CN102527971B - Online forecasting method for internal crack defect of casting blank - Google Patents

Abstract

An online forecasting method for an internal crack defect of a casting blank belongs to the field of metal casting and comprises a network composed of a computer of L3, a computer of L2 and a computer of L1 and data transmission among the computers, wherein on the basis of the existing computer of L2 or the same control level, a model computer is arranged to obtain the internal stress strain information of the casting blank by real-time online analog computation of cooling and solidifying process of the casting blank, and then forecasts the internal crack defect of the casting blank in real time according to the variation tendency of the strain; then the quality information of the casting blank in the production process is timely transmitted to the cutting computer of L1 which is used for optimizing and controlling the cutting process of the casting blank with the defect; the quality control accuracy of a product and the product percent of pass are increased, the percent of pass and the commercial grade of the product are increased, and the whole economic benefit of an enterprise is further increased. Therefore, the online forecasting method for the internal crack defect of the casting blank can be widely applied to optimizing/control field of the cutting process of the casting blank during the production of slab continuous casting.

Description

A kind of online forecasting method of casting blank crack defect

Technical field

The invention belongs to field of metal casting technology, relate in particular to a kind of online forecasting/control method for continuous casting production process Inner Quality of Billet defect.

Background technology

In casting process, strand can be subject to the effect of various external force and deform, and when slab process bending section, can be subject to the effect of bending stress; When slab process is rectified value section, can be subject to rectifying the effect of value stress; Under the effect of ferrostatic pressure, strand there will be bulge deformation.Under the effect of these external force, continuous casting steel billet can produce certain strain, if accumulation strain has surpassed critical strain, just there will be underbead crack (abbreviation implosion).

Implosion is common slab quality defect, once form, very large on the combination property impact of postorder processing and product.Serious casting blank crack, may cause being with even broken belt of steel layering in course of hot rolling.

Generally, once strand inherent vice forms just, exist all the time, be difficult to eliminate in postorder process.

Therefore, the control of Inner Quality of Billet can only be carried out in continuous casting production process, by improving technique and operant level, constantly reduces the incidence of strand inherent vice.

Yet the origin cause of formation of continuous casting defect is very complicated, relate to factor very many, these factor weave ins, more make defect cause be difficult to accurately define sometimes.

When implosion defect forms, generally by the optimization cutting of strand, reduce the impact of defect on following process and properties of product.For instance, if there is inherent vice at the regional area of an end of strand, can cut defect area is excised by optimization, thereby guarantee the total quality of residue slab.

Yet when implosion appears in strand, the base shell that rejected region has been solidified surrounds, and cannot detect online.

Conventional method of operating is that, after strand cuts completely, head or afterbody sampling at strand, then check by macroscopic test whether strand has implosion.Because the macroscopic test cycle is generally about 2 days, this logistics that can have a strong impact between continuous casting and hot rolling is connected, and, also can not be to all inspections by sampling of all strands in reality.

For these reasons, people start to explore and how casting blank defect are forecast.

Defect forecast has two effects: when known casting blank defect occurs, forecast result provides information to execute-in-place slip-stick artist or technologist, possible in the situation that, adjust in time production run and controls parameter, shortens the duration of casting blank defect as far as possible; According to the forecast information of casting blank defect, the cutting process of defect strand is controlled to optimization, improve product percent of pass.

Forecasting procedure about casting blank defect, the day for announcing is on August 16th, 2006, notification number is in the Chinese patent of CN1269595C, " a kind of forecasting procedure due to the cooling casting billet surface lobe extremely causing of crystallizer " disclosed, it imbeds crosswise sequence of number, longitudinal at least three heat extraction galvanic couples below crystallizer liquid steel level position, by data acquisition system (DAS), these temperature are read in, go forward side by side line number according to one's analysis.The step of its data analysis at least comprises: under the stable condition of pulling rate, declining appears suddenly in certain electric thermo-couple temperature in certain row, 3 ℃ of rates/more than s; Under this thermopair successively also there is the downtrending of 3 ℃ of rates/more than s in two of same column electric thermo-couple temperatures, and adjacent two electric thermo-couple temperatures start the spacing that mistiming of declining and the product of instant pulling rate just in time equal these two thermopairs; Rule is consistent over time for three electric thermo-couple temperatures of these row, and the time of lower row's electric thermo-couple temperature continuous decrease is not less than the time of row's electric thermo-couple temperature continuous decrease.Visible, this technical scheme is by the thermopair of some, the temperature fluctuation in Real-Time Monitoring crystallizer being installed in certain position of crystallizer.When temperature fluctuation surpasses certain limit, and while meeting certain condition, can judge strand lobe defect occurs.

Said method is only applicable to the surface defect of bloom causing because of mould temperature unusual fluctuations.

According to correlative study, show, the main cause that causes implosion defect is: on the solid-liquid interface of the strand inside of not solidifying completely, owing to being subject to deformation that the effect of external force occurs, surpassed due to critical strain values, wherein strain comprises bending strain, aligning strain, bulgs stress etc.Yet although strain and casting blank crack defect is in close relations, strain itself is also difficult to detect.

Summary of the invention

Technical matters to be solved by this invention is to provide a kind of online forecasting method of casting blank crack defect, it is by cooled and solidified process real-time, that online simulation calculates strand, obtain the stressed strain information of strand inside, then according to the variation tendency of strain, the implosion defect of strand is forecast.And the slab quality information in production run is delivered to cutting L1 computing machine in time, it is for the cutting process of optimal control defect strand, the control accuracy of can improving the quality of products and product percent of pass.

Technical scheme of the present invention is: a kind of online forecasting method that casting blank crack defect is provided, comprise by L3 level computing machine, the network that L2 level computing machine and L1 level computing machine form and data transmission each other, wherein, L3 level computing machine is responsible for assigning production schedule instruction, L2 level computing machine is responsible for determining the various control parameters in production run according to the production schedule, and control parameter is issued to the execution of L1 level computing machine, L1 level computing machine is carried out steering order that L2 level computing machine issues or operative employee's input, directly or indirectly control the relevant device of casting machine, described L1 level computing machine at least comprises public L1 computing machine, casting L1 computing machine and cutting L1 computing machine, it is characterized in that described online forecasting method at least comprises the following steps:

A, in existing L2 level process computer or on L2 level process computer, a normatron is set;

B, described L2 level process computer pass through L1 level computing machine, and the various technique in casting process and control parameter collection is complete, then according to certain gap periods, send to normatron;

C, described normatron receive in real time, online the technique in strand production run and control parameter, determine the boundary condition that strand heat radiation is calculated;

D, the described normatron mathematical model based on strand diabatic process is described, and dynamic calculation strand and extraneous heat radiation process obtain the temperature field of strand inside and outside;

The cooled and solidified process of E, described normatron dynamic calculation strand, obtains the concreting thickness information of each slice position of strand;

F, described normatron performance analysis strand stressed variation in moving process, calculate strand due to the bulgs stress, aligning strain and the dislocation strain that are subject to External Force Acting and produce, and the overall strain that obtains strand distributes;

G, described normatron be according to the variation tendency of strain, and by judging whether strain surpasses critical strain values, and then whether prediction have implosion to occur, and the implosion defect of strand is carried out to real-time estimate;

If implosion appears in H judgement strand, normatron calculates concrete defect information, and information is associated with slab more specific location information, by L2 level process computer, is issued to cutting L1 computing machine, and the cutting process of strand is optimized to control;

I, cutting L1 computing machine are adjusted the cutting position of strand, for centre burst, appear near the slab predetermined slab head or tail position, after cutting, directly defect base are excised by optimization; For the defect that has occurred predetermined slab middle part, according to precalculated position, cut, but the slab after cutting is enclosed to flaw labeling, demote as requested or change and do its use;

J, above-mentioned steps, in to the casting process of strand in real time, carry out online.

Concrete, described normatron is PC, industrial computer, single-chip microcomputer or the virtual machine that is arranged in L2 level computing machine.

In described casting process, various technique and control parameter, be the relevant procedure parameter of strand diabatic process, and it at least comprises steel grade, molten steel temperature in tundish, thickness, pulling rate, width and cooling water flow; Described normatron, according to these data, is determined the boundary condition that strand Heat Transfer Meter is calculated, and determines the total amount of heat that in the unit interval, strand transmits to the external world; Along with the movement of strand physical location, normatron fixed cycle upgrades initial value and the boundary condition of strand Calculation of Heat Transfer.

Described normatron at each constantly, according to cooling water inflow, tolerance, first calculate the coefficient of heat transfer of casting billet surface, extrapolate based on this again the heat that in the unit interval, casting billet surface sheds to the external world, and then according to steel grade physical parameter, obtain the temperature field of strand inside and outside.

Further, in cooling procedure, the heat that described strand sheds to the external world adopts following expression to calculate:

φ=h(U _s-U _w)(w/m ²)

In formula, φ is the intensity of outwards dispelling the heat in unit area, U _sthe surface temperature of strand, U _wbe cooling water temperature, h is the coefficient of heat transfer of casting billet surface.

Further, the coefficient of heat transfer of described casting billet surface calculates by following expression:

h=kw ^rwa ^ra

Wherein, w is jet density, and rw is water yield coefficient; A is tolerance density, and ra is tolerance coefficient, and k is constant.

Further, the computing method of described jet density are, take cooling zone as unit, calculate the water spray total amount of certain cooling zone on strand upper surface, and divided by the area of cooling zone, what obtain is jet density; Described in it, the computing method of tolerance density are identical with the computing method of jet density.

Further, first described normatron calculates the solidification rate of each position on strand, then calculates concreting thickness according to solidification rate;

Described in it, the calculation expression of solidification rate is:

$\mathrm{fs}=\frac{T_l-T_c}{T_l-T_s}$

Wherein, the solidification rate that fs is strand, T _lfor the liquidus temperature of steel, T _sfor the solidus temperature of steel, T _cfor the temperature on slab center line;

Described normatron uses above-mentioned calculation expression, calculates the solidification rate of each position on the xsect of strand, then calculates the concreting thickness of each position.

Normatron described in it distributes and concreting thickness information based on calculating the strand temperature field obtaining, calculate respectively bulgs stress, aligning strain and dislocation strain, these Strain superimpositions are got up, take strand section as unit, calculate the overall strain size on each strand slice position.

Further, described bulgs stress adopts following expression formula to calculate:

${\mathrm{\ϵ}}_b\left(i\right)=\frac{1600s_i{\mathrm{\δ}}_i}{{l_i}^2}\×100\%$

In formula:

ε _b(i): the bulgs stress at casting blank solidification interface, i roller place, s _i: i the casting blank solidification thickness that casting roll position is corresponding, l _i: i roller spacing, δ _i: bulge deformation amount.

Further, the computing formula of the bulge deformation of described strand is:

${\mathrm{\δ}}_i=\frac{\mathrm{\η\αp}{l}^4\sqrt{\frac{l_i}{v_g}}}{32E{s}_i^3}$

For slab, η α=1; P: the ferrostatic pressure that casting roll is born; v _g: casting speed; E: elasticity coefficient.

Described in it, the computing formula of elasticity coefficient E is:

$E=\frac{T_s-T_m}{T_S-100}\&times;{10}^5\mathrm{kg}/{\mathrm{<figure-callout\; id="2"\; label="cm"\; filenames="HDA0000138735920000023.png"\; state="\{\{state\}\}">cm</figure-callout>}}^2$

Wherein, T _sfor the solidus temperature of steel, T _sfor temperature of solidification, T _mfor medial temperature;

Medial temperature T described in it _mcomputing formula be:

$T_m=\frac{T_s+{\mathrm{<figure-callout\; id="2"\; label="T\; f"\; filenames="HDA0000138735920000023.png"\; state="\{\{state\}\}">T</figure-callout>}}_f}{2}$

Wherein, T _sfor the solidus temperature of steel, T _ffor surface temperature.

Further, the computing formula of the aligning strain of described strand is:

${\mathrm{\&epsiv;}}_u\left(i\right)=(\frac{D}{2}-s_i)(\frac{1}{R_i-\frac{D}{2}}-\frac{1}{R_{i+1}-\frac{D}{2}})\&times;100\%$

In formula:

S _i: i smoothing roll place strand scull thickness;

D: slab thickness;

R _i: strand outer arc radius before i smoothing roll;

R _i+1: strand outer arc radius after i smoothing roll;

ε _u(i): i the strand aligning strain that smoothing roll is corresponding.

Dislocation strain described in it is the strain on continuous casting nip rolls casting blank solidification interface that arc is forbidden to cause, and its computing formula is:

${\mathrm{\&epsiv;}}_{m\left(i\right)}=\frac{300s_i{\mathrm{\&delta;}}_m}{l_i^2}\&times;100\%$

Wherein, ε _{m (i)}it is the strain that i roller place produces on freezing interface because of roller dislocation; δ _mmagnitude of misalignment for roller place; s _iit is the slab thickness at i roller place.

The overall strain that normatron described in it obtains strand through the following steps distributes:

Using casting roll position as indexing parameter, in all sections, search for, find out the identical section in position, directly the temperature information of section and concreting thickness information are brought into strain formula and calculated, after whole strain indexs are calculated, according to casting roll, calculate total strain, obtain current strand overall strain and distribute.

Or the overall strain that described normatron obtains strand through the following steps distributes:

Using casting roll position as indexing parameter, in all sections, search for, find out two sections of the most close casting roll position, two slice positions adjacent with front and back according to casting roll position, and temperature information and the concreting thickness information of the upper record of former and later two sections, linear interpolation obtains the corresponding strand temperature in casting roll position and concreting thickness information, and be brought into strain formula and calculate, after whole strain indexs are calculated, according to casting roll, calculate total strain, obtain current strand overall strain and distribute.

More specifically, according to steel grade, classify in advance, and be respectively every class steel grade and set critical strain values, and by these Parameter storages in the database of described normatron forecasting model, when normatron calculates, can from database index, go out corresponding critical strain according to cast steel grade, then, normatron is from casting machine outlet, to crystallizer direction, to each casting roll, corresponding strand overall strain judges successively, judge whether overall strain surpasses total critical strain, if strain has surpassed critical strain values, judgement has implosion defect to occur, the occurrence positions information of normatron recording defect.

Described in it, the critical strain span of strand is between 0.5%～0.8%, and the ultimate strain that strand can bear is relevant to steel grade, and its concrete numerical value obtains by engineer testing.

When described normatron judgement strand has implosion to occur, not only to record the particular location that implosion occurs, also to follow the tracks of the lasting time of implosion, finally to determine the region that implosion is covered on strand.

When described normatron completes after calculating, result of calculation information exchange is crossed to network delivery to the process computer of controlling strand cutting, according to features such as the order of severity of defect and area size, strand is optimized to cutting, improve the recovery rate of product.

Whole strand is divided into a series of thin slices that equate with casting blank section area along casting direction, forms described strand section; The Calculation of Heat Transfer of described strand is all carried out in strand section, fixed time period calculates; Described normatron, according to the temperature field information in each strand section, by interpolation calculation, can obtain any locational temperature information of strand; The calculating of described casting blank solidification thickness, temperature field information and solid-state temperature is also all carried out in strand section.

The mathematical model of described strand diabatic process is described, and comprises the boundary condition that solidifies calculating, Temperature Distribution equation of continuous casting steel billet and solves Temperature Distribution equation starting condition.

Described continuous casting steel billet solidify calculating, calculate from crystallizer, before going out casting machine, finish, for slab, only consider the heat conduction of thickness direction, do not consider the heat conduction of casting direction and strand Width; The liquid phase initial temperature of molten steel equals tundish temperature; At the same cooling section of continuous casting, intensity of cooling remains unchanged.

Described continuous casting steel billet solidify calculating, by following expression, describe:

$\mathrm{c\&rho;}\frac{\&PartialD;U(x,t)}{\&PartialD;t}=\frac{\&PartialD;}{\&PartialD;x}\left(K\frac{\&PartialD;U(x,t)}{\&PartialD;x}\right)$

Wherein: x is the distance apart from casting billet surface; T is the casting start time; U (x, t) is the Temperature Distribution of casting blank section; ρ is density; C is specific heat; K is pyroconductivity.

The boundary condition of described Temperature Distribution equation is:

Slab surface temperature U (0, t)=U _s,

Wherein, U _sfor slab surface temperature, h is heat-conduction coefficient, U _wcooling water temperature, U _extbe environment temperature, σ is Si Difen-Boltzmann constant, and ε is coefficient of blackness.

The described Temperature Distribution equation starting condition that solves is:

Suppose that it is t=0, U (x, 0)=T constantly that crystallizer injects molten steel _tD;

Concreting thickness initial value: x _s| _t=0=0

Surface temperature initial value: U _s| _x=0=TS

Wherein: T _tDfor tundish temperature, TS is solid-state temperature.

When the heat-conduction equation of the strand heat radiation process described in actual computation and boundary condition, by this equation discretize, the heat-conduction equation that is converted into difference scheme solves.

Described gap periods is a second chronomere for level, and its scope is conventionally between 5～10s.

Compared with the prior art, advantage of the present invention is:

By normatron according to current process data, the strain regime of strand is carried out dynamically following the tracks of and calculating, and then according to the variation tendency of strain, judges the quality condition of strand inside, result of calculation is real-time, on-line synchronous performance is good;

2. for centre burst, appear near the slab predetermined slab head or afterbody, by optimization, cut directly defect base is excised.For the defect that has occurred predetermined slab middle part, according to predetermined, cut, but the slab after cutting is enclosed to flaw labeling, demote as requested or change and do its use, therefore by using this method, can effectively improve product percent of pass and quality stability.

3. adopt after this method, contribute to improve the product quality of manufacturing enterprise, improve qualification rate and the commercial grade of product, and then improve whole economic benefit.

Accompanying drawing explanation

Fig. 1 is control method block diagram of the present invention;

Fig. 2 is the formation schematic diagram of the technical program control system;

Fig. 3 is the be related to schematic diagram of section with strand;

Fig. 4 is slice coordinates location schematic diagram.

Embodiment

Below in conjunction with drawings and Examples, the present invention will be further described.

In Fig. 1, the technical program provides a kind of method of online forecasting casting blank crack defect, it calculates the cooled and solidified process of strand by online simulation, obtain the stressed strain information of strand inside, then according to the variation tendency of strain, the implosion defect of strand is forecast.

The technical program, by the internal soundness state of online forecasting strand, is delivered to cutting calculations machine in time the slab quality information in production run, for the cutting process of optimal control defect strand, and the control accuracy of improving the quality of products and product percent of pass.

The committed step of present techniques scheme comprises:

A, in existing L2 level process computer or on L2 level process computer, a normatron is set;

B, described L2 level process computer pass through L1 level computing machine, and the various technique in casting process and control parameter collection is complete, then according to certain gap periods, send to normatron;

C, described normatron receive in real time, online the technique in strand production run and control parameter, determine the boundary condition that strand heat radiation is calculated;

D, the described normatron mathematical model based on strand diabatic process is described, and dynamic calculation strand and extraneous heat radiation process obtain the temperature field of strand inside and outside;

The cooled and solidified process of E, described normatron dynamic calculation strand, obtains the concreting thickness information of each slice position of strand;

F, described normatron performance analysis strand stressed variation in moving process, calculate strand due to the bulgs stress, aligning strain and the dislocation strain that are subject to External Force Acting and produce, and the overall strain that obtains strand distributes;

G, described normatron be according to the variation tendency of strain, and by judging whether strain surpasses critical strain values, and then whether prediction have implosion to occur, and the implosion defect of strand is carried out to real-time estimate;

If implosion appears in H judgement strand, normatron calculates concrete defect information, and information is associated with slab more specific location information, by L2 level process computer, is issued to cutting L1 computing machine, and the cutting process of strand is optimized to control;

I, cutting L1 computing machine are adjusted the cutting position of strand, for centre burst, appear near the slab predetermined slab head or tail position, after cutting, directly defect base are excised by optimization; For the defect that has occurred predetermined slab middle part, according to precalculated position, cut, but the slab after cutting is enclosed to flaw labeling, demote as requested or change and do its use;

J, above-mentioned steps, in to the casting process of strand in real time, carry out online.

Above-mentioned normatron is PC, industrial computer, single-chip microcomputer or the virtual machine that is arranged in L2 level computing machine.

Specifically, this method be take numerical simulation as basis, by the diabatic process to strand, quantitatively calculates, and calculates in real time change of temperature field and the concreting thickness of continuous casting steel billet in casting process and changes, and further calculate each locational strain variation of strand.By judging whether strain surpasses critical strain values and predict whether have implosion to occur, if have, calculate the specifying information of implosion defect, and information is associated with slab, pass to the process computer of controlling strand cutting, for the cutting process of optimal control strand.

First the computational problem of solidifying of continuous casting steel billet being discussed, being calculated from crystallizer, before going out casting machine, finish, is the process of a heat conducting and radiating substantially, and this process metallurgy mechanism can be described with following formula:

$\mathrm{c\&rho;}\frac{\&PartialD;U(x,t)}{\&PartialD;t}=\frac{\&PartialD;}{\&PartialD;x}\left(K\frac{\&PartialD;U(x,t)}{\&PartialD;x}\right)---\left(1\right)$

Wherein: x is the distance (m) apart from casting billet surface;

T is the casting start time (min);

U (x, t) is the Temperature Distribution of casting blank section;

ρ is density (kg/m ³);

C is specific heat (J/ (kg ℃);

K is pyroconductivity (KCal/ (mh ℃)).

When calculating, we need to do some reasonable assumptions according to actual conditions.

For slab, only consider the heat conduction of thickness direction, do not consider the heat conduction of casting direction and strand Width; The liquid phase initial temperature of molten steel equals tundish temperature (molten steel is in the medial temperature of tundish); At the same cooling section of continuous casting, intensity of cooling remains unchanged.

(1) boundary condition of Temperature Distribution equation is:

Slab surface temperature U (0, t)=U _s,

Wherein,

U _sfor slab surface temperature,

H is heat-conduction coefficient,

U _wcooling water temperature

U _extit is environment temperature

σ is Si Difen-Boltzmann constant, and ε is coefficient of blackness.

At x=Thick/2 place, slab thickness center:

$K\frac{\&PartialD;U(x,t)}{\&PartialD;x}{|}_{x=\mathrm{Thick}/2}=0---\left(3\right)$

(2) solve Temperature Distribution equation starting condition:

Suppose that it is t=0, U (x, 0)=T constantly that crystallizer injects molten steel _tD;

Concreting thickness initial value: x _s| _t=0=0

Surface temperature initial value: U _s| _x=0=TS

Wherein: T _tDfor tundish temperature, TS is solid-state temperature.

Provided heat-conduction equation and the boundary condition of describing strand heat radiation process above, need to be this equation discretize during actual computation, the heat-conduction equation that is converted into difference scheme solves.

In addition, because we will calculate all locational temperature informations of continuous casting steel billet and concreting thickness information, for the ease of computer solving, we are divided into a series of sections whole strand along casting direction.

So-called " section ", can be regarded as the thin slice equating with casting blank section area, and position relationship and the coordinate setting schematic diagram of itself and strand are shown in Figure of description 3 and Fig. 4.

Because above-mentioned two figure all adopt the conventional method for expressing in this area and mark pattern, those skilled in the art is appreciated that its implication and represented information completely, therefore no longer narrate at this.

Section produces at the meniscus place of crystallizer, thereafter, with strand, moves, and goes out auto-destruct after casting machine.The Calculation of Heat Transfer of strand is all carried out in section, and fixed cycle calculates.During casting, casting billet surface and internal temperature are all continually varyings.Therefore, the temperature field information according in each strand section, by interpolation calculation, can obtain any locational temperature information of strand.In addition, casting blank solidification thickness can calculate by temperature field information and solid-state temperature, and all calculating is also all carried out in the section of strand.

Strand status information based on above-mentioned calculating gained, can calculate all kinds of strains that strand occurs when being subject to External Force Acting according to strain formula, and then according to the variation tendency of strain, judges the quality condition of strand inside.

The concrete computation process that provides defect forecast below, this process fixed cycle repeats:

A) collect strand production process parameters, determine the boundary condition that strand heat radiation is calculated.

B) mathematical model based on strand diabatic process is described, and dynamic calculation strand and extraneous heat radiation process obtain the temperature field of strand inside and outside.

C) the cooled and solidified process of dynamic calculation strand, obtains the concreting thickness information of each position of strand.

D) performance analysis strand stressed variation in moving process, calculates strand due to the bulgs stress, aligning strain and the dislocation strain that are subject to External Force Acting and produce, and the overall strain that obtains strand distributes.

E), according to the variation tendency of strain, the implosion defect of strand is predicted.

F) if implosion appears in judgement strand, calculate concrete defect information, and pass to cutting calculations machine, the cutting process of strand is optimized to control.

For step a), collect the procedure parameter relevant to strand diabatic process, comprise steel grade, molten steel temperature in tundish, thickness, pulling rate, width, cooling water flow etc.According to these data, determine the boundary condition of strand heat transfer calculations, determine the total amount of heat that in the unit interval, strand transmits to the external world.Along with the movement of strand, fixed cycle upgrades initial value and the boundary condition of strand Calculation of Heat Transfer.

For step b), in cooling procedure, the heat that strand sheds to the external world can calculate with following formula:

φ=h(U _s-U _w)(w/m ²) (4)

In formula, φ is the intensity of outwards dispelling the heat in unit area, U _sthe surface temperature of strand, U _wbe cooling water temperature, h is the coefficient of heat transfer of casting billet surface, with following formula, calculates:

h=kw ^rwa ^ra (5)

Wherein, w is jet density, and rw is water yield coefficient; A is tolerance density, and ra is tolerance coefficient, and k is constant.

Wherein the computing method of water yield density are, take cooling zone as unit, calculate the water spray total amount of certain cooling zone on strand upper surface, and divided by the area of cooling zone, what obtain is jet density.The computing method of tolerance density are identical with it, at this repeated description no longer.

At each constantly, according to cooling water inflow, tolerance, first calculate the coefficient of heat transfer of casting billet surface, then extrapolate based on this heat that in the unit interval, casting billet surface sheds, and then obtain temperature field, strand inside and outside according to steel grade physical parameter.

For step c), it should be noted that, it is solid-state that the solidifying of molten steel is not from liquid state, to become simply, but along with the reduction of temperature, have two-phase region, and we are referred to as " mushy zone " at ordinary times.Below the solidus temperature of steel grade, molten steel could change solid completely into.While calculating the concreting thickness of strand, first calculate the solidification rate of each position on strand, then calculate concreting thickness according to solidification rate.

The computing formula of solidification rate is:

$\mathrm{fs}=\frac{T_l-T_c}{T_l-T_s}---\left(6\right)$

Wherein, the solidification rate that fs is strand, T _lfor the liquidus temperature of steel, T _sfor the solidus temperature of steel, T _cfor the temperature on slab center line.

Use above-mentioned computing formula, on the xsect of strand, calculate the solidification rate of each position, then calculate the concreting thickness of each position.

For steps d), the strand temperature field obtaining based on previous calculations distributes and concreting thickness information, calculates respectively bulgs stress, aligning strain and dislocation strain, and calculates each locational overall strain of strand.In calculating, use the following bulgs stress that calculates:

${\mathrm{\&epsiv;}}_b\left(i\right)=\frac{1600s_i{\mathrm{\&delta;}}_i}{{l_i}^2}\&times;100\%---\left(7\right)$

In formula:

ε _b(i): the bulgs stress (%) at casting blank solidification interface, i roller place

S _i: i the casting blank solidification thickness (mm) that casting roll position is corresponding

L _i: i roller spacing (mm)

δ _i: bulge deformation amount (mm)

In these parameters, the computing formula of the bulge deformation of strand is:

${\mathrm{\&delta;}}_i=\frac{\mathrm{\&eta;\&alpha;p}{l}^4\sqrt{\frac{l_i}{v_g}}}{32E{s}_i^3}---\left(8\right)$

For slab, η α=1;

P: the ferrostatic pressure that casting roll is born, kg/cm2;

V _g: casting speed, cm/min;

E: elasticity coefficient,

t _sfor the solidus temperature of steel, T _sfor solidifying temperature, T _mfor medial temperature.

Wherein,

t _mfor medial temperature, T _ffor surface temperature,

When strand passes through the aligning district of conticaster, under the effect of aligning stress, can make the strand of this position produce aligning strain.Aligning strain is relevant with conticaster aligning point radius-of-curvature, and computing formula is:

${\mathrm{\&epsiv;}}_u\left(i\right)=(\frac{D}{2}-s_i)(\frac{1}{R_i-\frac{D}{2}}-\frac{1}{R_{i+1}-\frac{D}{2}})\&times;100\%---\left(9\right)$

In formula:

S _i: i smoothing roll place strand scull thickness (mm)

D: slab thickness (mm);

R _i: strand outer arc radius (mm) before i smoothing roll

R _i+1: strand outer arc radius after i smoothing roll, mm

ε _u(i): i the strand aligning strain (%) that smoothing roll is corresponding

Dislocation strain is the strain on continuous casting nip rolls casting blank solidification interface that arc is forbidden to cause, and its computing formula is:

${\mathrm{\&epsiv;}}_{m\left(i\right)}=\frac{300s_i{\mathrm{\&delta;}}_m}{l_i^2}\&times;100\%---\left(10\right)$

ε wherein _{m (i)}it is the strain that i roller place produces on freezing interface because of roller dislocation; δ _mmagnitude of misalignment (mm) for roller place; s _iit is the slab thickness (mm) at i roller place.

The temperature field of the strand that the strain computing formula based on above and previous calculations obtain and concreting thickness information, calculate respectively bulgs stress, aligning strain and dislocation strain.

Its concrete grammar is: using casting roll position as indexing parameter, in all sections, search for, find out the identical section in position, or find out two sections of the most close casting roll position.

For the previous case, directly the temperature information of section and concreting thickness information are brought into strain formula and calculated; For latter event, need two slice positions adjacent with front and back according to casting roll position, and temperature information and the concreting thickness information of the upper record of former and later two sections, linear interpolation obtains the corresponding strand temperature in casting roll position and concreting thickness information, and is brought into strain formula and calculates.

After whole strain indexs are calculated, according to casting roll, calculate total strain, obtain current strand overall strain and distribute.

For step e), the critical strain that strand occurs crackle due to stress deformation is generally between 0.5%～0.8%.Yet due to the difference of chemical composition and mechanical characteristic, the ultimate strain that strand can bear is relevant to steel grade, concrete numerical value obtains by engineer testing.

During practical application, according to steel grade, classify in advance, and be respectively every class steel grade and set critical strain values, and by these Parameter storages in the database of forecasting model.

When forecasting model calculates, can from database index, go out corresponding critical strain according to cast steel grade.Next, from casting machine outlet, to crystallizer direction, to each casting roll, corresponding strand overall strain judges successively, judge whether overall strain surpasses total critical strain, if strain has surpassed critical strain values, judgement has implosion defect to occur, the occurrence positions information of recording defect.

For step f), because the casting blank crack due to strain can affect one section of region, therefore, when model judgement has implosion to occur, not only to record the particular location that implosion occurs, also will follow the tracks of the lasting time of implosion, finally to determine the region that implosion is covered on strand.Complete after calculating, these information exchanges are crossed network delivery to the process computer of controlling strand cutting, according to features such as the order of severity of defect and area size, strand are optimized to cutting, improve the recovery rate of product.

Embodiment:

The vertical bending type slab caster of certain steel mill, machine two streams, product specification is mainly 220mm * 1930mm.

The formation of its production control system is as shown in Figure of description 2.

In process of production, L3 level computing machine is responsible for assigning production schedule instruction, and L2 level computing machine is responsible for determining the various control parameters in production run according to the production schedule, and control parameter is issued to the execution of L1 level computing machine; On the other hand, L2 level computing machine is by L1 level computing machine, and the various technique in casting process and control parameter collection is complete, then according to certain cycle, sends to normatron.

Normatron, according to current process data, carries out dynamically following the tracks of and calculating to the strain regime of strand.

When normatron is determined while there is casting blank crack defect, the positional information of defect is corresponded on predetermined slab, then these information are sent to L2 level computing machine, by L2 level computing machine, these information are issued to cutting L1 computing machine, and according to the position at defect place, adjust the cutting position of strand, to improve the qualification rate of strand product.

Everybody is set up the dynamic strain of generation to calculate strand, first will obtain in real time temperature field, inside and outside and the concreting thickness information of strand, and they obtain by calculating the diabatic process of strand.

Calculation of Heat Transfer is used the heat-conduction equation of formula (1), calculates the parameter information using as follows:

For the computation formula for thermal conduction of empty cooled region, coefficient of blackness ε selects 0.85; The coefficient of heat transfer of casting billet surface is used formula below to calculate

h=280.56w ^0.382a ^0.1373

Wherein, w use is sprayed onto the lip-deep actual amount of water of strand inner arc and the actual strand area covering of spraying water calculates, and a adopts identical method to calculate.

According to interval 50mm, be that standard is divided and cut on strand, the gap periods that model calculates is selected 8s.

During the technical scheme of application described in the application, fixed cycle is carried out following flow process:

The cooling water inflow, tolerance (aerosol control), casting speed, tundish temperature, crystallizer cooling water inflow, crystallizer that L2 level computing machine (being designated hereinafter simply as L2) fixed cycle is collected each cooling circuit from L1 level computing machine imported and exported the process data information such as cooling water temperature is poor.

According to the production process information of L2 input, utilize heat transfer equation dynamic calculation strand each locational temperature field and concreting thickness, wherein concreting thickness refers on the xsect of strand, the fs=1.0 position of calculating according to formula (6) and the vertical range of casting billet surface.

Based on formula (7)～(10), the bulgs stress in all strand sections of dynamic calculation, aligning strain and dislocation strain.

Because every kind of strain is to be all subject to independently External Force Acting to occur, therefore, these strains can stack up.So, take and cut into slices as unit, calculate the overall strain size in each strand section.

From casting machine outlet, forward to crystallizer outlet, the strain in each section is judged, the threshold value of judgement obtains according to current cast steel grade index from database.

If the overall strain in strand section is greater than threshold limit value, a strain abnormality sign is set.

Successively each strand section is judged, and in each strand section, set corresponding sign according to the comparative result of overall strain and threshold limit value.

If in a upper computation period, the strain of model judgement strand has surpassed critical value, and at current period, the strain of strand is all normal, now, according to the abnormality mark in strand section, can determine the particular location that strain abnormality occurs.

The defect starting position of strand and end position and stream number are sent to cutting L1 computing machine, the abnormality mark in strand section is removed simultaneously.

According to model, determine defect occurrence positions, cutting L1 computing machine is adjusted the cutting position of strand.

Use after this method, for centre burst, appear near the slab predetermined slab head or afterbody, after cutting by optimization, directly defect base is excised.For the defect that has occurred predetermined slab middle part, according to predetermined, cut, but the slab after cutting is enclosed to flaw labeling, demote as requested or change and do its use.

Technical scheme of the present invention, by cooled and solidified process real-time, that online simulation calculates strand, obtains the stressed strain information of strand inside, more in real time the implosion defect of strand is forecast according to the variation tendency of strain.Then the slab quality information in production run is delivered to cutting L1 computing machine in time, it is for the cutting process of optimal control defect strand, the control accuracy of can improving the quality of products and product percent of pass.By using this method, improved product percent of pass and quality stability, improve qualification rate and the commercial grade of product, and then improved the economic benefit of enterprise's integral body.

The present invention can be widely used in the optimization/control field of the cutting process of strand in sheet billet continuous casting production run.

Claims (27)

1. the online forecasting method of a casting blank crack defect, comprise by L3 level computing machine, the network that L2 level computing machine and L1 level computing machine form and data transmission each other, wherein, L3 level computing machine is responsible for assigning production schedule instruction, L2 level computing machine is responsible for determining the various control parameters in production run according to the production schedule, and control parameter is issued to the execution of L1 level computing machine, L1 level computing machine is carried out steering order that L2 level computing machine issues or operative employee's input, directly or indirectly control the relevant device of casting machine, described L1 level computing machine at least comprises public L1 computing machine, casting L1 computing machine and cutting L1 computing machine, it is characterized in that described online forecasting method at least comprises the following steps:

A, in existing L2 level process computer or on L2 level process computer, a normatron is set;

B, described L2 level process computer pass through L1 level computing machine, and the various technique in casting process and control parameter collection is complete, then according to certain gap periods, send to normatron;

C, described normatron receive in real time, online the technique in strand production run and control parameter, determine the boundary condition that strand heat radiation is calculated;

D, the described normatron mathematical model based on strand diabatic process is described, and dynamic calculation strand and extraneous heat radiation process obtain the temperature field of strand inside and outside;

The cooled and solidified process of E, described normatron dynamic calculation strand, obtains the concreting thickness information of each slice position of strand;

F, described normatron performance analysis strand stressed variation in moving process, calculate strand due to the bulgs stress, aligning strain and the dislocation strain that are subject to External Force Acting and produce, and the overall strain that obtains strand distributes;

G, described normatron be according to the variation tendency of strain, and by judging whether strain surpasses critical strain values, and then whether prediction have implosion to occur, and the implosion defect of strand is carried out to real-time estimate;

If implosion appears in H judgement strand, normatron calculates concrete defect information, and information is associated with slab more specific location information, by L2 level process computer, is issued to cutting L1 computing machine, and the cutting process of strand is optimized to control;

I, cutting L1 computing machine are adjusted the cutting position of strand, for centre burst, appear near the slab predetermined slab head or tail position, after cutting, directly defect base are excised by optimization; For the defect that has occurred predetermined slab middle part, according to precalculated position, cut, but the slab after cutting is enclosed to flaw labeling, demote as requested or change and do its use;

J, above-mentioned steps, in to the casting process of strand in real time, carry out online.

2. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that described normatron is PC, industrial computer, single-chip microcomputer or the virtual machine that is arranged in L2 level computing machine.

3. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that various technique and control parameter in described casting process, for the relevant procedure parameter of strand diabatic process, it at least comprises steel grade, molten steel temperature in tundish, thickness, pulling rate, width and cooling water flow; Described normatron, according to these data, is determined the boundary condition that strand Heat Transfer Meter is calculated, and determines the total amount of heat that in the unit interval, strand transmits to the external world; Along with the movement of strand physical location, normatron fixed cycle upgrades initial value and the boundary condition of strand Calculation of Heat Transfer.

4. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that described normatron at each constantly, according to cooling water inflow, tolerance, first calculate the coefficient of heat transfer of casting billet surface, extrapolate based on this again the heat that in the unit interval, casting billet surface sheds to the external world, and then according to steel grade physical parameter, obtain the temperature field of strand inside and outside.

5. according to the online forecasting method of casting blank crack defect claimed in claim 4, it is characterized in that in cooling procedure, the heat that described strand sheds to the external world adopts following expression to calculate:

φ=h(U _s-U _w)(w/m ²)

In formula, φ is the intensity of outwards dispelling the heat in unit area, U _sthe surface temperature of strand, U _wbe cooling water temperature, h is the coefficient of heat transfer of casting billet surface.

6. according to the online forecasting method of casting blank crack defect claimed in claim 5, it is characterized in that the coefficient of heat transfer of described casting billet surface calculates by following expression:

h=kw ^rwa ^ra

Wherein, w is jet density, and rw is water yield coefficient; A is tolerance density, and ra is tolerance coefficient, and k is constant.

7. according to the online forecasting method of casting blank crack defect claimed in claim 4, the computing method that it is characterized in that described jet density are, take cooling zone as unit, calculate the water spray total amount of certain cooling zone on strand upper surface, divided by the area of cooling zone, what obtain is jet density; Described in it, the computing method of tolerance density are identical with the computing method of jet density.

8. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that first described normatron calculates the solidification rate of each position on strand, then calculate concreting thickness according to solidification rate;

Described in it, the calculation expression of solidification rate is:

$\mathrm{fs}=\frac{T_l-T_c}{T_l-T_s}$

Wherein, the solidification rate that fs is strand, T _lfor the liquidus temperature of steel, T _sfor the solidus temperature of steel, T _cfor the temperature on slab center line;

Described normatron uses above-mentioned calculation expression, calculates the solidification rate of each position on the xsect of strand, then calculates the concreting thickness of each position.

9. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that described normatron distributes and concreting thickness information based on calculating the strand temperature field obtaining, calculate respectively bulgs stress, aligning strain and dislocation strain, these Strain superimpositions are got up, take strand section as unit, calculate the overall strain size on each strand slice position.

10. according to the online forecasting method of casting blank crack defect claimed in claim 9, it is characterized in that described bulgs stress adopts following expression formula to calculate:

${\mathrm{\&epsiv;}}_b\left(i\right)=\frac{1600s_i{\mathrm{\&delta;}}_i}{{l_i}^2}\&times;100\%$

In formula:

ε _b(i): the bulgs stress at casting blank solidification interface, i roller place, s _i: i the casting blank solidification thickness that casting roll position is corresponding, l _i: i roller spacing, δ _i: the bulge deformation amount at i roller place.

11. according to the online forecasting method of casting blank crack defect claimed in claim 10, it is characterized in that the computing formula of the bulge deformation of described strand is:

${\mathrm{\&delta;}}_i=\frac{\mathrm{\&eta;\&alpha;p}{l}^4\sqrt{\frac{l_i}{v_g}}}{32E{s}_i^3}$

For slab, η α=1; P: the ferrostatic pressure that casting roll is born; v _g: casting speed; E: elasticity coefficient.

12. according to the online forecasting method of the casting blank crack defect described in claim 11, it is characterized in that the computing formula of described elasticity coefficient E is:

$E=\frac{T_s-T_m}{T_S-100}\&times;{10}^5\mathrm{kg}/{\mathrm{cm}}^2$

Wherein, T _sfor the solidus temperature of steel, T _sfor temperature of solidification, T _mfor medial temperature;

Medial temperature T described in it _mcomputing formula be:

$T_m=\frac{T_s+T_f}{2}$

Wherein, T _sfor the solidus temperature of steel, T _ffor surface temperature.

13. according to the online forecasting method of casting blank crack defect claimed in claim 9, it is characterized in that the computing formula of the aligning strain of described strand is:

${\mathrm{\&epsiv;}}_u\left(i\right)=(\frac{D}{2}-s_i)(\frac{1}{R_i-\frac{D}{2}}-\frac{1}{R_{i+1}-\frac{D}{2}})\&times;100\%$

In formula:

S _i: i smoothing roll place strand scull thickness;

D: slab thickness;

R _i: strand outer arc radius before i smoothing roll;

R _i+1: strand outer arc radius after i smoothing roll;

ε _u(i): i the strand aligning strain that smoothing roll is corresponding.

14. according to the online forecasting method of casting blank crack defect claimed in claim 9, it is characterized in that described dislocation strain is the strain on continuous casting nip rolls casting blank solidification interface that arc is forbidden to cause, and its computing formula is:

${\mathrm{\&epsiv;}}_{m\left(i\right)}=\frac{300s_i{\mathrm{\&delta;}}_m}{l_i^2}\&times;100\%$

Wherein, ε _{m (i)}it is the strain that i roller place produces on freezing interface because of roller dislocation; δ _mmagnitude of misalignment for roller place; s _iit is the slab thickness at i roller place.

15. according to the online forecasting method of casting blank crack defect claimed in claim 9, it is characterized in that described normatron obtains the overall strain distribution of strand through the following steps:

Using casting roll position as indexing parameter, in all sections, search for, find out the identical section in position, directly the temperature information of section and concreting thickness information are brought into strain formula and calculated, after whole strain indexs are calculated, according to casting roll position, calculate total strain, obtain current strand overall strain and distribute.

16. according to the online forecasting method of casting blank crack defect claimed in claim 9, it is characterized in that described normatron obtains the overall strain distribution of strand through the following steps:

Using casting roll position as indexing parameter, in all sections, search for, find out two sections of the most close casting roll position, two slice positions adjacent with front and back according to casting roll position, and temperature information and the concreting thickness information of the upper record of former and later two sections, linear interpolation obtains the corresponding strand temperature in casting roll position and concreting thickness information, and be brought into strain formula and calculate, after whole strain indexs are calculated, according to casting roll position, calculate total strain, obtain current strand overall strain and distribute.

17. according to the online forecasting method of casting blank crack defect claimed in claim 9, it is characterized in that according to steel grade, classifying in advance, and be respectively every class steel grade and set critical strain values, and by these Parameter storages in the database of described normatron forecasting model, when normatron calculates, can from database index, go out corresponding critical strain according to cast steel grade, then, normatron is from casting machine outlet, to crystallizer direction, to each casting roll, corresponding strand overall strain judges successively, judge whether overall strain surpasses total critical strain, if strain has surpassed critical strain values, judgement has implosion defect to occur, the occurrence positions information of normatron recording defect.

18. according to the online forecasting method of casting blank crack defect claimed in claim 9, it is characterized in that the critical strain span of described strand is between 0.5%～0.8%, the ultimate strain that strand can bear is relevant to steel grade, and its concrete numerical value obtains by engineer testing.

19. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that when described normatron judgement strand has implosion to occur, not only to record the particular location that implosion occurs, also to follow the tracks of the lasting time of implosion, finally to determine the region that implosion is covered on strand.

20. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that completing after calculating when described normatron, result of calculation information exchange is crossed to network delivery to the cutting L1 computing machine of controlling strand cutting, according to features such as the order of severity of defect and area size, strand is optimized to cutting, improves the recovery rate of product.

21. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that whole strand to be divided into a series of thin slices that equate with casting blank section area along casting direction, forms described strand section; The Calculation of Heat Transfer of described strand is all carried out in strand section, fixed time period calculates; Described normatron, according to the temperature field information in each strand section, by interpolation calculation, can obtain any locational temperature information of strand; The calculating of described casting blank solidification thickness, temperature field information and solid-state temperature is also all carried out in strand section.

22. according to the online forecasting method of casting blank crack defect claimed in claim 1, the mathematical model that it is characterized in that described strand diabatic process is described, and comprises the boundary condition that solidifies calculating, Temperature Distribution equation of continuous casting steel billet and solves Temperature Distribution equation starting condition.

23. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that the calculating of solidifying of described continuous casting steel billet, calculating is from crystallizer, before going out casting machine, finish, for slab, only consider the heat conduction of thickness direction, do not consider the heat conduction of casting direction and strand Width; The liquid phase initial temperature of molten steel equals tundish temperature; At the same cooling section of continuous casting, intensity of cooling remains unchanged.

24. according to the online forecasting method of the casting blank crack defect described in claim 23, it is characterized in that the calculating of solidifying of described continuous casting steel billet, by following expression, describes:

$\mathrm{c\&rho;}\frac{\&PartialD;U(x,t)}{\&PartialD;t}=\frac{\&PartialD;}{\&PartialD;x}\left(K\frac{\&PartialD;U(x,t)}{\&PartialD;x}\right)$

Wherein: x is the distance apart from casting billet surface; T is the casting start time; U (x, t) is the Temperature Distribution of casting blank section; ρ is density; C is specific heat; K is pyroconductivity.

25. according to the online forecasting method of the casting blank crack defect described in claim 22, it is characterized in that the boundary condition of described Temperature Distribution equation is:

Slab surface temperature U (0, t)=U _s,

Wherein, U _sfor slab surface temperature, h is heat-conduction coefficient, U _wcooling water temperature, U _extbe environment temperature, σ is Si Difen-Boltzmann constant, and ε is coefficient of blackness.

26. according to the online forecasting method of the casting blank crack defect described in claim 22, solves Temperature Distribution equation starting condition to be described in it is characterized in that:

Suppose that it is t=0, U (x, 0)=T constantly that crystallizer injects molten steel _tD;

Concreting thickness initial value: x _s| _t=0=0

Surface temperature initial value: U _s| _x=0=TS

Wherein: T _tDfor tundish temperature, TS is solid-state temperature.

27. according to the online forecasting method of casting blank crack defect claimed in claim 1, it is characterized in that described gap periods is for a second chronomere for level, and its concrete numerical range is between 5～10s.

Images