Method for monitoring the operation of an electric arc furnace, associated electronic monitoring device

Abstract

A method for monitoring the operation of an electric arc furnace, operation during which a first load of material to be melted is charged in the electric arc furnace, the method comprising: acquiring: o an arc stability signal (Stb), representative of the stability or instability of an electric arc present in the electric arc furnace, o one or more thermal losses signals (Tin, Tout, CWF), relative to thermal losses (TL) of the electric arc furnace, determining a recharging readiness signal (RR), which specifies whether the electric arc furnace is in condition for charging a second load of material to be melted, or not, the recharging readiness signal being determined based on the arc stability signal and on the one or more thermal losses signals, emitting the recharging readiness signal.

Description

Method for monitoring the operation of an electric arc furnace, associated electronic monitoring device

[0001] The technical field is that of steel making using an electric arc furnace, in particular monitoring and controlling the operation of an electric arc furnace.Technical background

[0002] One route for producing liquid steel is to melt steel scraps in an Electric Arc Furnace (also designated as an “EAF” in this document), which allows for recycling steel and reducing carbon footprint. During this process, usually, two or more buckets of scraps (or other kinds of loads) are successively charged and melted in the EAF. A first bucket of steel scraps is charged in the steelmaking vessel of the EAF (where some liquid steel usually remains from a previous heat), and an electric current is supplied to the graphite electrodes, which enables to heat and melt the scraps. Then, the EAF is charged with a second bucket of scraps which is then melted. Then, a next bucket of scraps is possibly charged again and so on. After a number of buckets of scraps have been charged and melted, the liquid steel is tapped.

[0003] The moment the second (or next) bucket of scraps is charged has to be well chosen. If too early, the first (or previous) load of scraps will not be melted yet, and there is a risk of liquid steel overflow. If too late (i.e.: noticeably later than the moment the first steel scraps are completely melted), the global productivity decreases and the energy consumption per bucket increases.Summary

[0004] In this context, a method for monitoring an EAF, according to claim 1 , is provided.

[0005] This method may comprise one or several additional features, defined in claims 1 to 13, considered alone or in combination.

[0006] The instant technology also concerns a method for operating an electric arc furnace according to claim 14, an electronic monitoring device according to claim 15, an electric arc furnace installation according to claim 16. The instant technology also concerns a computer program whose execution on a computer (possibly connected to relevant sensors, actuators or controller), makes the computer to execute the method for monitoring presented above. The instant technology also concerns a non-transitory computer-readable medium storing such a computer program.Detailed description

[0007] The instant technology will now be described in more detail and illustrated by examples without introducing limitations, with reference to the appended figures.

[0008] Figure 1 is a schematically represents an EAF installation.

[0009] Figure 2 represents, in the form of a block-diagram, operations carried out in an embodiment of the method for monitoring the EAF of the installation of figure 1 .

[0010] Figure 3 schematically represents an exemplary recording of an arc stability signal, against time.

[0011] Figure 4 schematically represents a membership function used in a fuzzy logic module of the block diagram of figure 2, to fuzzify one of the inputs.

[0012] Figure 5 schematically represents an output of the fuzzy logic module, as a function of inputs of this module.

[0013] Figure 6 schematically represents a histogram of a recharging recommendation time determined in the course of the monitoring method, as well as target statistical properties for recharging recommendation time.

[0014] As above mentioned, the instant technology concerns, inter-alia, a method for determining if an appropriate time for charging a second load of material has come, after a first load of material to me melted had been charged in an EAF. This method is based in particular on detecting that the melt of the first load of material is completed. This determination is based at least on:- an arc stability signal, representative of the stability or instability of an electric arc present in the electric arc furnace, and on- one or more thermal losses signals, relative to thermal losses of the electric arc furnace.

[0015] An EAF installation, suitable for implementing that method, is presented first in reference to figure 1. The method itself is presented then in reference to figures 2 to 6. Exemplary test results are presented then.EAF installation

[0016] Figure 1 schematically represents an EAF installation 1 that comprises an EAF 2, an electric supply (not represented in figure 1 ) for supplying the electrodes of the EAF with electricity, and an electronic system 3 for monitoring and controlling the EAF operation. As represented, the electronic system 3 comprises:An electronic monitoring unit 10,A human-machine interface 5, which comprises for instance a display screen and / or visual indicators, and possibly input devices like a keyboard, buttons and / or selectors.- An industrial production database 6,An operation controller 7.

[0017] The electronic monitoring unit 10 comprises at least a processor and a memory. It comprises also one or more data interfaces (like network or communication cards or chips) for receiving and emitting data and / or signals, in particular for: receiving Furnace Monitoring Signals FMS, received from one or more sensors of the EAF installation 1 such as a temperature, a flow, a weight, an electric current, an electric power or a voltage sensor; the Furnace Monitoring Signals FMR may also comprise one or morecontrol signals transmitting setpoints to low-level automates of the EAF, for instance a total electric power setpoint, transmitted to the electric supply;- sending a recharging readiness signal RR (relative to a scrap melting status) to one or more, here to all of: the human-machine interface 5, the industrial production database 6, the operation controller 7.

[0018] Here, the electronic monitoring unit 10 is an industrial computer like a Programmable Logic Controller or like a module of a Distributed Control System. It is configured, here programmed for executing the method for monitoring the operation of the EAF described further below. In particular, it is configured for determining the recharging readiness signal RR, which specifies whether the electric arc furnace is in condition for charging the second load of material to be melted, or not, after the first load of material had been charged in the EAF.

[0019] The operation controller 7 is an electronic device comprising control logic (in the form of one or more programmable circuits, for instance) and possibly control capabilities like control loops and / or drivers for controlling or regulating one or more actuators of the EAF. The operation controller 7 is configured in particular for controlling one or more loading devices of the EAF (like an electrically actuated feed hatch, an overhead crane bucket or scoop, or a scrap claw or electromagnet) for charging loads of material to me melted in the EAF. The operation controller 7 mat be configured also for controlling the electric supply of the EAF depending on setpoints, commands and / or monitoring signals received by the operation controller 7.

[0020] Furnace Monitoring Signals FMS received by the electronic system 3, and possibly the recharging readiness signal RR or other signals exchanged within the electronic system 3 or emitted by it may be transmitted using wire or wireless communication and / or through a local network or bus, for instance of the CAN (Controller Area Network), CAN+ or fieldbus type. One or more of these signals may also be transmitted through a public network, like the internet.

[0021] During the operation of the EAF, the first load of material, to be melted, is introduced in the EAF, for instance through a dedicated opening which is then closed. In practice, some liquid steel (called the ‘hot heel’) often remains, from previous heats, in the bottom of the EAF. The electrodes of the EAF are supplied with an electric current (causing the formation of an electric arc within the EAF), which allows for heating and melting the material introduced. The material to be melted may be some steel scraps, or some Direct Reduced Iron (for instance in the form of briquettes of pellets). The first load of material may correspond to a bucket or scoop of said material, for instance, or to the content of a loading chamber of the EAF (a loading chamber for material pre-heating, for instance). After the first load has been charged and melted, preferably entirely, the second load of material is charged, and melted similarly. One or more additional loads of materials may further be charged afterwards, and melted. Then, all or part of the liquid steel is tapped.

[0022] It is noted that the two loads of material respectively called the “first load” and the “second load” are designated as such with reference one to another, as they are charged successively in the EAF, one after the other. But it does not mean that the first load would necessarily be an initial load charged in the EAF at the beginning of a heat (that is just after a preceding tapping operation). Indeed, the heat may comprise charging and melting more than two successive loads of materials before tapping the liquid steel, for instance charging three or even four successive loads of material. And the instant method for determining if the EAF is in condition for charging a next load of material can be applied for any (and possibly for all) of these loads.

[0023] The electronic system 3 is configured so that, during the operation of the EAF, the recharging readiness signal RR, or a visual indication representation thereof is displayed by the human-machine interface to notify an operator of the EAF that the EAF is in condition for charging the second load of material, as soon as it is in such a condition. The second load is charged once the operator has triggered (manually) the charging, for instance using the human-machine interface 5. Still, in other embodiments, the electronic monitoring unit 10 and / or the operation controller 7 may be configured so that, when the recharging readiness signal RR starts indicating that the EAF is in condition for charging the second load of material, then, the second load is charged without requesting an action from the operator (i.e.: is charged automatically), except if the operator inhibits charging the second load by entering an inhibit command.

[0024] It is noted that the functionalities of the electronic system 3 could be distributed or organised differently than in figure 1. In particular, the electronic monitoring unit 10, one or more elements of the operation controller 7 and possibly the human-machine interface 5 could be implemented together, in the form of a single, standalone computer device (instead of different computer devices operatively connected to each other). Conversely, the determination of the recharging readiness signal RR could be determined using remote (and possibly distributed) computing resources such a cloud computing resources, instead of using a dedicated electronic device.Method for monitoring the operation of the EAF

[0025] Some general aspects of the method are presented first, before describing in more details the exemplary embodiment corresponding to figure 2.

[0026] In this method, the fact that the electric arc furnace is in condition for charging a second load of material to be melted, or that it is not in condition for that (in particular, the fact the melt of the preceding, first bucket of scraps is completed, or not) is determined based at least on the arc stability signal Stb and on the one or more thermal losses signals. It may, like here, be determined taking also into account an electrical energy Eeiecconsumed by the EAF betweenan initial instant ti, at which the first bucket of scraps is charged in the EAF, and a current instant t.

[0027] When a load of material is charged in the EAF, the electric arc is first rather unstable; it fluctuates (in terms of amplitude, phase, or harmonic content) and has significant harmonic distortion and / or parasitic components. The electric arc then becomes more stable once the load is melted, and its variations are then usually closer to sinusoidal ones. Having a stable electric arc and / or one with a low harmonic distortion is thus a good indication that the melt of the previous load has been completed.

[0028] In addition to this indication, thermal losses of the furnace are taken into account to determine the recharging readiness signal RR. Indeed, once the melt of the charged material is complete, the energy supplied to the EAF is no more used for melting the material, and the heat released by the EAF thus increases (its thermal losses increase). The thermal losses TL thus provide usefully complementary information, and allow for a more reliable estimation of the readiness of the furnace for charging a second load than when taking only into account the arc stability / harmonicity.

[0029] In practice, the signals relative to the thermal losses of the EAF, and the arc stability signal Stb may be noisy ones. And for monitoring the melt of the first load of material, a high degree of confidence is desirable. Indeed, charging the second load too early may cause a liquid steel overflow. Taking also into account the electrical energy Eeiec, to determine if the electric arc furnace is in condition for charging a second load of material, increases the reliability of this determination (without having to introduce additional safety delays). Eeiecis for instance taken into account by comparing it with an energy expected for melting the first load (for instance an energy that is estimated to the minimum energy required for melting the first load), given the weight of this first load, to have a complementary check that it is likely that the first melt is completed.

[0030] The recharging readiness signal RR, or an intermediary end-of-melt signal endM employed for further computing the recharging readiness signal RR, may be determined based on two or more of the above-mentioned inputs (Stb, TL, Eeiec), using a fuzzy logic module. Using this fuzzy logic module allows for basing this determination on conventional logic rules, in particular predetermined rules such as operation rules derived by an expert or derived from experts’ knowledge. Indeed, such predetermined rules can be conveniently entered by an expert (or otherwise loaded) and then translated into fuzzy logic rules (based on the systematic translation rules of the fuzzy logic configuration selected for the fuzzy module). A determination close to an expert’s knowledge and recommendation, while thinner and more gradual than if based directly on conventional logic rules, is thus obtained. An operation rule such as the above-mentioned ones may be for instance that: “the first load of scraps is considered melted if a duration of arc stability TGS is high, whatever the thermal losses”. Or it may be that “thefirst load of scraps is considered completely melted if the duration of arc stability is normal and the thermal losses are high”, for example.

[0031] The method may comprise more particularly:- determining a duration of arc stability TGS, spanning from an instant tsfrom which the electric arc is stable, to a current instant t (see figure 3), determining the thermal losses TL of the EAF, here for a duration spanning from the initial instant ti to the current instant t, and computing the intermediary end-of-melt signal endM using the fuzzy logic module 16, based at least on: the duration of arc stability TGS, the thermal losses TL and also (optional input) a position or a speed of displacement of the electrodes of EAF.

[0032] Figure 2 represents, in the form of a block-diagram, operations carried out for determining the recharging readiness signal RR, for the exemplary embodiment considered here.

[0033] In this embodiment, the arc stability signal Stb is a Total Harmonic Distortion index of a signal picked up using a Rogowski coil (arranged to measure an electric current flowing through one or several of the electrodes of the EAF). The higher this signal, the less stable the electric arc. The electric arc is considered stable when the stability signal Stb passes a given threshold Ths. An exemplary recording of the arc stability signal Stb is represented in figure 3, Stb being represented in arbitrary units, against time (in arbitrary units too). In figure 3, the first load of material is charged at the initial instant ti. After the instant ts, the electric arc is considered stable. The current instant is time noted t. The second load of material is loaded at time t2.

[0034] In block 1 1 of the block-diagram of figure 2, the arc stability signal Stb is thresholded (as above explained) and time-integrated to compute the duration of arc stability TGS (TGS being equal to t-ts). In a subsequent block 12, an adjusted duration of arc stability TGS_a is computed from TGS, to implement an auto-adjusting optional feature (which is presented further below).

[0035] In this embodiment, the one or more thermal losses signals comprise:- an input temperature Tinof a cooling water, at an inlet of a cooling system employed for cooling a roof of the EAF,- an output temperature Toutof the cooling water, at an outlet of that cooling system, and- a flow rate CWF of the cooling water circulating in that cooling system.

[0036] A cooling power is computer from these signals (by multiplying CWF with the difference Tout " Tin ) and time integrated in block 13 to compute the thermal losses TL. This timeintegration spans from the initial instant ti to time t. In a subsequent block 14, an adjusted thermal loss TL_a is computed from TL to implement the auto-adjusting optional feature.

[0037] The adjusted duration of arc stability TGS_a, the adjusted thermal loss TL_a, and data representative of a vertical position pos of the electrodes of the EAF are input into the fuzzy logic module 16. The fuzzy logic module 16 outputs the end-of-melt signal endM, which is all the higher than it is likely that the first load of material is completely melted.

[0038] The fuzzy logic module 16 may be configured, as explained below, based on the so- called Mamdani system (or possibly the Mamdani-Sugeno system, for instance). In this case, the operations executed comprise: fuzzifying the inputs (here TGS_a, TL_a and pos), using fuzzy membership functions,- computing the results of the fuzzy rules planned for analysing the inputs (rules that are translations of conventional logic rules, into the fuzzy logic system considered), combining and de-fuzzifying the results of these rules.

[0039] In this embodiment, for each input, the membership functions are a set of three trapezoidal functions respectively associated to “low”, “normal” and “high” cases for the input considered. Figure 4 schematically represents the membership functions for the adjusted thermal loss input. More precisely, it represents the membership value for the “low”, “normal” and “high” cases (denoted L, N and H in figure 4), as a function of TL_a, expressed in kWh.

[0040] Regarding the rules, in this embodiment, all the possible combinations of cases (“low”, “normal” and “high”) for the three inputs (TGS_a, TL_a, pos) are each associated to a corresponding rule (exhaustive implementation of rules, for all the possible cases that could be encountered). And so, 3x3x3=27 rules are implemented, here. Regarding the translation system, from the conventional to fuzzy logic, the AND operator is coded as the min (minimum) value of the two operands, the OR operator as the max (maximum) value.

[0041] For combing the results of the different rules, the results are aggregated based on the max value of these results, and the defuzzification is achieved by computing the centroid of this aggregate. The output endM has a value from 1 to 1000, here. Figure 5 represents the output endM as a function of TGS_a and TL_a (TGS_a being represented in arbitrary units and TL_a in kWh), for a fixed value of the position pos, for an exemplary parametrization of the rules and fuzzifying membership functions.

[0042] The end-of-melt signal endM is transmitted to the block 17, whose inputs also comprise the electrical energy Eeiecconsumed from the initial time ti to the current instant t, and the weight W of the first load of material. In block 17, it is tested if Eeiecis above an energy Eminexpected for melting the first load, given the weight W of this first load.

[0043] In practice, a min and a max preset values of electrical consumption per unit weight of charged material, Cons_min and Cons_max (in kWh per ton, for instance), are each multiplied by the weight W of the first load to obtain a min and a max value of expected electrical energy consumption Eminand Emax. A preset value for an average electrical consumption, per unit weight of charged material, is also multiplied by the weight W to compute a mean value Emean.

[0044] Then, in block 17, if the end-of-melt signal endM is above a given confidence threshold The and the electrical energy Eeiec(consumed at time t) is from Emin tO Emax, then the recharging readiness signal RR output by block 17 indicates that the EAF is in condition for charging the second load of material; and otherwise, the binary signal RR indicates that the EAF is not in condition for charging the second load.

[0045] Besides, in block 17, if the end-of-melt signal endM is below the confidence threshold The and the electrical energy Eeiecis above Emean, then a pre-alarm signal pRR, output by block 17, indicates that the EAF is almost in condition for charging the second load of material. And otherwise, the binary signal pRR indicates that the EAF is not yet in condition for charging the second load.

[0046] The min, max and average preset values of electrical consumption per unit weight of charged material may be values entered by the operator. They may also be derived automatically from a historic of past values of electrical energy consumed for melting loads of material (using an automated statistical processing of such past values).

[0047] The confidence threshold Thcmay have a value that is from 60% to 90% of the maximum value the end-of-melt signal could reach (for instance 80% of this value, which corresponds here to 800).

[0048] The parametrization of the fuzzy rules and fuzzifying membership functions may be achieved by first pre-setting manually the corresponding parameters based on an expert or operator’s knowledge, or based on observations of a history of the EAF operation, and then fine-tunning these parameters.

[0049] Such a fine-tunning, and more generally the parametrization of the fuzzy rules and fuzzifying membership functions, may be based on reference values relative to the operation of the EAF. The reference values are for instance values that are observed when the EAF is controlled by an operator who decides directly at what time the second load of material is to be charged (i.e.: who decides it without using the instant method), based on his expert knowledge and experience, and based on monitoring signals (such as the arc stability signal, or the thermal losses signals). These reference values may in particular comprise:- reference values Ats,ref for the lag time Ats, which is the total duration, before charging the second load, during which the electric arc is stable, that is Ats=t2-ts, and / or- reference values TLi,ref for the thermal losses TLi, which are the thermal losses at the time the second load is charged (i.e.: cumulated over time until ta).

[0050] Here, the several reference values for Ats,ref, and TLi,ref are acquired for several successive heats, so that statically representative values (average and standard deviation) for these quantities can be deduced thereof. Still, in alternative embodiments, the reference values in question could be directly a recommended average values for the lag time, and arecommended average value for the total thermal losses, instead of being a set of historical values acquired as above explained.

[0051] The numerical parametrization of the fuzzy logic module may comprise:- determining target statistical properties (such as a target average and a target standard deviation) for:- a recharging recommendation time AtreCo, which the duration between tsand the instant IRR at which the recharging readiness signal RR starts indicating that the EAF is in condition for charging the second load of material,- and also, here, for the thermal loss at the recharging recommendation time IRR, noted TLl reco,- parametrizing the fuzzy rules and / or fuzzifying membership functions (that is, adjusting numerical parameters thereof) so that the statistical properties obtained for the recharging recommendation time AtreCo and for TLireCo, as determined by the electronic monitoring unit 10, match the target statistical properties.

[0052] Here, the target average, and also the target standard deviation are each set to be equal, or possibly slightly smaller than the average and standard deviations of the reference values Ats,ref, and TLi,ref. For instance, the target average pn for the recharging recommendation time Atreco is set as equal to x times the average of the reference values Ats,ref for the lag time Ats(observed when the EAF is controlled directly by the operator), with x from 0.5 to 1 , or from 0.8 to 1 . It means that the parametrization is then achieved so that, in average, the recharging recommendation time is equal to or smaller than the lag time observed when the EAF is operated directly by an operator. In other words, the settings for the fuzzy logic are such that the electronic monitoring unit 10 either mimics the recommendations of the operator or recommends recharging the EAF slightly faster than what the operator would do (while respecting safety precautions).

[0053] In the embodiment of figure 2, the auto-adjusting feature is implemented. It aims at keeping on obtaining desired performances in spite of possible disturbances, or possible evolutions of the material properties (for instance, evolution of the steel scraps type, which can be more or less easy to melt).

[0054] The auto-adjusting comprises determining the adjusted duration of arc stability TGS_a by applying an adjustment operation to TGS. The adjustment operation is based on:- the target statistical properties for the recharging recommendation time AtreCo, namely the target average .tand a target standard deviation ct, and- the statistical properties actually observed, for the recharging recommendation time AtreCo during some (immediately) preceding heats.

[0055] The adjustment operation is such that, if applied to AtreCo, the statistical properties of Atreco (once corrected) would then match (i.e.: would then be equal to) the target statistical properties .t, ot. Still, it is noted that the adjustment operation is applied to the input signal TGS of the fuzzy logic module, not to AtreCo or to another output thereof.

[0056] Figure 6 represents an histogram Ht considered appropriate for values of the recharging recommendation time AtreCo (AtreCo being in arbitrary units). The histogram Ht is for instance the histogram of the reference values Ats,ref of the lag time, each multiplied by x (with x equal to or smaller than 1 ). The average value for the histogram Ht is the target average .t, and its standard deviation is ct. Figure 6 also represents an example of a histogram HAtreCo for the recharging recommendation time AtreCo, observed during successive heats. The average and standard deviation for this set of observed values of AtreCo are noted respectively |_iObs and C^obs.

[0057] The adjustment operation can be achieved for instance by computing TGS_a according to eqn 1 below:T GS_a = a. TGS + b eqn 1 with a=ct / cObs and b=p.t- Hobs (which, if applied to AtreCo, would indeed lead to a distribution having the target statistical properties).

[0058] In the embodiment described here, the auto-adjusting for the thermal losses TL is achieved similarly as the one for the duration of arc stability TGS, except that the linear adjustment operation applied to TL is based on target, compared to observed statistical properties for the thermal losses at the recharging recommendation time, TLireCo, instead of being based on statistical properties for the recharging recommendation time AtreCo.

[0059] The auto-adjusting may be executed automatically and regularly, for instance every day, or every time a given number of new heats have been executed.

[0060] The auto-adjusting implemented in this way do not require complex computations and can thus be implemented conveniently in the electronic monitoring unit even when this unit takes the form of a programmable controller, or of a component of a Distributed Control System. Besides, it does not affect the initial parametrization of the fuzzy logic module, as being executed upstream of it. It has been observed that auto-adjusting the electronic monitoring unit in this way, by correcting the inputs TGS and TL, allows for maintaining good performances over time, in spite of the variations of the scraps meltability.

[0061] The instant method for monitoring the operation of the EAF can be implemented differently than described above with reference to figure 2, and different modifications can be made to the embodiment of figure 2.

[0062] For instance, the arc stability signal Stb may be of a different kind than presented above. It may be obtained, for example, by computing a quantity different from the TotalHarmonic Distortion index, representative of an amplitude of fluctuations of the signal picked up using the Rogowski coil. The arc stability signal Stb could also be derived from another quantity, for instance from a magnetic field in a surrounding of the electrodes, or in a surrounding of the EAF. The arc stability signal Stb could also be a signal, for instance a binary signal, specifying that the arc is either stable or not, provided directly by an external, independent module (for instance a commercially available module outputting such a signal).

[0063] Besides, in the example described above, the thermal losses TL are thermal losses at the roof of the EAF. Yet, in other embodiments, the thermal losses, taken into account to determine the recharging readiness signal RR, could be thermal losses for another part of the EAF, for instance thermal losses through side walls of the EAF. They could also be estimated at the level of heat shields surrounding the EAF.

[0064] Regarding the detailed way to determine the recharging readiness signal RR, from the arc stability and thermal losses signals, several alternatives to the embodiment of figure 2 are possible. For instance, the electrical consumption signal could be taken into account at the level of the fuzzy logic module, instead of being taken into account afterwards, in block 17. Besides, other signals could be input to the fuzzy logic module, in addition to or as an alternative to the electrical consumption signal. Besides, instead of a fuzzy logic module, a classical logic module could be employed.Exemplary results

[0065] The monitoring method, in its embodiment corresponding to figure 2, has been implemented, in several steel making plants equipped with different models of EAF, and tested therein in actual production conditions. In average, the recommendation for recharging the EAF, given by the recharging readiness signal RR, is respected by the operators for more than 50% of the heats, and even, in many plants, for more than 2 / 3 of the heats; its means that for more than two thirds of the heats, as soon as the recharging readiness signal RR indicates that the EAF is in condition for recharging, the operator triggers the charging of the second load of material (within a few seconds). It illustrates that the recommendations computed according to this method are actually followed by the operators.

[0066] Besides, these tests have shown that, when this method is implemented (that is when the signal RR is available, for the operators), the total electric consumption of the EAF, per ton of material to be melted, is reduced by 1% in average, compared to a triggering actuated by the operators without the help of the recommendation given by the signal RR.

[0067] No overflow or EAF malfunction was observed during these tests, when the rechargings were triggered as recommended by the signal RR.

[0068] As an example, during a test in one of these plants:- for 72% of the heats, the recommendation for recharging the EAF given by the signal RR was respected by the operator,- for 22% of the heats, the operator triggered the charging of the second load before the signal RR indicates that the EAF is in condition for this recharging, and- for 6% of the heats, the operator triggered the charging of the second load after (more than 1 min later than) the signal RR indicates that the EAF is in condition for this recharging.

[0069] For this test, a reduction of about 4% was observed for the electric energy, per unit weight (in kWh / ton of steel, for instance), consumed for melting the first load (consumed between ti and ts), compared to a to a triggering actuated by the operators without the help of the recommendation given by the signal RR. And a reduction of 1% was observed for the total electric energy (in kWh, for instance) consumed by the EAF, for the overall melting process.

[0070] The efficiency of the auto-adjusting feature was also tested during these trials. In practice, when this feature is not implemented, the proportion of heats for which the recommendation given by the signal RR is respected may gradually decrease and fall to less than 50% (due to a variation over time of the scraps type, for instance). Executing the abovedescribed auto-adjustment then enables to restore a proportion above 50%, and even above 2 / 3. For example, during one of these tests, the proportion of respected recommendations had fallen to 44% (while, for 49% of the heats, the operator triggered the recharging before the signal RR had recommended it). Applying the auto-adjustment procedure enabled then to restore a satisfying determination of the signal RR, with a proportion of respected recommendations of 72%.

Claims

CLAIMS1 . A method for monitoring the operation of an electric arc furnace (2), operation during which a first load of material to be melted is charged in the electric arc furnace (2), the method comprising: acquiring: o an arc stability signal (Stb), representative of the stability or instability of an electric arc present in the electric arc furnace, o one or more thermal losses signals (Tin, Tout, CWF), relative to thermal losses (TL) of the electric arc furnace, determining a recharging readiness signal (RR), which specifies whether the electric arc furnace is in condition for charging a second load of material to be melted, or not, the recharging readiness signal (RR) being determined based on the arc stability signal (Stb) and on the one or more thermal losses signals (Tin, Tout, CWF),- emitting the recharging readiness signal (RR).

2. A method according to claim 1 ,- further comprising acquiring an electrical consumption signal, representative of an electrical power consumed by the electric arc furnace or representative of an electrical energy (Eeiec) consumed by the electric arc furnace between an initial instant (t), at which the first load of material is charged in the electric arc furnace, and a current instant (t),- and wherein the electrical consumption signal is also taken into account to determine the recharging readiness signal (RR).

3. A method according to claim 1 or 2, comprising:- determining a duration of arc stability (TGS), spanning from an instant from which the electric arc is stable (ts) to a current instant (t),- and wherein a fuzzy logic module (16) computes: o the recharging readiness signal (RR), or an intermediary end-of-melt signal (endM) employed for further computing the recharging readiness signal (RR), o based at least on the duration of arc stability (TGS) and the thermal losses (TL) of the electric arc furnace.

4. A method according to claims 2 and 3, wherein the recharging readiness signal (RR) specifies that the electric arc furnace is in condition for charging the second load of materialprovided that a given condition is fulfilled by the end-of-melt signal (endM) and by the electrical energy (Eeiec) consumed by the electric arc furnace between said initial instant and the current instant.

5. A method according to claim 4, wherein said condition is that: the end-of-melt signal (endM) passes a given confidence threshold and said electrical energy (Eeiec) is above a consumption threshold.

6. A method according to claim 5, wherein the consumption threshold is determined by multiplying a weight (W) of the first load of material by a preset value (Cons_min) of electrical consumption per unit weight of charged material.

7. A method according to claim 5 or 6, wherein the recharging readiness signal (RR) specifies that the electric arc furnace is in condition for charging the second load of material on the further condition that said electrical energy (Eeiec) is below a maximum consumption threshold.

8. A method according to anyone of claims 3 to 7, wherein an adjustment operation is applied to the duration of arc stability (TGS) before transmitting it to the fuzzy logic module, the adjustment operation being parametrized depending on:- one or more statistical properties ( .Obs, nObs) of a set of values of a recharging recommendation time (AtreCo), acquired during preceding heats, the recharging recommendation time (AtreCo) being the duration between the instant (ts) from which the electric arc is stable and an instant from which the recharging readiness signal (RR) indicates that the electric arc furnace is in condition for charging the second load of material, and on- one or more target statistical properties ( .t, ot).

9. A method according to claim 8, wherein the adjustment operation is such that: if applied to the set of values of the recharging recommendation time (AtreCo), said set of values, once adjusted, would have one or more statistical properties matching the one or more target statistical properties ( .t, ot).

10. A method according to claim 8 or 9, wherein the target statistical properties ( .t, ct) are determined from references values of a lag time, the lag time being a duration between the initial instant (ti) and an instant (t2) at which the second load is charged in the electric arcfurnace, the reference values being acquired when an operator triggers charging the second load of material with no information regarding the recharging readiness signal (RR).1 1. A method according to anyone of claims 8 to 10, wherein the parametrization of the adjustment operation, achieved according to 8, is repeated automatically, for instance at regular time intervals.

12. A method according to anyone of claims 3 to 11 , wherein an additional adjustment operation is applied to the thermal losses (TL) before transmitting this signal to the fuzzy logic module (16), the additional adjustment operation being parametrized depending on:- one or more statistical properties of a set of values of the thermal losses at a recharging recommendation time, TLireCo, acquired during preceding heats, the thermal losses at the recharging recommendation time being the thermal losses (TL) at the instant (IRR) from which the recharging readiness signal (RR) indicates that the electric arc furnace is in condition for charging the second load of material, and on- one or more corresponding target statistical properties.

13. A method according to anyone of claims 1 to 12, wherein the recharging readiness signal (RR) is transmitted: to a human-machine interface (5) which, when the recharging readiness signal (RR) indicates that the electric arc furnace is in condition for charging the second load of material, displays corresponding visual information, and / or- to an operation controller (7), configured for controlling a loading device so that the loading device charges the second load of material in the electric arc furnace (2) when the recharging readiness signal (RR) indicates that the electric arc furnace is in condition for charging the second load of material.

14. A method for operating an electric arc furnace (2), the method comprising:- Charging a first load of material to be melted in electric arc furnace,- Monitoring the operation of the electric arc furnace according to the method of anyone of the preceding claims,- Charging a second load of material to be melted in electric arc furnace.

15. Electronic monitoring device (10; 3) comprising at least a processor and a memory, configured for executing the method according to anyone of claims 1 to 13.

16. Electric arc furnace installation (1 ) comprising an electric arc furnace (2), sensors or signal acquisition devices for acquiring one or more thermal losses signal (Tin, Tout, CWF) and an arc stability signal (Stb), and the electronic monitoring device (10; 3) of the preceding claim, operatively connected to said sensors or signal acquisition devices.

17. Computer program comprising instructions whose execution by a computer (10) makes the computer to execute the method according to anyone of claims 1 to 13.

Images