CN104134120A - System and method for monitoring ore-dressing production indexes - Google Patents

System and method for monitoring ore-dressing production indexes Download PDF

Info

Publication number
CN104134120A
CN104134120A CN201410370724.5A CN201410370724A CN104134120A CN 104134120 A CN104134120 A CN 104134120A CN 201410370724 A CN201410370724 A CN 201410370724A CN 104134120 A CN104134120 A CN 104134120A
Authority
CN
China
Prior art keywords
index
ore
production target
magnetic
data matrix
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
Application number
CN201410370724.5A
Other languages
Chinese (zh)
Other versions
CN104134120B (en
Inventor
俞胜平
郑秀萍
初延刚
王昭
徐泉
胡毅
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northeastern University China
Original Assignee
Northeastern University China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Northeastern University China filed Critical Northeastern University China
Priority to CN201410370724.5A priority Critical patent/CN104134120B/en
Publication of CN104134120A publication Critical patent/CN104134120A/en
Application granted granted Critical
Publication of CN104134120B publication Critical patent/CN104134120B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The invention discloses a system and method for monitoring ore-dressing production indexes, and belongs to the technical field of ore-dressing production processes. The method includes the steps of data obtaining, index influencing factor screening, index monitoring, case library maintaining, index anomaly analyzing and data storing. Deep analysis of influence factors of ore-dressing important production indexes and analysis of abnormal conditions of the ore-dressing important production indexes are achieved, and the efficiency degree of monitoring on the ore-dressing important production indexes is improved; the problems that only statistic displaying of all ore-dressing production indexes is carried out in current ore-dressing production index monitoring, the way how the important production indexes most cared by an enterprise are effectively monitored by analyzing mutual influence relations between the various production indexes is not researched, causes of anomalies of the indexes in the ore-dressing process are not analyzed and processed, and therefore the important production indexes in the ore-dressing process can not be effectively monitored in an on-line mode are solved.

Description

A kind of mineral processing production indicators monitor system and method
Technical field
The invention belongs to dressing Production Process technical field, be specifically related to a kind of mineral processing production indicators monitor system and method.
Background technology
Mineral resources are a kind of important foundation raw materials of economy development requirement, at the aspect such as the national economic development, the development of defense-related science and technology, in multiple fields such as metallurgy, building, traffic, chemical industry, play a part very important, mineral resources are again the natural resourcess that is difficult to regeneration simultaneously, so each state is all advocating the strategy of sustainable development energetically.Under these circumstances, mining processing industry enterprise can not merely pursue economic benefit as before, and should more focus on the quality of product, reduce production costs, consumption economizes on resources, reduce environmental pollution, only in this way could meet the more and more higher quality requirements of enterprise, the survival and development better of Cai Nengshi enterprise.In dressing Production Process, exist a large amount of production targets, comprise quality index, measuring index, equipment operating statistic, target energy, the indicator of costs, ore storage bin material level and technic index.Above-mentioned various production targets are carried out to effective monitoring, particularly important to ensureing product quality.
At present only there is to a small amount of patent mineral processing production index monitoring aspect, as " 201410053940 (method of tin ore tabling automatic monitoring) ", by automatic camera module monitors, by PLC regulating and controlling valve, realizes best washings flow control; By automatic camera module monitors, add single shaft control servomotor driving device member by PLC and realize best lathe bed slope control; By automatic camera module monitors, control mechanical component by PLC and realize tin concentrate and access position control.Thereby guarantee that shaking table feed ore concentration, volume are stabilized in optimum condition, improve sorting index." 201310704334.2 (the ore dressing comprehensive production index optimization method under a kind of capacity of equipment change condition) " is in the time that overhaul of the equipments or equipment failure cause capacity of equipment to occur changing, for ore dressing comprehensive production index optimization aim and constraint condition, each ore handling capacity is adjusted to optimization, and then realize the optimization of ore dressing comprehensive production index." 201310647027.5 (a kind of ore dressing process operating index optimization method) " forecasts quality index and production index, obtain quality index predicted value and production index predicted value, the default definite value of proofreading and correct operating index desired value, obtains operating index optimal value." 201310723320.5 (visual mineral processing production full-flow process index optimization decision system and method) " to algorithm encapsulate, configuration data interface, configuration data type, or packaged algoritic module is modified; Configuration forms mineral processing production whole process control strategy; To trend grouping and variable data packet configuration dynamic-configuration display interface.Above-mentioned patent mainly for be how mineral processing production target setting value is optimized, and the important production target that how research is not concerned about enterprise by the relation that influences each other of studying between various production targets most effectively monitors, more index abnormal cause in dressing Production Process is not analyzed and processed, be therefore difficult to the important production target of ore dressing to carry out online effective monitoring.
Summary of the invention
For the shortcoming of prior art, the present invention proposes a kind of mineral processing production indicators monitor system and method, realize in-depth analysis to the important production target influence factor of ore dressing to reach, analysis during to the important production target abnormal conditions of ore dressing, improve the object of the important production target effective monitoring of ore dressing degree.
A kind of mineral processing production indicators monitor system, comprises data capture unit, Factors Affecting Parameters screening unit, index monitor unit, case library maintenance unit, Indexes Abnormality analytic unit and data storage cell, wherein,
Data capture unit: for obtaining mineral processing production whole process production target historical data, comprise quality index, measuring index, equipment operating statistic index, ore storage bin material level index, technic index, target energy and the indicator of costs, and the data of obtaining are sent in Factors Affecting Parameters screening unit and data storage cell;
Factors Affecting Parameters screening unit: for determining according to the actual requirements monitor control index and having multiple production targets that affect of the relation of impact with monitor control index, and the mineral processing production index screening method of employing based on PLS-VIP, further screen from multiple the impact production target of choosing, determine crucial production target, and crucial production target is sent to case library maintenance unit;
Case library maintenance unit: safeguard for the case library that analysis is used to Indexes Abnormality, comprise interpolation case, revise case, delete case and check case;
Index monitor unit: for according to the monitoring image of monitor control index, the numerical value of the crucial production target of on line real-time monitoring, and be sent to Indexes Abnormality analytic unit;
Indexes Abnormality analytic unit: for occurring when monitor control index when abnormal, adopt based on reasoning by cases method, according to the historical case of recording in case library, determine case under current Indexes Abnormality situation;
Data storage cell: for store the mineral processing production whole process production target historical data of obtaining, the monitor control index of choosing, choose multiplely affect the production target historical datas that production target, the in real time crucial production target actual numerical value that detects and case library are recorded with monitor control index has a relation of impact.
The method for supervising that adopts mineral processing production indicators monitor system to carry out, comprises the following steps:
Step 1, obtain mineral processing production whole process production target historical data, comprise quality index, measuring index, equipment operating statistic index, ore storage bin material level index, technic index, target energy and the indicator of costs;
Step 2, is according to the actual requirements chosen required monitor control index from mineral processing production whole process production target;
Step 3, is according to the actual requirements chosen the multiple production targets that affect that have the relation of impact with monitor control index from mineral processing production whole process production target;
Step 4, the mineral processing production index screening method of employing based on PLS-VIP, further screen from multiple the impact production target of choosing, and determines crucial production target, and concrete grammar is as follows:
Step 4-1, according to historical data, determine that many groups affect the concrete numerical value of production target, using this production target as independent variable, and build argument data matrix;
The line number of argument data matrix is the group number that affects production target, and matrix column number is the number that affects production target, and entry of a matrix element is for affecting the concrete numerical value of production target;
Step 4-2, according to historical data, determine the concrete numerical value of many group monitor control indexs, be dependent variable by monitor control index, and build dependent variable data matrix;
The line number of dependent variable data matrix is the group number of monitor control index, the number that matrix column number is monitor control index, and entry of a matrix element is the concrete numerical value of monitor control index;
Step 4-3, argument data matrix and dependent variable data matrix are carried out to standardization;
Step 4-4, employing NIPALS method are carried out Principle component extraction to argument data matrix and dependent variable data matrix;
Step 4-5, determine the variable drop importance index of each independent variable;
VI P j = p Rd ( Y ; t 1 , . . . , t s ) Σ h = 1 s Rd ( Y ; t h ) w hj 2 - - - ( 1 )
Wherein, VIP jbe the variable drop importance index of j independent variable, j=1,2 ..., p, p is independent variable number;
S is the total degree of Principle component extraction;
T hfor the major component that the h time is extracted from argument data matrix;
W hjfor matrix the corresponding unit character vector w of eigenvalue of maximum institute hj component, E h-1represent the argument data matrix obtaining in the h-1 time Principle component extraction, F h-1represent the dependent variable data matrix obtaining in the h-1 time Principle component extraction; F is the matrix after dependent variable data matrix Y standardization;
rd (y j; t h) be t hto the interpretability of dependent variable data matrix Y;
Rd (y j; t h)=r 2(y j; t h), r (y j; t h) be t hwith dependent variable y jsimple correlation coefficient;
for t 1..., t sto the accumulation interpretability of dependent variable data matrix Y;
Step 4-6, judge whether the variable drop importance index of each independent variable is more than or equal to variable drop importance index threshold value, if, retain this independent variable, retain the corresponding production target that affects, execution step 4-8 after all independents variable have all judged, otherwise, this independent variable deleted, deleting correspondence affects production target, execution step 4-7 after all independents variable have all judged;
Step 4-7, remaining independent variable is reconstituted to argument data battle array, be about to many group lingering effect production targets and reconstitute argument data battle array, and return to execution step 4-4 to step 4-6, be all more than or equal to variable drop importance index threshold value until remain the variable drop importance index of each independent variable;
Step 4-8, by final remaining independent variable, the final remaining production target that affects is as crucial production target;
Step 5, according to obtain crucial production target, inquiry case library, determine when monitor control index is abnormal the numerical value that crucial production target is corresponding, and the analysis result of storing in case library in this situation and adjustment scheme;
Step 6, field personnel are according to the monitoring image of monitor control index, the numerical value of the crucial production target of on line real-time monitoring, when monitor control index occurs when abnormal, adopt based on reasoning by cases method, according to the historical case of recording in case library, determine the affiliated case of current Indexes Abnormality situation, concrete grammar is as follows:
Step 6-1, obtain the concrete numerical value of current crucial production target;
Step 6-2, the abnormal crucial production target of definite generation;
Make it by the actual numerical value of each crucial production target and its setting value poor, the actual numerical value of poor absolute value and crucial production target is divided by and is obtained indicator difference rate, indicator difference rate is arranged to threshold with it, set indicator difference rate threshold value if be greater than, this key production target occurs abnormal;
Step 6-3, adopt nearest neighbor method, determine the similarity of each case in current monitor control index abnormal conditions and case library:
SIM ( G in , G k ) = Σ i = 1 m ω i × sim ( v i , v i , k ) Σ i = 1 m ω i - - - ( 2 )
Wherein, G infor current Indexes Abnormality situation;
G kfor k case in case library;
M is the number of crucial production target;
ω ibe the weights of i crucial production target with respect to monitor control index impact;
SIM (G in, G k) be the similarity of current Indexes Abnormality situation and k case;
V i, kbe the value of i feature in k case, i.e. i crucial production target;
Sim (v i, v i, k) be i crucial production target in current Indexes Abnormality situation, the similarity with i in k case crucial production target, is calculated as follows:
sim ( v i , v i , k ) = 1 - | v i - v i , k | max ( v i , v i , k ) - - - ( 3 )
Step 7, select current monitor control index abnormal conditions and the highest case of historical case similarity, as case under current Indexes Abnormality situation, according to analysis result and adjustment scheme that in case library, this case is recorded, instructing measure to be issued to corresponding production process concrete production adjustment carries out scene adjustment, ensures that monitor control index meets the demands.
Quality index described in step 1 comprises the comprehensive fine work of ore dressing position, smart moisture is combined in ore dressing, smart scaling loss Ig is combined in ore dressing, smart S is combined in ore dressing, smart CaO is combined in ore dressing, smart SiO2 is combined in ore dressing, measuring and calculating grade of sinter, measuring and calculating sintering deposit SiO2, combine smart granularity, combine accurate measurement and calculate compliance rate, combine smart SiO2 compliance rate, comprehensive lump ore rate, the primary overflow recovery, roasted ore grade, barren rock grade, weak magnetic magnetic fine work position, weak magnetic magnetic essence SiO2, weak magnetic magnetic essence CaO, weak magnetic magnetic essence Ig, weak magnetic enters to grind grade, weak fine work position, weak tail grade, weak magnetic magnetic tail grade, three magnetic fine work positions, weak fine work position qualification rate, flotation is to ore deposit grade, flotation is to ore deposit SiO2, the floating tail grade of weak magnetic, the floating smart SiO2 of weak magnetic, strong magnetic enters to grind grade, strong fine work position, the strong comprehensive tailings grade of magnetic, flat ring is combined fine work position, high gradient is combined fine work position, high gradient tailings grade, strong fine work position qualification rate, strong smart SiO2, strong magnetic feed particle size, the selected concentration of strong magnetic, 1-2 revolves excessive concentration, 2-2 revolves excessive concentration, weak three concentration of magnetic, magnetic tailing grade, magnetic tailing concentration, full choosing ratio, primary overflow yield qualification rate, barren rock bucket number, barren rock amount, barren rock productive rate, the theoretical metal recovery rate of low intensity magnetic separation, the theoretical ratio of concentration of low intensity magnetic separation, flotation choosing ratio, the theoretical metal recovery rate of high intensity magnetic separation, the grade of the theoretical ratio of concentration of high intensity magnetic separation and various raw ores, S content, SiO2 content and CaO content and moisture content,
Described measuring index comprises that smart output is combined in the ore dressing that comprises moisture content, smart output, inferior fine magnetite concentrate output, high intensity magnetic mineral output are combined in the ore dressing of removing moisture content, selects 3# ore deposit amount, dry weight and the weight in wet base of piece 1# ore deposit amount, powder 2# ore deposit amount, finished product-2# ore deposit amount, barren rock-1# ore deposit amount, 2X essence 4# ore deposit amount, 1-1# ball milling ore deposit amount, 2-1# ball milling ore deposit amount, 3-1# ball milling ore deposit amount, 4-1# ball milling ore deposit amount, strong magnetic mine-supplying quantity, weak magnetic mine-supplying quantity, go down the hill ore deposit amount and various raw ores;
When described equipment operating statistic index comprises raw ore stove when fortune, strong magnetic bowl mill fortune, when weak magnetic bowl mill fortune, bowl mill operating rate, strong magnetic bowl mill operating rate and weak magnetic bowl mill operating rate, when operating rate is bowl mill fortune and the ratio of T.T.;
Described ore storage bin material level index comprises accumulating storage ore storage bin material level, once sieves ore storage bin material level, regrading ore storage bin material level, strong magnetic ore storage bin material level and weak magnetic ore storage bin material level;
Described technic index comprises shaft furnace heating gas amount, reduction shaft furnace coal gas amount, intensity magnetic separator electric current, strong magnetic machine drift ice is washed electric current, the dynamo-electric stream of vertical ring, and vertical ring machine drift ice is washed electric current, floating agent, concentration, frequency, the flow of concentrated large well, the pressurization storehouse pressure of pressing filter;
Described target energy comprises the unit consumption of electricity, Zhong Shui, Xin Shui, coke-oven gas, blast furnace gas, steam, Living Water and always consumes;
The described indicator of costs comprises raw material unit cost, raw material total cost, energy unit cost and energy total cost.
Employing NIPALS method described in step 4-4 is carried out Principle component extraction to argument data matrix and dependent variable data matrix, and concrete grammar is as follows:
Step 4-4-1, acquisition matrix the corresponding unit character of eigenvalue of maximum vector w h, and then h major component t of acquisition independent variable h:
t h=E h-1w h (4)
Wherein, E h-1represent the argument data matrix obtaining in the h-1 time Principle component extraction, F h-1represent the dependent variable data matrix obtaining in the h-1 time Principle component extraction; be the transposed matrix of the dependent variable data matrix that obtains in the h-1 time Principle component extraction, it is the argument data transpose of a matrix matrix obtaining in the h-1 time Principle component extraction; In the time of h=1, E 0represent the argument data matrix building, F 0represent the dependent variable data matrix building;
Step 4-4-2, acquisition matrix the corresponding unit character of eigenvalue of maximum vector c h, and then h major component u of acquisition dependent variable h:
u h=F h-1c h (5)
Step 4-4-3, according to the h major component of independent variable and the h major component of dependent variable, obtain the residual matrix of dependent variable data matrix and dependent variable data matrix:
E h = E h - 1 - t h p h T F h = F h - 1 u h r h T - - - ( 6 )
Wherein, p hrepresent t hload on argument data matrix,
R hrepresent u hload on dependent variable data matrix,
Step 4-4-4, judgement major component t now hwhether the intersection validity value to dependent variable data matrix Y is less than setting value, and if so, complete argument data matrix and dependent variable data matrix are carried out to Principle component extraction, otherwise, execution step 4-4-5;
The validity value computing formula of intersecting is as follows:
Q h 2 = 1 - PRESS ( h ) SS ( h - 1 ) - - - ( 7 )
Wherein, represent the major component t extracting for the h time hto the intersection validity of dependent variable data matrix Y;
j=1,2 ..., q, q is dependent variable number, PRESS j(h) be dependent variable y jprediction sum squares: y ijrepresent the value of j dependent variable in i group data, represent y ijat independent variable sample point x (i)on predicted value; I=1,2 ..., n, the group number that n is required monitor control index;
sS j(h-1) represent y jerror of fitting quadratic sum; y ijrepresent the value of j dependent variable in i group data, represent y jat independent variable sample point x (i)on match value;
Step 4-4-5, return execution step 4-4-1 to step 4-4-4, until major component t hintersection validity value to dependent variable Y is less than setting value, completes argument data matrix and dependent variable data matrix are carried out to Principle component extraction.
Advantage of the present invention:
Only that the statistics of all mineral processing production indexs is shown for current mineral processing production index monitoring, the important production target that how research is not concerned about enterprise by the relation that influences each other of analyzing between various production targets most effectively monitors and index abnormal cause in ore dressing process is not analyzed and processed, thereby cause being difficult to the important production target of ore dressing process to be carried out the problem of online effective monitoring, the present invention proposes a kind of mineral processing production indicators monitor system and method, comprise data acquisition, Factors Affecting Parameters screening, index monitors, case library is safeguarded, Indexes Abnormality is analyzed and data storage, realize in-depth analysis to the important production target influence factor of ore dressing and analysis during to the important production target abnormal conditions of ore dressing, improved the degree of functioning of the important production target monitoring of ore dressing.
Brief description of the drawings
Fig. 1 is the structured flowchart of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Fig. 2 is the process flow diagram of a kind of mineral processing production index method for supervising of the specific embodiment of the invention;
Fig. 3 is the process flow diagram of the Factors Affecting Parameters screening of a kind of mineral processing production index method for supervising of the specific embodiment of the invention;
Fig. 4 is that smart grade index monitoring image is combined in the ore dressing of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Fig. 5 is the primary overflow recovery index monitoring image of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Fig. 6 is that the weak magnetic of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention enters to grind grade index monitoring image;
Fig. 7 is the strong smart grade index monitoring image of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Fig. 8 is the strong magnetic feed particle size index monitoring image of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Fig. 9 is that the strong magnetic of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention enters to grind grade index monitoring image;
Figure 10 is the weak smart grade index monitoring image of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 11 is three size indicator monitoring images of weak magnetic of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 12 is the pre-ore dressing grade index monitoring image of the historical time section of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 13 is that smart grade index monitoring image is combined in the online ore dressing in real time of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 14 is the online primary overflow recovery index monitoring image in real time of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 15 is that the online in real time weak magnetic of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention enters to grind grade index monitoring image;
Figure 16 is the online strong smart grade index monitoring image in real time of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 17 is the online in real time strong magnetic feed particle size index monitoring image of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 18 is that the online in real time strong magnetic of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention enters to grind grade index monitoring image;
Figure 19 is the online in real time weak smart grade index monitoring image of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 20 is online in real time weak three size indicator monitoring images of magnetic of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 21 is the online in real time pre-ore dressing grade index monitoring image of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention;
Figure 22 is the monitoring image of the ore dressing of a kind of mineral processing production indicators monitor system of the specific embodiment of the invention while combining fine work position Indexes Abnormality.
Embodiment
Below in conjunction with accompanying drawing, an embodiment of the present invention is described further.
In the embodiment of the present invention, mineral processing production indicators monitor system as shown in Figure 1, comprises data capture unit, Factors Affecting Parameters screening unit, index monitor unit, case library maintenance unit, Indexes Abnormality analytic unit and data storage cell, wherein,
Data capture unit: for obtaining mineral processing production whole process production target historical data, comprise quality index, measuring index, equipment operating statistic index, ore storage bin material level index, technic index, target energy and the indicator of costs, and the data of obtaining are sent in Factors Affecting Parameters screening unit and data storage cell;
Factors Affecting Parameters screening unit: for determining according to the actual requirements q monitor control index and having p of the relation of impact to affect production target with monitor control index, and the mineral processing production index screening method of employing based on PLS-VIP, further screen from multiple the impact production target of choosing, (m < is individual crucial production target p), and crucial production target is sent to case library maintenance unit to determine m;
Case library maintenance unit: for according to m the crucial production target obtaining, inquiry case library, determine when monitor control index is abnormal, numerical value corresponding to crucial production target, and the analysis result of storing in case library in this situation and adjustment scheme, and be sent to Indexes Abnormality analytic unit and data storage cell;
In the embodiment of the present invention, case library maintenance unit is safeguarded the case library of using in Indexes Abnormality analysis, comprises interpolation case, revises case, deletes case, checks the function such as case, help;
Index monitor unit: set up the monitoring image of the mineral processing production index that needs monitoring, show in real time in this picture that the mineral processing production index with needs monitoring that needs q of monitoring mineral processing production desired value and filter out has m crucial production target value of major effect relation simultaneously;
In the embodiment of the present invention, the C# under employing Microsoft Visual Studio 2008 carries out the structure of monitoring image;
Indexes Abnormality analytic unit: for occurring when monitor control index when abnormal, adopt based on reasoning by cases method, according to the historical case of recording in case library, determine case under current Indexes Abnormality situation;
Data storage cell: for store the mineral processing production whole process production target historical data of obtaining, the monitor control index of choosing, choose multiplely affect the production target historical datas that production target, the in real time crucial production target actual numerical value that detects and case library are recorded with monitor control index has a relation of impact.
In the embodiment of the present invention, the method for supervising that adopts mineral processing production indicators monitor system to carry out, method flow diagram as shown in Figure 2, comprises the following steps:
Step 1, obtain mineral processing production whole process production target historical data, comprise quality index, measuring index, equipment operating statistic index, ore storage bin material level index, technic index, target energy and the indicator of costs;
Described quality index comprises the comprehensive fine work of ore dressing position, smart moisture is combined in ore dressing, smart scaling loss Ig is combined in ore dressing, smart S is combined in ore dressing, smart CaO is combined in ore dressing, smart SiO2 is combined in ore dressing, measuring and calculating grade of sinter, measuring and calculating sintering deposit SiO2, combine smart granularity, combine accurate measurement and calculate compliance rate, combine smart SiO2 compliance rate, comprehensive lump ore rate, the primary overflow recovery, roasted ore grade, barren rock grade, weak magnetic magnetic fine work position, weak magnetic magnetic essence SiO2, weak magnetic magnetic essence CaO, weak magnetic magnetic essence Ig, weak magnetic enters to grind grade, weak fine work position, weak tail grade, weak magnetic magnetic tail grade, three magnetic fine work positions, weak fine work position qualification rate, flotation is to ore deposit grade, flotation is to ore deposit SiO2, the floating tail grade of weak magnetic, the floating smart SiO2 of weak magnetic, strong magnetic enters to grind grade, strong fine work position, the strong comprehensive tailings grade of magnetic, flat ring is combined fine work position, high gradient is combined fine work position, high gradient tailings grade, strong fine work position qualification rate, strong smart SiO2, strong magnetic feed particle size, the selected concentration of strong magnetic, 1-2 revolves excessive concentration, 2-2 revolves excessive concentration, weak three concentration of magnetic, magnetic tailing grade, magnetic tailing concentration, full choosing ratio, primary overflow yield qualification rate, barren rock bucket number, barren rock amount, barren rock productive rate, the theoretical metal recovery rate of low intensity magnetic separation, the theoretical ratio of concentration of low intensity magnetic separation, flotation choosing ratio, the theoretical metal recovery rate of high intensity magnetic separation, the grade of the theoretical ratio of concentration of high intensity magnetic separation and various raw ores, S content, SiO2 content and CaO content and moisture content,
Described measuring index comprises that smart output is combined in the ore dressing that comprises moisture content, smart output, inferior fine magnetite concentrate output, high intensity magnetic mineral output are combined in the ore dressing of removing moisture content, selects 3# ore deposit amount, dry weight and the weight in wet base of piece 1# ore deposit amount, powder 2# ore deposit amount, finished product-2# ore deposit amount, barren rock-1# ore deposit amount, 2X essence 4# ore deposit amount, 1-1# ball milling ore deposit amount, 2-1# ball milling ore deposit amount, 3-1# ball milling ore deposit amount, 4-1# ball milling ore deposit amount, strong magnetic mine-supplying quantity, weak magnetic mine-supplying quantity, go down the hill ore deposit amount and various raw ores;
When described equipment operating statistic index comprises raw ore stove when fortune, strong magnetic bowl mill fortune, when weak magnetic bowl mill fortune, bowl mill operating rate, strong magnetic bowl mill operating rate and weak magnetic bowl mill operating rate, when operating rate is bowl mill fortune and the ratio of T.T.;
Described ore storage bin material level index comprises accumulating storage ore storage bin material level, once sieves ore storage bin material level, regrading ore storage bin material level, strong magnetic ore storage bin material level and weak magnetic ore storage bin material level;
Described technic index comprises shaft furnace heating gas amount, reduction shaft furnace coal gas amount, intensity magnetic separator electric current, strong magnetic machine drift ice is washed electric current, the dynamo-electric stream of vertical ring, and vertical ring machine drift ice is washed electric current, floating agent, concentration, frequency, the flow of concentrated large well, the pressurization storehouse pressure of pressing filter;
Described target energy comprises the unit consumption of electricity, Zhong Shui, Xin Shui, coke-oven gas, blast furnace gas, steam, Living Water and always consumes;
The described indicator of costs comprises raw material unit cost, raw material total cost, energy unit cost and energy total cost.
Step 2, is according to the actual requirements chosen required monitor control index from mineral processing production whole process production target;
In the embodiment of the present invention, from mineral processing production whole process production target, determining the index that need to monitor is the comprehensive concentrate grade of ore dressing;
Step 3, is according to the actual requirements chosen the multiple production targets that affect that have the relation of impact with monitor control index from mineral processing production whole process production target;
In the embodiment of the present invention, the comprehensive concentrate grade of ore dressing of monitoring as required, just selects 30 production targets that have the relation of impact with the comprehensive concentrate grade of ore dressing: strong magnetic enters to grind grade, strong fine work position, the strong comprehensive tailings grade of magnetic, flat ring concentrate grade, flat ring tailings grade, high gradient tailings grade, combines smart granularity (200 order), strong magnetic feed particle size, 1-2 revolve excessive concentration, 2-2 revolves excessive granularity, 2-2 revolves excessive concentration, three granularities of weak magnetic (300 order), three concentration of weak magnetic, weak magnetic magnetic fine work position, weak magnetic magnetic essence SiO by artificial experience 2, weak magnetic magnetic essence CaO, weak magnetic enter to grind grade, weak fine work position, weak magnetic magnetic tail grade, three magnetic fine work positions, flotation to ore deposit grade, flotation to ore deposit SiO 2, the floating fine work of weak magnetic position, the floating smart SiO2 of weak magnetic, the primary overflow recovery, magnetic tailing grade, pre-ore dressing grade, Hei Gou ore deposit grade, periphery ore deposit grade, imported iron ore fines grade;
Step 4, the mineral processing production index screening method of employing based on PLS-VIP, affect further screening production target from 30 of choosing, and determines the individual crucial production target of m (m < 30), and as shown in Figure 3, concrete grammar is as follows:
Step 4-1, according to historical data, determine 151 groups of concrete numerical value that affect production target, using this production target as independent variable, and build argument data matrix; The line number of argument data matrix is 151, and matrix column number is 30, and entry of a matrix element is for affecting the concrete numerical value of production target;
Step 4-2, according to historical data, determine the concrete numerical value of 151 groups of monitor control indexs, be dependent variable by monitor control index, and build dependent variable data matrix; The line number of dependent variable data matrix is 151, and matrix column number is 1, and entry of a matrix element is the concrete numerical value of monitor control index;
In the embodiment of the present invention, set up 30 independent variable x 1..., x 30, the corresponding and comprehensive concentrate grade of ore dressing has 30 production targets of the relation of impact respectively; Set up 1 dependent variable y 1, to the comprehensive concentrate grade of ore dressing in requisition for monitoring;
According to 30 independents variable and 1 dependent variable set up, choose 151 groups of data in May, 2013 as analyzing samples (as shown in table 1), form respectively argument data battle array X=(x ij) 151 × 30(i=1 ..., 151; J=1 ..., 30) and dependent variable data matrix Y=(y ij) 151 × 1(i=1 ...., 151; J=1);
The original index data of table 1 sample
Time Strong magnetic enters to grind grade x1 Strong fine work position x2 Peripheral frame grade x29 Imported iron ore fines grade x30 Fine work position (Tfe) y is combined in ore dressing
2013/5/115:00:00 31.8 44.5 33.45 61.55 54.3
2013/5/117:00:00 32.7 46 33.86 60.5 53.56
2013/5/119:00:00 32.7 45.6 35.71 62.1 53.8
2013/5/121:00:00 31.15 44.9 36.45 60.4 53.6
2013/5/1313:00:00 30.6 45.9 34.86 59.65 53.05
2013/5/1315:00:00 32.1 46.9 32.05 60.5 53.03
2013/5/1317:00:00 30.6 46.6 34.79 60.5 53.45
2013/5/1319:00:00 31.55 46.8 31.95 61.35 53.21
2013/5/1321:00:00 33.55 47.7 32.2 58.6 53.21
Step 4-3, argument data matrix X and dependent variable data matrix Y are carried out to standardization;
In the embodiment of the present invention, course of standardization process is as follows:
e ij = x ij - x &OverBar; j S x j ( i = 1,2 , . . . , 151 ; j = 1,2 , . . . , 30 ) , E 0 = ( e ij ) 151 &times; 30 ;
f ij = y ij - y &OverBar; j S y j ( i = 1,2 , . . . , 151 ; j = 1 ) , F 0 = ( f ij ) 151 &times; 1 ;
Wherein, e ijfor the argument data matrix E after standardization 0the capable j column element of i value; f ijfor the dependent variable data matrix F after standardization 0the capable j column element of i value; for the mean value of the j column data of argument data matrix; for the mean value of the j column data of dependent variable data matrix Y; for the standard deviation of the j column data of argument data matrix; for the standard deviation of the j column data of dependent variable data matrix; Data after standardization are as shown in table 2:
Data after table 2 standardization
e1 e2 e29 e30 f0
0.8243 -0.5540 -0.2898 0.6957 1.3693
1.5881 0.7264 -0.1137 -0.2555 0.0923
1.5881 0.3850 0.6813 1.1940 0.5065
0.2726 -0.2125 0.9993 -0.3461 0.1614
-0.1942 0.6410 0.3160 -1.0256 -0.7877
1.0789 1.4946 -0.8914 -0.2555 -0.8222
-0.1942 1.2385 0.2860 -0.2555 -0.0975
0.6121 1.4092 -0.9344 0.5145 -0.5116
2.3096 2.1775 -0.8270 -1.9765 -0.5116
Step 4-4, employing NIPALS method are carried out Principle component extraction to argument data matrix and dependent variable data matrix;
Adopt NIPALS method to carry out Principle component extraction to argument data matrix and dependent variable data matrix, concrete grammar is as follows:
Step 4-4-1, acquisition matrix the corresponding unit character of eigenvalue of maximum vector w h, and then h major component t of acquisition independent variable h:
t h=E h-1w h (4)
Wherein, E h-1represent the argument data matrix obtaining in the h-1 time Principle component extraction, F h-1represent the dependent variable data matrix obtaining in the h-1 time Principle component extraction; be the transposed matrix of the dependent variable data matrix that obtains in the h-1 time Principle component extraction, it is the argument data transpose of a matrix matrix obtaining in the h-1 time Principle component extraction; In the time of h=1, E 0represent the argument data matrix building, F 0represent the dependent variable data matrix building;
Step 4-4-2, acquisition matrix the corresponding unit character of eigenvalue of maximum vector c h, and then h major component u of acquisition dependent variable h:
u h=F h-1c h (5)
Step 4-4-3, according to the h major component of independent variable and the h major component of dependent variable, obtain the residual matrix of dependent variable data matrix and dependent variable data matrix:
E h = E h - 1 - t h p h T F h = F h - 1 u h r h T - - - ( 6 )
Wherein, p hrepresent t hload on argument data matrix,
R hrepresent u hload on dependent variable data matrix,
Step 4-4-4, judgement major component t now hwhether the intersection validity value to dependent variable data matrix Y is less than setting value 0.0975, and if so, complete argument data matrix and dependent variable data matrix are carried out to Principle component extraction, otherwise, execution step 4-4-5:
The validity value computing formula of intersecting is as follows:
Q h 2 = 1 - PRESS ( h ) SS ( h - 1 ) - - - ( 7 )
Wherein, represent the major component t extracting for the h time hto the intersection validity of dependent variable data matrix Y;
PRESS (h) gives up certain sample point x in argument data matrix from all n sample point (i)(i=1,2 ..., n) afterwards, simulate the regression equation containing h major component with a remaining n-1 sample point, then to x (i)(i=1,2 ..., the Prediction sum squares of n) predicting; j=1,2 ..., q, q is dependent variable number, PRESS j(h) be dependent variable y jprediction sum squares: y ijrepresent the value of j dependent variable in i group data, represent y ijat the sample point x certainly becoming (i)on predicted value; I=1,2 ..., n, the group number that n is required monitor control index;
SS (h-1) is the error of fitting quadratic sum containing the regression equation of h-1 major component simulating with all n sample point: SS ( h - 1 ) = &Sigma; j = 1 q S S j ( h - 1 ) ; SS j(h-1) represent y jerror of fitting quadratic sum: SS ( h - 1 ) = &Sigma; j = 1 q S S j ( h - 1 ) , SS j(h-1) represent y jerror of fitting quadratic sum; y ijrepresent the value of j dependent variable in i group data, represent y jat independent variable sample point x (i)on match value;
Step 4-4-5, return execution step 4-4-1 to step 4-4-4, until major component t hintersection validity value to dependent variable Y is less than setting value, completes argument data matrix and dependent variable data matrix are carried out to Principle component extraction.
In the embodiment of the present invention, Principle component extraction number of times is 3, finally extracts 3 major component: t 1, t 2, t 3.
Step 4-5, determine the variable drop importance index of each independent variable;
VI P j = p Rd ( Y ; t 1 , . . . , t s ) &Sigma; h = 1 s Rd ( Y ; t h ) w hj 2 - - - ( 1 )
Wherein, VIP jbe the variable drop importance index of j independent variable, j=1,2 ..., p, p is independent variable number;
S is the total degree of Principle component extraction;
T hfor the major component that the h time is extracted from argument data matrix;
W hjfor matrix the corresponding unit character vector w of eigenvalue of maximum institute hj component, E h-1represent the argument data matrix obtaining in the h-1 time Principle component extraction, F h-1represent the dependent variable data matrix obtaining in the h-1 time Principle component extraction; F is the matrix after dependent variable data matrix Y standardization;
rd (y j; t h) be t hto the interpretability of dependent variable data matrix Y;
Rd (y j; t h)=r 2(y j; t h), r (y j; t h) be t hwith dependent variable y jsimple correlation coefficient;
for t 1..., t sto the accumulation interpretability of dependent variable data matrix Y;
In the embodiment of the present invention, VIP value respective value is as shown in table 3;
Table 3
VIP1 VIP2 VIP3 VIP4 VIP5 VIP6 VIP7 VIP8 VIP9 VIP10
1.1429 1.6407 0.3584 0.7038 0.7366 0.8411 0.4653 1.4239 0.7269 0.2308
VIP11 VIP12 VIP13 VIP14 VIP15 VIP16 VIP17 VIP18 VIP19 VIP20
0.9706 1.4896 0.8272 1.5537 1.1864 0.3337 1.1429 1.0532 0.2217 1.0162
VIP21 VIP22 VIP23 VIP24 VIP25 VIP26 VIP27 VIP28 VIP29 VIP30
1.2922 1.0402 0.7837 0.8845 1.9651 0.5056 1.1204 0.4164 0.5034 0.7217
Step 4-6, judge whether the variable drop importance index of each independent variable is more than or equal to variable drop importance index threshold value VIP *=0.8, if so, retain this independent variable, retaining correspondence affects production target, execution step 4-8 after all independents variable have all judged, otherwise, delete this independent variable, deleting correspondence affects production target, execution step 4-7 after all independents variable have all judged;
Step 4-7, remaining independent variable is reconstituted to argument data battle array, be about to many group lingering effect production targets and reconstitute argument data battle array, and return to execution step 4-4 to step 4-6, be all more than or equal to variable drop importance index threshold value until remain the variable drop importance index of each independent variable;
In the embodiment of the present invention, residue independent variable is: x 1, x 2, x 6, x 8, x 11, x 12, x 13, x 14, x 15, x 17, x 18, x 20, x 21, x 22, x 24, x 25, x 27, and residue independent variable and y are reconstituted to argument data battle array and dependent variable data matrix;
Step 4-8, by final remaining independent variable, the final remaining production target that affects is as crucial production target;
In the embodiment of the present invention, remaining independent variable is the influence factor that finishing screen is selected, and finishing screen is selected 8 major influence factors: x 1, x 2, x 8, x 12, x 17, x 18, x 25, x 27, the key index filtering out is that pre-ore dressing grade, the primary overflow recovery, weak magnetic enter to grind grade, three granularities of weak magnetic, weak fine work position, strong magnetic and enter to grind grade, strong magnetic feed particle size and eight of positions of strong fine work ore dressing data acquisition index;
Step 5, according to obtain crucial production target, inquiry case library, determine when monitor control index is abnormal the numerical value that crucial production target is corresponding, and the analysis result of storing in case library in this situation and adjustment scheme;
In the embodiment of the present invention, according to filter out 8 crucial production targets, set up respectively the case library while adjusting in Indexes Abnormality situation for the comprehensive concentrate grade of ore dressing, the attribute in case library is 8 crucial production targets, and its value is 8 crucial production target values; Case solution feature in case library comprises analysis result and adjustment scheme; Case library structure is as shown in table 4:
Table 4
Step 6, field personnel are according to the monitoring image of monitor control index, the numerical value of the crucial production target of on line real-time monitoring, when monitor control index occurs when abnormal, adopt based on reasoning by cases method, according to the historical case of recording in case library, determine the affiliated case of current Indexes Abnormality situation, concrete grammar is as follows:
Step 6-1, obtain the concrete numerical value of current crucial production target;
In the embodiment of the present invention, there are 8 crucial production targets of major effect relation according to the comprehensive concentrate grade of ore dressing with needs monitoring filtering out, set up the monitoring image of the comprehensive concentrate grade of ore dressing, in this picture, show simultaneously the comprehensive concentrate grade desired value of ore dressing and filter out have 8 crucial production target values of major effect relation with the comprehensive concentrate grade of ore dressing; Fig. 4~Figure 12 is the monitoring image of the comprehensive concentrate grade of ore dressing of historical time section, and Figure 13~Figure 21 is the LINE REAL TIME MONITORING picture of the comprehensive concentrate grade of ore dressing;
In the time that field staff finds that by the comprehensive concentrate grade monitoring image of ore dressing abnormal conditions appear in the comprehensive concentrate grade in ore deposit, as shown in figure 22, the comprehensive concentrate grade actual value of ore dressing on July 29 during lower than planned value, adopts and based on reasoning by cases method, the comprehensive concentrate grade abnormal conditions of ore dressing is analyzed;
Step 6-2, the abnormal crucial production target of definite generation;
Determine that formula is as follows:
| target setting value-index actual value |/index actual value) < α, wherein: α is the indicator difference rate threshold value that occurs abnormal conditions, α=0.86;
Step 6-3, adopt nearest neighbor method, determine the similarity of each case in current monitor control index abnormal conditions and case library:
SIM ( G in , G k ) = &Sigma; i = 1 m &omega; i &times; sim ( v i , v i , k ) &Sigma; i = 1 m &omega; i - - - ( 2 )
Wherein, G infor current Indexes Abnormality situation;
G kfor k case in case library;
M is the number of crucial production target;
ω ibe the weights of i crucial production target with respect to monitor control index impact;
In the embodiment of the present invention, 8 crucial production targets with respect to the weights of monitor control index impact are: ω 1=0.6, ω 2=1.2, ω 3=0.6, ω 4=0.8, ω 5=1.8, ω 6=1.5, ω 7=1.5, ω 8=2;
SIM (G in, G k) be the similarity of current Indexes Abnormality situation and k case;
V i, kbe the value of i feature in k case, i.e. i crucial production target;
Sim (v i, v i, k) be i crucial production target in current Indexes Abnormality situation, the similarity with i in k case crucial production target, is calculated as follows:
sim ( v i , v i , k ) = 1 - | v i - v i , k | max ( v i , v i , k ) - - - ( 3 )
Step 7, select current monitor control index abnormal conditions and the highest case of historical case similarity, as case under current Indexes Abnormality situation, according to analysis result and adjustment scheme that in case library, this case is recorded, instructing measure to be issued to corresponding production process concrete production adjustment carries out scene adjustment, ensures that monitor control index meets the demands.
In the embodiment of the present invention, the SIM (G of the case similarity value maximum that retrieval obtains in, G k)=0.9, therefore the method for adjustment while selecting this case to combine fine work position abnormal conditions as this ore dressing, concrete adjustment measure is the adjustment scheme in this case solution feature, the operating mode that obtains former because: strong magnetic enters to grind the on the low side and mineral washability of grade larger variation occurs; Obtaining producing adjusting instructs measure to be: strengthen supervision, the monitoring of reinforcement to material quality to fine ore, instruct measure to be issued to corresponding production process production adjustment and carry out scene adjustment, ensure that the comprehensive concentrate grade index of ore dressing meets the demands.

Claims (4)

1. a mineral processing production indicators monitor system, is characterized in that, comprises data capture unit, Factors Affecting Parameters screening unit, index monitor unit, case library maintenance unit, Indexes Abnormality analytic unit and data storage cell, wherein,
Data capture unit: for obtaining mineral processing production whole process production target historical data, comprise quality index, measuring index, equipment operating statistic index, ore storage bin material level index, technic index, target energy and the indicator of costs, and the data of obtaining are sent in Factors Affecting Parameters screening unit and data storage cell;
Factors Affecting Parameters screening unit: for determining according to the actual requirements monitor control index and having multiple production targets that affect of the relation of impact with monitor control index, and the mineral processing production index screening method of employing based on PLS-VIP, further screen from multiple the impact production target of choosing, determine crucial production target, and crucial production target is sent to case library maintenance unit;
Case library maintenance unit: safeguard for the case library that analysis is used to Indexes Abnormality, comprise interpolation case, revise case, delete case and check case;
Index monitor unit: for according to the monitoring image of monitor control index, the numerical value of the crucial production target of on line real-time monitoring, and be sent to Indexes Abnormality analytic unit;
Indexes Abnormality analytic unit: for occurring when monitor control index when abnormal, adopt based on reasoning by cases method, according to the historical case of recording in case library, determine case under current Indexes Abnormality situation;
Data storage cell: for store the mineral processing production whole process production target historical data of obtaining, the monitor control index of choosing, choose multiplely affect the production target historical datas that production target, the in real time crucial production target actual numerical value that detects and case library are recorded with monitor control index has a relation of impact.
2. the method for supervising that adopts mineral processing production indicators monitor system claimed in claim 1 to carry out, is characterized in that, comprises the following steps:
Step 1, obtain mineral processing production whole process production target historical data, comprise quality index, measuring index, equipment operating statistic index, ore storage bin material level index, technic index, target energy and the indicator of costs;
Step 2, is according to the actual requirements chosen required monitor control index from mineral processing production whole process production target;
Step 3, is according to the actual requirements chosen the multiple production targets that affect that have the relation of impact with monitor control index from mineral processing production whole process production target;
Step 4, the mineral processing production index screening method of employing based on PLS-VIP, further screen from multiple the impact production target of choosing, and determines crucial production target, and concrete grammar is as follows:
Step 4-1, according to historical data, determine that many groups affect the concrete numerical value of production target, using this production target as independent variable, and build argument data matrix;
The line number of argument data matrix is the group number that affects production target, and matrix column number is the number that affects production target, and entry of a matrix element is for affecting the concrete numerical value of production target;
Step 4-2, according to historical data, determine the concrete numerical value of many group monitor control indexs, be dependent variable by monitor control index, and build dependent variable data matrix;
The line number of dependent variable data matrix is the group number of monitor control index, the number that matrix column number is monitor control index, and entry of a matrix element is the concrete numerical value of monitor control index;
Step 4-3, argument data matrix and dependent variable data matrix are carried out to standardization;
Step 4-4, employing NIPALS method are carried out Principle component extraction to argument data matrix and dependent variable data matrix;
Step 4-5, determine the variable drop importance index of each independent variable;
VI P j = p Rd ( Y ; t 1 , . . . , t s ) &Sigma; h = 1 s Rd ( Y ; t h ) w hj 2 - - - ( 1 )
Wherein, VIP jbe the variable drop importance index of j independent variable, j=1,2 ..., p, p is independent variable number;
S is the total degree of Principle component extraction;
T hfor the major component that the h time is extracted from argument data matrix;
W hjfor matrix the corresponding unit character vector w of eigenvalue of maximum institute hj component, E h-1represent the argument data matrix obtaining in the h-1 time Principle component extraction, F h-1represent the dependent variable data matrix obtaining in the h-1 time Principle component extraction; F 0for the matrix after dependent variable data matrix Y standardization;
rd (y j; t h) be t hto the interpretability of dependent variable data matrix Y;
Rd (y j; t h)=r 2(y j; t h), r (y j; t h) be t hwith dependent variable y jsimple correlation coefficient;
for t 1..., t sto the accumulation interpretability of dependent variable data matrix Y;
Step 4-6, judge whether the variable drop importance index of each independent variable is more than or equal to variable drop importance index threshold value, if, retain this independent variable, retain the corresponding production target that affects, execution step 4-8 after all independents variable have all judged, otherwise, this independent variable deleted, deleting correspondence affects production target, execution step 4-7 after all independents variable have all judged;
Step 4-7, remaining independent variable is reconstituted to argument data battle array, be about to many group lingering effect production targets and reconstitute argument data battle array, and return to execution step 4-4 to step 4-6, be all more than or equal to variable drop importance index threshold value until remain the variable drop importance index of each independent variable;
Step 4-8, by final remaining independent variable, the final remaining production target that affects is as crucial production target;
Step 5, according to obtain crucial production target, inquiry case library, determine when monitor control index is abnormal the numerical value that crucial production target is corresponding, and the analysis result of storing in case library in this situation and adjustment scheme;
Step 6, field personnel are according to the monitoring image of monitor control index, the numerical value of the crucial production target of on line real-time monitoring, when monitor control index occurs when abnormal, adopt based on reasoning by cases method, according to the historical case of recording in case library, determine the affiliated case of current Indexes Abnormality situation, concrete grammar is as follows:
Step 6-1, obtain the concrete numerical value of current crucial production target;
Step 6-2, the abnormal crucial production target of definite generation;
Make it by the actual numerical value of each crucial production target and its setting value poor, the actual numerical value of poor absolute value and crucial production target is divided by and is obtained indicator difference rate, indicator difference rate is arranged to threshold with it, set indicator difference rate threshold value if be greater than, this key production target occurs abnormal;
Step 6-3, adopt nearest neighbor method, determine the similarity of each case in current monitor control index abnormal conditions and case library:
SIM ( G in , G k ) = &Sigma; i = 1 m &omega; i &times; sim ( v i , v i , k ) &Sigma; i = 1 m &omega; i - - - ( 2 )
Wherein, G infor current Indexes Abnormality situation;
G kfor k case in case library;
M is the number of crucial production target;
ω ibe the weights of i crucial production target with respect to monitor control index impact;
SIM (G in, G k) be the similarity of current Indexes Abnormality situation and k case;
V i, kbe the value of i feature in k case, i.e. i crucial production target;
Sim (v i, v i, k) be i crucial production target in current Indexes Abnormality situation, the similarity with i in k case crucial production target, is calculated as follows:
sim ( v i , v i , k ) = 1 - | v i - v i , k | max ( v i , v i , k ) - - - ( 3 )
Step 7, select current monitor control index abnormal conditions and the highest case of historical case similarity, as case under current Indexes Abnormality situation, according to analysis result and adjustment scheme that in case library, this case is recorded, instructing measure to be issued to corresponding production process concrete production adjustment carries out scene adjustment, ensures that monitor control index meets the demands.
3. method for supervising according to claim 2, is characterized in that, the quality index described in step 1 comprises that ore dressing combines fine work position, smart moisture is combined in ore dressing, smart scaling loss Ig is combined in ore dressing, smart S is combined in ore dressing, smart CaO is combined in ore dressing, smart SiO2 is combined in ore dressing, measuring and calculating grade of sinter, measuring and calculating sintering deposit SiO2, combine smart granularity, combine accurate measurement and calculate compliance rate, combine smart SiO2 compliance rate, comprehensive lump ore rate, the primary overflow recovery, roasted ore grade, barren rock grade, weak magnetic magnetic fine work position, weak magnetic magnetic essence SiO2, weak magnetic magnetic essence CaO, weak magnetic magnetic essence Ig, weak magnetic enters to grind grade, weak fine work position, weak tail grade, weak magnetic magnetic tail grade, three magnetic fine work positions, weak fine work position qualification rate, flotation is to ore deposit grade, flotation is to ore deposit SiO2, the floating tail grade of weak magnetic, the floating smart SiO2 of weak magnetic, strong magnetic enters to grind grade, strong fine work position, the strong comprehensive tailings grade of magnetic, flat ring is combined fine work position, high gradient is combined fine work position, high gradient tailings grade, strong fine work position qualification rate, strong smart SiO2, strong magnetic feed particle size, the selected concentration of strong magnetic, 1-2 revolves excessive concentration, 2-2 revolves excessive concentration, weak three concentration of magnetic, magnetic tailing grade, magnetic tailing concentration, full choosing ratio, primary overflow yield qualification rate, barren rock bucket number, barren rock amount, barren rock productive rate, the theoretical metal recovery rate of low intensity magnetic separation, the theoretical ratio of concentration of low intensity magnetic separation, flotation choosing ratio, the theoretical metal recovery rate of high intensity magnetic separation, the grade of the theoretical ratio of concentration of high intensity magnetic separation and various raw ores, S content, SiO2 content and CaO content and moisture content,
Described measuring index comprises that smart output is combined in the ore dressing that comprises moisture content, smart output, inferior fine magnetite concentrate output, high intensity magnetic mineral output are combined in the ore dressing of removing moisture content, selects 3# ore deposit amount, dry weight and the weight in wet base of piece 1# ore deposit amount, powder 2# ore deposit amount, finished product-2# ore deposit amount, barren rock-1# ore deposit amount, 2X essence 4# ore deposit amount, 1-1# ball milling ore deposit amount, 2-1# ball milling ore deposit amount, 3-1# ball milling ore deposit amount, 4-1# ball milling ore deposit amount, strong magnetic mine-supplying quantity, weak magnetic mine-supplying quantity, go down the hill ore deposit amount and various raw ores;
When described equipment operating statistic index comprises raw ore stove when fortune, strong magnetic bowl mill fortune, when weak magnetic bowl mill fortune, bowl mill operating rate, strong magnetic bowl mill operating rate and weak magnetic bowl mill operating rate, when operating rate is bowl mill fortune and the ratio of T.T.;
Described ore storage bin material level index comprises accumulating storage ore storage bin material level, once sieves ore storage bin material level, regrading ore storage bin material level, strong magnetic ore storage bin material level and weak magnetic ore storage bin material level;
Described technic index comprises shaft furnace heating gas amount, reduction shaft furnace coal gas amount, intensity magnetic separator electric current, strong magnetic machine drift ice is washed electric current, the dynamo-electric stream of vertical ring, and vertical ring machine drift ice is washed electric current, floating agent, concentration, frequency, the flow of concentrated large well, the pressurization storehouse pressure of pressing filter;
Described target energy comprises the unit consumption of electricity, Zhong Shui, Xin Shui, coke-oven gas, blast furnace gas, steam, Living Water and always consumes;
The described indicator of costs comprises raw material unit cost, raw material total cost, energy unit cost and energy total cost.
4. method for supervising according to claim 2, is characterized in that, the employing NIPALS method described in step 4-4 is carried out Principle component extraction to argument data matrix and dependent variable data matrix, and concrete grammar is as follows:
Step 4-4-1, acquisition matrix the corresponding unit character of eigenvalue of maximum vector w h, and then h major component t of acquisition independent variable h:
t h=E h-1w h (4)
Wherein, E h-1represent the argument data matrix obtaining in the h-1 time Principle component extraction, F h-1represent the dependent variable data matrix obtaining in the h-1 time Principle component extraction; be the transposed matrix of the dependent variable data matrix that obtains in the h-1 time Principle component extraction, it is the argument data transpose of a matrix matrix obtaining in the h-1 time Principle component extraction; In the time of h=1, E 0represent the argument data matrix building, F 0represent the dependent variable data matrix building;
Step 4-4-2, acquisition matrix the corresponding unit character of eigenvalue of maximum vector c h, and then h major component u of acquisition dependent variable h:
u h=F h-1c h (5)
Step 4-4-3, according to the h major component of independent variable and the h major component of dependent variable, obtain the residual matrix of dependent variable data matrix and dependent variable data matrix:
E h = E h - 1 - t h p h T F h = F h - 1 u h r h T - - - ( 6 )
Wherein, p hrepresent t hload on argument data matrix,
R hrepresent u hload on dependent variable data matrix,
Step 4-4-4, judgement major component t now hwhether the intersection validity value to dependent variable data matrix Y is less than setting value, and if so, complete argument data matrix and dependent variable data matrix are carried out to Principle component extraction, otherwise, execution step 4-4-5;
The validity value computing formula of intersecting is as follows:
Q h 2 = 1 - PRESS ( h ) SS ( h - 1 ) - - - ( 7 )
Wherein, represent the major component t extracting for the h time hto the intersection validity of dependent variable data matrix Y;
j=1,2 ..., q, q is dependent variable number, PRESS j(h) be dependent variable y jprediction sum squares: y ijrepresent the value of j dependent variable in i group data, represent y ijat independent variable sample point x (i)on predicted value; I=1,2 ..., n, the group number that n is required monitor control index;
sS j(h-1) represent y jerror of fitting quadratic sum; y ijrepresent the value of j dependent variable in i group data, represent y jat independent variable sample point x (i)on match value;
Step 4-4-5, return execution step 4-4-1 to step 4-4-4, until major component t hintersection validity value to dependent variable Y is less than setting value, completes argument data matrix and dependent variable data matrix are carried out to Principle component extraction.
CN201410370724.5A 2014-07-30 2014-07-30 System and method for monitoring ore-dressing production indexes Active CN104134120B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410370724.5A CN104134120B (en) 2014-07-30 2014-07-30 System and method for monitoring ore-dressing production indexes

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410370724.5A CN104134120B (en) 2014-07-30 2014-07-30 System and method for monitoring ore-dressing production indexes

Publications (2)

Publication Number Publication Date
CN104134120A true CN104134120A (en) 2014-11-05
CN104134120B CN104134120B (en) 2017-05-24

Family

ID=51806793

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410370724.5A Active CN104134120B (en) 2014-07-30 2014-07-30 System and method for monitoring ore-dressing production indexes

Country Status (1)

Country Link
CN (1) CN104134120B (en)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104570739A (en) * 2015-01-07 2015-04-29 东北大学 Ore dressing multi-production-index optimized decision making system and method based on cloud and mobile terminal
CN106200602A (en) * 2016-09-07 2016-12-07 东北大学 A kind of based on Internet of Things and the industry preparation equipment moving monitoring system of cloud and method
CN106408192A (en) * 2016-09-23 2017-02-15 东北大学 Mineral processing equipment operating state monitoring system and method
CN108108880A (en) * 2017-12-07 2018-06-01 东北大学 A kind of mineral processing production index Visualized Analysis System and method
CN108287534A (en) * 2018-03-16 2018-07-17 马鞍山钢铁股份有限公司 A kind of sintering process gas permeability partitive optimization method
CN108446845A (en) * 2018-03-16 2018-08-24 马鞍山钢铁股份有限公司 A kind of sintering process gas permeability sectional evaluation method
CN108469797A (en) * 2018-04-28 2018-08-31 东北大学 A kind of grinding process modeling method based on neural network and evolutionary computation
CN109961205A (en) * 2017-12-26 2019-07-02 北大方正集团有限公司 Engineering design data processing method and processing device
CN110090727A (en) * 2019-04-26 2019-08-06 北矿智云科技(北京)有限公司 Processing method, device, equipment and the medium of operation data in a kind of time in grinding production
CN110180683A (en) * 2019-06-21 2019-08-30 安徽国兰智能科技有限公司 A kind of floatation system based on big data analysis
WO2020172913A1 (en) * 2019-02-28 2020-09-03 东北大学 Production index visual monitoring system and method based on configuration

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1680945A (en) * 2004-04-09 2005-10-12 东北大学 Method for proportioning ores for beneficiation
US20100017010A1 (en) * 2008-07-17 2010-01-21 International Business Machines Corporation Monitoring a process sector in a production facility
CN103257644A (en) * 2013-05-21 2013-08-21 青岛理工大学 Method for online monitoring of tailings pond safe state
CN103745406A (en) * 2013-12-23 2014-04-23 东北大学 Visual ore dressing production full-flow technic index optimized decision system and method thereof

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1680945A (en) * 2004-04-09 2005-10-12 东北大学 Method for proportioning ores for beneficiation
US20100017010A1 (en) * 2008-07-17 2010-01-21 International Business Machines Corporation Monitoring a process sector in a production facility
CN103257644A (en) * 2013-05-21 2013-08-21 青岛理工大学 Method for online monitoring of tailings pond safe state
CN103745406A (en) * 2013-12-23 2014-04-23 东北大学 Visual ore dressing production full-flow technic index optimized decision system and method thereof

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
刘长鑫: "面向Web的选矿生产全流程工艺指标决策系统软件平台的设计与开发", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
沈翠良: "选矿过程监控与自动控制的应用", 《包钢科技》 *
黄肖玲等: "选矿制造执行系统的研究与应用", 《计算机集成制造系统》 *

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104570739B (en) * 2015-01-07 2017-01-25 东北大学 Ore dressing multi-production-index optimized decision making system and method based on cloud and mobile terminal
CN104570739A (en) * 2015-01-07 2015-04-29 东北大学 Ore dressing multi-production-index optimized decision making system and method based on cloud and mobile terminal
CN106200602A (en) * 2016-09-07 2016-12-07 东北大学 A kind of based on Internet of Things and the industry preparation equipment moving monitoring system of cloud and method
CN106200602B (en) * 2016-09-07 2018-08-28 东北大学 A kind of preparation equipment moving monitoring system and method based on Internet of Things and industrial cloud
CN106408192B (en) * 2016-09-23 2019-07-23 东北大学 A kind of preparation equipment running state monitoring system and method
CN106408192A (en) * 2016-09-23 2017-02-15 东北大学 Mineral processing equipment operating state monitoring system and method
CN108108880A (en) * 2017-12-07 2018-06-01 东北大学 A kind of mineral processing production index Visualized Analysis System and method
CN109961205A (en) * 2017-12-26 2019-07-02 北大方正集团有限公司 Engineering design data processing method and processing device
CN108287534A (en) * 2018-03-16 2018-07-17 马鞍山钢铁股份有限公司 A kind of sintering process gas permeability partitive optimization method
CN108446845A (en) * 2018-03-16 2018-08-24 马鞍山钢铁股份有限公司 A kind of sintering process gas permeability sectional evaluation method
CN108446845B (en) * 2018-03-16 2021-07-27 马鞍山钢铁股份有限公司 Method for evaluating breathability in segmentation mode in sintering process
CN108469797A (en) * 2018-04-28 2018-08-31 东北大学 A kind of grinding process modeling method based on neural network and evolutionary computation
WO2020172913A1 (en) * 2019-02-28 2020-09-03 东北大学 Production index visual monitoring system and method based on configuration
CN110090727A (en) * 2019-04-26 2019-08-06 北矿智云科技(北京)有限公司 Processing method, device, equipment and the medium of operation data in a kind of time in grinding production
CN110090727B (en) * 2019-04-26 2020-12-22 北矿智云科技(北京)有限公司 Method, device, equipment and medium for processing operation data in ore grinding production
CN110180683A (en) * 2019-06-21 2019-08-30 安徽国兰智能科技有限公司 A kind of floatation system based on big data analysis
CN110180683B (en) * 2019-06-21 2021-01-26 安徽国兰智能科技有限公司 Flotation system based on big data analysis

Also Published As

Publication number Publication date
CN104134120B (en) 2017-05-24

Similar Documents

Publication Publication Date Title
CN104134120A (en) System and method for monitoring ore-dressing production indexes
Sbárbaro et al. Advanced control and supervision of mineral processing plants
CN1680945B (en) Method for proportioning ores for beneficiation
CN105413842B (en) The ore-dressing technique of Ultra-low-grade magnetite and system
CN103412489B (en) A kind of grinding particle size online forecasting system and method
CN101183260A (en) Mineral concentration full flow process automatic control method
CN103439889B (en) Based on the burning through point forecasting method of discharge flue gas analysis
CN103617456B (en) A kind of ore dressing process operating index optimization method
CN103617470A (en) Optimization method for mineral separation comprehensive production index under equipment capability changing condition
CN104299045B (en) System and method for forecasting yield of concentrate in whole mineral dressing process
CN104785358B (en) Iron ore of low ore grade beneficiation method and device
CN103056020A (en) Maximum optional size grading process for beneficiation and grinding
CN108108880A (en) A kind of mineral processing production index Visualized Analysis System and method
CN104318313B (en) Intelligent decision system and method for ore dressing total flow operation index on the basis of case-based reasoning
Lishchuk et al. Classification of geometallurgical programs based on approach and purpose
CN108469797A (en) A kind of grinding process modeling method based on neural network and evolutionary computation
CN108380380A (en) Product dry type fine crushing sorts Iron concentrate technique
Ding et al. Data mining based feedback regulation in operation of hematite ore mineral processing plant
Jämsä-Jounela Future automation systems in context of process systems and minerals engineering
CN103377247B (en) The intelligent extract method of grind mineral control case
CN104607311A (en) Ore dressing method of iron ore
Lynch, AJ & Morrison Simulation in mineral processing history, present status and possibilities
CN107583755A (en) A kind of predictor method on semi-autogenous mill power
Kuvaiev et al. Promising approach to the identification and classification of steady-state operation of an iron-ore dressing section
Lian et al. Research on slag grinding process control

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant