CN109885012B - Real-time optimization compensation method for gold hydrometallurgy full flow - Google Patents

Real-time optimization compensation method for gold hydrometallurgy full flow Download PDF

Info

Publication number
CN109885012B
CN109885012B CN201910236152.4A CN201910236152A CN109885012B CN 109885012 B CN109885012 B CN 109885012B CN 201910236152 A CN201910236152 A CN 201910236152A CN 109885012 B CN109885012 B CN 109885012B
Authority
CN
China
Prior art keywords
optimization
data
optimal
compensation
leaching
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
Application number
CN201910236152.4A
Other languages
Chinese (zh)
Other versions
CN109885012A (en
Inventor
常玉清
刘亚东
牛大鹏
王姝
王福利
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 CN201910236152.4A priority Critical patent/CN109885012B/en
Publication of CN109885012A publication Critical patent/CN109885012A/en
Application granted granted Critical
Publication of CN109885012B publication Critical patent/CN109885012B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02PCLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/02Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

The invention relates to a real-time optimization compensation method for the whole flow of gold hydrometallurgy; the method comprises the following steps: s1, performing online analysis on the real-time optimization result of the whole gold hydrometallurgy flow by using a process running state evaluation method to obtain an evaluation result; s2, selecting a matched compensation method for the evaluation result to process; s21, aiming at the condition that the evaluation result is suboptimal, a compensation method of self-optimization control is adopted; s22, aiming at the condition that the evaluation result is non-optimal, adopting a data-based operation quantity optimization setting compensation method; s23, aiming at the condition that the data similar to the current working condition can not be found in the historical database, the previous working condition data is subjected to a gold hydrometallurgy full-process re-optimization method to obtain the optimal operation; by establishing and solving the compensation model, the invention avoids the problems that the uncertain disturbance or uncertain variable exists in the production process, the mechanism model cannot be established, and the optimal operation cannot be obtained, and has important significance for improving the production efficiency and the economic benefit of enterprises.

Description

Real-time optimization compensation method for gold hydrometallurgy full flow
Technical Field
The invention belongs to the field of real-time optimization of gold hydrometallurgy processes, and particularly relates to a real-time optimization compensation method for a whole gold hydrometallurgy process.
Background
Hydrometallurgical processes have begun to receive high levels of attention from countries around the world as the high grade ore has been depleted. Compared with the traditional pyrometallurgy, the hydrometallurgy technology has the advantages of high efficiency, cleanness, suitability for recycling low-grade complex metal mineral resources and the like. Especially aiming at the characteristics of rich lean ore, complex symbiosis and high impurity content of mineral resources in China, the industrialization of the hydrometallurgy process has great significance for improving the comprehensive utilization rate of the mineral resources, reducing the yield of solid wastes and reducing the environmental pollution. In recent years, the research on the hydrometallurgical process is rapidly advanced. However, the hydrometallurgy reaction mechanism is complex, the process conditions are severe, such as high temperature, high pressure, strong corrosion and the like, the process flow is long, and the types of equipment are various, so that the safe, stable and continuous operation of production can be ensured only by continuously improving the large-scale industrialized intelligent control level of hydrometallurgy enterprises, and the quality and the yield of products are further ensured. The research of the hydrometallurgical process technology in China is in the international advanced level, some aspects of the technology are even in the leading position, but the automation level is relatively low due to the reasons of more hydrometallurgical process types, large process condition difference, relatively small scale and the like. And the independent optimization control of each process is far from meeting the requirement of industrial production. In order to improve the technical and economic indexes such as yield, metal recovery rate and comprehensive utilization rate of mineral resources to the maximum extent, reduce the operation cost and the yield of solid wastes, reduce environmental pollution, achieve the aims of high yield, high quality, energy conservation, consumption reduction and the like, the real-time optimization of the whole wet metallurgy process is urgently needed to be researched, and finally the aim of improving the economic benefit of enterprises is achieved.
The whole gold hydrometallurgy process comprises the hydrometallurgy process flows of ore grinding, flotation, dehydration and size mixing, cyaniding and leaching, filter pressing and washing, replacement and the like. Firstly, raw ore is subjected to pretreatment processes such as ore grinding, separation flotation and the like to obtain certain ore pulp, and then a medicament carried in the pretreatment process is separated from the ore through a dense filter pressing process to obtain a filter cake with a very small amount of liquid. And then stirring the filter cake and the size mixing water in the size mixing process to obtain ore pulp with a certain concentration, and pumping the mixed ore pulp into a leaching tank of a subsequent leaching process by an ore discharge pump. The leaching process comprises two leaching processes, wherein insoluble gold in the concentrate is reacted with a leaching agent (NaCN) to generate water-soluble ions, ore pulp after the two leaching processes is introduced into a filter press to be filter-pressed and washed to generate pregnant solution, and finally, metal gold is obtained through a replacement process. The main process of gold hydrometallurgy is shown in figure 1.
The reasonable hydrometallurgical process flow is a basic premise for ensuring effective recycling of gold in the ore and high return of income for enterprises. At present, the modeling, optimization and basic control of the whole wet metallurgy process are mostly concentrated at the level of each subprocess (such as leaching, filter pressing washing/dense washing, extraction/replacement) at home and abroad, and no application and research related to the real-time optimization of the whole wet metallurgy process exists, and many researches are not high in precision or lack of consideration on each subprocess and the physical characteristics of each subprocess in the whole process, so that the models cannot reflect the whole wet metallurgy process and the practical application capability of the models is limited. The whole hydrometallurgy flow generally has the characteristics of multiple flows, strong coupling, large hysteresis, nonlinearity and the like. Therefore, the established model can embody the complexity and lay a solid foundation for realizing the real-time optimization of the whole flow of the hydrometallurgy. In addition, in many practical industrial processes, there are errors or uncertainties associated with production conditions, measurement variations, raw material characteristics, and the like. At home and abroad, the research on model establishment mostly focuses on a simple mechanism model or a simple data model, and because uncertainty factors exist, models of some local links in the production process cannot be obtained, and at the moment, the full-process optimization control cannot be performed based on the process model. Therefore, reasonably establishing a mixed model with coexisting process qualitative models and quantitative models has important practical significance for improving the production efficiency and economic benefits of enterprises and facilitating production adjustment.
At present, the real-time optimization research on the whole flow of the hydrometallurgy at home and abroad is few, the automation level is not high, and the theoretical research only stays on the optimization level of each procedure. As the whole hydrometallurgy flow is a complex process consisting of a series of typical processes, the optimization of a single process can not meet the production requirements of the whole hydrometallurgy flow along with the continuous development of the industry. With such challenges, hydrometallurgical full flow optimization has received extensive attention and has become an important developmental goal of the mineral processing industry. However, the hydrometallurgical full flow process is oversized, the process and variables are excessive, and the process model may be subject to uncertain random perturbations. These characteristics of hydrometallurgical process production may cause the established process model to be mismatched with the actual production process, so that the optimization results obtained by the model-based optimization method are not the optimal solution of the actual process when applied to the actual process. The characteristics make the real-time optimization problem of the whole wet metallurgy process more complex, and the research on the real-time optimization method of the whole wet metallurgy process is urgently needed. Therefore, an appropriate full-process modeling method and a real-time optimization method must be found, and the invention provides a method suitable for the wet metallurgy full-process real-time optimization.
Disclosure of Invention
Technical problem to be solved
The invention provides a real-time optimization compensation method for a whole gold hydrometallurgy process, which aims to solve the technical problems that a model established by the existing gold hydrometallurgy method is not matched with an actual production process and an optimal solution cannot be obtained when the model is applied.
(II) technical scheme
In order to achieve the purpose, the invention adopts the main technical scheme that:
the online evaluation and operation compensation part comprises the following steps:
s1, performing online analysis on the real-time optimization result of the whole gold hydrometallurgy flow by using the process running state evaluation method to obtain an evaluation result;
s2, selecting a compensation method matching the evaluation result for the evaluation result, which specifically includes:
s21, aiming at the condition that the evaluation result is suboptimal, selecting a linear combination of the output variables of the whole process as a controlled variable by adopting a compensation method of self-optimization control, and realizing the self-optimization control of the whole process of the hydrometallurgy under the uncertainty disturbance by controlling the tracking set value of the controlled variable;
s22, aiming at the condition that the evaluation result is non-optimal, a data-based operation quantity optimization setting compensation method is adopted, and compensation operation is carried out on the current operation quantity based on a timely learning thought and a partial least square method, so that self-optimization of process operation is realized;
s23, aiming at the condition that the data similar to the current working condition can not be found in the historical database, adopting a gold hydrometallurgy full-process re-optimization method for the previous working condition data, and further obtaining the optimal operation under the current working condition.
Optionally, an offline modeling section is included;
the offline modeling part specifically comprises: the method comprises the steps of establishing a gold hydrometallurgy full-process optimization model, establishing a process running state evaluation model, establishing an optimal operation case base and establishing a self-optimization control model.
Optionally, the establishing of the gold hydrometallurgy full-process optimization model comprises:
the economic cost of the whole-flow production process of gold hydrometallurgy is minimized as an optimization target, and the gold is leached by two cyanidationsSodium cyanide addition in the process Qji,cnJ 1 and 2 are two leaching processes, i 1,2 and 3 are the number of leaching tanks in one leaching process and the addition quantity Q of zinc powder in a replacement processznForming a gold hydrometallurgy full-process operation vector u ═ Q as an operation variable11,cn,...,Q23,cn,Qzn]T
Finally establishing a gold hydrometallurgy full-process optimization model which satisfies the quantitative relation of each procedure in the gold hydrometallurgy full-process mechanism model by satisfying the practical requirement range constraints of the quality indexes and the operation quantities of each sub-process, wherein the gold hydrometallurgy full-process optimization model is expressed as follows:
wherein J is economic cost function of the whole process of gold hydrometallurgy, Pcn,Pzn,Pcnd,PAuFor the unit price of each material, EC is the total energy consumption for production, Q23,sShowing the ore flow in the last tank of the second leaching process, C13,cn、C23,cnRespectively showing the concentration of cyanide ions in the last tank in the primary leaching process and the secondary leaching process, C23,sRepresenting the grade value of the solid gold in the last groove of the second leaching process; f. ofpA mechanism model representing the p-th sub-process; x is the number oftp(p ═ 1,2,3) represents the leaching rate in the primary leaching process, the leaching rate in the secondary leaching process, and the substitution rate in the substitution process,is the minimum requirement constraint of the first leaching rate, the second leaching rate and the replacement rate; u. ofpThe operation vector of the p procedure is shown, and z is a working condition, including ore flow, solid gold grade and ore particle size; cwThe ore pulp concentration represents a random disturbance variable; c1i,lAu、C1(i-1),lAu、C2i,lAu、C2(i-1),lAuRespectively representing the liquid phase gold grade of the ith tank in the primary leaching process, the (i-1) th tank in the primary leaching process, the ith tank in the secondary leaching process and the (i-1) th tank in the secondary leaching process, wherein i is 1,2 and 3; c1i,cn、C1(i-1),cn、C2i,cn、C2(i-1),cnRespectively showing the concentrations of liquid phase cyanide ions in the ith tank in the primary leaching process, the (i-1) th tank in the primary leaching process, the ith tank in the secondary leaching process and the (i-1) th leaching tank in the secondary leaching process; qsIs the ore flow, V is the reactant volume;the leaching rate index required in the production process is shown; u. of3Denotes the amount of zinc powder added in the substitution process, wherein u3,min,u3,maxRespectively representing the minimum value and the maximum value of the addition amount of the zinc powder required in the actual production process; u. ofjiDenotes the amount of sodium cyanide added in the ith cell in the jth procedure, uji,min、uji,maxThe minimum and maximum required amounts of NaCN in each tank are indicated.
Optionally, the process operation state evaluation model building includes:
establishing a hydrometallurgy full-process evaluation model according to the full-process industrial characteristics and the data characteristics of each sub-process, and specifically implementing the following steps:
a1, preprocessing the quantitative data and the qualitative data acquired on line in the metal metallurgy process to obtain the quantitative data and the qualitative data to be analyzed, wherein the data preprocessing steps are as follows:
a11, selecting economic cost in the whole-flow production process as an evaluation index, and selecting process variables capable of influencing the process running state;
a12, sequentially expressing different state grades by a series of positive integers for qualitative variables; for the quantitative variable, carrying out simple smoothing treatment, selecting a sliding window with a certain length, and representing the information of the whole window by using the variable mean value in the window;
a2, establishing an operation state evaluation model for evaluating the operation state grade of the whole process according to the quantitative data and the qualitative data in the metal metallurgy process in the historical time period, and specifically implementing the following steps:
a21, determining the grade of the whole process; let the modeling data beH is the number of samples, J is the number of variables, including qualitative variables and quantitative variables; according to the economic cost of the full-process evaluation index, the process running state is divided into three grades of optimal, suboptimal and non-optimal, and is marked as X1,X2,X3Wherein, X comprises 6 subblock data which are respectively a size mixing process subblock, a primary cyanidation leaching process subblock, a primary filter-pressing washing process subblock, a secondary cyanidation leaching process subblock, a secondary filter-pressing washing process subblock and a replacement process subblock; the corner marks 1,2 and 3 respectively represent that the performance grade is excellent, suboptimal and non-excellent;
a22, establishing a full-process evaluation model; obtaining data X of each gradelAfter (1, 2 and 3), establishing a decision table for the full-process model by using a fuzzy probability rough set;
wherein the condition attribute is Xl(l 1,2,3), the decision attribute is the full-process operating state level l, and the domain contains Xl(l ═ 1,2, 3).
Optionally, the creating of the optimal operating case library includes:
historical optimization-based process operating data { xn}n=1,...,NN is the number of data with the optimal process running state grade reserved in the history, and an optimal operation case library is establishedWherein z isnAs a condition of operation, unUnder the working condition of znOptimum operating variable of time, NcThe number of data in the optimal case library is determined.
Optionally, the self-optimization control model establishing comprises:
the method comprises the following steps of taking economic cost and energy efficiency of the operation of the whole gold hydrometallurgy process as cost functions, selecting linear combination of whole process output variables as controlled variables, and establishing a self-optimization control model of the whole gold hydrometallurgy process, wherein the specific self-optimization control off-line modeling process comprises the following steps:
b1 distribution space C for uncertain disturbance variable ore pulp concentrationw∈[32%,40%]Generating N using Monte Carlo samplingsGroup sequence
B2 case d for each group of pulp concentration(n)Performing off-line optimization solution by adopting a full-flow optimization method, and calculating to obtain optimal input uopt=[Q11,cn,...,Q23,cn,Qzn]T
B3, recording the corresponding optimal output variable under each disturbance situation according to the previous solution resultWherein the output variable isComprises the optimum input variables u of the sodium cyanide addition amount of three leaching tanks in the first leaching process and the second leaching process and the zinc powder addition amount in the replacement processopt=[Q11,cn,Q12,cn,Q13,cn,Q21,cn,Q22,cn,Q23,cn,Qzn]Grade C of solid gold in two leaching processesji,sWith concentration of cyanide particles Cji,cnGrade C of cyanogen slag in replacement Processcnd(ii) a The measured noise is Gaussian noise, and the standard deviation is 5% of the nominal value; forming a matrix by the optimal output variables obtained under all disturbance conditionsWhereinIs a covariance matrix of the measured noise;
b4, taking a nominal working point CwCalculating gain matrix G of output variable value by finite difference method as 36%y,refAnd hessian matrix J of the loss function J with respect to the manipulated variable uuu,ref
B5, calculating the optimal combination matrix H in combination C ═ Hy so that the average loss is minimal, i.e.The combination matrix H can be obtained by solving the following equation:wherein the content of the first and second substances,representing a matrix of optimal output variables containing measurement noise, Gy,refA gain matrix representing the input and output at nominal operation, Juu,refA hessian matrix representing the economic cost function J of the gold hydrometallurgy full flow relative to the operational variable u; optimal output variable based on nominal working pointUsing formulasObtaining the set value C of the controlled variables
Optionally, the online evaluation and operation compensation part specifically includes: the method comprises the steps of online evaluation of process running states, online compensation of self-optimization control and establishment of an operation quantity optimization setting compensation model based on data, and specifically comprises the following steps:
solving and obtaining an optimal operation variable under a nominal working point through full-process optimization, and bringing the operation variable into an actual gold hydrometallurgy production process, wherein uncertainty disturbance exists in the actual process;
firstly, judging the current process running performance on line based on a process running state evaluation method;
secondly, respectively implementing compensation strategies corresponding to a self-optimization control method and a data-based operation amount optimization setting compensation method aiming at different performance levels;
thirdly, the compensated operation variables are brought into the actual production process, so that the optimal operation is obtained; the specific steps of online evaluation and operation compensation are as follows:
101. operating variables determined from nominal operating pointsBrought into the actual gold hydrometallurgical production process to obtain a process variable xtAnd an output variable yt
102. Evaluating the process running state on line; aiming at the current process variable and the output variable, processing the current process variable by adopting a full-process running state evaluation model established off-line, and determining the grade of the full-process running state in the metal metallurgy process
The online evaluation comprises the following specific steps:
c1, obtaining online data xtMatching the data in the decision table established off line with the fuzzy probability rough set to obtain xtLambda intercept of fuzzy equivalence class ofWherein Representing data xtλ is a given threshold;
c2, calculating the probability of belonging to the ith grade as:
wherein d islDenotes the l-th rank, Xl(1, 2,3) determining the operation state level according to historical data; judging the running state grade of the whole process as follows:
103. according to different evaluation results, a corresponding compensation strategy is adopted, and the specific compensation form is divided into the following steps:
case 1: when the evaluation result is excellent, the current operation variable is indicated to be the optimal operation, and compensation operation is not needed until new process disturbance occurs;
case 2: when the evaluation result is suboptimal, namely the process running state deviates from the optimal operation, but the deviation is not large; aiming at the situation, self-optimization control online compensation operation is adopted, the controlled variable is controlled to track a set value to realize that the influence of uncertainty and change disturbance on the optimization of the economic target is minimum, and then the current operation is compensated to realize the self-optimization control of the hydrometallurgy process under the uncertainty disturbance;
the specific compensation steps are as follows:
d1, connection matrix H obtained based on off-line calculation and output variable y under current working conditiontUsing the formula Ct=HytTo obtain the controlled variable Ct
D2, calculating the current operating variable based on the following formulaIs compensated for by a compensation value delta u1.k
Δy=Gy·Δu
H′=BH,B=(H·Gy)-1
C′=B·C=H′·y
B·ΔC=H′·Δy=(H·Gy)-1·H·Gy·Δu=I·Δu
Where Δ y denotes the difference in the output values at nominal operation, determined by means of finite differences, GyIs Gy,refThe compensation value Deltau, i.e. the gain matrix representing the input and output under nominal operationIs Δ u1.kΔ C is a controlled variable set value CsWith the currently sought controlled variable CtA deviation of (a); h' represents the general solution of the connection matrix H;
d3, adding new operation variablesCarrying into the actual production process, and executing the next stage;
case 3: when the evaluation result is non-optimal, namely the process running state deviates from the optimal operation, and the deviation is very large; aiming at the situation, a data-based operation quantity optimization setting compensation method is adopted, so that the production process returns to an optimal or suboptimal operation state; the method comprises the steps of searching several groups of data which are most similar to the current working condition in a historical optimal database based on a timely learning thought, establishing a data model between the working condition and an operation amount compensation value by adopting a partial least square method, compensating the current operation, and further realizing process operation self-optimization;
the specific compensation steps are as follows:
e1 for the current process variable xtSearching the current working condition z in a case library with excellent history established off-linetSimilar case data ziCalculating the sample similarity S based on the following formulaz
Sz=ρexp(-Di)+(1-ρ)cos(ωi)
Wherein Sz∈[0,1]The larger the sample is, the stronger the sample similarity is; diAnd cos (ω)i) Respectively represents ziAnd ztThe cosine values of the Euclidean distance and the included angle between the two cosine values; rho (rho is more than or equal to 0 and less than or equal to 1) is an index of information weight of the balance distance and the included angle;
e2, defining a sample similarity threshold (0 < 1), and judging the sample similarity SzWhether greater than a threshold;
in the first case: if S iszNot less than all, indicating the excellent case of historyThe database has data similar to the current working condition; if the number of the similar samples is larger than or equal to L, performing operation compensation based on a timely learning idea, and otherwise, performing second-situation operation;
the specific operation steps for the first case are as follows:
f1, calculating to obtain L similar samples serving as modeling data of the local regression model based on a sample similarity calculation formula;
f2, based on L similar samples { zi,ui}i=1,2,LCalculating the manipulated variable ui(i ═ 1, 2.. multidot.l) and the current manipulated variableThe difference between them, is recorded as Δ ui(i ═ 1,2,. ·, L); building a modeling dataset { zi,Δui}i=1,2,L
F3, establishing a local regression model delta u based on a partial least square methodi=fr(ziθ), where θ is an update parameter that varies from modeling data to modeling data; for the current operating condition ztBased on a local regression model frCalculating a compensation value Deltau for the current manipulated variable2,k
F4, adding new operation variableCarrying into the actual production process, and executing the next stage;
in the second situation: if S isz< or SzNumber of samples equal to or greater than<And L, indicating that a large amount of data similar to the current working condition cannot be found in the historical optimal case library, and thus, adopting the traditional real-time optimization method, namely identifying disturbance variables, re-optimizing the full-process optimization model and further obtaining the optimal operation.
Optionally, when the evaluation result is non-optimal, several groups of data closest to the current working condition are searched in the historical optimal database based on the timely learning idea, and the method is divided into two situations according to the number of the similar data:
the first situation is as follows: if the number of the similar data is larger than a certain threshold value, it is indicated that enough data similar to the current working condition exists in the historical preferred case library; a data model between the working condition and the compensation value of the operation amount is established by adopting a partial least square method, the current operation is compensated, and further, the process operation self-optimization is realized;
case two: if the number of the similar data is smaller than a certain threshold value, it is indicated that a large amount of data similar to the current working condition cannot be found in the historical preferred case library; and (3) identifying disturbance variables and re-optimizing the full-flow optimization model by adopting a traditional real-time optimization method so as to obtain optimal operation.
Optionally, the system further comprises a process detection system, wherein the process detection system comprises a PLC (programmable logic controller), concentration detection, pressure detection and flow detection;
the PLC controller adopts a Simens 400 series CPU 414-2 and is provided with a Profibus DP port connected with a distributed IO; an Ethernet communication module is equipped for the PLC, and is used for an upper computer to access PLC data; the PLC controller and the Ethernet communication module are arranged in a PLC cabinet in the central control room;
the upper computer selects an i7 association computer and adopts a WINDOWS XP operating system;
the gold hydrometallurgy full-flow real-time optimization system adopts C #2008 programming software on an i7 association computer, and the gold hydrometallurgy full-flow real-time optimization algorithm adopts Matlab 2016a programming software;
the signal transmission software for the PLC and the full-flow real-time optimization adopts C #2008 programming software;
the method comprises the steps that a detection instrument is installed on the site of the hydrometallurgy process, the detection instrument transmits collected signals to a PLC through Profibus-DP, the PLC transmits the collected signals to an upper computer through Ethernet at regular time, and the upper computer transmits received data to a gold hydrometallurgy full-flow real-time optimization system to perform full-flow real-time optimization compensation operation and provide operation guidance suggestions.
(III) advantageous effects
The invention has the beneficial effects that: firstly, the method establishes a full-process model by using reliable expert knowledge and a mechanism of a process technology, and provides guarantee for obtaining an accurate and reliable optimization model; secondly, the process running state is evaluated in real time by using variable information which can be measured quantitatively or estimated qualitatively on line, so that the production process is more efficient, and the economic benefit of an enterprise is ensured; finally, aiming at different process disturbance influences, the self-optimization control compensation and the data-based operation amount optimization setting compensation provide reasonable and reliable operation guidance suggestions for operators, the optimization performance of the control system is improved, and the total production benefit of the whole process is further improved.
Drawings
FIG. 1 is a flow chart of a main supply flow of gold hydrometallurgy according to an embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating a real-time optimization control of a gold hydrometallurgy process according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating variation of ore pulp concentration variation in a gold hydrometallurgical pulp conditioning process according to an embodiment of the present invention;
FIG. 4 is a schematic view of the overall gold hydrometallurgy process flow for online evaluation and compensation of operation according to an embodiment of the present invention;
figure 5a shows the optimum operating variable Q for different pulp concentration disturbances provided by an embodiment of the present invention11,cnA schematic diagram;
figure 5b shows the optimum operating variable Q for different pulp concentration disturbances provided by an embodiment of the present invention12,cnA schematic diagram;
figure 5c shows the optimum operating variable Q for different pulp concentration disturbances provided by an embodiment of the present invention13,cnA schematic diagram;
figure 5d shows the optimum operating variable Q for different pulp concentration disturbances provided by an embodiment of the present invention21,cnA schematic diagram;
figure 5e shows the optimum operating variable Q for different pulp concentration disturbances provided by an embodiment of the present invention22,cnA schematic diagram;
figure 5f shows the optimum operating variable Q for different pulp concentration disturbances provided by an embodiment of the present invention23,cnA schematic diagram;
figure 5g shows the optimum operating variable Q for different pulp concentration disturbances provided by an embodiment of the present inventionznA schematic diagram;
figure 6 is a schematic diagram of the simulation results of the full process optimization for different slurry concentrations according to an embodiment of the present invention;
fig. 7 is a schematic diagram of a process operation state evaluation result according to an embodiment of the present invention;
FIG. 8a is a schematic diagram of the evaluation result and economic loss based on the nominal operating point under 65 random disturbances provided by an embodiment of the present invention;
FIG. 8b is a schematic diagram of the evaluation result and economic loss based on the self-optimized control compensation method under 65 random disturbances according to an embodiment of the present invention;
FIG. 9 is a schematic diagram of the deviation of the controlled quantity based on the self-optimized control compensation method under 65 random disturbances according to an embodiment of the present invention;
fig. 10 is a diagram illustrating simulation results for three different operations of economic loss according to an embodiment of the present invention.
Detailed Description
For the purpose of better explaining the present invention and to facilitate understanding, the present invention will be described in detail by way of specific embodiments with reference to the accompanying drawings.
The whole process flow of hydrometallurgy is shown in figure 1, raw ore is subjected to pretreatment processes such as ore grinding, separation flotation and the like to obtain certain ore pulp, and then a medicament carried in the pretreatment process is separated from the ore through a dense filter pressing process to obtain a filter cake with a very small amount of liquid; then stirring the filter cake and size mixing water in a size mixing process to obtain ore pulp with a certain concentration, allowing the ore pulp subjected to size mixing to enter a cyaniding leaching process, washing the leached ore pulp by filter pressing, and allowing the washed filter cake to enter a replacement process after size mixing; and (4) purifying and deoxidizing the pregnant solution, and then performing zinc powder replacement to generate gold mud. The hydrometallurgy process detection system mainly comprises concentration detection, pressure detection and flow detection.
The embodiment provides a real-time optimization compensation method for a gold hydrometallurgy full flow, wherein a PLC controller adopts a Simens 400 series CPU 414-2 and is provided with a Profibus DP port connected with a distributed IO; an Ethernet communication module is equipped for the PLC and used for an upper computer to access PLC data, and the PLC controller and the Ethernet communication module are placed in a PLC cabinet in a central control room.
The pH value is detected on line by a BPHM-II acidimeter developed by Beijing mining and metallurgy research institute, and the change of the pH value of the solution is converted into the change of a mV signal. The glass electrode PH measurement system blows the end of a glass tube of a glass film sensitive to pH into a bubble shape, and the tube is filled with 3mol/l KCL buffer solution containing saturated AgCl, and the PH value is 7. The potential difference reflecting the PH value and existing on the two sides of the glass film is derived by an Ag/AgCl conduction system, and then the mA number is converted into the PH value by a mA collecting instrument to be displayed.
The pressure is detected on line by a DSIII pressure detector produced by SIEMENS company, the pressure of medium directly acts on a sensitive diaphragm, a Wheatstone bridge consisting of resistors distributed on the sensitive diaphragm realizes the conversion from the pressure to an electric signal by using piezoresistive effect, and a millivolt signal generated by a sensitive element is amplified into an industrial standard current signal by an electronic circuit.
The dissolved oxygen concentration was measured on-line by an inpro6870+ M400 type oxygen amount measuring sensor manufactured by mettleltoreq corporation. The oxygen measuring sensor consists of cathode, counter electrode with current and reference electrode without current, the electrode is immersed in electrolyte, the sensor is covered with diaphragm, the diaphragm separates the electrode and electrolyte from the measured liquid, only dissolved gas can permeate the diaphragm, so that the sensor is protected, the electrolyte can be prevented from escaping, and the pollution and poisoning caused by invasion of foreign matter can be prevented. The current signal is sent to the transducer, and the oxygen content is calculated by using the relationship curve between the oxygen content and the oxygen partial pressure and temperature stored in the transducer, and then converted into a standard signal to be output.
The upper computer selects an i7 associative computer and adopts a WINDOW XP operating system.
The full-flow process real-time optimization system adopts C #2008 programming software on an i7 association computer, and the full-flow process real-time optimization algorithm adopts Matlab 2016a programming software.
The signal transmission software of the PLC and the process real-time optimization system adopts C #2008 programming software. The method comprises the steps that a detection instrument is installed on the site of the hydrometallurgy process, the detection instrument transmits collected signals to a PLC through a Profibus-DP, the PLC transmits the collected signals to an upper computer at regular time through an Ethernet, and the upper computer transmits received data to a process real-time optimization system for real-time optimization and provides operation guidance suggestions.
The block diagram of the whole process real-time optimization of the gold hydrometallurgy high copper ore is taken as an example in the invention is shown in figure 2. The real-time optimization method for the whole process of gold hydrometallurgy (high copper ore) comprises the following concrete implementation steps:
step 1: and (4) obtaining the actual ore pulp concentration value in the pulp mixing process within a period of time by combining expert knowledge and the experience of field operators.
Step 2: based on the nominal variables in the actual gold hydrometallurgy full-flow production process, as shown in table 1, 500 groups of samples are randomly generated for the ore pulp concentration variables by a Monte Carlo sampling method, and the optimal operation variables under different ore pulp concentrations are obtained by performing optimal solution by using a full-flow optimization model.
Specific examples of the optimization results are shown in fig. 5a to 5 g; from fig. 5a to 5g, it can be seen that the optimal manipulated variable Q is obtained by the full-process optimization method under different pulp concentrations11,cn、Q12,cn、Q13,cn、Q21,cn、Q22,cn、Q23,cnAnd QznThe method changes, so the ore pulp concentration is used as random disturbance to research the effectiveness of the real-time optimization method of the gold hydrometallurgy full flow.
TABLE 1 nominal values of process variables of gold hydrometallurgy
And step 3: to further illustrate the problem that the change of the concentration of the ore pulp can change the optimization result of the whole process, the invention specifically lists three different typesAnd researching an optimization result according to the ore pulp concentration change under the working condition. The procedure of the simulation experiment is as follows. Respectively selecting ore pulp concentration as Cw0.32, nominal value Cw0.36 and CwRespectively solving by using a full-process optimization method to obtain corresponding optimal operating variables u-Q11,cn,...,Q23,cn,Qzn]T. Then, the pulp concentration C is respectively measuredw0.36 and CwThe optimization result of 0.40 is brought into a simulation model, wherein the concentration of ore pulp in the simulation model is CwThe optimization results are shown in fig. 6 and table 2, 0.40.
TABLE 2 Total process optimization results at different pulp concentrations
Variables of Optimization result Cw=0.32(g/g) Optimization result Cw=0.36(g/g) Optimization result Cw=0.40(g/g)
xt1(%) 0.9529 0.9549 0.9573
xt2(%) 0.8732 0.8734 0.8737
xt3(%) 0.9997 0.9997 0.9997
Q11,cn(kg/h) 11.9835 12.0058 12.0495
Q12,cn(kg/h) 0.1001 0.1001 0.3980
Q13,cn(kg/h) 0.0553 0.0059 0.0587
Q21,cn(kg/h) 24.6387 22.5764 22.8302
Q22,cn(kg/h) 10.4865 11.4495 11.9523
Q23,cn(kg/h) 0.1001 1.9160 3.3507
Qzn(kg/h) 0.2779 0.2781 0.2784
xc1(Yuan/Shi) 155.3776 155.0312 160.0790
xc2(Yuan/Shi) 450.8838 460.0551 488.1045
J (Yuan/Shi) 2977.2293 3047.0352 3119.2983
From fig. 5a to 5g, it can be seen that the economic cost J (yuan/hour) of the whole process is different for different pulp concentrations. The higher the pulp concentration is, the larger the leaching rate is, however, in order to meet the leaching rate requirement, the more material consumption is, and therefore the more economic cost is. Concentration C of ore pulpw0.36 and CwAs can be seen by putting the optimization results of 0.40 into the simulation models, the actual economic cost is 3119.2983 (yuan/time). However, the nominal operating point CwThe optimum manipulated variable found at 0.36 is not the actual optimum operation, and the economic cost of bringing into the simulation model is 3175.7859 (in/min), which is significantly greater than the actual economic cost.
As can be seen from table 3, when the operation variables obtained from the nominal operating points are brought into the actual process, the leaching rate is smaller than the leaching rate obtained from the actual process, and the actual production requirements cannot be met. Therefore, the variation in pulp concentration should be considered as a random disturbance.
TABLE 3 comparison of the Leaching rates in two different operations
Variables of Total leaching rate One time leaching rate Secondary leaching rate
Cw=0.36-0.40 0.9945 0.9569 0.8730
Cw=0.40-0.40 0.9946 0.9573 0.8737
And 4, step 4: and compensating and implementing the operation variables in the suboptimal operation state by using a self-optimization control method. First, 450 samples were randomly produced in a pulp concentration disturbance space using the Monte Carlo method. Carrying out optimization solution by using a full-flow optimization method to obtain an optimal operating variable u ═ Q11,cn,...,Q23,cn,Qzn]TAnd a process variable. The invention selects the output variable y ═ Q in the whole process production process11,cn,...,Q23,cn,Qzn,C11,s,...,C23,s,C11,cn,...,C23,cn,Ccnd]TThe measured noise is 0 as a mean and 2% of the nominal standard deviation. Obtaining a set value C of a controlled variable by using a self-optimization control off-line modeling methods=Hy=[0.0083 -0.0039 0.0310 0.01292 0.0064 -0.0091 0.0005]T
The simulation experiment adopts two methods for experimental verification: the first method is implemented by using a nominal working point. The optimal operation variable is the optimal operation obtained by selecting a nominal working point, and the actual production process is randomly disturbed and changed by the ore pulp concentration, namely d ═ Cw]=[0.320.40](ii) a And the second method is implemented by adopting a self-optimization control method. The operation variable is changed in real time according to the self-optimization control method, and the actual production process is disturbed and changed randomly by the ore pulp concentration, namely d ═ Cw]=[0.320.40]。
The method carries out random simulation tests for 65 times, firstly, before economic loss results of two methods are compared, performance results of the 65 processes are evaluated through an operation state evaluation method, the performance grades are divided into 1,2 and 3, and the performance grades respectively represent a superior grade, a suboptimal grade and a non-superior grade. The evaluation results are shown in fig. 7. Because 65 disturbances are randomly generated for simulation in the invention, three performance levels are possible. The ore pulp concentration disturbance variable close to the nominal value C with the optimal running state gradewWith a nominal operating point of 0.36, the process performance obtained is excellent.
And for suboptimal and suboptimal performance, the ore pulp concentration is obviously deviated from a nominal value, and the deviation is larger when the suboptimal degree is larger. As can be seen from the economic loss of fig. 8a, the economic loss obtained with the nominal operating point is large, with a maximum loss of 34.81 (n/hr) and an average loss of 9.98 (n/hr). After the compensation operation by the self-optimization control compensation method, the economic loss is 0.699 (yuan/hr) as shown in fig. 8b, and the average loss is 0.0130 (yuan/hr). It is clear that the economic losses are greatly improved.
In the simulation test, the deviation of the controlled variable obtained when the operation is carried out by using the nominal working point is large, the maximum deviation is 2.0601, and the average deviation is 0.9881. The deviation of the controlled variable obtained by the self-optimization control compensation method is shown in fig. 9, and the maximum value is 0.3849, and the average deviation is 0.1397. This further illustrates that when the nominal operating point is applied to an actual production process, the operating variables and process variables are not actually optimal, thus reducing production efficiency and bringing about huge economic losses.
And 5: the method for optimizing and setting the compensation of the operation amount based on the data is utilized to enable the non-optimal production process to return to an optimal or suboptimal operation state; the invention randomly carries out 5 groups of simulation experiments, and obtains the optimal operating variable under the working condition of the group and the economic loss corresponding to the optimal operating variable in advance aiming at each group of experiments, and each group of experiments adopts a nominal working point for experimental comparison.
Firstly, aiming at each group of working conditions, searching several groups of data which are most similar to the current working conditions in a historical optimal database by adopting a just-in-time learning idea (JIT for short);
secondly, the difference between similar sample data and the operation variable under the current working condition is solved to obtain a modeling data set { zi,Δui}i=1,2,...,7Wherein z isiFor similar sample conditions, Δ uiIs the difference in manipulated variables. Establishing a data model delta u between a working condition and an operation amount compensation value by adopting a Partial Least Squares (PLS) methodi=fr(zi,θ);
Finally, aiming at each group of experiments, working condition conditions are brought into the well-established data model and are based on the regression model delta u2,k=fr(ztTheta) compensating the current operation to obtain a compensation value delta u2,kAnd further obtain new operation variablesAnd (4) bringing new operation variables into the actual production process to obtain the economic cost J of each group of experiments. In order to better perform experimental comparison, the invention adopts three methods to perform simulation verification, which respectively comprise the following steps: method one, nominal working point operation; the second method is that the operation of the compensation method is optimally set based on the operation amount of the data; the third method comprises the optimal operation of the actual process; the results of the experiment are shown in FIG. 10.
Since the nominal operating point is not actually optimal, the economic cost obtained when bringing it into the actual production process is not optimal, with large deviations. After the operation of the data-based operation quantity optimization setting compensation method is adopted, the compensated operation variables are close to the actual optimal operation, so that the deviation between the obtained economic cost and the actual optimal economic cost is very small when the operation is brought into the actual production process. Table 4 lists the Root Mean Square Error (RMSE) and the Mean Absolute Error (MAE) of the economic losses after different process operations. As can be seen from the table, for comparison of five groups of random simulation experiments, the economic loss after the operation is greatly reduced by adopting the data-based operation amount optimization setting compensation method, and the RMSE and the MAE are both smaller than the error obtained by the operation based on the nominal working point.
TABLE 4 error comparison under two different operations
Through the above examples, the effectiveness of the gold hydrometallurgy full-process real-time optimization method is shown, the real-time optimization compensation operation of the hydrometallurgy full-process from suboptimal to non-optimal from the perspective of different evaluation results is realized, the problems that a quantitative model cannot be established due to uncertainty factors existing in the actual production process and the optimization result based on a mechanism model is not actually optimal are solved, an effective method is provided for solving the real-time optimization of a complex industrial process, and the method has a wide application prospect.
Finally, it should be noted that: the above-mentioned embodiments are only used for illustrating the technical solution of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.

Claims (9)

1. A real-time optimization compensation method for a gold hydrometallurgy full flow comprises processes of filter pressing, dewatering, size mixing, leaching, washing and replacement, and is characterized by comprising an online evaluation and operation compensation part, wherein the online evaluation and operation compensation part comprises the following steps:
s1, performing online analysis on the real-time optimization result of the whole gold hydrometallurgy flow by using the process running state evaluation method to obtain an evaluation result;
s2, selecting a compensation method matching the evaluation result for the evaluation result, which specifically includes:
s21, aiming at the condition that the evaluation result is suboptimal, selecting a linear combination of the output variables of the whole process as a controlled variable by adopting a compensation method of self-optimization control, and realizing the self-optimization control of the whole process of the hydrometallurgy under the uncertainty disturbance by controlling the tracking set value of the controlled variable;
s22, aiming at the condition that the evaluation result is non-optimal, a data-based operation quantity optimization setting compensation method is adopted, and compensation operation is carried out on the current operation quantity based on a timely learning thought and a partial least square method, so that self-optimization of process operation is realized;
s23, aiming at the condition that the data similar to the current working condition can not be found in the historical database, adopting a gold hydrometallurgy full-process re-optimization method for the previous working condition data, and further obtaining the optimal operation under the current working condition.
2. The method of claim 1, comprising an offline modeling portion;
the offline modeling part specifically comprises: the method comprises the steps of establishing a gold hydrometallurgy full-process optimization model, establishing a process running state evaluation model, establishing an optimal operation case base and establishing a self-optimization control model.
3. The method of claim 2, wherein the gold hydrometallurgical full process optimization modeling comprises:
wetting with goldThe economic cost of the whole process production process of metallurgy is the minimum, and the addition quantity Q of sodium cyanide in the two cyanidation leaching processes is used as the optimization targetji,cnJ 1 and 2 are two leaching processes, i 1,2 and 3 are the number of leaching tanks in one leaching process and the addition quantity Q of zinc powder in a replacement processznForming a gold hydrometallurgy full-process operation vector u ═ Q as an operation variable11,cn,...,Q23,cn,Qzn]T
Finally establishing a gold hydrometallurgy full-process optimization model which satisfies the quantitative relation of each procedure in the gold hydrometallurgy full-process mechanism model by satisfying the practical requirement range constraints of the quality indexes and the operation quantities of each sub-process, wherein the gold hydrometallurgy full-process optimization model is expressed as follows:
wherein J is economic cost function of the whole process of gold hydrometallurgy, Pcn,Pzn,Pcnd,PAuFor the unit price of each material, EC is the total energy consumption for production, Q23,sShowing the ore flow in the last tank of the second leaching process, C13,cn、C23,cnRespectively showing the concentration of cyanide ions in the last tank in the primary leaching process and the secondary leaching process, C23,sRepresenting the grade value of the solid gold in the last groove of the second leaching process; f. ofpA mechanism model representing the p-th sub-process; x is the number oftp(p ═ 1,2,3) represents the leaching rate in the primary leaching process, the leaching rate in the secondary leaching process, and the substitution rate in the substitution process,is the minimum requirement constraint of the first leaching rate, the second leaching rate and the replacement rate; u. ofpThe operation vector of the p procedure is shown, and z is a working condition, including ore flow, solid gold grade and ore particle size; cwThe ore pulp concentration represents a random disturbance variable; c1i,1Au、C1(i-1),1Au、C2i,1Au、C2(i-1),1AuRespectively representing the liquid phase gold grade of the ith tank in the primary leaching process, the (i-1) th tank in the primary leaching process, the ith tank in the secondary leaching process and the (i-1) th tank in the secondary leaching process, wherein i is 1,2 and 3; c1i,cn、C1(i-1),cn、C2i,cn、C2(i-1),cnRespectively showing the concentrations of liquid phase cyanide ions in the ith tank in the primary leaching process, the (i-1) th tank in the primary leaching process, the ith tank in the secondary leaching process and the (i-1) th leaching tank in the secondary leaching process; qsIs the ore flow, V is the reactant volume;the leaching rate index required in the production process is shown; u. of3Denotes the amount of zinc powder added in the substitution process, wherein u3,min,u3,maxRespectively representing the minimum value and the maximum value of the addition amount of the zinc powder required in the actual production process; u. ofjiDenotes the amount of sodium cyanide added in the ith cell in the jth procedure, uji,min、uji,maxThe minimum and maximum required amounts of NaCN in each tank are indicated.
4. The method of claim 3, wherein the process operating state evaluation model building comprises:
establishing a hydrometallurgy full-process evaluation model according to the full-process industrial characteristics and the data characteristics of each sub-process, and specifically implementing the following steps:
a1, preprocessing the quantitative data and the qualitative data acquired on line in the metal metallurgy process to obtain the quantitative data and the qualitative data to be analyzed, wherein the data preprocessing steps are as follows:
a11, selecting economic cost in the whole-flow production process as an evaluation index, and selecting process variables capable of influencing the process running state;
a12, sequentially expressing different state grades by a series of positive integers for qualitative variables; for the quantitative variable, carrying out simple smoothing treatment, selecting a sliding window with a certain length, and representing the information of the whole window by using the variable mean value in the window;
a2, establishing an operation state evaluation model for evaluating the operation state grade of the whole process according to the quantitative data and the qualitative data in the metal metallurgy process in the historical time period, and specifically implementing the following steps:
a21, determining the grade of the whole process; let the modeling data beH is the number of samples, J is the number of variables, including qualitative variables and quantitative variables; according to the economic cost of the full-process evaluation index, the process running state is divided into three grades of optimal, suboptimal and non-optimal, and is marked as X1,X2,X3Wherein, X comprises 6 subblock data which are respectively a size mixing process subblock, a primary cyanidation leaching process subblock, a primary filter-pressing washing process subblock, a secondary cyanidation leaching process subblock, a secondary filter-pressing washing process subblock and a replacement process subblock; the corner marks 1,2 and 3 respectively represent that the performance grade is excellent, suboptimal and non-excellent;
a22, establishing a full-process evaluation model; obtaining data X of each gradelAfter (1, 2 and 3), establishing a decision table for the full-process model by using a fuzzy probability rough set;
wherein the condition attribute is Xl(l 1,2,3), the decision attribute is the full-process operating state level l, and the domain contains Xl(l ═ 1,2, 3).
5. The method of claim 4, wherein the optimal operating case base establishment comprises:
historical optimization-based process operating data { xn}n=1,...,NN is the number of data with the optimal process running state grade reserved in the history, and an optimal operation case library is establishedWherein z isnAs a condition of operation, unUnder the working condition of znOptimum operating variable of time, NcThe number of data in the optimal case library is determined.
6. The method of claim 5, wherein self-optimizing control model establishment comprises:
the method comprises the following steps of taking economic cost and energy efficiency of the operation of the whole gold hydrometallurgy process as cost functions, selecting linear combination of whole process output variables as controlled variables, and establishing a self-optimization control model of the whole gold hydrometallurgy process, wherein the specific self-optimization control off-line modeling process comprises the following steps:
b1 distribution space C for uncertain disturbance variable ore pulp concentrationw∈[32%,40%]Generating N using MonteCarlo samplingsGroup sequence
B2 case d for each group of pulp concentration(n)Performing off-line optimization solution by adopting a full-flow optimization method, and calculating to obtain optimal input uopt=[Q11,cn,...,Q23,cn,Qzn]T
B3, recording the corresponding optimal output variable under each disturbance situation according to the previous solution resultWherein the output variable isComprises the optimum input variables u of the sodium cyanide addition amount of three leaching tanks in the first leaching process and the second leaching process and the zinc powder addition amount in the replacement processopt=[Q11,ch,Q12,cn,Q13,cn,Q21,cn,Q22,cn,Q23,cn,Qzn]Grade C of solid gold in two leaching processesji,sWith concentration of cyanide particles Cji,cnGrade C of cyanogen slag in replacement Processcnd(ii) a The measured noise is Gaussian noise, and the standard deviation is 5% of the nominal value; forming a matrix by the optimal output variables obtained under all disturbance conditionsWhereinIs a covariance matrix of the measured noise;
b4, taking a nominal working point CwCalculating gain matrix G of output variable value by finite difference method as 36%y,refAnd hessian matrix J of the loss function J with respect to the manipulated variable uuu,ref
B5, calculating the optimal combination matrix H in combination C ═ Hy so that the average loss is minimal, i.e.The combination matrix H can be obtained by solving the following equation:wherein the content of the first and second substances,representing a matrix of optimal output variables containing measurement noise, Gy,refA gain matrix representing the input and output at nominal operation, Juu,refA hessian matrix representing the economic cost function J of the gold hydrometallurgy full flow relative to the operational variable u; optimal output variable based on nominal working pointUsing formulasObtaining the set value C of the controlled variables
7. The method according to claim 6, wherein the online evaluation and operation compensation section specifically comprises: the method comprises the steps of online evaluation of process running states, online compensation of self-optimization control and establishment of an operation quantity optimization setting compensation model based on data, and specifically comprises the following steps:
solving and obtaining an optimal operation variable under a nominal working point through full-process optimization, and bringing the operation variable into an actual gold hydrometallurgy production process, wherein uncertainty disturbance exists in the actual process;
firstly, judging the current process running performance on line based on a process running state evaluation method;
secondly, respectively implementing compensation strategies corresponding to a self-optimization control method and a data-based operation amount optimization setting compensation method aiming at different performance levels;
thirdly, the compensated operation variables are brought into the actual production process, so that the optimal operation is obtained; the specific steps of online evaluation and operation compensation are as follows:
101. operating variables determined from nominal operating pointsBrought into the actual gold hydrometallurgical production process to obtain a process variable xtAnd an output variable yt
102. Evaluating the process running state on line; aiming at the current process variable and the output variable, processing the current process variable by adopting a full-process running state evaluation model established off-line, and determining the grade of the full-process running state in the metal metallurgy process
The online evaluation comprises the following specific steps:
c1, obtaining online data xtAnd performing comparison with data in a decision table established offline according to the fuzzy probability rough setMatching to obtain xtLambda intercept of fuzzy equivalence class ofWherein Representing data xtλ is a given threshold;
c2, calculating the probability of belonging to the ith grade as:
wherein d islDenotes the l-th rank, Xl(1, 2,3) determining the operation state level according to historical data; judging the running state grade of the whole process as follows:
103. according to different evaluation results, a corresponding compensation strategy is adopted, and the specific compensation form is divided into the following steps:
case 1: when the evaluation result is excellent, the current operation variable is indicated to be the optimal operation, and compensation operation is not needed until new process disturbance occurs;
case 2: when the evaluation result is suboptimal, namely the process running state deviates from the optimal operation, but the deviation is not large; aiming at the situation, self-optimization control online compensation operation is adopted, the controlled variable is controlled to track a set value to realize that the influence of uncertainty and change disturbance on the optimization of the economic target is minimum, and then the current operation is compensated to realize the self-optimization control of the hydrometallurgy process under the uncertainty disturbance;
the specific compensation steps are as follows:
d1, connection matrix H obtained based on off-line calculation and output variable y under current working conditiontUsing the formula Ct=HytTo obtain the controlled variable Ct
D2, calculating the current operating variable based on the following formulaIs compensated for by a compensation value delta u1.k
Δy=Gy·Δu
H′=BH,B=(H·Gy)-1
C′=B·C=H′·y
B·ΔC=H′·Δy=(H·Gy)-1·H·Gy·Δu=I·Δu
Where Δ y denotes the difference in the output values at nominal operation, determined by means of finite differences, GyIs Gy,refThe compensation value delta u is delta u representing the gain matrix of the input and the output under the nominal work1.kΔ C is a controlled variable set value CsWith the currently sought controlled variable CtA deviation of (a); h' represents the general solution of the connection matrix H;
d3, adding new operation variablesCarrying into the actual production process, and executing the next stage;
case 3: when the evaluation result is non-optimal, namely the process running state deviates from the optimal operation, and the deviation is very large; aiming at the situation, a data-based operation quantity optimization setting compensation method is adopted, so that the production process returns to an optimal or suboptimal operation state; the method comprises the steps of searching several groups of data which are most similar to the current working condition in a historical optimal database based on a timely learning thought, establishing a data model between the working condition and an operation amount compensation value by adopting a partial least square method, compensating the current operation, and further realizing process operation self-optimization;
the specific compensation steps are as follows:
e1 for the current process variable xtSearching the current working condition z in a case library with excellent history established off-linetSimilar case data ziCalculating the sample similarity S based on the following formulaz
Sz=ρexp(-Di)+(1-ρ)cos(ωi)
Wherein Sz∈[0,1]The larger the sample is, the stronger the sample similarity is; diAnd cos (ω)i) Respectively represents ziAnd ztThe cosine values of the Euclidean distance and the included angle between the two cosine values; rho (rho is more than or equal to 0 and less than or equal to 1) is an index of information weight of the balance distance and the included angle;
e2, defining a sample similarity threshold (0 < 1), and judging the sample similarity SzWhether greater than a threshold;
in the first case: if S iszThe data similar to the current working condition exists in the historical excellent case library; if the number of the similar samples is larger than or equal to L, performing operation compensation based on a timely learning idea, and otherwise, performing second-situation operation;
the specific operation steps for the first case are as follows:
f1, calculating to obtain L similar samples serving as modeling data of the local regression model based on a sample similarity calculation formula;
f2, based on L similar samples { zi,ui}i=1,2,LCalculating the manipulated variable ui(i ═ 1, 2.. multidot.l) and the current manipulated variableThe difference between them, is recorded as Δ ui(i ═ 1,2,. ·, L); building a modeling dataset { zi,Δui}i=1,2,L
F3, establishing a local regression model delta u based on a partial least square methodi=fr(ziθ), where θ is an update parameter that varies from modeling data to modeling data; for the current operating condition ztBased on a local regression model frIs calculated whenCompensation value Deltau of front operation variable2,k
F4, adding new operation variableCarrying into the actual production process, and executing the next stage;
in the second situation: if S isz< or SzAnd if the number of the similar samples is more than or equal to L, indicating that a large amount of data similar to the current working condition cannot be found in the historical optimal case library, so that the traditional real-time optimization method is adopted, namely, disturbance variables are identified, the full-process optimization model is re-optimized, and optimal operation is further obtained.
8. The method according to claim 1, wherein when the evaluation result is non-optimal, several groups of data closest to the current working condition are searched in the historical optimal database based on the timely learning thought, and the method is divided into two situations according to the number of similar data:
the first situation is as follows: if the number of the similar data is larger than a certain threshold value, it is indicated that enough data similar to the current working condition exists in the historical preferred case library; a data model between the working condition and the compensation value of the operation amount is established by adopting a partial least square method, the current operation is compensated, and further, the process operation self-optimization is realized;
case two: if the number of the similar data is smaller than a certain threshold value, it is indicated that a large amount of data similar to the current working condition cannot be found in the historical preferred case library; and (3) identifying disturbance variables and re-optimizing the full-flow optimization model by adopting a traditional real-time optimization method so as to obtain optimal operation.
9. The method of any one of claims 1-8, further comprising a process detection system, the process detection system comprising a PLC controller, concentration detection, pressure detection, and flow detection;
the PLC controller adopts a Simens 400 series CPU 414-2 and is provided with a Profibus DP port connected with a distributed IO; an Ethernet communication module is equipped for the PLC, and is used for an upper computer to access PLC data; the PLC controller and the Ethernet communication module are arranged in a PLC cabinet in the central control room;
the upper computer selects an i7 association computer and adopts a WINDOWS XP operating system;
the gold hydrometallurgy full-flow real-time optimization system adopts C #2008 programming software on an i7 association computer, and the gold hydrometallurgy full-flow real-time optimization algorithm adopts Matlab 2016a programming software;
the signal transmission software for the PLC and the full-flow real-time optimization adopts C #2008 programming software; the method comprises the steps that a detection instrument is installed on the site of the hydrometallurgy process, the detection instrument transmits collected signals to a PLC through Profibus-DP, the PLC transmits the collected signals to an upper computer through Ethernet at regular time, and the upper computer transmits received data to a gold hydrometallurgy full-flow real-time optimization system to perform full-flow real-time optimization compensation operation and provide operation guidance suggestions.
CN201910236152.4A 2019-04-12 2019-04-12 Real-time optimization compensation method for gold hydrometallurgy full flow Active CN109885012B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910236152.4A CN109885012B (en) 2019-04-12 2019-04-12 Real-time optimization compensation method for gold hydrometallurgy full flow

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910236152.4A CN109885012B (en) 2019-04-12 2019-04-12 Real-time optimization compensation method for gold hydrometallurgy full flow

Publications (2)

Publication Number Publication Date
CN109885012A CN109885012A (en) 2019-06-14
CN109885012B true CN109885012B (en) 2020-12-25

Family

ID=66934633

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910236152.4A Active CN109885012B (en) 2019-04-12 2019-04-12 Real-time optimization compensation method for gold hydrometallurgy full flow

Country Status (1)

Country Link
CN (1) CN109885012B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110918266B (en) * 2019-12-10 2020-09-29 中南大学 Control device and method for improving flotation froth quality
CN111142488B (en) * 2019-12-30 2021-05-04 浙江中控技术股份有限公司 Process industrial optimization method based on working condition comparison driving

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101526814B (en) * 2009-02-03 2011-05-04 东北大学 Leaching rate prediction and optimization operation method in wet metallurgical leaching process
CN104062953B (en) * 2014-06-12 2017-03-01 东北大学 Golden hydrometallurgy process running status on-line evaluation method
CN104597755B (en) * 2014-06-12 2017-02-15 东北大学 Hydrometallurgical gold cyanide leaching optimization method
CN106886154B (en) * 2017-04-26 2019-04-12 东北大学 Hydrometallurgy whole process modeling method and optimization method based on interval number
CN106950946B (en) * 2017-05-16 2019-01-22 东北大学 A kind of hydrometallurgy exception control method based on optimization principles
CN108921352B (en) * 2018-07-06 2021-10-22 东北大学 Hydrometallurgy leaching process optimization method with interval uncertainty

Also Published As

Publication number Publication date
CN109885012A (en) 2019-06-14

Similar Documents

Publication Publication Date Title
CN109885012B (en) Real-time optimization compensation method for gold hydrometallurgy full flow
CN101526814B (en) Leaching rate prediction and optimization operation method in wet metallurgical leaching process
CN101813932B (en) Method for component content prediction and optimization operation in wet-process metallurgic extraction process
CN104062953A (en) Gold hydrometallurgy process running state online evaluation method
Xie et al. An integrated predictive model with an on-line updating strategy for iron precipitation in zinc hydrometallurgy
CN104199389B (en) The optimization method of golden hydrometallurgy replacement process
CN104597755A (en) Hydrometallurgical gold cyanide leaching optimization method
CN107121490B (en) The evaluation method of operating status in a kind of hydrometallurgy process
CN108921352B (en) Hydrometallurgy leaching process optimization method with interval uncertainty
CN101748186A (en) FSVM-based lysine fermentation process key state variable soft measuring method and system
CN108037747B (en) Golden hydrometallurgy whole process hierarchical block operating status on-line evaluation method
CN102004444A (en) Multi-model predictive control method for component content in process of extracting rare earth
CN106886154B (en) Hydrometallurgy whole process modeling method and optimization method based on interval number
CN102778843A (en) Operation control method of high magnetic grading process
CN102676818B (en) Prediction method for zinc powder effectiveness coefficient in wet zinc smelting and copper removal process
CN110866659A (en) On-line prediction method for caustic ratio in dissolution process of high-pressure pipeline in Bayer-process alumina production
CN107563656B (en) Method for evaluating running state of gold hydrometallurgy cyaniding leaching process
CN109033524A (en) A kind of chemical process concentration variable On-line Estimation method based on robust mixed model
Liu et al. Industrial process operating optimality assessment based on Gaussian mixture model
Liu et al. A novel real-time optimization compensation method based on POPOA for the gold hydrometallurgy process
CN104263960A (en) Online prediction method of replacement rate and gold mud grade in hydro-metallurgy and replacement processes of gold
CN109840722A (en) A kind of gold hydrometallurgy whole process operating status on-line evaluation method and system
Liu et al. Operating performance assessment based on GMM-GPR for gold hydrometallurgy processes
Chang et al. Online operating performance evaluation for the plant‐wide industrial process based on a three‐level and multi‐block method
CN112034709A (en) RBF ANN-based gold cyanidation leaching rate interval prediction method and device

Legal Events

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