CN103377247B - The intelligent extract method of grind mineral control case - Google Patents
The intelligent extract method of grind mineral control case Download PDFInfo
- Publication number
- CN103377247B CN103377247B CN201210129285.XA CN201210129285A CN103377247B CN 103377247 B CN103377247 B CN 103377247B CN 201210129285 A CN201210129285 A CN 201210129285A CN 103377247 B CN103377247 B CN 103377247B
- Authority
- CN
- China
- Prior art keywords
- case
- ore
- grinding
- data
- concentration
- 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.)
- Active
Links
Landscapes
- Disintegrating Or Milling (AREA)
- Manufacture And Refinement Of Metals (AREA)
Abstract
The present invention relates to the intelligent extract method of a kind of grind mineral control case.Comprising the steps: in raw ore field, ore grinding material inlet, ore grinding discharge gate, laboratory is respectively mounted information acquisition terminal；Information acquisition terminal by gathering the Ore kind come, the grade of ore, Ore disseminated grain size, grindability, grinding particle size, ball mill entrance add the signal of discharge, overflow concentration in classifier, hydrocyclone feed ore concentration and hydrocyclone feed force and be sent in the data base that case extracts computer, case is extracted computer and is analyzed signal processing, extract effective, the grind mineral control case optimized, is then sent to the grind mineral control case extracted in the Process Control System of grinding process.Advantages of the present invention effect: the present invention is avoided that and manually determines grind mineral control initial case library redundancy or the inadequate shortcoming of coverage, it is possible to is effectively improved the qualification rate of time in grinding production granularity, stablizes time in grinding production, has significant economic benefit.
Description
Technical field
The present invention relates to the extracting method of a kind of case, particularly relate to one and utilize a large amount of historical data to extract complexity
The intelligent extract method of grind mineral control case under operating mode.
Background technology
Grinding process is the important component part of mineral processing production, and its task is to make valuable mineral and gangue mineral and various have
Leave by mutually solution between mineral, and provide qualified selected material for follow-up sorting operation.The quality that grinding process controls is directly
It is related to the production target such as metal recovery rate, concentrate grade, eventually has influence on the macroeconomic technical specification in ore dressing plant.In recent years
Coming, be difficult to ensure that uniformity for ore properties in mineral processing production, the boundary condition fluctuation range of production is big, equipment attrition and
Repeatedly after maintenance, there is change in device characteristics, poor controllability, the operating condition change feature such as frequently and pushing away based on case of using
The optimal control of reason achieves preferable effect.Good initial case is the correct key effectively run of whole control system,
Initial case library is typically to be determined input by domain expert in actual applications, but due to grinding process had comprehensive complicated
Characteristic so that manually determine and initially control case, the redundancy or the coverage that may result in control case are inadequate, for avoiding this
A little problems, The present invention gives the intelligent extract method of a kind of grind mineral control case.
Summary of the invention
The intelligent extract method of a kind of grind mineral control case is provided to solve the above-mentioned technical problem present invention, it is therefore an objective to
Have accumulated the Optimal Control case that can automatically extract grinding process in the data base of substantial amounts of historical data.
For reaching the intelligent extract method of above-mentioned purpose grind mineral control of the present invention case, comprise the steps: in raw ore field, mill
Ore deposit material inlet, ore grinding discharge gate, laboratory is respectively mounted information acquisition terminal；Information acquisition terminal will gather the Ore kind of coming
Class, the grade of ore, Ore disseminated grain size, grindability, grinding particle size, ball mill entrance add discharge, overflow concentration in classifier,
The signal of hydrocyclone feed ore concentration and hydrocyclone feed force is sent in the data base that case extracts computer, case
Example is extracted computer and is analyzed signal processing, and extracts effective, the grind mineral control case of optimization, then will extract
Grind mineral control case is sent in the Process Control System of grinding process.
Described case extraction computer is analyzed process and comprises determining that control case structure, case extract number signal
According to pretreatment, case extract the classification of data and be combined into case and then extract the case of all categories, then constitute initial
Case library.
Described determination controls case structure and includes two parts, i.e. includes that the operating mode of case describes and the solution of case；Case
Operating mode describe be to affect the factor of ore grinding, including Ore kind, head grade, Ore disseminated grain size fineness, can grind
Property and grinding particle size as case operating mode describe；The solution of case is the factor needing ore grinding to control, including new mine-supplying quantity, ball
Mill entrance adds discharge, overflow concentration in classifier, hydrocyclone feed ore concentration and hydraulic cyclone feed force as case
The solution of example.
It is according to the most true by the historical data in data base in a period of time that described case extracts the pretreatment of data
Fixed case structure is organized, i.e. by Ore kind, head grade, Ore disseminated grain size fineness, grindability, ore grinding grain
Degree, new mine-supplying quantity, ball mill entrance add discharge, overflow concentration in classifier, hydrocyclone feed ore concentration and hydraulic cyclone
Feed force constitutes a pending data set；And filter abnormal record, completion missing values and utilizing at moving average
Reason denoising, underproof for grinding particle size on the same day new mine-supplying quantity, ball mill entrance are added discharge, overflow concentration in classifier,
Hydrocyclone feed ore concentration and hydraulic cyclone feed force data record remove.
It is that data record case being extracted the pretreatment of data good utilizes from group that described case extracts the classification of data
Knit Feature Mapping network, by self training, automatically input pattern is classified.
The described case that is combined into is to extract the record that the classification of data has been classified automatically, at every oneclass classification according to case
In good record, find out the record that time in grinding production effect is best, the Ore kind of this record correspondence, head grade, ore deposit
Stone disseminated grain size fineness, grindability and grinding particle size describe as the operating mode of case, new mine-supplying quantity that this record is corresponding,
Ball mill entrance adds discharge, overflow concentration in classifier, hydrocyclone feed ore concentration and hydraulic cyclone feed force conduct
The solution of case, the solution that the operating mode of case describes with case corresponds to a new case, is stored in case library.
Advantages of the present invention effect: the present invention is avoided that and manually determines grind mineral control initial case library redundancy or coverage
Inadequate shortcoming, it is possible to be effectively improved the qualification rate of time in grinding production granularity, stablize time in grinding production, has significant economic benefit.
Detailed description of the invention
The intelligent extract method of grind mineral control case of the present invention, comprises the steps: in raw ore field, ore grinding material inlet, mill
Ore deposit discharge gate, laboratory is respectively mounted information acquisition terminal；Information acquisition terminal by gather come Ore kind, the grade of ore,
Ore disseminated grain size, grindability, grinding particle size, ball mill entrance add discharge, overflow concentration in classifier, hydrocyclone to
The signal of ore deposit concentration and hydrocyclone feed force is sent in the data base that case extracts computer, and case extracts computer
It is analyzed signal processing, extracts effective, the grind mineral control case of optimization, the grind mineral control case that then will extract
It is sent in the Process Control System of grinding process.
Signal is analyzed processing and comprises determining that control case structure, case extract data by case extraction computer
Pretreatment, case are extracted the classification of data and are combined into case and then extract the case of all categories, then constitute initial case
Storehouse.
Determine that control case structure includes two parts, i.e. include that the operating mode of case describes and the solution of case；The operating mode of case
Description is to affect the factor of ore grinding, including Ore kind, head grade, Ore disseminated grain size fineness, grindability and mill
Ore particle degree describes as the operating mode of case；The solution of case is the factor needing ore grinding to control, and enters including new mine-supplying quantity, ball mill
Mouth adds discharge, overflow concentration in classifier, hydrocyclone feed ore concentration and the hydraulic cyclone feed force solution as case.
Such as by Ore kind K, head grade, Ore disseminated grain size fineness, grindabilityWith grinding particle size Rs as case
The operating mode of example describes, respectively with f1, f2 ..., f5, ore grinding is needed the factor controlled, including new mine-supplying quantity, ball mill entrance
Add discharge, overflow concentration in classifier, hydrocyclone feed ore concentrationWith hydraulic cyclone feed force P as case
The solution of example, respectively with fs1, fs2 ... fs5 represents, then the structure of grind mineral control case is as shown in the table:
It is according to the case being determined above by the historical data in data base in a period of time that case extracts the pretreatment of data
Example structure is organized, i.e. by Ore kind, head grade, Ore disseminated grain size fineness, grindability, grinding particle size, new
Mine-supplying quantity, ball mill entrance are added discharge, overflow concentration in classifier, hydrocyclone feed ore concentration and hydraulic cyclone and are pressed to ore deposit
Power constitutes a pending data set；And filter abnormal record, completion missing values and utilize moving average to process denoising
Sound, adds discharge, overflow concentration in classifier, waterpower rotation by underproof for grinding particle size on the same day new mine-supplying quantity, ball mill entrance
Stream device feed ore concentration and hydraulic cyclone feed force data record remove.To ensure that the case excavated can ensure that ore grinding is raw
That produces is qualified.
It is that data record case being extracted the pretreatment of data good utilizes self-organizing feature that case extracts the classification of data
Mapping network, by self training, classifies to input pattern automatically.
Being combined into case is to extract the record that the classification of data has been classified automatically, in the note that every oneclass classification is good according to case
In record, find out the record that time in grinding production effect is best, Ore kind K of this record correspondence, head grade, Ore embedding
Cloth granularity fineness, grindabilityDescribing as the operating mode of case with grinding particle size Rs, what this record was corresponding newly gives ore deposit
Amount, ball mill entrance add discharge, overflow concentration in classifier, hydrocyclone feed ore concentrationAnd hydraulic cyclone
Feed force P corresponds to a new case as the solution of case, the solution that the operating mode of case describes with case, is stored in case library
In.Then extract the case of all categories, then constitute initial case library.
Claims (1)
1. the intelligent extract method of grind mineral control case, it is characterised in that comprise the steps: in raw ore field, ore grinding material inlet,
Ore grinding discharge gate, laboratory is respectively mounted information acquisition terminal；Information acquisition terminal will gather the Ore kind, the Ore product that come
Position, Ore disseminated grain size, grindability, grinding particle size, ball mill entrance add discharge, overflow concentration in classifier, hydraulic cyclone
The signal of device feed ore concentration and hydrocyclone feed force is sent in the data base that case extracts computer, and case extracts meter
Signal is analyzed processing by calculation machine, extracts effective, the grind mineral control case of optimization, the grind mineral control that then will extract
Case is sent in the Process Control System of grinding process；Described case extraction computer is analyzed process to signal and includes:
Determine and control case structure, case extracts the pretreatment of data, case extracts the classification of data and be combined into case then to extract
The case of all categories, then constitute initial case library；Described determination controls case structure and includes two parts, i.e. includes case
Operating mode describe and the solution of case；It is to affect the factor of ore grinding that the operating mode of case describes, including Ore kind, head grade,
Ore disseminated grain size fineness, grindability and grinding particle size describe as the operating mode of case；The solution of case is by ore grinding needs
The factor controlled, including new mine-supplying quantity, that ball mill entrance adds discharge, overflow concentration in classifier, hydrocyclone is dense to ore deposit
Degree and hydraulic cyclone feed force are as the solution of case；It is by data in a period of time that described case extracts the pretreatment of data
Historical data in storehouse is organized according to the case structure being determined above, i.e. by Ore kind, head grade, Ore embedding cloth grain
Degree fineness, grindability, grinding particle size, new mine-supplying quantity, ball mill entrance add discharge, overflow concentration in classifier, waterpower
Cyclone feed ore concentration and hydraulic cyclone feed force constitute a pending data set；And filter abnormal record, completion
Missing values and utilize moving average process denoising, by underproof for grinding particle size on the same day new mine-supplying quantity, ball mill entrance mend
Add discharge, overflow concentration in classifier, hydrocyclone feed ore concentration and hydraulic cyclone feed force data record to remove；Described
Case to extract the classification of data be that data record case being extracted the pretreatment of data good utilizes self-organizing feature map net
Network, by self training, classifies to input pattern automatically；The described case that is combined into is to extract dividing of data according to case
The record that class has been classified automatically, in the record that every oneclass classification is good, finds out the record that time in grinding production effect is best, this
Ore kind, head grade, Ore disseminated grain size fineness, grindability and the grinding particle size of record correspondence is as the work of case
Condition describes, and new mine-supplying quantity, ball mill entrance that this record is corresponding add discharge, overflow concentration in classifier, hydrocyclone
Feed ore concentration and hydraulic cyclone feed force as the solution of case, the operating mode of case describe the solution with case correspond to one new
Case, is stored in case library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210129285.XA CN103377247B (en) | 2012-04-28 | 2012-04-28 | The intelligent extract method of grind mineral control case |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210129285.XA CN103377247B (en) | 2012-04-28 | 2012-04-28 | The intelligent extract method of grind mineral control case |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103377247A CN103377247A (en) | 2013-10-30 |
| CN103377247B true CN103377247B (en) | 2016-11-02 |
Family
ID=49462373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210129285.XA Active CN103377247B (en) | 2012-04-28 | 2012-04-28 | The intelligent extract method of grind mineral control case |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103377247B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103713619B (en) * | 2013-12-20 | 2016-05-25 | 东北大学 | A kind of method of testing of grinding process progress control method |
| CN109214415B (en) * | 2018-07-19 | 2021-11-02 | 西安建筑科技大学 | Method for judging ore crushing and grinding effect |
| CN112598618B (en) * | 2020-11-16 | 2023-11-17 | 鞍钢集团矿业有限公司 | Image recognition technology-based ore feeding amount prediction method for mill |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1598534A (en) * | 2004-07-28 | 2005-03-23 | 东北大学 | Soft investigating method for overflow grain index of ore grinding system based on case inference |
| CN1603986A (en) * | 2004-10-29 | 2005-04-06 | 东北大学 | Grinding system intelligent optimization initialization method based on case ratiocination |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009077940A1 (en) * | 2007-12-14 | 2009-06-25 | University Of Kwazulu-Natal | A method of grinding a mineral-containing ore |
-
2012
- 2012-04-28 CN CN201210129285.XA patent/CN103377247B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1598534A (en) * | 2004-07-28 | 2005-03-23 | 东北大学 | Soft investigating method for overflow grain index of ore grinding system based on case inference |
| CN1603986A (en) * | 2004-10-29 | 2005-04-06 | 东北大学 | Grinding system intelligent optimization initialization method based on case ratiocination |
Non-Patent Citations (2)
| Title |
|---|
| 选矿生产过程综合自动化系统;柴天佑等;《首届全国有色金属自动化技术与应用学术年会论文集》;20031101;第1-5页 * |
| 选矿过程工艺指标确定方法及应用;杨新等;《控制工程》;20090520;第16卷(第3期);第371-374、382页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103377247A (en) | 2013-10-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101559403B (en) | Method for concentrating extremely low-grade refractory magnetite | |
| CN104394993B (en) | A kind of preprocess method of ilmenite iron selection tailings | |
| CN102698871B (en) | Ore dressing technology for treating vanadium titano-magnetite | |
| CN108380380A (en) | Product dry type fine crushing sorts Iron concentrate technique | |
| CN107350157B (en) | A kind of stage grinding of microfine low-grade magnetite, stage grading new method | |
| CN104384020B (en) | A kind of Ultra-low-grade magnetite stone depth crushing ore-dressing technique | |
| CN104134120A (en) | System and method for monitoring ore-dressing production indexes | |
| CN105107616A (en) | Low-cost high-recovery environment-friendly mineral separation method of low-grade vanadium titano-magnetite | |
| CN112330653B (en) | Ore granularity online detection method based on image recognition technology | |
| CN102716793A (en) | Method and system for producing ultra-pure powdered iron | |
| CN204448257U (en) | A kind ofly select iron device for super low-grade vanadium titano-magnetite | |
| CN103377247B (en) | The intelligent extract method of grind mineral control case | |
| CN103056020A (en) | Maximum optional size grading process for beneficiation and grinding | |
| CN102824956B (en) | Poor hematite grading level and narrow level sorting process | |
| CN105107617A (en) | Method for increasing comprehensive utilization ratio of ilmenite | |
| CN107413517A (en) | A kind of beneficiation method for improving micro fine particle magnetite stone stage grinding and stage separation sorting index | |
| CN104056714A (en) | Beneficiation process for refractory micro-fine particle-graded iron-copper ore | |
| CN106362863A (en) | Strongly magnetic mineral accurate magnetic separation method | |
| CN110052321A (en) | A kind of two sections of Continuous Grinding grading technologies of non-standard configuration | |
| CN109894267A (en) | The magnetic of the red compound iron ore of magnetic-- weight method for separating | |
| CN103736603B (en) | A kind of ore-dressing technique of grind grading | |
| CN108080139A (en) | A kind of uranium-bearing paigeite sorting process | |
| CN110694937A (en) | Low-grade dip-dyed skarn type scheelite pre-waste-throwing process | |
| CN206519238U (en) | A kind of novel magnetite beneficiation system | |
| CN201804265U (en) | Intelligent grading control equipment for ore grinding of ore dressing plants |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |