelting reduction installation occur and only the accelerometer data in that relevant frequency range is going to be processed. This will reduce the required data storage and computer power to process the accelerometer data.
The method provided to process the accelerometer data comprises the steps of:
The meaning of the expression of “corresponds to the threshold value” is to be interpreted as including that the peak value nears, equals or passes the threshold value. Another possibility is to define a range for the threshold value and to take action whenever the peak value is within the defined threshold range.
The invention will be further explained on hand of the example shown in the drawing, in which:
In
The smelting reduction vessel 2 includes a forehearth 11 connected to the smelting chamber 8 via a connection that allows continuous metal product 12 outflow from the vessel. The forehearth 11 operates as a molten metal-filled siphon seal, which allows the molten metal level in the smelting chamber 8 to be known and controlled to within a small tolerance. Molten slag 13 produced in the process is discharged from the smelting chamber 8 through a slag tap hole 14.
The level of the slag 13 under stable operation conditions is about as indicated in the figure. The slag 13 will splash around at about this level and is stable in height and in heat transfer to the copper cooled panels provided in the wall of the smelting reduction vessel.
With a slag foaming event the slag level will raise considerable and when the level of the oxygen lances 10 is reached the slag will become even more gaseous and will raise further in an accelerated manner. The foaming slag will reach the smelt cyclone 3 and beyond and at the same time could push the molten metal out of the smelting chamber 8 and out of the forehearth 11. With that the maximum possible damage is done and a time consuming and costly clean-up of the installation will be necessary.
With the present installation useful results were obtained by measuring the acceleration every minute for a period of 0.5 seconds over a frequency range of 1-4000 Hz and converting the data to a spectrum of velocity (mm/s) and frequency. By looking at a single spectrum there is in most cases not a clear pattern visible, but by plotting a number of successively measured spectra in the same graph a clear pattern becomes visible showing a peak velocity value around 45 Hz.
At the lower frequency a large peak is shown which has to with the limitation of accelerometers at lower frequencies which is about 5 Hz and lower. For that reason all data below 5 Hz should not be used. It can further be seen that values above about 140 Hz do not show significant changes over time and for that reason should also not to be taken into account.
Good results have been obtained by only using data in a frequency range which is relevant for the method, which in this case is the range of 5-100 Hz. In order to damp the effect of sudden large deviations in peak velocity values a moving average of peak velocity values is determined and compared with the threshold value. A further correction is to apply an exponential smoothing to damp the peaks.
As an example: for the method an algorithm for a prediction model was build which comprises the following:
1. the installation should be in production mode for at least 15 minutes.
2. sample accelerometer data every minute in measurement periods of 0.5 seconds,
3. take the sum of all peaks between 5-100 Hz within one measurement period.
4. take the maximum peak in the 5-100 Hz range and divide the maximum peak by the sum of all peaks. This gives the contribution of the maximum peak with respect to the others.
5. Apply exponential smoothing with α=0,2 to damp the peaks
6. Take the 5 minute moving average of the calculated contribution factor to smooth the curve.
7. If there are 3 contribution factors within subsequent 5 minutes below 0.045 (threshold value), give an alarm.
Testing the prediction model based on the above algorithm on the historical data gave an alarm 20 minutes in advance before a slag foaming event and no alarms during stable production.
Instead of the above measurement periods of 0.5 sec every minute other measurements periods at different time intervals could be used which will also provide good results. For instance, with the given setup of the smelting reduction installation good results can be obtained if accelerometer data is sampled in successive sets of 0.1-2 seconds at an interval of 0.2-2 minutes.
Number | Date | Country | Kind |
---|---|---|---|
18156983.1 | Feb 2018 | EP | regional |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2019/050864 | 1/15/2019 | WO | 00 |