CN1224720C - Blast furnace smelt controlling method with intelligent control system - Google Patents

Abstract

The present invention relates to a method for controlling blast furnace smelting by using an intelligent control system. A No. 1 optimization station and a No. 2 optimization station are arranged in a master control room, and the No. 1 optimization station, the No. 2 optimization station and a plurality of operation stations form a microcomputer local area network. A server is arranged in an iron work and is networked with the terminal of each control procedure to form a local area network of the iron work. The local area network of the iron work is networked with the local area network of the master control room to form a network system with the characteristics of the real-time collection and the automatic transfer of ironmaking procedural information, an intelligent control software module is arranged in the No. 1 optimization station, a module for automatically displaying an intelligent control chart and a fault indication chart according to a smelting program flow are arranged in the No. 2 optimization station, the No. 1 optimization station and the No. 2 optimization station automatically operate in the work process through data which is collected in a real-time and on-line way by the intelligent control system software module in company with a blast furnace smelting process and data which is automatically transferred on the local area networks according to certain time requirements, and results are output in the mode of man-machine conversations and simple and clear model charts. Operations are provided for a blast furnace foreman, and the foreman is prompted to take adequate measures. The present invention can effectively avoid generating furnace condition failure, and keep the blast furnace operating under the optimal thermal condition.

Description

A kind of method of utilizing intelligence control system control blast-furnace smelting

Technical field

The present invention relates to utilize the method for intelligence control system control blast-furnace smelting.

Background technology

In the various technological process of productions of Steel Complex, the blast furnace iron-making process flow process is the most complicated.

Existing blast-furnace smelting system comprises the transmitter of online dynamic acquisition blast furnace data, be equipped in the Computerized monitor system that several active stations that the blast furnace main control room is used to judge blast-furnace body working order, heat state of blast furnace and blast-furnace smelting process direct motion state are formed, each control operation and information conveying flow of iron-smelting process be as shown in Figure 1: the shift foreman organizes the production process of blast furnace ironmaking at the blast furnace master control room: require that the batch weighing system installs in the truck according to the various crude fuels of charger sheet weighing under the groove; Require elevator material loading cloth system furnace charge to be installed in the blast furnace according to the charging system of determining; Require the hotblast stove workshop according to the quantity of determining to blast furnace air feed, oxygen enrichment and humidification; Require the coal powder injection workshop according to given coal powder injection speed to pulverized coal injection in blast furnace.The regulation and control of these complexity be the section chief by brainwork, make according to oneself experience and easy process calculation.And each is controlled sub-operation and all has data gathering of oneself and control subsystem to implement the control that the section chief requires, and they abbreviate the basic automatization of sub-operation as.

The shift foreman by observing the picture of instrument or microcomputer monitoring, holds the changing conditions of hundreds of parameters of blast-furnace body temperature, pressure flow detection at the blast furnace master control room; The various data of record comprise automatic analytical data of furnace roof mixed gas and the molten iron of being offered by the laboratory phone, slag, go into stove crude fuel chemical composition data in the blast furnace operation daily record; The molten iron station molten iron weighing data offered of phone of weighing; The field data that front slagging is tapped a blast furnace, molten iron thermometric and molten iron on-line weighing data etc.In each stove pig iron smelting process, the section chief makes comprehensive judgement and regulation and control decision with the personal experience to the working of a furnace and furnace temperature, and the sub-operation that phone is relevant is regulated and control then.The ironmaking expert claims that the mode of operation of this " watch the instruments, copy data, by rule of thumb " is a s-generation blast furnace operation technology.

Obviously, numerous owing to the iron manufacturing process influence factor, situation is intricate, and complicated various to the judgement and the control of furnace temperature, therefore carelessness and error take place unavoidably in operation by rule of thumb, cause blast wandering, and output descends, consumption is increased, the situation of quality fluctuation takes place.And each class section chief's experience is also inconsistent, and operant level also has difference, causes blast furnace ironmaking process not keep in the best condition.

As from the foregoing: though the existing blast furnace technological process of production has realized local automation's control of each sub-operation, but the section chief of blast furnace master control room the integral body of blast-furnace smelting process being planned strategies for exists 2 major defects aspect the control: 1) the data information transfer means fall behind; 2) smelting process data message processing mode falls behind, and operates with the personal experience.Therefore the blast furnace production potential can not be given full play to.

From the seventies in 20th century, just the begin one's study automatization control of blast furnace ironmaking process of domestic and international metallurgical boundary.But, because the blast furnace process automatization relates to the multidisciplinary crossing research in aspect such as iron-smelting process technology, computer information technology, automatic control technology and mathematical modeling technique, technical difficulty is big, and therefore, blast furnace ironmaking process fully automated control does not so far realize." Blast Furnace Expert System " adopted in international at present advanced large blast furnace process control, and section chief's operation is unified on the expert systems platform.

The patent of NKK's application that Patent Office of the People's Republic of China announced on December 23rd, 1987 " method of control operation of blast furnace " (CN87103633A), the patent " blast furnace operation management method and device " of company of the Nippon Steel application of announcing July 11 nineteen ninety (CN1043745A) and the patent " utilizing the method for artificial intelligence expert systems blast-furnace smelting " of Capital Iron ﹠ Steel General Co's application of announcing on April 22nd, 1998, can see that these expert systemss set up expert knowledge library and inference engine from different angles separately, help operational management and the control that the blast furnace section chief strengthens blast furnace with giant-powered computer, carry out working of a furnace diagnosis, in time the prompting section chief takes adequate measures, the development of control furnace temperature is avoided the blast furnace fault and is ensured normally carrying out of producing.

More than 3 background technology files all do not set up the optimum regime and the Stochastic Optimal Control mathematical model of blast furnace process from the principle angle of Comtemporary Control Theory and intelligent control opinion, set up blast furnace process Model for Multi-Objective Optimization and silicon content of hot metal [Si] intelligent control partial differential equation; Also do not set up the personal computer net of iron work, realization information networking and " intelligence control system " are used in operation-technology-managerial " trinity ".

Summary of the invention

The objective of the invention is deficiency at the above-mentioned background technology, a kind of method of utilizing intelligence control system control blast-furnace smelting is provided, to realize that blast-furnace smelting is in operation-technology-managerial " trinity " intelligent control, reach raising output, reduce consuming, improve the target of molten steel quality and maintenance working of a furnace stable smooth operation, blast furnace longevity.

The method of utilizing intelligence control system control blast-furnace smelting of the present invention, comprise and utilize the online dynamic acquisition blast furnace of transmitter data, utilization is equipped in the Computerized monitor system of several active stations compositions of blast furnace main control room and judges the blast-furnace body working condition, heat state of blast furnace and blast-furnace smelting process direct motion state, this method also is included in the blast furnace master control room and No. 2 of optimizing station and double as server for No. 1 are set optimize two microcomputers of standing, these two microcomputers and several active stations are formed the master control room personal computer net, constitute the hardware platform of section chief's intellectualized operation; At iron work server is set, terminal with the sub-operation of batching under the groove, the terminal of elevator cloth material for making clothes operation, the terminal of the sub-operation of oxygen enriched blast, the terminal of the sub-operation of coal powder injection, laboratory terminal, and the terminal networking of administrative authority such as factory director office, engineering department, motor-driven section, production statistics section, operation department, form the iron work local area network, and network, form iron manufacturing process real time information sampling and the network system that transmits automatically with the master control room local area network; And be provided with and comprise database module optimizing the station for No. 1, cooling system Device Diagnostic module, the furnace condition anterograde fault diagnosis module, batching computation optimization module, optimal heat state and Stochastic Optimal Control computing module, the intelligent control module of silicon content of hot metal and sulphur content, and the intelligence control system software module of managerial report automatically-generating module are provided with by smelting program circuit and show the module of intelligent control chart and cooling system, furnace condition anterograde fault indication chart automatically optimizing the station for No. 2; Optimize the station for No. 1 and optimize stations for No. 2 and follow blast-furnace smelting process real-time period automatic data collection, the shift foreman calls No. 1 by human-computer interaction interface and optimizes each software module of standing, and obtains also guiding operation in view of the above of real-time evaluation; Optimize the station No. 2, with second the level process be automatically to gather the cooling apparatus data cycle to carry out trouble diagnosis, show at picture, with second the level process to be the cycle carry out fault to air quantity, blast, three sensitive parameters of ventilation property judges, show at picture, be the cycle to gather working of a furnace data automatically with a minute level process, diagnosis working of a furnace failure symptom, show at picture, the intelligent control picture of 30 minutes online demonstration silicon content of hot metal of process and sulphur content, its output data is the foundation of section chief's regulating and controlling furnace temperature; In the laboratory terminal, the chemical composition data of laboratory technician's typing molten iron, slag, the chemical ingredients of agglomerate, pellet, rawore and coke and check data, and be sent to the blast furnace master control room automatically by network system, become the latest data source that batching is checked single calculating automatically, the calculating of silicon content of hot metal governing equation, sulphur content prediction and calculation, the hot state computation of best stove; Show thermal equilibrium and basicity balance deviation optimization range if carry out charge calculation automatically according to up-to-date crude fuel composition, then point out the section chief to become material and do new charge calculation, it is as follows to it is characterized in that having set up the Model for Multi-Objective Optimization formula in the blast furnace optimal heat state of optimizing the station for No. 1 and Stochastic Optimal Control module:

$U\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_1(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\→\mathrm{Max}$

$K\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_2(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\→\min$

$\left[\mathrm{Si}\right]\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_3(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\∈\left(\right[\mathrm{Si}{]}_0-a,[\mathrm{Si}{]}_0+a)$

$\left[S\right]\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_4(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\≤[S{]}_0$

In the above-mentioned formula, U (t) is the capacity factor of a blast furnace, and Max represents maximization, and K (t) is a coke ratio, and min represents to minimize, and [Si] is silicon content of hot metal (t), requires to be controlled at optimum range ([Si] ₀-a, [Si] ₀+ a) in, [S] is sulfur content of hot metal (t), satisfies steelmaking quality and requires to be lower than limit value [S] ₀The lower limit t of the sign of integration ₀Represent the time opening and the concluding time in a certain smelting stage, the F under the sign of integration respectively with upper limit t ₁, F ₂, F ₃, F ₄Expression complicated function association separately, wherein Y is the crude fuel parameter, and Z is a device parameter, and X is a state parameter, and u is a controlled variable, Be respectively their velocity of variation, d τ represents the integral and calculating to the time, to U (t), K (t), [Si] (t) and [S] (t) these 4 variablees divide sample according to time series, carrying out multivariate statistics calculates, the important parameter optimal control scope and the combination thereof of determine to improve utilization coefficient, reduce coke ratio, silicon content of hot metal is suitable and sulfur content of hot metal meets the demands, its calculation result is divided into the output of 9 width of cloth pictures on No. 1 optimization station: 1) silicon content of hot metal-basicity of slag binary is preferably schemed; 2) air quantity-permeability index binary is preferably schemed; 3) blast energy-coke load match map; 4) utilization coefficient-combustion intensity is preferably schemed; 5) utilization coefficient-silicon content of hot metal is preferably schemed; 6) utilization coefficient-basicity of slag is preferably schemed; 7) sulfur content of hot metal-silicon content of hot metal is preferably schemed; 8) sulfur content of hot metal-basicity of slag is preferably schemed; 9) batching basicity-basicity of slag correlation diagram; Obtain the optimum regime and the Stochastic Optimal Control scope of blast furnace ironmaking process important parameter thus, and along with the variation of blast furnace condition, by self study correction parameters optimization scope;

It is as follows to have set up the partial differential equation formula of using neural network algorithm calculating in the silicon content of hot metal intelligent control module:

$\frac{\∂\left[\mathrm{Si}\right]}{\∂t}=A\frac{\∂\mathrm{LS}}{\∂t}+F\frac{\∂\mathrm{FF}}{\∂t}+C\frac{\∂\mathrm{TLC}(t-{\mathrm{\τ}}_1)}{\∂t}+Q\frac{\∂\mathrm{MQ}}{\∂t}+L\frac{\∂\mathrm{FL}}{\∂t}+W\frac{\∂\mathrm{FW}}{\∂t}+B\frac{\∂\mathrm{PMB}}{\∂t}+S\frac{\∂\mathrm{FS}}{\∂t}$

In the formula Be the variable quantity of silicon content of hot metal,

Be the fast variation index of material, Be the ventilation property variation index, Be iron amount difference index, τ ₁Be amount time lag,

Be the gas utilization rate variation index, Be the air quantity variation index,

Be the wind-warm syndrome variation index,

Be coal powder injection speed variation index, Be the rheumatism variation index, A, F, C, Q, L, W, the every constant coefficient of B, S then are the data with the size of a sample of a certain setting, the regression coefficient that obtains according to the multiple regression self study; The result that this Equation for Calculating obtains is presented at No. 2 with the form of furnace temperature intelligence control system figure and optimizes on the screen of station, and follow the operation process of smelting, carried out prediction and calculation and frame update in per 30 minutes, dynamically control the indicating instrument of furnace temperature as the section chief, the control furnace temperature develops towards optimum regime.

Above-mentioned intelligence control system software module has the 3-tier architecture of database layer, program layer and human-computer interaction interface layer, database layer comprises: go up the data of collection automatically and the data of the manual typing of setting by blast furnace programmable logic controller PLC, the raw data base that divides 50 data sheet to set up different levels such as a minute level, hour level on server is by iron-smelting process principle and the expert knowledge library of experience foundation and the parameters optimization storehouse that is obtained by calculated with mathematical model; Program layer comprises: the quantity computation program of each mathematical model, each logic determining program and knowledge reasoning program, and the human-computer interaction interface layer comprises: various simple and clear intelligent chart and data display that service data and prediction of result are provided for the section chief;

Adopt the inventive method, its intelligence control system software module is being followed the data of transmitting automatically on automatic data of gathering of blast-furnace smelting process real-time online and the personal computer net and is being required automatic operation according to the regular hour, mode with man-machine conversation and straight-forward mode chart is exported the result, the section chief provides operation to blast furnace, the prompting section chief takes appropriate measures, avoid taking place working of a furnace fault, keep blast furnace under the optimal heat state, to move.

The present invention has the advantage of following several respects:

(1) owing to the singularity of iron-smelting process, production process relates to a large amount of process calculation and statistics.Use the present invention and can improve the level of section chief's operation process intellectual work automatization, alleviate section chief's brainwork intensity, increase work efficiency.

(2) to relate to sub-operation many for iron-making production, and data transfer is originally undertaken by telephone set, and efficient is low and make mistakes easily.After the present invention set up the iron work local area network, data transfer had realized networking, thereby can realize the automatization of information processing.

(3) after each the relevant section office networking of iron work factory department, the slip-stick artist of engineering department, factory director etc. can both use " intelligence control system " to carry out real-time technique analysis and optimum decision in office separately, realized intelligence control system " trinity " application in operation control-technical Analysis-management decision.The level of IT application and the level of scientific management of iron-making production have been improved greatly.

(4) because the foundation of intelligent mathematical model, particularly Model for Multi-Objective Optimization is determined the calculating of the intelligent control partial differential equation of smelting process optimal heat state and Stochastic Optimal Control and furnace temperature, for section chief's optimal control furnace run provides scientific basis, therefore use the present invention and can obtain the remarkable benefit that improves output, cuts down the consumption of energy, guarantees molten steel quality requirement and production steady development.

(5) the present invention adapts to the equipment condition of Chinese ordinary blast." intelligence control system " not only overcome introduction " Blast Furnace Expert System " can not be applied to general medium and small blast furnace, can not all-around service in the shortcoming of section chief's operation, and in " Blast Furnace Expert System " that on blast furnace optimal heat state and the Stochastic Optimal Control model, in the design of silicon content of hot metal control partial differential equation, be better than introducing.Can substitute the Blast Furnace Expert System of import fully, and its expense only is to introduce the 1/2-1/3 of expert systems.

Description of drawings

Fig. 1 is blast furnace ironmaking process information acquisition and control flow synoptic diagram;

Fig. 2 is the personal computer net configuration schematic diagram of intelligence control system;

Fig. 3 is the processing sequence schema of intelligence control system;

Fig. 4 judges figure for the cooling range condition intelligent;

Fig. 5 is a blast-furnace body temperature field diagnostic graph;

Fig. 6 is working of a furnace stable smooth operation condition diagnosing figure;

Fig. 7 is " the furnace charge batching is checked single " of batching computation optimization;

Fig. 8 preferably schemes (furnace temperature figure) for the hot optimum regime of smelting process stove;

Fig. 9 is smelting process air quantity-ventilation property preferred figure of control (wind is schemed thoroughly);

Figure 10 is smelting process coke load and blast energy control match map;

Figure 11 preferably schemes for the combustion intensity scope of utilization coefficient the best;

Figure 12 is the preferred figure of the silicon content of hot metal [Si] of utilization coefficient the best;

Figure 13 is the preferred figure of the basicity of slag (R) of utilization coefficient the best;

Figure 14 is the preferred figure of [Si] of sulfur content of hot metal [S] best decision;

Figure 15 is the preferred figure of (R) of sulfur content of hot metal [S] best decision;

Figure 16 is batching basicity and actual basicity of slag non-linear correlation figure;

Figure 17 is furnace temperature intelligence control system figure.

Embodiment

The method of utilizing intelligence control system control blast-furnace smelting of the present invention, comprise foundation intelligence control system as shown in Figure 2, be equipped with judgement blast-furnace body working order, No. 1 active station of heat state of blast furnace and blast-furnace smelting process direct motion state (Weighing system under the monitoring groove), the master control room of the Computerized monitor system that No. 2 active stations (monitoring hoisting system) and No. 3 active stations (monitoring blast-furnace body parameter) are formed, No. 2 of optimizing station and double as server (database server 1) for No. 1 are set optimize two microcomputers of standing, these two microcomputers and 3 active stations are formed the master control room local area network, constitute the hardware platform of section chief's intellectualized operation; Simultaneously, be provided with LAN server 5 in the iron work operation department, terminal (hotblast stove terminal 6) with the sub-operation of oxygen enriched blast, groove is the terminal (terminal 7 under the groove) of the sub-operation of batching down, the terminal (elevator terminal 8) of elevator cloth material for making clothes operation, the terminal of the sub-operation of coal powder injection (coal powder injection terminal 9), laboratory terminal 10, weigh station terminal 11 and factory director office, engineering department, motor-driven section, terminal 12 networkings of administrative authoritys such as production statistics section, form the iron work local area network, and, form iron manufacturing process real time information sampling and the network system that transmits automatically with the networking of master control room local area network.Be provided with to the section chief and control the intelligence control system software module that furnace run provides operation optimizing the station for No. 1, said intelligence control system software module comprises: database module (contains raw data base, expert knowledge library, parameters optimization storehouse and inquiry, add, revise, a function such as deletion), cooling system Device Diagnostic module, the furnace condition anterograde fault diagnosis module, batching computation optimization module, optimal heat state and Stochastic Optimal Control computing module, the intelligent control module of silicon content of hot metal and sulphur content, and the managerial report automatically-generating module, the process function design of these software modules is to carry out in section chief's the operation and the ironmaking target call of realization " high-quality; low consumption; high yield; longevity " according to all-around service.Be provided with the module that shows intelligent control chart and cooling system, furnace condition anterograde fault indication chart by the smelting program circuit automatically optimizing station for No. 2.Above-mentioned server is industrial server, optimizes station, active station and terminal and all adopts the above microcomputer of Pentium III type.

The mode of operation of section chief's operation is divided into man-machine interaction and points out two kinds automatically.Working process as shown in Figure 3, the section chief is in each software module of optimizing for No. 1 on the station according to the man-machine interaction mode calling system, the intelligent control module and the task management form automatically-generating module that comprise cooling apparatus fault diagnosis module, furnace condition anterograde fault diagnosis module, batching computation optimization module, the hot state of best stove and Stochastic Optimal Control model computation module, silicon content of hot metal and sulphur content, launch function menu step by step, setting-up time and parameter as required, just can obtain the various information of blast furnace process intelligent control, instruct section chief's operation.On the other hand, owing to realized information networking, therefore, each module can onlinely be moved according to the time requirement of setting automatically, according to the latest data of up-to-date channel and online and real time data acquisition and terminal typing, calculate and judge, and optimize at No. 2 and to show intelligent control chart, cooling system failure, furnace condition anterograde fault indication chart on the station automatically, the prompting section chief in time takes some countermeasures or measure, realizes blast furnace process optimization and intelligent control.Optimize on the station at No. 2, the program automatic flow comprises: carry out the cooling apparatus automatic fault diagnosis with the relevant data that second, the level process was gathered automatically as 5 second cycle, if be diagnosed as malfunction, then show the diagnosis picture automatically, the prompting section chief takes measures to eliminate fault, otherwise does not show; Second, a level process was judged air quantity, blast, 3 sensitive parameters of ventilation property, can show automatically when judging working of a furnace failure symptom and diagnose picture, and the prompting section chief takes measures to eliminate failure symptom; Being a minute level process then, was the cycle relevant data of working of a furnace fault to be diagnosed with 3 minutes, if be diagnosed as working of a furnace failure symptom, then showed the diagnosis picture automatically, and the prompting section chief eliminates working of a furnace failure symptom, otherwise does not show; Be 30 minutes processes then, the intelligent control picture of online demonstration silicon content of hot metal and sulphur content to carrying out 3-4 control forecast during every stove iron smelting, shows 30 minutes-60 minutes-90 minutes work picture, and its output data is the foundation of section chief's regulating and controlling furnace temperature; In the laboratory terminal, the chemical composition data of laboratory technician's typing molten iron, slag, the chemical ingredients of agglomerate, pellet, rawore and coke and check data, and be sent to the blast furnace master control room automatically by network system, become the latest data source that batching is checked single calculating automatically, the calculating of silicon content of hot metal governing equation, sulphur content prediction and calculation, the hot state computation of best stove; Show thermal equilibrium and basicity balance deviation optimization range if carry out charge calculation automatically, then point out the section chief to become material and do new charge calculation according to up-to-date crude fuel composition.Every day, operation was finished, and after various work datas all entered database, the section chief can operational management form automatically-generating module, the production data of statistical summaries the day before yesterday.

Below several software modules are illustrated respectively:

1) cooling system equipment fault diagnosis module

Cooling system equipment fault diagnosis module was gathered the H of cooling stave temperature, inflow temperature, leaving water temperature(LWT), mixed gas automatically with 5 second cycle ₂The data of content, furnace body temperature, corresponding data comprehensively contrast and judge in these data and the expert knowledge library, obtain the automatic diagnosis or the warning of cooling apparatus fault.Optimize " the cooling range condition intelligent is judged figure " that the station shows automatically at No. 2, as Fig. 4, color and " condition prompting " with distinctness among the figure tell the section chief which portion water temperature difference in 9 sections cooling staves is normal, and which portion water temperature difference is too high or too low, and which part is the highest or minimum; The thermocouple which part may leak or detect is bad.When judging that furnace body temperature is error state (ERST), optimize the station at No. 2 and show " blast-furnace body temperature field diagnostic graph " automatically, as shown in Figure 5, the upper right side data sheet is each position temperature value of current actual measurement among the figure, the lower right is the optimization range of each position temperature, obtain by the self study statistics, the length of lateral axis is the ratio decision of actual value and optimal value among the state of temperature figure of the left side, the oversize then temperature of lateral axis is too high, it is too short that then temperature is low excessively, the section chief unusual place that comes into plain view and to see tens temperature datas clearly thus, thus in time take measures to be eliminated.The Intelligentized method of Fig. 4, Fig. 5 checks one by one that compared with section chief oneself whether tens item numbers certificate has normally improved working efficiency greatly, avoids the incidental error of human negligence.

2) working of a furnace fault diagnosis module

Working of a furnace fault diagnosis module is to the data of automatic data of gathering of 3 minute cycle and laboratory input, and utilization mathematical logic operational model and incident occurrence Probability Model are carried out working of a furnace trouble diagnosis; When judging that certain sign in furnace wall knot thick-dross-intractable dross, cupola well accumulation, hanging and 4 kinds of main type working of a furnace faults of pipeline may take place blast furnace, optimize the station for No. 2 and show " working of a furnace stable smooth operation condition diagnosing figure " automatically, as Fig. 6, mode with unusual working of a furnace probability of occurrence is carried out early warning, the prompting section chief notes in time taking measures, and avoids fault to take place.The judgment rule that the ironmaking expert provides directly enrolls diagnostor, and the quantity of judging is read in corresponding field in the expert knowledge library according to then in the variable mode.If the reference data in the change expert knowledge library, Zhen Duan conclusion also just changes thereupon so.Among the figure 12 parameter upset conditions are judged that the knowledge according to the ironmaking expert is divided into: rising, decline, play rise, sharp fall is indicated, and standard state does not then show.Like this, the section chief is easy to judge the parameter place of determining the generation problem the conclusion from the intellectuality of computer, thereby takes appropriate measures the elimination fault.

3) batching computation optimization module

Batching computation optimization module is changed into the mode of operation that the section chief becomes hand computation in the material " furnace charge is checked single " on No. 1 optimization station and is finished the batching computation optimization automatically.Optimize on the station at No. 1, when up-to-date-stove molten iron and the arrival of slag chemical examination compositional data are transmitted by network in the laboratory, then move the charge calculation program automatically, carry out the check of coke load-batching basicity, carry out correlation analysis with actual basicity of slag and silicon content of hot metal, judge whether current silicon content of hot metal [Si] and basicity of slag (R) depart from optimum regime.Stove heat is continuous to rise or downtrending continuously if be judged as, and departs from optimum regime, then optimizes automatic " the furnace charge batching the is checked single " calculation result that shows under the corresponding crude fuel composition in station for No. 1, as Fig. 7, points out the section chief to rerun batching computation optimization module.The section chief can enter the computation program that batching is optimized module, selects in 3 kinds of proportion schemes of choice menus, finishes " furnace charge is checked single " automatically and calculates.If select intelligence numerical procedure automatically, so then by batching basicity and coke load under the automatically definite current working of a furnace of computer.The section chief only need import these 2 parameters of charge gross weight and agglomerate ratio, and whole " furnace charge is checked single " calculated and promptly finished output result such as Fig. 7 automatically.Batching is optimized automatic calculation mathematic model formula and is comprised:

∑X _i＝W

1-X ₃/(X ₁+X ₂+X ₃)＝SLB

∑X _i×(TFe) _I×1.0645＝LW

LZR＝(∑X _i×(CaO) _I+X ₄×0.0065)/(∑X _i×(SiO ₂) _I+X ₄×0.063

+PML×0.098/LPS-LW×E([Si])×60/28)

LZR＝F(R)

W/X ₄＝FH

In the formula: X ₁For agglomerate is criticized heavily X ₂For pellet is criticized heavily X ₃For rawore is criticized heavily X ₄For dried coke (per) charge weighs.Each chemical ingredients footnote is corresponding with mineral.W is the charge gross weight, and SLB is an agglomerate ratio, and LW is the theoretical iron amount of every batch of material.PML is a coal powder injection speed, and 0.098 is the average ash of coal dust, and LPS is a hour charge number; E ([Si]) is [Si] filter value.FH is for doing burnt load, and LZR is batching basicity, and the nonlinear relationship F (R) between it and the actual basicity of slag (R) is obtained by Figure 16.The utilization mathematic programming methods can solve X ₁～X ₄

4) blast furnace optimal heat state and Stochastic Optimal Control computing module

According to the iron-smelting process principle, the model formation of setting up the blast furnace process multiple-objection optimization is as follows:

$U\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_1(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\→\mathrm{Max}$

$K\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_2(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\→\min$

$\left[S\right]\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_4(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\≤[s{]}_0$

$\left[\mathrm{Si}\right]\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_3(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\∈\left(\right[\mathrm{Si}{]}_0-a,[\mathrm{Si}{]}_0+a)$

Above-mentioned 4 formula are comprising raising output, are reducing coke ratio, sulfur content of hot metal [S] meets that steelmaking quality requires and the steady control of silicon content of hot metal [Si].In the formula, U (t) is the capacity factor of a blast furnace, Max represents maximization, K (t) is a coke ratio, and min represents to minimize, and [Si] is silicon content of hot metal (t), requirement is controlled at optimum range ([Si] 0-a, [Si] 0+a) in, [S] is sulfur content of hot metal (t), satisfies steelmaking quality and requires to be lower than limit value [S] ₀The lower limit t of the sign of integration ₀Represent the time opening and the concluding time in a certain smelting stage respectively with upper limit t, the F1 under the sign of integration, F2, F3, F4 represent complicated function association separately respectively, and wherein Y is the crude fuel parameter, and Z is a device parameter, and X is a state parameter, and u is a controlled variable,

Be respectively their velocity of variation, d τ represents the integral and calculating to the time, to U (t), K (t), [Si] (t) and [S] (t) these 4 variablees divide sample according to time series, carrying out multivariate statistics calculates, the important parameter optimal control scope and the combination thereof of determine to improve utilization coefficient, reduce coke ratio, silicon content of hot metal is suitable and sulfur content of hot metal meets the demands, its calculation result are divided into the output of 9 width of cloth pictures on No. 1 optimization station:

1) the preferred figure (see figure 8) of silicon content of hot metal-basicity of slag binary, use clustering method obtain than [Si]-(R) the optimal control scope of high utilization factor be:

Ω ₁＝{[Si]∈([Si] ₀-a；[Si] ₀+a).and.(R)∈((R) ₀-b；(R) ₀+b)}；

2) the preferred figure (see figure 9) of air quantity-permeability index binary; The application clustering method obtains than the optimal control scope of the air quantity FL-ventilation property FF of high utilization factor:

Ω ₂={ (r ₁≤ r≤r ₂) ∩ (θ ₁≤ θ≤θ ₂) in the formula $r=\sqrt{{(\mathrm{FL}-{\mathrm{FL}}_0)}^2+{(\mathrm{FF}-{\mathrm{FF}}_0)}^2}$

$\mathrm{\θ}=\arctan \frac{\mathrm{FL}-{\mathrm{FL}}_0}{\mathrm{FF}-{\mathrm{FF}}_0}$

3) blast energy-coke load match map (see figure 10) is used for preferred coke load and blast energy optimization range.Obtain than the optimal control scope of high utilization factor be:

Ω ₃={ (OPC ₀≤ OPC≤OPC ₁) ∩ (BE ₀≤ BE) } so can obtain the parameter optimization span of control of blast furnace ironmaking process optimum regime and be:

Ω＝Ω ₁∩Ω ₂∩Ω ₃

4) utilization coefficient-combustion intensity is preferably schemed (seeing Figure 11), the nonlinear relationship between its reflection utilization coefficient and combustion intensity, and as can be known when combustion intensity is controlled at a certain scope, utilization coefficient can reach maximum value by the peak value among the figure.

5) utilization coefficient-silicon content of hot metal is preferably schemed (seeing Figure 12), the nonlinear relationship between its reflection utilization coefficient and silicon content of hot metal [Si], and as can be known as [Si] when being controlled at a certain scope, utilization coefficient will reach maximum value by the peak value among the figure.

6) utilization coefficient-basicity of slag is preferably schemed (seeing Figure 13), the nonlinear relationship of its reflection utilization coefficient and basicity of slag (R), and hence one can see that when basicity is controlled at a certain scope, and utilization coefficient will reach maximum value.

7) sulfur content of hot metal-silicon content of hot metal is preferably schemed (seeing Figure 14), the quality index-sulfur content of hot metal [S] of its reflection molten iron and the nonlinear relationship of silicon content of hot metal [Si].Calculate the optimal control scope of the silicon content of hot metal [Si] that satisfies the molten steel quality requirement thus.

8) sulfur content of hot metal-basicity of slag is preferably schemed (seeing Figure 15), the quality index-sulfur content of hot metal [S] of its reflection molten iron and the nonlinear relationship of basicity of slag (R).Calculate the optimal control scope of the basicity of slag (R) that satisfies the molten steel quality requirement thus.

Synthesizing map 8～Figure 15, we can calculate and both satisfy the molten steel quality requirement, excavate the optimum control scope of silicon content of hot metal [Si] Yu the basicity of slag (R) of blast furnace production potential again as far as possible.

9) batching basicity-basicity of slag correlation diagram (seeing Figure 16), it reflects the nonlinear relationship of theoretical basicity of slag LZR and actual basicity of slag (R).Can learn thus: after best basicity of slag (R) was determined, optimum proportion basicity LZR also just determined thereupon.

Obtain the optimum regime and the Stochastic Optimal Control scope of blast furnace ironmaking process important parameter thus, and along with the variation of blast furnace condition, by self study correction parameters optimization scope.

By the optimum regime of above multiple calculated with mathematical model blast furnace process and the Stochastic Optimal Control scope of significant parameter, its result offers the blast furnace section chief with graph mode, instructs section chief's regulation and control.These seismic responses calculated are all manual can't to be realized.The section chief gets final product the optimum parameter that online rapid calculation obtains blast furnace ironmaking process as long as calling module carries out the man-machine conversation according to operation steps now, and the section chief regulates and control according to unified parameters optimal scope, can improve the technico-economical comparison of blast furnace certainly.

5) intelligent control module of blast furnace ironmaking process silicon content of hot metal

The partial differential equation that the application neural network algorithm is set up smelting process silicon content of hot metal [Si] control are cores of this technology.The intelligent control partial differential equation of being set up are:

$\frac{\∂\left[\mathrm{Si}\right]}{\∂t}=A\frac{\∂\mathrm{LS}}{\∂t}+F\frac{\∂\mathrm{FF}}{\∂t}+C\frac{\∂\mathrm{TLC}(t-{\mathrm{\τ}}_1)}{\∂t}+Q\frac{\∂\mathrm{MQ}}{\∂t}+L\frac{\∂\mathrm{FL}}{\∂t}+W\frac{\∂\mathrm{FW}}{\∂t}+B\frac{\∂\mathrm{PMB}}{\∂t}+S\frac{\∂\mathrm{FS}}{\∂t}$

In the formula

Be the variable quantity of silicon content of hot metal,

Be the fast variation index of material,

Be the ventilation property variation index,

Be iron amount difference index, τ ₁Be amount time lag, Be the gas utilization rate variation index, Be the air quantity variation index,

Be the wind-warm syndrome variation index, Be coal powder injection speed variation index,

Be the rheumatism variation index, A, F, C, Q, L, W, the every constant coefficient of B, S then are the data with the size of a sample of a certain setting, the regression coefficient that obtains according to the multiple regression self study.The result that this Equation for Calculating obtains is presented at No. 2 with the form of " furnace temperature intelligence control system figure " and optimizes on the screen of station, and follow the operation process of smelting, carried out prediction and calculation and frame update in per 30 minutes, dynamically control the indicating instrument of furnace temperature as the section chief, the control furnace temperature develops towards optimum regime.

Figure 17 is furnace temperature intelligence control system figure.The data of this figure per 30 minutes data and laboratories according to up-to-date automatic collection, the station terminal typing of weighing refresh once automatically, for section chief's regulating and controlling furnace temperature provides Operating Guideline.The top of Figure 17 is silicon content [the Si]-state of sulphur content [S] of nearest 4 stove molten iron and the prediction of next stove.The left side of Figure 17 is the change histogram of 3 state variabless; The right is the change histogram of 3 controlled variable.Middle then be the state and control change of judging material speed-ventilation property-air quantity-coal powder injection according to back 30 minutes-60 minutes-90 minutes branches of tapping a blast furnace for 3 times automatically, the developing direction of furnace temperature and to the suggestion of controlled variable (coal powder injection-air quantity-wind-warm syndrome) Control Countermeasure comprises autonomous conclusion of judging of section chief and control effect.

According to the characteristics of iron-smelting process, the 30 fens clock times in every interval are optimized the calculating of automatically carrying out a silicon content of hot metal [Si] partial differential equation on the station at No. 2.Like this, during a stove iron is smelted (about 2 hours), can carry out predictive control 3 times.It has changed the passive situation that original various furnace temperature forecasting model for once forecasts and verifies during a stove iron is smelted.In 3 predictive control, if the control of front time change does not meet the optimal control requirement of furnace temperature in partial differential equation calculation result, so can Correction and Control in half an hour subsequently.One stove iron smelting process can have the chance of 3 Correction and Control, can guarantee that furnace temperature develops towards optimization range.

Test shows, use the present invention the steady controlled levels of blast furnace temperature is improved, comprise that the number of times of working of a furnace fault such as hanging significantly reduces, the hit rate of silicon content of hot metal [Si] predictive control reaches more than 85%, obtain the raising utilization coefficient, reduced the remarkable technical economic benefit of coke ratio.

Claims (3)

1. method of utilizing intelligence control system control blast-furnace smelting, comprise and utilize the online dynamic acquisition blast furnace of transmitter data, utilization is equipped in the Computerized monitor system of the active station composition of blast furnace main control room and judges the blast-furnace body working condition, heat state of blast furnace and blast-furnace smelting process direct motion state, this method also is included in the blast furnace master control room and No. 2 of optimizing station and double as server for No. 1 are set optimize two microcomputers of standing, these two microcomputers and active station are formed the master control room personal computer net, constitute the hardware platform of section chief's intellectualized operation; At iron work server is set, terminal with the sub-operation of batching under the groove, the terminal of elevator cloth material for making clothes operation, the terminal of the sub-operation of oxygen enriched blast, the terminal of the sub-operation of coal powder injection, laboratory terminal, and the terminal networking of factory director office, engineering department, motor-driven section, production statistics section, operation department administrative authority, form the iron work local area network, and network, form iron manufacturing process real time information sampling and the network system that transmits automatically with the master control room local area network; And be provided with and comprise database module optimizing the station for No. 1, cooling system Device Diagnostic module, the furnace condition anterograde fault diagnosis module, batching computation optimization module, optimal heat state and Stochastic Optimal Control computing module, the intelligent control module of silicon content of hot metal and sulphur content, and the intelligence control system software module of managerial report automatically-generating module are provided with by smelting program circuit and show the module of intelligent control chart and cooling system, furnace condition anterograde fault indication chart automatically optimizing the station for No. 2; Optimize the station for No. 1 and optimize stations for No. 2 and follow blast-furnace smelting process real-time period automatic data collection, the shift foreman calls No. 1 by human-computer interaction interface and optimizes each software module of standing, and obtains also guiding operation in view of the above of real-time evaluation; Optimize the station No. 2, with second the level process be automatically to gather the cooling apparatus data cycle to carry out trouble diagnosis, show at picture, with second the level process to be the cycle carry out fault to air quantity, blast, three sensitive parameters of ventilation property judges, show at picture, be the cycle to gather working of a furnace data automatically with a minute level process, diagnosis working of a furnace failure symptom, show at picture, the intelligent control picture of 30 minutes online demonstration silicon content of hot metal of process and sulphur content, its output data is the foundation of section chief's regulating and controlling furnace temperature; In the laboratory terminal, the chemical composition data of laboratory technician's typing molten iron, slag, the chemical ingredients of agglomerate, pellet, rawore and coke and check data, and be sent to the blast furnace master control room automatically by network system, become the latest data source that batching is checked single calculating automatically, the calculating of silicon content of hot metal governing equation, sulphur content prediction and calculation, the hot state computation of best stove; Show thermal equilibrium and basicity balance deviation optimization range if carry out charge calculation automatically according to up-to-date crude fuel composition, then point out the section chief to become material and do new charge calculation, it is as follows to it is characterized in that having set up the Model for Multi-Objective Optimization formula in the blast furnace optimal heat state of optimizing the station for No. 1 and Stochastic Optimal Control module:

$U\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_1(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\→\mathrm{Max}$

$K\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_2(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\→\min$

$\left[\mathrm{Si}\right]\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_3(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\∈({\left[\mathrm{Si}\right]}_0-a,[\mathrm{Si}{]}_0+a)$

$\left[S\right]\left(t\right)=\underset{t_0}{\overset{t}{\∫}}{F}_4(Y,\frac{\∂Y}{\∂\mathrm{\τ}},Z,\frac{\∂Z}{\∂\mathrm{\τ}},X,\frac{\∂X}{\∂\mathrm{\τ}},u,\frac{\∂u}{\∂\mathrm{\τ}})\mathrm{d\τ}\≤{\left[S\right]}_0$

In the above-mentioned formula, U (t) is the capacity factor of a blast furnace, and Max represents maximization, and K (t) is a coke ratio, and min represents to minimize, and [Si] is silicon content of hot metal (t), requires to be controlled at optimum range ([Si] ₀-a, [Si] ₀+ a) in, [S] is sulfur content of hot metal (t), satisfies steelmaking quality and requires to be lower than limit value [S] ₀The lower limit t of the sign of integration ₀Represent the time opening and the concluding time in a certain smelting stage, the F under the sign of integration respectively with upper limit t ₁, F ₂, F ₃, F ₄Expression complicated function association separately, wherein Y is the crude fuel parameter, and Z is a device parameter, and X is a state parameter, and u is a controlled variable,

Be respectively their velocity of variation, d τ represents the integral and calculating to the time, to U (t), K (t), [Si] (t) and [S] (t) these 4 variablees divide sample according to time series, carrying out multivariate statistics calculates, the important parameter optimal control scope and the combination thereof of determine to improve utilization coefficient, reduce coke ratio, silicon content of hot metal is suitable and sulfur content of hot metal meets the demands, its calculation result is divided into the output of 9 width of cloth pictures on No. 1 optimization station: 1) silicon content of hot metal-basicity of slag binary is preferably schemed; 2) air quantity-permeability index binary is preferably schemed; 3) blast energy-coke load match map; 4) utilization coefficient-combustion intensity is preferably schemed; 5) utilization coefficient-silicon content of hot metal is preferably schemed; 6) utilization coefficient-basicity of slag is preferably schemed; 7) sulfur content of hot metal-silicon content of hot metal is preferably schemed; 8) sulfur content of hot metal-basicity of slag is preferably schemed; 9) batching basicity-basicity of slag correlation diagram; Obtain the optimum regime and the Stochastic Optimal Control scope of blast furnace ironmaking process important parameter thus, and along with the variation of blast furnace condition, by self study correction parameters optimization scope;

It is as follows to have set up the partial differential equation formula of using neural network algorithm calculating in the silicon content of hot metal intelligent control module:

$\frac{\∂\left[\mathrm{Si}\right]}{\∂t}=A\frac{\∂\mathrm{LS}}{\∂t}+F\frac{\∂\mathrm{FF}}{\∂t}+C\frac{\∂\mathrm{TLC}(t-{\mathrm{\τ}}_1)}{\∂t}+Q\frac{\∂\mathrm{MQ}}{\∂t}$

$+L\frac{\∂\mathrm{FL}}{\∂t}+W\frac{\∂\mathrm{FW}}{\∂t}+B\frac{\∂\mathrm{PMB}}{\∂t}+S\frac{\∂\mathrm{FS}}{\∂t}$

In the formula

Be the variable quantity of silicon content of hot metal, Be the fast variation index of material, Be the ventilation property variation index,

Be iron amount difference index, τ ₁Be amount time lag, Be the gas utilization rate variation index, Be the air quantity variation index, Be the wind-warm syndrome variation index,

Be coal powder injection speed variation index,

Be the rheumatism variation index, A, F, C, Q, L, W, the every constant coefficient of B, S then are the data with the size of a sample of a certain setting, the regression coefficient that obtains according to the multiple regression self study; The result that this Equation for Calculating obtains is presented at No. 2 with the form of furnace temperature intelligence control system figure and optimizes on the screen of station, and follow the operation process of smelting, carried out prediction and calculation and frame update in per 30 minutes, dynamically control the indicating instrument of furnace temperature as the section chief, the control furnace temperature develops towards optimum regime.

2. a kind of method of utilizing intelligence control system control blast-furnace smelting according to claim 1 is characterized in that cooling system Device Diagnostic module is with the automatic H that gathers cooling stave temperature, inflow temperature, leaving water temperature(LWT), mixed gas of 5 second cycle ₂The data of content, furnace body temperature, corresponding data comprehensively contrast and judge in these data and the expert knowledge library, obtain the automatic diagnosis or the warning of cooling apparatus fault, optimize the station at No. 2 and show the cooling stave trouble-shooting charts; When judging that furnace body temperature is error state (ERST), optimize the station at No. 2 and show furnace body temperature field diagnostic graph automatically; The prompting section chief in time takes measures, and fixes a breakdown.

3. a kind of method of utilizing intelligence control system control blast-furnace smelting according to claim 1, it is characterized in that the data of furnace condition anterograde fault diagnosis module to automatic data of gathering of 3 minute cycle and laboratory input, utilization mathematical logic operational model and incident occurrence Probability Model are carried out working of a furnace trouble diagnosis; When judging that blast furnace furnace wall may take place is tied certain sign in thick dross, cupola well accumulation, hanging, 4 types of working of a furnace faults of pipeline, optimize the station for No. 2 and show working of a furnace trouble-shooting chart automatically, mode with the fault probability of occurrence is carried out early warning, the prompting section chief notes in time taking measures, and avoids fault to take place.

Images