CN1097804A - Computerized blast furnace smelting expert system method - Google Patents
Computerized blast furnace smelting expert system method Download PDFInfo
- Publication number
- CN1097804A CN1097804A CN 93108671 CN93108671A CN1097804A CN 1097804 A CN1097804 A CN 1097804A CN 93108671 CN93108671 CN 93108671 CN 93108671 A CN93108671 A CN 93108671A CN 1097804 A CN1097804 A CN 1097804A
- Authority
- CN
- China
- Prior art keywords
- blast furnace
- data
- model
- furnace
- reasoning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The present invention is a kind of method of comprehensive control blast furnace process, image data from the blast furnace transmitter; According to expert's the experience and knowledge and the knowledge base of fuzzy relation matrix formation; Data to online acquisition compare, reasoning, judge the direct motion state and the hot state of blast furnace state, furnace run.The result realizes that blast furnace process is regulated and control according to weather report.
Description
The present invention relates to the computer expert system of the condition predicting of blast-furnace smelting pig iron process, be used for the Operating Guideline and the management of blase furnace cast iron smelting process.
The seventies, blase furnace cast iron was smelted quite maturation of mathematical model.Its weakness is a bad adaptability.The working of a furnace just often, hit rate is higher, but the working of a furnace changes or when unusual condition occurring, hit rate is just very low.This is because blast furnace ironmaking process is a very complicated combined process.Smelting is to carry out in the body of heater of sealing, and the operator can only judge situation in the stove by instrument and experience.And the data that mathematical model of the computer can only rely on instrument to detect.The instrument data presented is the situation of the transmitter regional area that can reflect just often.When working of a furnace variation or generation unusual condition, the variation of local parameter is difficult to reflect exactly the overall condition in the stove.Therefore the hit rate of mathematical model reduces greatly.Use advanced computer and artificial intelligence technology to set up blase furnace cast iron and smelt expert systems, become anxious problem to be solved.It is exactly that a kind of blase furnace cast iron is smelted expert systems (871A09036) that the NKK that on December 23rd, 1987, Chinese Patent Office announced applies for a patent " method of control operation of blast furnace ".The data of sending here with the giant-powered computer processes sensor with determining the factor (CF) value, are inferred the blast furnace unusual conditions such as accident, hanging and pipeline that whether will meet accident.July 1 nineteen ninety Patent Office to disclose " blast furnace operation management method and the device " of Nippon Steel also be a kind of Blast Furnace Expert System (CN1043745A).This application is divided into the rule in the knowledge base: whether need take the passive rule of passive action for avoiding the blast furnace fault, with the active rule of the initiatives of taking for the reduction running cost, and judge whether that the distribution that needs to change the charge distribution behavior improves rule.Computer is at first carried out the deduction of passive rule.If need take passive action, then indicate corresponding passive action, and stop inferring; If do not need to take passive behavior, then carry out the deduction of relevant active rule.If need take passive action, then indicate corresponding passive action, and stop inferring; If need not take the initiative behavior, then carry out the deduction that relevant batching is improved rule.Improve action if need take to distribute, then indicate corresponding improvement action, and stop inferring; If need not take action, then send the indication that keeps present situation.
More than two background technology files the technical conceive of blast furnace dynamic mathematical models is only disclosed roughly, do not possess the function scheme of artificial intelligence, and various concrete technical parameters openly do not come out.Expert of this area or advanced techniques personnel will implement this technology and need many technical characterictics be proposed could implement after the concrete scheme through the creative work of oneself.Those skilled in the art without the creative work of oneself, can not realize.
Purpose of the present invention, it is exactly deficiency at the above-mentioned background technology, provide a kind of those skilled in the art to realize, judge that the blast-furnace smelting shape is direct motion, still take place hanging, pipeline maybe will take place, be difficult to walk, judge that blast furnace temperature is overheated, thermotropism or cold to cold or play, judge that the blast furnace knot is thick, still bonding situation or burn the blase furnace cast iron smelting artificial intelligence expert systems that waits, to strengthen the operational management of blast furnace, correctly diagnose the working of a furnace, in time remind the section chief to take adequate measures, avoid having an accident, the production of assurance blast furnace is normally carried out.
It is the computer-controlled important component part of blast furnace that blase furnace cast iron is smelted expert systems.For working conveniently, operation of blast furnace parameter detecting and control are divided into four groups: blast-furnace body, hotblast stove, material loading and coal powder injection.As shown in Figure 1, each group all is provided with commercial run controlling machine able to programme (PLC), each organize check point with main production run parameter with curtage etc., send process controller (PLC) to through automatically controlled instrument (DDC).Process controller by distributing system is to automatically controlled, instrument control signal detects, calculate, print, show, processing such as warning, show production operation parameter and point-to-point communication, computer management is carried out in production, simultaneously that mathematical model is the required simulate amount and the data of switching value are sent to medium-mini computer (following title upper computer), the task of upper computer, except that serving as the work of blast furnace digital-to-analogue, also the data with blast furnace regularly are delivered to factory department, by the terminal microcomputer, by the requirement of computer management network, the general data of blast furnace production is sent into company computer's managerial grid.In order to be production and scientific research service, upper computer has also disposed microcomputer, the online production data can be accessed to carry out off-line application processing.Upper computer is except that the automatic detection information of accepting from blast furnace, also accept the essential information sent into by terminating machine, as blast furnace main control room on duty, the chemical ingredients of slag, molten iron and all kinds of feed stock for blast furnace that auxiliary posies such as the information that watch-keeping cubicle such as hotblast stove, material loading, coal powder injection are sent into, laboratory, raw material station are sent into.
Expert systems of the present invention is divided into the judgement of direct motion situation, the body of heater state is judged and the hot state of stove is judged three subsystems.The alleged smooth operation of furnace situation of the present invention is judged subsystem, it is the person that utilizes the blast furnace operating and ironmaking expert's experience, judge the whether direct motion of operation of blast furnace state, and to hanging (being difficult to walk), the material that collapses, pipeline, the development of edge air-flow, the intelligent program that the unusual working of a furnace such as central gas stream development is predicted and judged; The blast furnace state judges that subsystem is furnace wall to be corroded monitoring and burns monitoring, the judgement intelligent program of the forecast that bearth damage and furnace wall knot are thick; Heat state of blast furnace judges that subsystem is to judge that blast furnace temperature is a thermotropism, overheated, normal, still to cool, acute cold situation.No matter be which kind of to carry out judge, all use inference engine (as shown in Figure 2), system is gathered and pretreated real time data automatically, compare, make inference and judge, and the result is stored and output with the knowledge or the experience that are stored in the knowledge base.Blast furnace operating person or smelting expert can be from the situations of man-machine interface understanding system operation.This model user window mainly is with the lower section:
1, deposits data, the conclusion of using in the reasoning process in the disk (see figure 3), so that the expert retrieves, verifies whether the rule of being set up is correct, and whether conclusion is reliable.
2, large color screen graphic presentation (see figure 8), this part mainly are that the ultimatum that the expert systems reasoning draws is shown the types of facial makeup in Beijing operas with the form that personalizes on large color screen.Represent respectively that by smiling face, serious, the face of crying blast furnace is various in good condition, normal and undesired sign occurs.
3, curve display (seeing that production management, blast furnace operating among Fig. 3 show) is drawn as curve to final judged result, the variation of various states in directly reflection is come out of the stove.
Table 1, provided under the blast furnace different states, the response that each parameter is made, i.e. the knowledge frame of expert systems, it be conclude, arrangement blast furnace operating expert's experience and the working of a furnace set up and the logical relation of parameter.
Fig. 3 is expert systems flow chart of data processing figure.
The Blast Furnace Expert System mathematical model uses the data overwhelming majority on-the-spotly to collect by the process controller real-time online from smelting, and does not gather then because of technology or equipment reason computer before the sub-fraction data to be required in accordance with regulations desired data to be imported computer by the people on time.Fig. 4 express upper computer to process controller beat the collection data owner to be divided into four big systems.Their coal injection system, blast-furnace body system, hot blast stove system and feeding system.Coal injection system mainly is to give upper computer with upper and lower jar weight, pressure and coal powder injection total pressure by one minute periodic transfer.And the blast-furnace body system need give upper computer by five seconds, thirty second, one minute and five minutes periodic transfer according to production management and digital-to-analogue, charge signal for example: about the material chi be to adopt with 5 seconds cycle lifies, and main parameters such as the main manufacturing parameter blast of blast furnace, air quantity, top temperature, oxygen amount, hotblast stove are adopted with 30 second cycle number.Same raw material signal also is to gather with 30 second cycle as ore, coke, assorted ore deposit weight signal, and mass data such as molten iron signal, cross temperature, mixed gas CO, CO
2, H
2Content, stove in static pressure, soft melt zone temperature, hotblast stove main parameters adopt with one minute cycle number, and change comparatively slowly wall temperature behind the water tank, temperature of the furnace hearth etc. are then adopted with five minute cycle number.Above-mentioned data major part leaves in the internal memory shared region of upper computer after the collection of process machine at once.And signals such as some data such as raw material, coke, material kind, weight must be earlier screen, smoothly data, check that signal whether in the trusted scope, then handles again, judges whether in the blanking process, just in blanking.Blanking finishes, and follows the trail of and judge additional coke weight, material kind, weight signal constantly, and has the data special data area of shared region in these signals are left in.Equally, wall temperature etc. and coal injection system data also must be carried out data screening respectively behind blast-furnace body data such as the water tank, remove unreasonable signal, and some weight signal is sued for peace and waited processing, yet require minute one hour and four hours periods are deposited according to digital-to-analogue.And then it is sued for peace calculating such as maximizing (MAX), minimum value, variance.The data that selecting each model program again needs are handled by rule request in the expert system knowledge base, leave the digital-to-analogue special data area in the internal memory shared region in accordance with regulations in.Share data and judgement for subsystem model provides, use information.In addition, the mass data of blast-furnace body hotblast stove, feeding system is that signals such as air quantity, blast, wind-warm syndrome, roof pressure, permeability index, top temperature are also through carrying out the production management of computer behind the data screening.Production management is to provide display data to the blast furnace production operation, and need print production report by producing.The data that leave the special-purpose internal memory shared region of digital-to-analogue in are shared data that all submodel operations need, they are through the operation of process separately, carry out logic judgement, reasoning according to knowledge base knowledge, then again with the sign of judged result, the data of enjoying for other model are returned to the internal memory shared region, in disk file, structured data repository is seen Fig. 4 with the deposit data that provides history lookup and model self study to use.
Fig. 4 is a Blast Furnace Expert System mathematical model structured data repository form, because this system is the real-time online operation, for guaranteeing all process operations, this system adopts the internal memory shared region to deposit shared, frequently-used data, need storage and call data, rule, then adopt file mode to deposit.
The production process of blast furnace is controlled by process controller (PLC).Blast furnace manufacturing parameter by the controlling machine collection is sent into upper computer by communication cable, leaves in the internal memory shared region.The data that process machines such as some data such as raw material composition, output, silicon content of hot metal can not be gathered are automatically then imported in the computer by hand by the section chief.By to processing, handle, form the production management database with the data of the relevant production management of artificial input in the internal memory shared region.Become material situation, files such as operating duty as stokehold cast iron ingredient, raw material composition, charging.According to the needs of producing needs and mathematical model original production data in the internal memory shared region was gathered once in per two minutes, kept nearest seven days data, form main manufacturing parameter data file.To wall temperature behind all kinds of temperature values such as the water tank in the internal memory shared region, temperature data file commonly used is collected, puts in order, formed to temperature of furnace wall, temperature of the furnace hearth etc.
As mentioned above, blast furnace smelting expert system is by direct motion state subgroup system, and hot state subgroup system and blast furnace state subgroup system form, and the data that each subsystem processes is crossed leave in the corresponding database with document form.In the direct motion slip condition database, include direct motion and give the processing file, direct motion Status Flag file, direct motion intermediate result file and direct motion state outcome file etc.Wherein direct motion is given and is handled file and mainly deposit parameters such as air quantity, blast, permeability index, material speed are given data after the processing, and signs such as direct motion Status Flag file is mainly deposited the blast furnace blowing-out, changed stove, multiple wind reach to blast furnace the set sign of some abnormal conditions takes place.Direct motion intermediate result file is mainly deposited the intermediate result of calculating or judging, direct motion state outcome file is mainly deposited the judged result and the various data relevant with the result of direct motion model.In hot slip condition database, include hot state and give the processing file, hot state intermediate result file and hot state outcome file etc.Wherein hot state gives to be handled file and mainly deposits hot state model desired parameters is given data after the processing, hot state intermediate result file is mainly deposited the intermediate result of calculating and judging, heavily reaches coal dust amount etc. as molten iron production rate, attached coke ratio, wind-warm syndrome changing conditions, ore deposit.Hot state outcome file is mainly deposited the judged result of thermal model, various signs and corresponding data.In the body of heater slip condition database, include body of heater state outcome file etc.Wherein the body of heater state is handled file and is mainly deposited the mean value that air quantity, permeability index, material chi, temperature of furnace wall are given birth to isoparametric hour, two hours or eight hours.The intermediate result file is mainly deposited the intermediate result of computation process, as degree of membership of direct motion index, parameter etc.Body of heater state outcome file is mainly deposited the end value of body of heater state model.Furnace wall is tied thick value and is burnt value, direction etc., is beneficial to the preservation and the inquiry of data.Fig. 5 is the expert system knowledge base schema.
Knowledge base is the important component part of expert systems, is that each model is in order to carry out the foundation of reasoning, judgement.Because the problem solving of expert systems, be that the mode of thinking that the utilization expert provides special knowledge to simulate the expert is carried out, so knowledge is the key factor that determines that an expert systems performance is whether superior.The ability level of an expert systems just depends on the quality and quantity of knowledge that its knowledge base is deposited.
In order to set up the Action Target of knowledge base according to expert systems, extensively collect operative knowledge and practical experience that blast-furnace smelting knowwhy and smelting expert and site operation personnel enrich, arrangement forms elementary knowledge base.According to computer acceptable mode these elementary knowledge are converted into then that computer can store and the knowledge that can act in accordance with it leaves in the internal memory, form knowledge base, this processing treatment mainly contains following several mode:
1, set up judgment rule:
Rule of the present invention is based on the production representation method of knowledge:
R#:IF RLS THEN RRS
In the formula:
R#: the sequence number of production rule in rule base;
RLS: the conditions of operation;
RRS: judged result truth and false.
The implication of following formula is if working of a furnace operational conditions RLS is met, can think that then conclusion RRS sets up.
According to the production representation method of knowledge, the present invention is processed into a large amount of rules with expertise, deposits in formation rule knowledge base in the file, in order to using when model carries out reasoning and judging.
For example:
RVLE 11010S IF SB≥9.5 AND GPI↓6%AND TGT≥270 THEN hanging?CF=1.0
The practical significance of this rule is: when material speed (SB) 〉=9.5 minute/batch, and permeability index (GPI) is than last time descending 6% and furnace top gas temperature (TGT) 〉=270 ℃, when these three conditions satisfied simultaneously, the possibility that hanging takes place was 1.0, i.e. system's judgement hanging must take place at this moment.
Undoubtedly, the accuracy of model judgement is closely related with the accuracy of the formulation of rule.
2, set up experimental formula
In order accurately to describe the level and the variation tendency of a certain status flag of blast furnace, the present invention is by theoretical analysis or use expertise, with relevant with it various smelting parameters, sets up one through appropriate combination and describes formula, calculate characteristic parameter, formulistic knowledge base that Here it is.
Formula constitutes can be divided into two classes: a class is empirical, promptly obtains from long term production; On the other hand, derive in conjunction with the actual conditions of blast furnace based on the blast-furnace smelting principle.The former plays a part bigger in the judgement of heat state of blast furnace in the practical application.The latter obtains unusual effect in the body of heater state is judged.
3, the Fuzzy processing of parameter
In the conditions of blast furnace diagnosis, exist many unconventional property problems, can't obtain a result with simple accurately reasoning.And must adopt the membership function in the fuzzy mathematics that uncertain knowledge is carried out quantification treatment.We have adopted the membership function of various ways in actual applications.For example: the direct motion index, burn index, knot thick index etc., in this model, many variablees have been set up membership function, carried out the quantification treatment of uncertain knowledge.
4, set up fuzzy relation matrix
In blast furnace process, the relation between many states and the detect parameters is difficult to make sure description with simple mode, and we cry fuzzy relation these relations.We represent with above-mentioned membership function for fuzzy relation.Need this is built up a fuzzy eigenvector after having off status to calculate degree of membership, set up the fuzzy relation matrix between a state and the parameter simultaneously, find the solution by fuzzy relation equation with membership function.If the index of reflection state is Ij, there is Several Parameters at a time the degree of membership vector of this state to be A, then:
A=(a1,a
2,a
3……a
n
In the formula, a1, a
2, a
3A
n, be respectively the element in the degree of membership vector of each corresponding parameter.
If R is a fuzzy relation matrix, then:
Ij=A·R
Like this, find the solution, uncertain relation is converted to definite estimation by fuzzifying equation.
With the thick forecast of furnace wall knot is example, and fuzzy reasoning process is described:
If it is Ijh that furnace wall is tied thick index, there be n parameter relevant with Ijh.With u(Xi) represent the degree of membership of each parameter, W(Xi) expression fuzzy relation then gets according to fuzzy relation:
n
Ijh=∑μ(Xi)·W(Xi)
i=1
Wherein, the Ijh-furnace wall is tied thick index
The degree of membership of μ (Xi)-each parameter amount
W(Xi)-fuzzy relation
By the following formula operation result, obtain tying thick predicted value.
The key of above-mentioned reasoning is determining of degree of membership threshold value or threshold value.We determine that the method for degree of membership has two:
(1), empirical value: the empirical value with expert's accumulation is a basis;
(2), be foundation with a large amount of statistical treatment production datas of Bayesian formula.
5, a small amount of relation of determining.Adopt finite element method to calculate variation tendency, for example to corrode be that the value calculated with finite element method is as main value, in conjunction with its dependent variable diagnosis forecast together for cupola well, furnace bottom.
At this moment cupola well, furnace bottom section are divided into some nodes, as 366 nodes, 642 unit, the relevant monitoring parameter by database storage is found the solution the Two-Dimensional Heat equation as design conditions:
Can obtain cupola well, furnace bottom and corrode cross hatching, and the cupola well that draws, furnace bottom change figure.
At present, it is 10 days that cupola well, furnace bottom corrode the variation tendency computation period, and this is that this expert systems is the longest one in all forecast cycles.Reason is the blow-on initial stage, corrodes still in safety range too short having little significance of cycle.In the later stage of blast-furnace production campaign, computation period will shorten accordingly.
For above-mentioned conclusion, the knowledge that gathers, process engineer according to the needs of model development with the Computer Engineer, weigh, put in order, sort out to divide and carry out mathematical derivation, calculating, and the knowledge abstraction of knowledge base is become to have derivation tree of logical relation or the like.All above-mentioned work all are to finish in the programming stage.
Then, the machine function that the Computer Engineer provides according to employed upper computer and the requirement of system design of expert systems, with the form input computer of above-mentioned knowledge with program, build up the knowledge base of a multiple structure of knowledgeization, needs according to direct motion state, body of heater state, hot state model, be divided into knowledge base 1, knowledge base 2 ... knowledge base n, and for identical, be public knowledge base just as inference engine part knowledge base.
Call the knowledge of knowledge base when the computer run Blast Furnace Expert System, utilize the inference engine (see figure 6) to carry out reasoning.The result of reasoning is the integral part that model running is judged.The result who draws the blast furnace judgement provides the program of blast furnace judgement graphic presentation to use (see figure 8), and comprehensively judges and Operating Guideline at the picture directly perceived of blast furnace master control room.
Process engineer and Computer Engineer can be according to blast furnace practical condition and model running and predicting condition, compare item by item, the suggestion of refinement is carried out in proposition to original expert systems, and the function by the artificial intelligence self-teaching in this Blast Furnace Expert System mathematical model, make amendment additional, improve work such as upgrading, thereby reach constantly perfect, improve service efficiency, improve the hit rate of blast furnace mathematical model forecast knowledge base.
Fig. 6 is the inference engine structure iron of blast furnace smelting expert system.Reasoning process is as follows:
At first, determine the current target of wanting reasoning.After the reasoning target is determined, enter reasoning.Inference engine comprises inference method and control strategy two portions.
Send out for first one and be divided into inference method.The inference method of expert systems generally can be divided into two kinds of accurate reasoning and inexact reasonings, owing to have many uncertain knowledge in the blast furnace process, the present invention has selected inexact reasoning.
In inexact reasoning, according to indetermination theory, its general type is: IF E1 AND E
2AND ... AND En THEN H(X)
Ei(i=1,2 wherein ... n) be argument, H is one or more conclusions.
Its regular connotation is: when evidence E1 when En exists really, the confidence value that conclusion H sets up is X.X is a CF, and the value of X changes in [1,1] scope.
Calculate confidence level CF, CF and calculate, wherein use the notion of trusting the growth degree and distrusting the growth degree by following method.
MB[H, E] (〉=0) for distrusting the growth degree, expression increases hypothesis H for really trusting the growth degree because of the appearance of evidence E, be P(H/E)>P(H), MD[H, E] (〉=0) for distrusting the growth degree, expression increases the distrust growth degree of hypothesis H for vacation because of the appearance of evidence E, be P(H/E)<P(H), seek out confidence level CF by the two.CF is defined as follows:
CF[H.E]=MB[H.E]-MD[H.E]
When obtaining ultimatum with the notion of confidence level, can be divided into evidence is the situation of a plurality of conditions and single condition.Therefore, judge when whether argument is the situation of single argument regular IF E THEN H(X earlier) employed.The calculation formula of confidence level is as follows:
CF[H]=CF[H·E]·MAX{O,CF(E)}
Get O and CF[E in the formula] peaked meaning be: if CF[E] be entitled as vacation less than 0(promptly), illustrate that this rule can not use the CF[H that obtains in other words] formerly equal 0.Otherwise the confidence level of conclusion H equals the confidence level that the confidence level of rule multiply by argument.
When argument is not single, program enters the judgement that evidence is the logical combination of a plurality of conditions.It is the situation that AND connects that light is declared argument.If then the AND of a plurality of conditions is connected to following form:
IF E1 AND E
2AND ... AND En THEN H(X) then
CF[E]=CF[E1 AND E
2AND……AND En]
=MIN{CF[E1],CF[E
2]……CF[En]}
The result send the conclusion of output reasoning.
Argument is that also can be divided into argument be that OR connects to the logical combination of a plurality of conditions.
When a plurality of conditions connections were not AND, program entered OR and judges.If then the OR type of attachment of a plurality of conditions is as follows:
E=E1 OR E
2OR……OR En
CF[E]=CF[E1 OR E
2OR……OR En]
=MAX{CF[E1],CF[E
2]……CF[En]}
The result send the conclusion of output reasoning, when a plurality of arguments are not AND and OR, carry out two rules and has identical conclusion, that is:
IF?E1?THEN?H[CF(H·E1)]
IF E
2THEN H[CF(HE
2)] judgement.If not then reasoning finishes.
When two rules and when same conclusions is arranged, obtain respectively with the formula of front:
CF1[H]=CF(H,E1)·MAX{O,CF[E
2]}
CF1[H]=CF(H,E
2)·MAX{O,CF[E
2]}
According to two confidence levels that top formula is calculated respectively, obtain total confidence level according to following several situations again, judge CF1[H] ﹠amp; CF
2[H] 〉=0 situation,
As CF1[H] ﹠amp; CF
2Calculate [H] 〉=0 o'clock
CF12[H]=CF1[H]+CF
2[H]-CF1[H] CF
2[H], and the conclusion of output reasoning.
As CF1[H] ﹠amp; CF
2[H] 〉=0 o'clock enters judgement
CF1[H] ﹠amp; CF
2[H]<0 situation,
As CF1[H] ﹠amp; CF
2[H]<0 o'clock, calculate:
CF12[H]=CF1[H]+CF1[H]+CF1[H] CF
2[H] also exports the reasoning conclusion.
When not being above-mentioned two kinds of situations, then calculate
CF12[H]=CF1[H]+CF
2[H] also exports the reasoning conclusion.
Second section is: control strategy.Control strategy is the one other component in the inference engine.Mainly be the control of inference direction and the selection strategy of inference rule.Forward reasoning is wherein arranged, backward reasoning and forward and reverse mixed inference, what this model adopted is the method for forward reasoning.
Forward reasoning is set out by raw data and is carried out, promptly by raw data 1, raw data 2 ... raw data n is according to reasoning.
According to raw data, expert system knowledge in the utilization knowledge base could carry out reasoning, can not carry out reasoning and then finish.Can carry out, carry out in the process of reasoning at the utilization expertise, need to calculate the degree of membership of desired parameters, the pass between them is that AND connects:
CF[E1]CF[E
2]……CF[En]
By the degree of membership that each calculation of parameter is come out, calculate CF[E according to the AND relation], promptly
CF[E]=MIN{CF[E1], CF[E
2] ... CF[En] } and export the reasoning conclusion.
The pass of calculating between they of desired parameters is that OR connects degree of membership
According to the degree of membership CF that calculates, connect to OR according to concerning between them again and ask total degree of membership:
CF[E]=MAX{CF[E1]CF[E
2]……CF[E
2]}
According to the result that above-mentioned inference method and control strategy draw, the conclusion of output reasoning.
If the conclusion that reasoning obtains is consistent with the actual working of a furnace, then reasoning finishes.
If the conclusion and the actual working of a furnace that reasoning obtains are inconsistent, then knowledge base is carried out refinement, carry out model self study, the threshold value of adjustment model parameters needed and rule simultaneously.
Fig. 7 is the judgment result displays schema.
The end value that color graphic display system draws each model running, handle by analysis, express the current actual working of a furnace with vivid picture and language, each submodel leaves end value in the shared region in, use in order to graphic display system, end value is divided into three parts in shared region, i.e. direct motion part, body of heater state part, hot state part.
The end value of direct motion part mainly is: blowing-out, air-supply recover slow wind operation, press the magnitude relation anxiety, partially chi, slide rule, the material that collapses, hanging, airflow line.
The end value of body of heater state part mainly is: the furnace wall knot thick with burn the furnace wall knot thick with burn the direction that and the thick value of knot of all directions.
The end value of the part of hot state mainly is: stove to cold, stove to cold, stove thermotropism, overheated etc.
Indicating system scanned shared region in per 2 minutes, carried out the judgement of the comprehensive working of a furnace then according to end value, and the result of the judgement of the comprehensive working of a furnace is shown with people's shape of face formula by colour video display unit.The types of facial makeup in Beijing operas that personalize divide three kinds substantially: smiling face/furnace condition anterograde, strictly/stove owes steady, face/the working of a furnace of crying is not good enough, and these three kinds of working of a furnaces judge that according to the end value of each submodel a certain unusual condition of each submodel is divided into 5 ranks, 1 expression just commonly used, there is anomaly trend to represent, may takes place to represent with 3, will take place to represent with 4 with 2, taken place to represent that with 55 other AND of level, OR draw the comprehensive working of a furnace excessively.
For example: 1, red light, the face working of a furnace of crying is not good enough
Direct motion provides the material that collapses, hot state provides the superheated end value, and material 5, AND have then collapsed.Overheated 5=5 send red light.
Perhaps, the body of heater state provides and at certain presentation the thick end value of knot takes place.Then tie thick 5=5, sent red light.
2, amber light, the serious working of a furnace owes steady
The body of heater state provides attention and will take place in certain one, ties thick end value, and then, attention will be tied thick 4=4 and be sent amber light.
The direct motion state provides slide rule or hanging takes place, and simultaneously, hot state is given the end value of come out of the stove cold or stove heat, and then amber light is sent in (slide rule 5, OR, hanging, 5), AND, (stove cold 3, OR, stove heat 3)=3.
3, green light, smiling face's furnace condition anterograde
Each submodel all provides normal end value, then send green light.
Fig. 8 is an expert systems operational scheme block diagram.Start the standard of operation as each relevant sub-routine with cycle working time.After the start, finished the preceding button-up operations of real-time process is installed, after the initialize, system brings into operation.
Five seconds process is at first calculated material speed and is five seconds data from the machine-readable fetch cycle of process.Gather blanking information from shared region, it is heavy to remove the ore deposit with blanking last time to the timed interval of this blanking, draws the required time of ore deposit per ton, takes advantage of the tonnage of every batch of material, calculates the material speed of every batch of material and deposits.
Second step was 15 seconds processes.The judgement of operation direct motion subsystem utilizes the hole index, pipelines such as permeability index, material speed, top temperature, be difficult to walk, whether hanging take place.The data of judging are sent into shared region.
The 3rd step was 30 seconds processes.(1) from the machine-readable fetch cycle of process is 30 seconds data; (2) operation blanking model; (3) the comprehensive determining program of operation graphic presentation is by the relevant figure of judgment result displays; (4) operation body of heater furnace wall is tied thick judgement; With 30 seconds be get in the cycle about material chi difference, the maximum difference of getting every batch of material is averaged to the aniseed chi difference of the every batch of material in 2 hours, ties thick model with furnace wall and judges.
The 4th step was 1 minute process.(1) from the machine-readable fetch cycle of process is 1 minute data; (2), finish the data processing that each submodel and production management need, as carry out data smoothing, the upper and lower limit screening according to the design of system; (3) silicone content in the molten iron is calculated in the silicon forecast in the operation blast furnace thermal model; (4) operation production management program is produced the desired data forecast.
The 5th step was 2 minutes processes.Operation direct motion judgment models 2, be difficult to walk (hanging), air-flow be (pipeline), the material (slide rule) that collapses forecast unusually, and send shared region the result, and data are deposited file.
The 6th step was 5 minutes processes.(1) operation body of heater state model 2, promptly furnace wall burns model, makes forecast by wall temperature after getting 7,8,9 sections cooling water tanks; (2) operation body of heater state model 3, and promptly cupola well burns the temperature that thermopair that model buries underground by the cupola well position records and judges.
The 7th step was 20 minutes processes.Operation blast furnace thermal model 2 gives processing according to the data of being gathered to it, calculates degree of confidence.
The 8th step was 1 hour process.(1) operation blast furnace thermal model 3, the i.e. self learning model of blast furnace temperature trend prediction; (2) operation body of heater state model 4.Be that furnace wall is tied thick model,, full blast rate level poor by temperature behind the cooling stave of collecting, material chi, permeability index and direct motion index are made the thick judgement of knot.
This system design is according to the needs of blast furnace production and mathematical model, but each detached process nested subroutines, Sun Jincheng reach the system function requirement.
Characteristics of the present invention are as follows:
1, the present invention uses formation knowledge base, inference engines such as fuzzy matrix, degree of membership, membership function, and combine with statistical model, mechanism model, start with by state, the situation of blast furnace process and to judge the trend of blast furnace process, based on expert systems, and fully draw the strong point of statistical model, mechanism model, be the development of blast furnace process control techniques;
2, the present invention is made of three subsystems, than single model, can more fully judge blast furnace process, and function is extensive, and is practical.
3, the present invention relies on the routine monitoring parameter, and comprehensive summing up blast furnace operating expert's knowledge and experience propose, and have good adaptability, are convenient to popularize in an all-round way.
Description of drawings:
Fig. 1 is a blast furnace computer hardware configuration schematic diagram.
Fig. 2 constitutes synoptic diagram for patent system.
Fig. 3 is the system data processing flow chart.
Fig. 4 is the structured data repository synoptic diagram.
Fig. 5 is the knowledge base processing flow chart.
Fig. 6 is the inference engine structure flow chart.
Fig. 7 is the judgment result displays schema.
Fig. 8 is the Blast Furnace Expert System operational flow diagram.
Table 1 is the blast furnace smelting expert system knowledge frame.
The present invention has carried out trial run and production application at No. 2 blast furnaces of Shoudu Iron and Steel Co, operation practice explanation, and each subsystem is instructing blast furnace operating, the burnt aspects of volume increase joint all to obtain satisfied effect.
Smooth operation of furnace is judged subsystem, strengthens under the condition of production run at No. 2 blast furnaces of Shoudu Iron and Steel Co, has judged that more exactly blast furnace is difficult to walk to hanging, and the working of a furnace is in time regulated in forecast in advance, has reached the purpose of smooth operation of furnace; And forecast that in time blast furnace air-flow anomaly trend reaches the generation to pipeline, in time regulated that avoided the working of a furnace unusual, hit rate reaches more than 90% by aspects such as charging systems.
After the blast furnace state is judged the subsystem on-line operation, forecast that knot is thick and it is repeatedly thick to tie,, avoided the thick generation of high furnace accretion effectively through in time handling.Before 91 year April 2 blast furnace overhaul, three times the forecast furnace wall burns, and is all identical with the actual working of a furnace.Hit rate reaches 100%.The furnace bottom encroachment line of forecast also coincide.
Since heat state of blast furnace judgement subsystem put into operation and goes, silicon content of hot metal forecast hit rate reached more than 85%.
In a word, the present invention produces on-line operation at No. 2 blast furnaces of Shoudu Iron and Steel Co, for stablize blast furnace operating, avoiding and reduces the blast furnace severe disorder, guarantees that blast furnace reinforcement, direct motion brought into play vital role.Create direct economic benefit more than 7,500,000 yuan, can save and introduce similar technical fee more than 1,000 ten thousand yuan, remarkable in economical benefits, and produce remarkable social benefit.
Annotate (1) ↑-rise, ↓-descend the ≈ fluctuation
The oxygen-consumption that adds combustion of coke during several batch of materials when (2) index SH=in hole adds the total oxygen demand that is blown into the air port combustion of coke during last a few batch of material-normal iron amount
(3) permeability index K=(Q wind)/(△ P)
Claims (5)
1, a kind of Computerized blast furnace smelting expert system; Be applicable to and judge whether blast furnace and furnace run be normal, native system is except that having General System about judging heat state of blast furnace and the blast-furnace smelting process direct motion state, judge the subsystem of blast furnace state in addition, form by data collecting system, database, knowledge base, inference engine and output system, all state criterions compare coupling, judge through rule, data with knowledge base all with online dynamic acquisition data on the blast furnace transmitter.Reasoning, reach a conclusion, by the given output result of interpre(ta)tive system, the person provides operation to blast furnace operating, it is characterized in that system's operating procedure is as follows: computer initialization, enter 5 second in second process, upper computer is gathered a secondary data to process controller, gathers blanking information and computer material speed to shared region; Process operation direct motion in 15 seconds is judged, and judged result is sent into shared region; 30 seconds processes, from per 30 seconds reading of data of process controller, operation blanking model, moving model is comprehensively judged graphic display program, the result is come out by graphic presentation, and operation body of heater furnace wall is tied thick judgement, carries out the maximum difference 30 seconds to be material chi difference, every batch of material about the getting of cycle, the aniseed chi difference of every batch of material in 2 hours is averaged, be used to tie thick judgement; One minute process, from process machine collection period 1 minute data, to each model and the required data of production management carry out smoothly, processing such as upper and lower limit screening, summation, average, maximizing, minimum value, deviation or weighting, and the cycle of operation be 1 minute thermal model and production management program; 2 minutes processes, the cycle of operation is 2 minutes a direct motion judgment models; 5 minutes processes, the cycle of operation is 5 minutes a blast furnace state model; 20 minutes processes, the cycle of operation is 20 minutes a blast furnace thermal model; 1 hour process, operation heat state of blast furnace model and blast furnace state model; The result data of each process operation is stored in direct motion slip condition database, hot slip condition database and body of heater slip condition database respectively, and graphic presentation shows that through gathering each subsystem operation result in real time the person provides operation to blast furnace operating.
2, according to the described Computerized blast furnace smelting expert system of claim 1, it is characterized in that knowledge base is that judgment rule and Operating Guideline rule base are set up in the basis with the blast furnace operating experience of accumulation for many years, knowledge base is the new knowledge representation form of skeleton construction and production rule.Constitute by rule, formula, parameter fuzzy processing and fuzzy relation square car, and carry out the oneself and learn again, to knowledge base make amendment, replenish, perfect.
3, according to the described Computerized blast furnace smelting expert system of claim 1, it is characterized in that the data of upper computer to collecting from process controller, according to production management and model needs, on carrying out, the lower limit screening, smoothly, summation, extreme value (max, min), processing such as variance, again according to expert systems, rule request is handled in the knowledge base, its result is stored into the internal memory shared region of model special use, each model is from internal memory shared region image data, carry out reasoning and judging, form smooth operation of furnace state outcome file, the heat state of blast furnace destination file, blast furnace state outcome file.
4, according to claim 1 or 3 described Computerized blast furnace smelting expert systems, it is characterized in that inference engine is by rule-based reasoning, the framework coupling, the mixed inference that mechanism model and fuzzy relation equation constitute, adopt the inexact reasoning method, as evidence E1 ... when En exists really, the confidence level that conclusion H sets up is X, degree of confidence CF(H, E) with trusting growth degree MB(H, Ei) with distrust growth degree MD(H, Ei) difference is obtained, compare analysis by the CF value, the conclusion of output reasoning, and carry out reasoning conclusion and the actual judgement that whether is consistent, if the then reasoning that conforms to finishes, if be not inconsistent, then carry out the oneself and learn again, revise the knowledge base data automatically.
5, according to the described Computerized blast furnace smelting expert system of claim 2, be characterised in that altogether that fuzzy characteristics is described and self-learning method to be useful Bayes decision-making principle carry out from correction the threshold value of the membership function of feature.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN93108671A CN1038146C (en) | 1993-07-21 | 1993-07-21 | Computerized blast furnace smelting expert system method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN93108671A CN1038146C (en) | 1993-07-21 | 1993-07-21 | Computerized blast furnace smelting expert system method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1097804A true CN1097804A (en) | 1995-01-25 |
CN1038146C CN1038146C (en) | 1998-04-22 |
Family
ID=4987254
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN93108671A Expired - Fee Related CN1038146C (en) | 1993-07-21 | 1993-07-21 | Computerized blast furnace smelting expert system method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN1038146C (en) |
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1039498C (en) * | 1995-11-23 | 1998-08-12 | 宝山钢铁(集团)公司 | Blast furnace comprhensive deterministic system |
CN1052758C (en) * | 1997-06-13 | 2000-05-24 | 冶金工业部自动化研究院 | Blast furnace operating consulting system |
CN1310106C (en) * | 2001-03-01 | 2007-04-11 | 费舍-柔斯芒特系统股份有限公司 | Remote analysis of process control plant data |
CN100358699C (en) * | 2003-11-12 | 2008-01-02 | 青岛高校软控股份有限公司 | Multivariate process optimizing and analyzing method |
CN100358698C (en) * | 2003-11-12 | 2008-01-02 | 青岛高校软控股份有限公司 | Self adaptive glue discharging control method |
CN100428162C (en) * | 2002-10-28 | 2008-10-22 | 王伟 | Manual intelligent machine and method of realizing natural speech understanding |
CN102676718A (en) * | 2011-03-07 | 2012-09-19 | 宝山钢铁股份有限公司 | System and method for alarming blast-furnace burden surface states |
CN102690916A (en) * | 2012-06-19 | 2012-09-26 | 冶金自动化研究设计院 | Real-time monitoring system and method of CO2 discharging amount of blast furnace |
CN102759419A (en) * | 2011-04-28 | 2012-10-31 | 宝山钢铁股份有限公司 | Determination method for heat redundancy in blast furnace |
CN102766708A (en) * | 2012-07-12 | 2012-11-07 | 中冶华天工程技术有限公司 | Blast furnace batching system with material variety precedence matrix |
CN102776303A (en) * | 2012-06-27 | 2012-11-14 | 浙江大学 | Method for estimating inner surface temperature of blast furnaces |
CN102799116A (en) * | 2011-05-24 | 2012-11-28 | 宝山钢铁股份有限公司 | Multi-variety coke screening nonlinear control device and method |
CN102888479A (en) * | 2012-10-23 | 2013-01-23 | 北京首钢自动化信息技术有限公司 | Automatic control system for changing combustion rhythm of hot blast furnace along with blast temperature of high furnace |
CN101513119B (en) * | 2006-09-18 | 2013-07-03 | 西马克·西马格公司 | Method for operating a melt-metallurgic furnace, and furnace |
CN103361454A (en) * | 2012-03-30 | 2013-10-23 | 鞍钢股份有限公司 | Data-filtering-based method for judging blast furnace hanging |
CN105441610A (en) * | 2015-12-10 | 2016-03-30 | 武汉钢铁(集团)公司 | Method for establishing operating model of blast furnace |
US9760651B2 (en) | 2002-04-15 | 2017-09-12 | Fisher-Rosemount Systems, Inc. | Web services-based communications for use with process control systems |
CN107817403A (en) * | 2017-10-31 | 2018-03-20 | 惠州鼎宝金属制品有限公司 | A kind of fault detection system for oven |
US9927788B2 (en) | 2011-05-19 | 2018-03-27 | Fisher-Rosemount Systems, Inc. | Software lockout coordination between a process control system and an asset management system |
CN108595380A (en) * | 2018-03-14 | 2018-09-28 | 山东科技大学 | A kind of Blast Furnace Abnormal Conditions Judgement detection method |
CN111492070A (en) * | 2017-12-19 | 2020-08-04 | 株式会社Posco | Furnace condition control apparatus and method |
CN111710135A (en) * | 2020-05-26 | 2020-09-25 | 广东韶钢松山股份有限公司 | Early warning and control process for brick collapse of blast furnace hot blast stove fire well wall |
CN112593032A (en) * | 2020-12-11 | 2021-04-02 | 安徽工业大学 | Key parameter processing method for blast furnace heat exchange air furnace |
CN112836096A (en) * | 2021-03-04 | 2021-05-25 | 中冶南方工程技术有限公司 | Blast furnace expert rule knowledge base management system based on data middleware |
CN113737015A (en) * | 2021-09-26 | 2021-12-03 | 云南锡业股份有限公司锡业分公司 | Intelligent crude tin smelting system and method |
WO2023070692A1 (en) * | 2021-10-26 | 2023-05-04 | 中冶南方工程技术有限公司 | Online evaluation system for blast furnace condition |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN100485047C (en) * | 2006-12-18 | 2009-05-06 | 冶金自动化研究设计院 | Intelligent diagnosis and determination support system for blast furnace |
EP3843016A1 (en) * | 2019-12-23 | 2021-06-30 | Commissariat à l'Energie Atomique et aux Energies Alternatives | Method for screening data implemented by computer |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS62270712A (en) * | 1986-05-20 | 1987-11-25 | Nippon Kokan Kk <Nkk> | System for detecting condition of blast furnace |
JPS62270708A (en) * | 1986-05-20 | 1987-11-25 | Nippon Kokan Kk <Nkk> | Control system for blast furnace heat |
EP0375282B1 (en) * | 1988-12-20 | 1996-04-17 | Nippon Steel Corporation | Blast furnace operation management method and apparatus |
-
1993
- 1993-07-21 CN CN93108671A patent/CN1038146C/en not_active Expired - Fee Related
Cited By (37)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1039498C (en) * | 1995-11-23 | 1998-08-12 | 宝山钢铁(集团)公司 | Blast furnace comprhensive deterministic system |
CN1052758C (en) * | 1997-06-13 | 2000-05-24 | 冶金工业部自动化研究院 | Blast furnace operating consulting system |
CN1310106C (en) * | 2001-03-01 | 2007-04-11 | 费舍-柔斯芒特系统股份有限公司 | Remote analysis of process control plant data |
CN1312548C (en) * | 2001-03-01 | 2007-04-25 | 费舍-柔斯芒特系统股份有限公司 | Remote-data analysis in process plant |
US9760651B2 (en) | 2002-04-15 | 2017-09-12 | Fisher-Rosemount Systems, Inc. | Web services-based communications for use with process control systems |
CN100428162C (en) * | 2002-10-28 | 2008-10-22 | 王伟 | Manual intelligent machine and method of realizing natural speech understanding |
CN100358699C (en) * | 2003-11-12 | 2008-01-02 | 青岛高校软控股份有限公司 | Multivariate process optimizing and analyzing method |
CN100358698C (en) * | 2003-11-12 | 2008-01-02 | 青岛高校软控股份有限公司 | Self adaptive glue discharging control method |
CN101513119B (en) * | 2006-09-18 | 2013-07-03 | 西马克·西马格公司 | Method for operating a melt-metallurgic furnace, and furnace |
CN102676718A (en) * | 2011-03-07 | 2012-09-19 | 宝山钢铁股份有限公司 | System and method for alarming blast-furnace burden surface states |
CN102676718B (en) * | 2011-03-07 | 2015-03-25 | 宝山钢铁股份有限公司 | System and method for alarming blast-furnace burden surface states |
CN102759419A (en) * | 2011-04-28 | 2012-10-31 | 宝山钢铁股份有限公司 | Determination method for heat redundancy in blast furnace |
US9927788B2 (en) | 2011-05-19 | 2018-03-27 | Fisher-Rosemount Systems, Inc. | Software lockout coordination between a process control system and an asset management system |
CN102799116A (en) * | 2011-05-24 | 2012-11-28 | 宝山钢铁股份有限公司 | Multi-variety coke screening nonlinear control device and method |
CN103361454A (en) * | 2012-03-30 | 2013-10-23 | 鞍钢股份有限公司 | Data-filtering-based method for judging blast furnace hanging |
CN103361454B (en) * | 2012-03-30 | 2015-03-11 | 鞍钢股份有限公司 | Data-filtering-based method for judging blast furnace hanging |
CN102690916B (en) * | 2012-06-19 | 2013-07-31 | 冶金自动化研究设计院 | Real-time monitoring system and method of CO2 discharging amount of blast furnace |
CN102690916A (en) * | 2012-06-19 | 2012-09-26 | 冶金自动化研究设计院 | Real-time monitoring system and method of CO2 discharging amount of blast furnace |
CN102776303A (en) * | 2012-06-27 | 2012-11-14 | 浙江大学 | Method for estimating inner surface temperature of blast furnaces |
CN102766708A (en) * | 2012-07-12 | 2012-11-07 | 中冶华天工程技术有限公司 | Blast furnace batching system with material variety precedence matrix |
CN102888479A (en) * | 2012-10-23 | 2013-01-23 | 北京首钢自动化信息技术有限公司 | Automatic control system for changing combustion rhythm of hot blast furnace along with blast temperature of high furnace |
CN105441610A (en) * | 2015-12-10 | 2016-03-30 | 武汉钢铁(集团)公司 | Method for establishing operating model of blast furnace |
CN107817403A (en) * | 2017-10-31 | 2018-03-20 | 惠州鼎宝金属制品有限公司 | A kind of fault detection system for oven |
CN111492070A (en) * | 2017-12-19 | 2020-08-04 | 株式会社Posco | Furnace condition control apparatus and method |
JP7050934B2 (en) | 2017-12-19 | 2022-04-08 | ポスコ | Reactor condition management device and method |
JP2021507115A (en) * | 2017-12-19 | 2021-02-22 | ポスコPosco | Reactor condition management equipment and method |
CN108595380B (en) * | 2018-03-14 | 2021-07-09 | 山东科技大学 | Method for detecting abnormal furnace conditions of blast furnace |
CN108595380A (en) * | 2018-03-14 | 2018-09-28 | 山东科技大学 | A kind of Blast Furnace Abnormal Conditions Judgement detection method |
CN111710135A (en) * | 2020-05-26 | 2020-09-25 | 广东韶钢松山股份有限公司 | Early warning and control process for brick collapse of blast furnace hot blast stove fire well wall |
CN112593032A (en) * | 2020-12-11 | 2021-04-02 | 安徽工业大学 | Key parameter processing method for blast furnace heat exchange air furnace |
CN112593032B (en) * | 2020-12-11 | 2024-03-26 | 安徽工业大学 | Key parameter processing method for blast furnace heat exchange air furnace |
WO2022183631A1 (en) * | 2021-03-04 | 2022-09-09 | 中冶南方工程技术有限公司 | Blast furnace expert rule knowledge base management system based on data middleware |
CN112836096B (en) * | 2021-03-04 | 2022-06-24 | 中冶南方工程技术有限公司 | Blast furnace expert rule knowledge base management system based on data middleware |
CN112836096A (en) * | 2021-03-04 | 2021-05-25 | 中冶南方工程技术有限公司 | Blast furnace expert rule knowledge base management system based on data middleware |
CN113737015A (en) * | 2021-09-26 | 2021-12-03 | 云南锡业股份有限公司锡业分公司 | Intelligent crude tin smelting system and method |
CN113737015B (en) * | 2021-09-26 | 2023-10-03 | 云南锡业股份有限公司锡业分公司 | Intelligent crude tin smelting system and method |
WO2023070692A1 (en) * | 2021-10-26 | 2023-05-04 | 中冶南方工程技术有限公司 | Online evaluation system for blast furnace condition |
Also Published As
Publication number | Publication date |
---|---|
CN1038146C (en) | 1998-04-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN1038146C (en) | Computerized blast furnace smelting expert system method | |
CN109918702A (en) | A kind of blast-furnace burden cooperates with Multipurpose Optimal Method with operation | |
CN108676955A (en) | A kind of BOF Steelmaking Endpoint carbon content and temprature control method | |
CN1021833C (en) | Blast furnace operation management method and apparatus | |
CN102915027B (en) | Blast furnace smelting expert system built based on pattern recognition technology and method thereof | |
CN103439999A (en) | Method for controlling abnormal furnace temperature of blast furnace according to temperature changes of cooling wall | |
CN101644924B (en) | Method for tracking and managing process of weighing and loading materials for blast furnace | |
CN109507961A (en) | A kind of semiconductor production line dynamic load uniform feeding control method | |
CN107937658A (en) | A kind of converter smelting automated intelligent control system | |
CN110238172A (en) | The method for dissolving the processing resource of solid waste containing zinc using blast furnace | |
WO2022183631A1 (en) | Blast furnace expert rule knowledge base management system based on data middleware | |
CN114066247A (en) | Parallel disassembly line setting method for merging priority relation matrix | |
CN102156405B (en) | Sintered ore chemical component prediction and intelligent control system under small sample poor information | |
CN109359320A (en) | Blast furnace index prediction technique based on multi-sampling rate ARDL model | |
CN1052758C (en) | Blast furnace operating consulting system | |
CN107976976A (en) | A kind of iron and steel enterprise's gas consumption equipment timing optimization method | |
CN116822984A (en) | Intelligent carbon digestion system and method for industrial park | |
US4857106A (en) | Method for controlling operation of a blast furnace | |
CN110531727A (en) | A kind of full-range multilevel optimization's control method of polyester fiber manufacturing | |
CN112861276B (en) | Blast furnace burden surface optimization method based on data and knowledge dual drive | |
JPH0673414A (en) | Method for controlling quality of molten iron in blast furnace | |
CN213951241U (en) | System for controlling emission source of sulfur dioxide in flue gas of iron-making hot blast stove | |
CN1594601A (en) | Method for stabilizing blast furnace operation based on equilibrium point | |
JPH0598325A (en) | Device for controlling distribution of charging materials in blast furnace | |
CN102830659A (en) | Blast furnace data filtering method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
C19 | Lapse of patent right due to non-payment of the annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |