CN101957942A - Accident planning expert system applied to steel mill - Google Patents
Accident planning expert system applied to steel mill Download PDFInfo
- Publication number
- CN101957942A CN101957942A CN201010249756.1A CN201010249756A CN101957942A CN 101957942 A CN101957942 A CN 101957942A CN 201010249756 A CN201010249756 A CN 201010249756A CN 101957942 A CN101957942 A CN 101957942A
- Authority
- CN
- China
- Prior art keywords
- knowledge
- accident
- prediction
- prediction scheme
- rule
- 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.)
- Pending
Links
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 34
- 239000010959 steel Substances 0.000 title claims abstract description 34
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 claims abstract description 62
- 238000000034 method Methods 0.000 claims abstract description 59
- 238000004519 manufacturing process Methods 0.000 claims abstract description 36
- 229910052742 iron Inorganic materials 0.000 claims abstract description 31
- 230000008569 process Effects 0.000 claims abstract description 26
- 230000007246 mechanism Effects 0.000 claims abstract description 13
- 238000007726 management method Methods 0.000 claims abstract description 12
- 238000003860 storage Methods 0.000 claims description 15
- 238000013461 design Methods 0.000 claims description 6
- 238000011160 research Methods 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000001143 conditioned effect Effects 0.000 claims description 4
- 125000004122 cyclic group Chemical group 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 208000037656 Respiratory Sounds Diseases 0.000 claims description 2
- 230000015572 biosynthetic process Effects 0.000 claims description 2
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 238000004891 communication Methods 0.000 claims description 2
- 230000008878 coupling Effects 0.000 claims description 2
- 238000010168 coupling process Methods 0.000 claims description 2
- 238000005859 coupling reaction Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 claims description 2
- 238000001914 filtration Methods 0.000 claims description 2
- 239000012467 final product Substances 0.000 claims description 2
- 230000035800 maturation Effects 0.000 claims description 2
- 230000008520 organization Effects 0.000 claims description 2
- 206010037833 rales Diseases 0.000 claims description 2
- 241001062472 Stokellia anisodon Species 0.000 claims 1
- 230000010354 integration Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 4
- 238000003723 Smelting Methods 0.000 description 3
- 238000009628 steelmaking Methods 0.000 description 3
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 229910052802 copper Inorganic materials 0.000 description 2
- 239000010949 copper Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000005245 sintering Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to an accident planning expert system applied to a steel mill, which belongs to the field of the management of iron and steel enterprises. The system consists of a client interface, a knowledge acquisition interface, a reasoning mechanism, an interpretative program, a database and a knowledge base, wherein the client interface is oriented to a user of the planning expert system; the knowledge acquisition interface is oriented to a knowledge engineer and used for adding, modifying and maintaining the knowledge base; the reasoning mechanism reasons a planning process according to knowledge in the knowledge base; the interpretative program explains and describes a plan reasoning process; the database is used for storing corresponding production data or history data; the knowledge base is the integration of stored planning rule knowledge; and the realization process of the system comprises a knowledge acquisition step, a semantic network expression step, a frame expansion step, a rule base establishment step and a mechanism reasoning step. In-time and correct expert treatment measure and method can be provided for a failure occurs in the production process of the steel mill and a human experience-based accident treatment mode of the iron and steel enterprises is replaced by a computer expert system treatment mode, so that the accident treatment efficiency of the production process is improved.
Description
Technical field
The invention belongs to enterprise management of iron and steel industry field, relate to iron and steel enterprise's industrial accident program management, particularly use expert system method accident prediction is carried out intelligent management.
Background technology
In iron and steel enterprise, flow process single devices operation control Expert System Study is more relatively, particularly sintering and blast furnace operation, and do not meet Expert System Study achievement to the enterprise accident prediction scheme.Document " iron and steel enterprise's production run management method research " (University of Science ﹠ Technology, Beijing's [PhD dissertation] 2007) has carried out systematic study to system modeling method, optimization method and manner of execution in the steel manufacture process management, and has inquired into system integration method based on infotech in conjunction with application example.Inquired into the ERP system of To enterprises business processs and the research work with novelty has been made in aspects such as ERP and MES function sharing and work compound.This literature research core is a process control method, does not relate to the accident prediction management method.Document " metallurgical on-line intelligence monitoring of flash and accident pre-alarming system research " (the journal .2007 of Institutes Of Technology Of Jiangxi, 28 (6)) be object with the copper flash smelting production run, on traditional PLC, DCS control system, advanced technologies such as integrated computer network, software development, expert system, digital-to-analogue emulation, neural network have been set up metallurgical on-line intelligence monitoring of flash and accident pre-alarming system.This document mainly carries out intelligent decision and early warning based on mathematical model to the molten smelting process of copper flash, is not the accident prediction management system towards whole corporate process.
Summary of the invention
In the iron and steel manufacturing process, fault or accident pattern that production run took place are various, and its corresponding accident prediction also varies.At present each iron and steel enterprise mainly adopts the prediction scheme handbook that accident prediction is concluded, put in order, and this traditional management method makes the reliability of accident prediction information, real-time poor, and each operation, inter-sectional prediction scheme information sharing are not smooth.In steel mill's production run, it is unpractical arranging the human expert to monitor the production fault in real time and corresponding prediction scheme is provided; If adopt traditional program that accident prediction is managed, then, solve the problem that needs the prediction scheme expert to solve again because the complicacy of accident and the diversity, the ambiguity that relate to problem make prediction scheme be difficult to describe accurately the accident treatment measure.
The invention provides the expert system management method of a kind of steel mill accident prediction, developed prediction scheme expert system with a large amount of special prediction scheme knowledge and accident treatment experience.This system can utilize the prediction scheme rule of storing in the knowledge base to carry out reasoning and judgement, and simulation prediction scheme expert's decision-making is instructed for iron and steel enterprise's accident provides the solution of expert's level and prediction scheme.The prediction scheme expert system is characterised in that: select semantic net method modularization to obtain accident prediction, utilization frame knowledge representation method expansion semantic net node knowledge, select rule-based knowledge representation in database, to represent prediction scheme knowledge, forward reasoning, backward reasoning and the mixed inference of integrated use knowledge reasoning strategy are finally realized the accident prediction expert system of steel mill.
Summary of the invention mainly comprise following some:
1. made up the accident prediction expert system structure
The prediction scheme expert system is made up of client end interface, knowledge acquisition interface, inference machine, interpretive routine, database and knowledge base, and its architecture as shown in Figure 1.Mainly towards the user of prediction scheme expert system, its authority only limits to use the prediction scheme expert system to client end interface; The knowledge acquisition interface is mainly used in increasing, revise and safeguarding of knowledge base towards the knowledge engineer; Inference machine then is the process that goes out prediction scheme according to the knowledge reasoning in the knowledge base; Interpretive routine is the explanation and the description of prediction scheme reasoning process; Database is used to store corresponding production data or historical data; Knowledge base is the set of storage prediction scheme rule knowledge.
Based on the general structure design of system, the implementation procedure of prediction scheme expert system comprises five steps: knowledge acquisition, semantic net are represented, framework is expanded, rule base is set up and inference mechanism, as shown in Figure 2.
2. developed accident prediction expert system knowledge acquisition methods
The accident prediction knowledge acquisition method, at first according to steel mill's production run characteristics, range of application in conjunction with accident prediction, determine knowledge source: 1. dispatch control expert: the domain expert is the main source of expert system knowledge, for steel mill's accident, the dispatch control expert has commander's experience and knowledge of specialty, is the indispensable knowledge source of prediction scheme expert system; 2. terminal user: the terminal user is a kind of valuable source of additional information, and system can obtain the accident prediction under each terminal different situations by to terminal user's consulting; 3. prediction scheme handbook: consult relevant prediction scheme handbook and list of references and help standardization and specialized field term, also can obtain the prediction scheme processing experience of historical accident, for knowledge base provides opinion and prediction scheme detail to problem.The knowledge of three kinds of knowledge sources after knowledge engineer's modularization, regularization are represented, enters the knowledge base storage, as shown in Figure 1 mainly by the knowledge acquisition interface.
Secondly, obtain knowledge from knowledge source.Obtaining of knowledge is a cyclic process (seeing accompanying drawing 3 for details), comprises following step: 1. collect: set up commitment in the accident prediction knowledge base, obtain the basic comprehension to prediction scheme earlier, as accident pattern, relate to station/department etc.; In the collection process, need effective interpersonal communication skill and strive for expert's cooperation technical ability, domain knowledge is fully collected; Set up the later stage in the accident prediction knowledge base, the knowledge engineer needs continuous repeated collection prediction scheme knowledge and filtering useless prediction scheme, obtains specific, useful prediction scheme knowledge; 2. explain: the information of collecting is commented and important prediction scheme is made excuses.Set up commitment in the accident prediction knowledge base, the information of collecting is quite general, need set up target, constraint and the scope of problem.Set up later stage in the accident prediction knowledge base, use formal method to explain needed knowledge in this task; 3. analyze: by explaining all critical learning of being found, formation theoretical for knowledge organization and problem solving strategy is offered suggestions, and determines conceptual relation and how to use these relations to solve problem; 4. design: after finishing collection, explanation and analysis task, the method for expressing and the inference mechanism of the knowledge of further research.
3. developed the knowledge representation method that semantic net combines with framework
Knowledge representation method of the present invention at first uses semantic net with the accident prediction type and to comprise the situation modulate expression clear, i.e. the node presentation-entity of semantic net, and accident pattern is as a class, and the various situations of accident are as the object of class.The line of semantic net is represented the relation between various situations and the accident, connects individual class and its parent with AKO, connects individual class and its object with IS-A.The semantic net representation of knowledge of an accident prediction sees accompanying drawing 4 for details.
Secondly, the extensibility of utilization framework expands to the framework with groove and slot value with the entity of semantic net node, and wherein, groove is represented the attribute (as name) of this node, and slot value be the occurrence (as blast furnace race iron) of this nodal community.If because the singularity and the complicacy of prediction scheme, can't describe clearly by the two-dimensional expansion of semantic net node, then can carry out the framework expansion to slot value again, set up three-dimensional framework knowledge hierarchy (seeing accompanying drawing 5 for details).Three-dimensional extended by node, make accident prediction set up a huge prediction scheme frame system, the structure of a certain pre-pattern frame can be the slot value of another pre-pattern frame in the system, and same prediction scheme framed structure can be simultaneously as the slot value of a plurality of pre-pattern frames.The three-dimensional extended of node can make some identical prediction schemes save storage space without repeated storage.
4. the rule-based prediction scheme expert system knowledge base of setting up
Representation of knowledge level height in the prediction scheme expert system knowledge base directly affects the efficient of reasoning and the convenience of knowledge base maintenance.Because the complicacy of accident and the diversity, the ambiguity that relate to problem, the present invention selects the production rule knowledge representation to represent accident prediction, and the method can not only directly and fully be expressed prediction scheme, but also can easily set up and the extension rule storehouse.Production rule is represented is between prerequisite and the prerequisite, the logical relation between conclusion and the conclusion, and the corresponding relation between prerequisite and the conclusion.Rule connects each other by variety of way, and when a certain each regular conclusion just in time was the prerequisite of another rule, these two rules i.e. series connection mutually.The representation of production representation method is: Rule (number of regulation, [condition list], [conclusion tabulation], rule type, CF).Number of regulation is the identification number of rule in the rule representation, condition list is the premise part of rule, the conclusion tabulation is the conclusion part of rule, rule type is the numbering according to accident prediction affiliated area (as smelting iron the zone) defined, CF is used to represent probabilistic rule, and the scope definition of confidence level is represented the possibility that certain conditioned disjunction rule is set up in (0～1) interval, the value of confidence level is big more, and the possibility that this conditioned disjunction rule is set up is high more.
The rule foundation of prediction scheme expert system knowledge base is on the basis of the semantic net of prediction scheme and frame representation, utilization production representation method is represented prediction scheme knowledge, can be expressed as a rule in the accompanying drawing 4: Rule (1, [A], [B], A, C), the programmed logic of this rule can be expressed as: IF A THEN B (C); A rule of semantic net node three-dimensional extended (accompanying drawing 5) can be expressed as: Rule (2, [A1, A11], [B1], A, C1), the programmed logic of this rule can be expressed as equally: IF A1 AND A11 THEN B1 (C1), B, B1 represent corresponding accident prediction in the rule, C, the corresponding confidence level of C1 delegate rules.
By after using the production representation method accident prediction being expressed as rule format, can utilize the ability of tissue, storage, inquiry and the management data of relevant database maturation, realize the bivariate table storage of prediction scheme rule in database.During rale store: number of regulation, rule type and CF all get final product with a field store, and condition list and conclusion tabulation are then looked its tabulation number and distributed its corresponding field can realize storage.
5. prediction scheme expert system database
The information such as number of times, accident prediction effect and system's operating position that be mainly used in the database of prediction scheme the expert system accident that storage takes place in history, all kinds of accident take place, and use suggestion with corresponding prediction scheme analysis result and prediction scheme.The data of database storing provide accident treatment to use for reference for spot dispatch expert and operating personnel on the one hand, are an apprentice of knowledge source for knowledge engineering on the other hand and obtain knowledge reference is provided.
6. set up prediction scheme inference engine of expert system system
Reasoning is the thought process of releasing the new fact (or other judgement) according to certain principle (axiom or rule) from the known fact (or judgement), the fact of wherein reasoning institute foundation is called prerequisite (or condition), and the new fact of being released by condition is called conclusion.In the prediction scheme expert system, reasoning is a basis with the existing knowledge in the knowledge base, releases corresponding prediction scheme knowledge, is a kind of based on knowledge and the reasoning that draws knowledge, i.e. fault → reasoning → prediction scheme.
The inference mechanism of prediction scheme expert system mainly comprises two aspects: inference strategy and search strategy.
1. inference strategy comprises: forward reasoning, backward reasoning and mixed inference.Forward reasoning is mainly used in the given definite accident title of client, and system's reasoning draws the pairing prediction scheme of this accident; Backward reasoning is used for when client wants is known the pairing accident of certain bar prediction scheme, and system adopts backward reasoning search knowledge base to draw certain accident or some accident that should prediction scheme; Mixed inference is used for when the client initial conditions is insufficient, system carries out forward reasoning with known conditions earlier, draw some corresponding prediction schemes, seek the pairing accident title of each bar prediction scheme according to the prediction scheme backward reasoning of releasing again, release the definite title or the corresponding prediction scheme of accident with this.
2. search strategy comprises: deep search (downward as much as possible a level search, see accompanying drawing 6 for details) and breadth first search (more carry out in one deck before the low level and search for dropping to the next one, see accompanying drawing 7 for details).Corresponding with inference strategy, the way of search of prediction scheme expert system in knowledge base, way of search adopts deep search when system adopts forward reasoning; Way of search adopts the breadth first search when system adopts backward reasoning; Way of search adopts the Hybrid Search that deep search and breadth first search combine when system adopts mixed inference.
7. prediction scheme expert system shell
The interpretive routine of prediction scheme expert system is exactly that the problem of Expert System Design person or user's proposition is explained and illustrated, is the key point that is different from traditional program.The prediction scheme expert system knowledge base has didactic characteristics, and the coupling of production is the basic operation of system's solution procedure, shows reasoning path and relevant production match condition, the explanation that has just produced prediction scheme expert system problem solving process.System of the present invention has added the daily record program in the explanation facility process, the behavior in this daily record program record expert system problem solving process on selected level comprises the match condition of knowledge under the various states and the conversion situation of state.Interpretive routine also is translated as the accessible explanation statement of user according to recorded content and explanation control structure with tracking results.
8. the operational mode of prediction scheme expert system
The operational mode of accident prediction expert system is a server/customer end, management and raising running efficiency of system for ease of accident prediction, the present invention with prediction scheme knowledge centralized stores in data in server storehouse table, manage by server, each client has only visit and search permission, and authority is revised in the storehouse that is ignorant.Knowledge base and database are made amendment by knowledge engineer or franchise personnel or are safeguarded.
Useful benefit:
The invention has the advantages that: exploitation is towards the prediction scheme expert system of iron and steel manufacturing process, promptly and accurately expert's treatment measures and method can be provided for the fault that steel mill's production run takes place, for the perfect not to the utmost artificial experience processing accident pattern of present iron and steel enterprise turns to scientific and reasonable computer expert system tupe that powerful guarantee is provided, production run accident treatment efficient is improved largely.
Description of drawings:
Fig. 1 is a prediction scheme expert system structure synoptic diagram.
Fig. 2 is the implementation procedure synoptic diagram of prediction scheme expert system.
Fig. 3 is a knowledge acquisition cyclic process synoptic diagram.
Fig. 4 represents synoptic diagram for the semantic net of accident prediction.
Fig. 5 is the three-dimensional extended synoptic diagram of accident prediction semantic net node.
Fig. 6 is the depth-first search mode synoptic diagram of prediction scheme knowledge base.
Fig. 7 is the BFS (Breadth First Search) mode synoptic diagram of prediction scheme knowledge base.
Fig. 8 represents synoptic diagram for the semantic net that blast furnace runs the iron accident.
Fig. 9 is increased to the three-dimensional extended synoptic diagram of three-dimensional knowledge by the node expansion for the two-dimentional knowledge of semantic net.
Figure 10 is the inference mechanism synoptic diagram of realization system.
(on behalf of accident, the A in steel mill zone or field, AA represent concrete accident, A1～A8 to represent eight kinds of situations, A11～A14 of A accident to represent four kinds of prediction schemes, A21～A24 of A1 to represent four kinds of prediction schemes of A2 among the figure)
Embodiment
Be example with steel mill's blast furnace race iron accident and corresponding prediction scheme thereof below, set forth specific implementation process of the present invention.Concrete condition is: two blast furnaces (No. 1 blast furnace and No. 2 blast furnaces), and every blast furnace has two cast houses (Nan Chang and Bei Chang), and each cast house respectively has two railway lines (line and two wires).
1, the semantic net of accident pattern and situation is represented
Be used for the basic structure between demonstration accident and the situation thereof, the node presentation-entity of semantic net, accident pattern is as a class, and the various situations of accident are as the object of class.Blast furnace runs the semantic net of iron accident and represents as shown in Figure 8:
2, the framework of semantic net expansion
The extensibility of utilization framework expands to the framework with groove and slot value with the entity of semantic net node.
1) expansion of node
The groove and the slot value of blast furnace race iron node are as shown in table 1
Table 1
| Groove | Slot value |
| name? | Blast furnace runs iron |
| case? | An one blast furnace north line runs iron, a blast furnace south line runs iron ... |
| influence? | Total accent, ironmaking portion, steel-making portion |
| ......? | ......? |
2) expansion of slot value
The two-dimentional knowledge of semantic net is increased to three-dimensional knowledge by allowing the node expansion.Running iron with a blast furnace south line is example, the three-dimensional extended of this node as shown in Figure 9:
By in slot value, using framework and succession, set up very powerful knowledge base system.On the basis of node expansion, again a blast furnace south line is run the metal trough value and expand, as shown in table 2
Table 2
| Groove | Slot value |
| name? | An one blast furnace south line runs iron |
| property? | Ironmaking portion prediction scheme |
| means? | One blast furnace Nan Chang stops to tap a blast furnace, a blast furnace checking ... |
| ......? | ......? |
3. Gui Ze extraction
On the basis of representing to expand in semantic net, extract 2 rules that blast furnace runs iron with framework.The iron of regulation blast furnace race herein accident pattern is represented with A.
Rule of the two-dimentional framework of semantic net is expressed as:
Rule (1, [blast furnace race iron], [influencing station: total accent, ironmaking portion, steel-making portion], A, 1.0)
Programmed logic is: the IF blast furnace runs iron THEN influences station: total accent, ironmaking portion, steel-making portion (1.0)
Rule of the three-dimensional framework of semantic net is expressed as:
Rule (2, [blast furnace south line runs iron, ironmaking portion prediction scheme], [a blast furnace Nan Chang stops to tap a blast furnace, a blast furnace checking ...], A, 0.7)
Programmed logic is: IF one a blast furnace south line runs the iron AND ironmaking prediction scheme THEN of portion one blast furnace Nan Chang and stops to tap a blast furnace a blast furnace checking ... (0.7)
4. the realization of inference mechanism
Suppose the prediction scheme of ironmaking portion when certain client need be seeked advice from field, blast furnace south one line race iron, then the inference mechanism of system as shown in figure 10.
Claims (8)
1. accident prediction expert system that is applied to steel mill, it is characterized in that the prediction scheme expert system is made up of client end interface, knowledge acquisition interface, inference mechanism, interpretive routine, database and knowledge base, client end interface is towards the user of prediction scheme expert system, and its authority only limits to use the prediction scheme expert system; The knowledge acquisition interface is used for increasing, revise and safeguarding of knowledge base towards the knowledge engineer; Inference machine then is the process that goes out prediction scheme according to the knowledge reasoning in the knowledge base; Interpretive routine is the explanation and the description of prediction scheme reasoning process; Database is used to store corresponding production data or historical data; Knowledge base is the set of storage prediction scheme rule knowledge;
Based on the general structure design of system, the implementation procedure of prediction scheme expert system comprises five steps: knowledge acquisition, semantic net are represented, framework is expanded, rule base is set up and inference mechanism.
2. a kind of according to claim 1 accident prediction expert system that is applied to steel mill is characterized in that having developed accident prediction expert system knowledge acquisition methods;
The accident prediction knowledge acquisition method, at first according to steel mill's production run characteristics, range of application in conjunction with accident prediction, determine knowledge source: 1. dispatch control expert: the domain expert is the main source of expert system knowledge, for steel mill's accident, the dispatch control expert has commander's experience and knowledge of specialty, is the indispensable knowledge source of prediction scheme expert system; 2. terminal user: the terminal user is a kind of valuable source of additional information, and system can obtain the accident prediction under each terminal different situations by to terminal user's consulting; 3. prediction scheme handbook: consult relevant prediction scheme handbook and list of references and help standardization and specialized field term, can also obtain the prediction scheme processing experience of historical accident, for knowledge base provides opinion and prediction scheme detail to problem; The knowledge of dispatch control expert, terminal user, three kinds of knowledge sources of prediction scheme handbook after knowledge engineer's modularization, regularization are represented, enters the knowledge base storage by the knowledge acquisition interface.
Secondly, obtain knowledge from knowledge source; Obtaining of knowledge is a cyclic process, comprises following step: 1. collect: set up commitment in the accident prediction knowledge base, obtain the basic comprehension to prediction scheme earlier, as accident pattern, relate to station/department; In the collection process, need effective interpersonal communication skill and strive for expert's cooperation technical ability, domain knowledge is fully collected; Set up the later stage in the accident prediction knowledge base, the knowledge engineer needs continuous repeated collection prediction scheme knowledge and filtering useless prediction scheme, obtains specific, useful prediction scheme knowledge; 2. explain: the information of collecting is commented and important prediction scheme is made excuses; Set up commitment in the accident prediction knowledge base, the information of collecting is quite general, need set up target, constraint and the scope of problem; Set up later stage in the accident prediction knowledge base, use formal method to explain needed knowledge in this task; 3. analyze: by explaining all critical learning of being found, formation theoretical for knowledge organization and problem solving strategy is offered suggestions, and determines conceptual relation and how to use these relations to solve problem; 4. design: after finishing collection, explanation and analysis task, the method for expressing and the inference mechanism of the knowledge of further research.
3. a kind of according to claim 1 accident prediction expert system that is applied to steel mill is characterized in that developing the knowledge representation method that semantic net combines with framework;
Knowledge representation method of the present invention at first uses semantic net with the accident prediction type and to comprise the situation modulate expression clear, i.e. the node presentation-entity of semantic net, and accident pattern is as a class, and the various situations of accident are as the object of class; The line of semantic net is represented the relation between various situations and the accident, connects individual class and its parent with AKO, connects individual class and its object with IS-A;
Secondly, the extensibility of utilization framework expands to the framework with groove and slot value with the entity of semantic net node, and wherein, groove is represented the attribute of this node, and as name, slot value be the occurrence of this nodal community, as blast furnace race iron; If because the singularity and the complicacy of prediction scheme, can't describe clearly by the two-dimensional expansion of semantic net node, then again slot value is carried out the framework expansion, set up three-dimensional framework knowledge hierarchy; Three-dimensional extended by node, make accident prediction set up a huge prediction scheme frame system, the structure of a certain pre-pattern frame can be as the slot value of another pre-pattern frame in the system, and same prediction scheme framed structure can be simultaneously as the slot value of a plurality of pre-pattern frames; The three-dimensional extended of node can make some identical prediction schemes without repeated storage, has saved storage space.
4. a kind of according to claim 1 accident prediction expert system that is applied to steel mill is characterized in that the rule-based prediction scheme expert system knowledge base of setting up;
Select the production rule knowledge representation to represent accident prediction, production rule is represented is between prerequisite and the prerequisite, the logical relation between conclusion and the conclusion, and the corresponding relation between prerequisite and the conclusion; The representation of production representation method is: Rule number of regulation, [condition list], [conclusion tabulation], rule type, CF; Number of regulation is the identification number of production rule in the production rule representation, condition list is the premise part of production rule, the conclusion tabulation is the conclusion part of production rule, the production rule type is according to the accident prediction affiliated area, as smelt iron the numbering of regional defined, CF is used to represent probabilistic rule, the scope definition of confidence level is in (0～1) interval, represent the possibility that certain conditioned disjunction rule is set up, the value of confidence level is big more, and the possibility that this conditioned disjunction rule is set up is high more;
The rule foundation of prediction scheme expert system knowledge base is on the basis of the semantic net of prediction scheme and frame representation, and utilization production representation method is represented prediction scheme knowledge, and a rule can be expressed as: Rule (1, [A], [B], A, C), the programmed logic of this rule can be expressed as: IF A THEN B (C); A rule of semantic net node three-dimensional extended can be expressed as: Rule (2, [A1, A11], [B1], A, C1), the programmed logic of this rule can be expressed as equally: IF A1 ANDA11 THEN B1 (C1), B, B1 represent corresponding accident prediction in the rule, C, the corresponding confidence level of C1 delegate rules;
By after using the production representation method accident prediction being expressed as rule format, can utilize the ability of tissue, storage, inquiry and the management data of relevant database maturation, realize the bivariate table storage of prediction scheme rule in database; During rale store: number of regulation, rule type and CF all get final product with a field store, and condition list and conclusion tabulation are then looked its tabulation number and distributed its corresponding field can realize storage.
5. a kind of according to claim 1 accident prediction expert system that is applied to steel mill, it is characterized in that the database of prediction scheme expert system is used to store number of times, accident prediction effect and system's operating position information of the accident that takes place in history, the generation of all kinds of accident, and use suggestion with corresponding prediction scheme analysis result and prediction scheme; The data of database storing provide accident treatment to use for reference for spot dispatch expert and operating personnel on the one hand, are an apprentice of knowledge source for knowledge engineering on the other hand and obtain knowledge reference is provided.
6. a kind of according to claim 1 accident prediction expert system that is applied to steel mill, it is characterized in that setting up prediction scheme inference engine of expert system system, reasoning is according to axiom or the regular thought process of releasing the new fact from known true or judgement, the fact of wherein reasoning institute foundation is called prerequisite or condition, and the new fact of being released by condition is called conclusion; In the prediction scheme expert system, reasoning is a basis with the existing knowledge in the knowledge base, releases corresponding prediction scheme knowledge, is a kind of based on knowledge and the reasoning that draws knowledge, i.e. fault → reasoning → prediction scheme;
The inference mechanism of prediction scheme expert system comprises two aspects: inference strategy and search strategy;
1. inference strategy comprises: forward reasoning, backward reasoning and mixed inference; Forward reasoning is used for the given definite accident title of client, and system's reasoning draws the pairing prediction scheme of this accident; Backward reasoning is used for when client wants is known the pairing accident of certain bar prediction scheme, and system adopts backward reasoning search knowledge base to draw certain accident or some accident that should prediction scheme; Mixed inference is used for when the client initial conditions is insufficient, system carries out forward reasoning with known conditions earlier, draw some corresponding prediction schemes, seek the pairing accident title of each bar prediction scheme according to the prediction scheme backward reasoning of releasing again, release the definite title or the corresponding prediction scheme of accident with this;
2. search strategy comprises: deep search and breadth first search, and deep search is the search of a downward level, the breadth first search more carries out search before the low level dropping to the next one in one deck; Corresponding with inference strategy, the way of search of prediction scheme expert system in knowledge base is: way of search adopts deep search when system adopts forward reasoning; Way of search adopts the breadth first search when system adopts backward reasoning; Way of search adopts the Hybrid Search that deep search and breadth first search combine when system adopts mixed inference.
7. a kind of according to claim 1 accident prediction expert system that is applied to steel mill, the interpretive routine that it is characterized in that the prediction scheme expert system are exactly that the problem of Expert System Design person or user's proposition is explained and illustrated, are the key points that is different from traditional program.The prediction scheme expert system knowledge base has didactic characteristics, and the coupling of production is the basic operation of system's solution procedure, shows reasoning path and relevant production match condition, the explanation that has just produced prediction scheme expert system problem solving process; The prediction scheme expert system has added the daily record program in the explanation facility process, the behavior in this daily record program record expert system problem solving process on selected level comprises the match condition of knowledge under the various states and the conversion situation of state; Interpretive routine also is translated as the accessible explanation statement of user according to recorded content and explanation control structure with tracking results.
8. a kind of according to claim 1 accident prediction expert system that is applied to steel mill, the operational mode that it is characterized in that the accident prediction expert system is a server/customer end, with prediction scheme knowledge centralized stores in data in server storehouse table, manage by server, each client has only visit and search permission, and authority is revised in the storehouse that is ignorant; Knowledge base and database are made amendment by knowledge engineer or franchise personnel or are safeguarded.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010249756.1A CN101957942A (en) | 2010-08-10 | 2010-08-10 | Accident planning expert system applied to steel mill |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010249756.1A CN101957942A (en) | 2010-08-10 | 2010-08-10 | Accident planning expert system applied to steel mill |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101957942A true CN101957942A (en) | 2011-01-26 |
Family
ID=43485260
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010249756.1A Pending CN101957942A (en) | 2010-08-10 | 2010-08-10 | Accident planning expert system applied to steel mill |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101957942A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106951558A (en) * | 2017-03-31 | 2017-07-14 | 广东睿盟计算机科技有限公司 | A kind of data processing method of the tax intelligent consulting platform based on deep search |
| CN109657796A (en) * | 2018-12-12 | 2019-04-19 | 北京天诚同创电气有限公司 | Judgment rule processing method, device and system for sewage disposal system |
| CN109872244A (en) * | 2019-01-29 | 2019-06-11 | 汕头大学 | A kind of task instructs type wisdom agricultural planting expert system |
| CN110033011A (en) * | 2018-12-14 | 2019-07-19 | 阿里巴巴集团控股有限公司 | Traffic accident Accident Handling Method and device, electronic equipment |
| CN111026046A (en) * | 2019-11-06 | 2020-04-17 | 重庆邮电大学 | Production line equipment fault diagnosis system and method based on semantics |
| CN111797200A (en) * | 2020-06-18 | 2020-10-20 | 北京亿宇嘉隆科技有限公司 | IT operation and maintenance method |
-
2010
- 2010-08-10 CN CN201010249756.1A patent/CN101957942A/en active Pending
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106951558A (en) * | 2017-03-31 | 2017-07-14 | 广东睿盟计算机科技有限公司 | A kind of data processing method of the tax intelligent consulting platform based on deep search |
| CN106951558B (en) * | 2017-03-31 | 2020-06-12 | 广东睿盟计算机科技有限公司 | Data processing method of tax intelligent consultation platform based on deep search |
| CN109657796A (en) * | 2018-12-12 | 2019-04-19 | 北京天诚同创电气有限公司 | Judgment rule processing method, device and system for sewage disposal system |
| CN110033011A (en) * | 2018-12-14 | 2019-07-19 | 阿里巴巴集团控股有限公司 | Traffic accident Accident Handling Method and device, electronic equipment |
| CN109872244A (en) * | 2019-01-29 | 2019-06-11 | 汕头大学 | A kind of task instructs type wisdom agricultural planting expert system |
| CN109872244B (en) * | 2019-01-29 | 2023-03-10 | 汕头大学 | Task guidance type intelligent agriculture planting expert system |
| CN111026046A (en) * | 2019-11-06 | 2020-04-17 | 重庆邮电大学 | Production line equipment fault diagnosis system and method based on semantics |
| CN111797200A (en) * | 2020-06-18 | 2020-10-20 | 北京亿宇嘉隆科技有限公司 | IT operation and maintenance method |
| CN111797200B (en) * | 2020-06-18 | 2021-02-09 | 北京亿宇嘉隆科技有限公司 | IT operation and maintenance method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sharp et al. | A survey of the advancing use and development of machine learning in smart manufacturing | |
| CN101957942A (en) | Accident planning expert system applied to steel mill | |
| CN102426525B (en) | Panoramic modeling method of multi-application system | |
| CN112462696A (en) | Intelligent manufacturing workshop digital twin model construction method and system | |
| CN102143202A (en) | Information integration method and information integration system for industrial production equipment | |
| CN101135891A (en) | Graphical programming language object editing and reporting tool | |
| CN101964081A (en) | Production information system based on factory management infrastructure platform | |
| CN104951590B (en) | Knowledge service system and method for die designing | |
| CN104407576A (en) | Realization method for general type production and manufacturing unit and control system thereof | |
| CN115293507A (en) | Situation knowledge graph construction method and system for power grid dispatching operation management | |
| Mardan et al. | Combining optimisation and simulation in an energy systems analysis of a Swedish iron foundry | |
| CN103065217A (en) | Digital model splicing method of electric power setting calculating system | |
| CN107084853A (en) | The lower equipment failure prediction method of cloud manufacture | |
| CN113793111A (en) | Power equipment asset management method, system, device and storage medium | |
| Nwasuka et al. | Industry 4.0: An Overview | |
| CN106844390A (en) | A kind of inter-sectional data resource cut-in method | |
| Kaźmierczak et al. | Technical infrastructure of “smart city”: needs of integrating various management tasks | |
| CN108121297B (en) | Trapezoid figure program managing device | |
| CN101719160A (en) | GIS parallelizing system based on cluster system | |
| CN104915523B (en) | A kind of regulation full-service unified modeling method based on time series | |
| CN105204826B (en) | The automatic shared platform of more scene service information towards electrical network business and method | |
| CN103955863A (en) | Method for processing power network monitoring device data | |
| CN114186755A (en) | Visual intelligent logistics dynamic optimization management and control method and system | |
| JP5268400B2 (en) | Process management system | |
| Dobrev et al. | An ontological approach for planning and scheduling in primary steel production |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20110126 |