CN102004444A - Multi-model predictive control method for component content in process of extracting rare earth - Google Patents
Multi-model predictive control method for component content in process of extracting rare earth Download PDFInfo
- Publication number
- CN102004444A CN102004444A CN 201010555634 CN201010555634A CN102004444A CN 102004444 A CN102004444 A CN 102004444A CN 201010555634 CN201010555634 CN 201010555634 CN 201010555634 A CN201010555634 A CN 201010555634A CN 102004444 A CN102004444 A CN 102004444A
- Authority
- CN
- China
- Prior art keywords
- model
- rare earth
- component concentration
- extraction process
- earth extraction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Feedback Control In General (AREA)
Abstract
The invention relates to a multi-model predictive control method for component content in the process of extracting rare earth. The method comprises the following steps of: establishing a multi-linear model with concentration extraction series of 1 at the component content detection point in the process of extracting and separating the rare earth; converting the established multi-linear model into a controlled autoregressive integral moving average (CARIMA) model by a generalized predictive control algorithm; and calculating to obtain the change of controlled quantity (flow quantity of an extracting agent or acid liquid) according to the deviation between a component content on-line estimated value and a preset value. The method is suitable for on-line detection and automatic control of the component content in the process of extracting and separating the rare earth.
Description
Technical field
The present invention relates to rare earth extraction detachment process component concentration multi-model forecast Control Algorithm, belong to rare earth extraction process component concentration on-line monitoring and automatic control technology field.
Background technology
Rare earth element is the indispensable important source material of many industrial circles, is developed to high-tech areas such as communication, space flight from conventional industries fields such as metallurgy, machinery, oil, chemical industry.In recent years, China's Rare Earth Separation industry has obtained fast development, and especially the Solvent Extraction Separation rare-earth process is reached the international leading level under the guidance of cascade extraction theory and new development thereof.The rare earth element kind is many, chemical property is similar, each other separation coefficient is less, and this makes that the industrial extract and separate flow process progression of being made up of cascade mixer-settler that generally adopts is many, mechanism is complicated, have multivariate, strong coupling, non-linear, the time characteristics such as change and large time delay.
Rare earth element component content is the key factor that influences the rare-earth products quality, and therefore component concentration has great importance to production run operation and automatic control in the online detection extraction process.By the flow in sensitivity level detection elements component concentration and timely adjustment extractant, feed liquid and acid solution, outlet obtains highly purified rare-earth products at two ends.At present, the instrument that is applied to the online detection of rare earth extraction process component concentration has a lot of shortcoming and defect, is difficult to prolonged application in industrial practice.Component concentration is set up correct mathematical model realize that soft measurement seems particularly important.The mathematical model of rare earth extraction process has a lot of forms, but initial some models of setting up all are based on traditional process modeling, it is static in itself, can not reflect the dynamic change of extraction process, perhaps some models are too complicated and be not suitable as the soft-sensing model of component concentration On-line Estimation.The appearance of all kinds of algorithms and development, for the foundation of rare earth extraction process component concentration soft-sensing model provides condition, someone adopts neural network to the extraction process modeling, become a new way of the online detection of component concentration, but because the data of gathering are limited, the training result of neural network can not cover whole working condition, influence to precision of prediction and network training effect is bigger, someone has proposed a class and has simplified the bilinearity dynamic model, its precision of prediction in subrange is higher, but robustness is relatively poor, but also the someone has proposed the multi-model soft-measuring technique, but model quantity is more, and calculated amount is bigger.Simultaneously, even the model of above-mentioned foundation can be realized the online detection of component concentration, but when the component concentration of monitoring point is not within the experience scope, do not provide controlled quentity controlled variable how much.
Summary of the invention
The objective of the invention is, rare earth extraction detachment process component concentration is set up the soft measurement forecasting procedure of correct multi-model, on the basis of the component concentration online soft sensor multi-model of setting up, in conjunction with generalized predictive control, the polyteny model is changed, obtained the relational expression of component concentration and extractant or acid solution flow control amount, by the deviation of On-line Estimation value and setting value, calculate the controlled quentity controlled variable of extractant or acid solution, realize the online detection and the robotization control of component concentration.
Technical scheme of the present invention is: according to the characteristics of rare earth extraction detachment process, based on material balance equation, and assemble the thought of modeling in conjunction with segmentation, set up to assemble extraction progression at component concentration check point place and be 1 polyteny model, utilize the GPC (Generalized Predictive Control) algorithm in the PREDICTIVE CONTROL, with the polyteny model conversion of setting up is controlled autoregressive integration running mean (CARIMA) model, by the deviation of online detected value of component concentration and setting value, calculate the variation of the flow control amount of extractant or acid solution.Described technical scheme concrete steps are:
(1) from the theoretical mobile equilibrium of cascade extraction,, determines the linear structure of submodel based on the material balance equation of extraction separation process; Secondly the sample data of gathering is carried out cluster analysis, obtain the steady-state operation point and the corresponding data set of extraction process, set up partial model according to the model structure of determining, by the initial parameter of each submodel of least square method of recursion identification; The model of selecting suitable multi-model scheduling mechanism (model switches or the model weighting) to get is to the end once more exported.Can realize the On-line Estimation of rare earth extraction process component concentration like this.
(2) on the component concentration On-line Estimation model based of above-mentioned foundation, in conjunction with generalized predictive control, the polyteny model is changed, obtain the relational expression of component concentration and extractant or acid solution flow control amount, deviation by online detected value and setting value, calculate the controlled quentity controlled variable of extractant or acid solution, realize the online detection and the robotization control of component concentration.
The multi-model PREDICTIVE CONTROL structured flowchart that switches based on model as shown in Figure 1.
Neutron model linear structure of the present invention is determined as follows:
Material balance can reflect the main dynamic perfromance that whole extraction process changes, so, set up the concentration dynamic equilibrium relation of i component at j level aqueous phase according to material balance:
Wherein, u
1, u
2, u
3Represent the flow of organic solvent, aqueous phase liquid and water cleansing solution respectively, work as j=1, during n+m, x
I, 0(t-τ)=x
I, 0, x
I, n+m+1(t-τ)=x
I, n+m+1Be known quantity, when j=n, (u
2+ u
3) x
I, j+1=u
3x
I, j+1+ u
2x
I, F, when j>n, u
2=0.R
jThe expression hold-up volume, τ represents retardation time, D
I, jRepresent two alternate partition factors.
Because the mobile more complicated of liquid in the basic extraction cells mixer-settler, and the model of setting up generally all is Nonlinear System of Equations, exist find the solution length consuming time, to the high shortcoming of computing machine requirement, can not be directly used in On-line Control and optimization.Therefore, the thought of utilizing segmentation to assemble modeling proposes to simplify to model, and promptly the dynamic perfromance according to the rare earth extraction process is divided into several extraction sections with extraction process, and each section is abstracted into a virtual extracting stage.According to above-mentioned thought, when supposing that whole extraction process reaches equilibrium state, set up check point in the l level, as shown in Figure 2.
In order to detect the component purity y at this some place, be based upon the concentration dynamic equilibrium relation of this some i of place component aqueous phase in the l level:
Wherein, x
I, FThe concentration of expression rare earth feed liquid, x
I, 0(t-τ)=x
I, 0, x
I, 2(t-τ)=x
I, 2Be known quantity.
The described relation of following formula is non-linear, in the working point
Have following relational expression to set up:
The working point of formula (2) is moved on to initial point, and carries out Taylor's linear expansion, obtain the local linear model of i component in the l level at initial point:
Can get formula (4) discretize:
x
i(k+1)=Ax
i(k)+Bu(k)+Dv(k) (5)
Wherein, u (k)=[u
1(k), u
2(k), u
3(k)]
TThe expression input vector, x
i(k)=x
I, 1(k) indicated concentration state vector, v (k)=x
I, F(k) expression perturbation vector, A, B, D representation model parameter matrix.
The output equation of system is
y
i(k)=Cx
i(k) (6)
Wherein, C is a parameter matrix, y
i(k)=y
I, l(k).
Can get the input/output relation of u and y by formula (5) and formula (6):
y(k+1)=Ay(k)+Gu(k)+Hv(k) (7)
Wherein, A, G, H are matrix of coefficients undetermined.
Can get by formula (7):
y(k+1)=θ
Tφ(k) (8)
Wherein, θ=[A, G, H]
T, φ (k)=[y (k)
T, u (k)
T, v (k)
T].Because system state can be surveyed,, then adopt least square method of recursion to carry out identification to θ so y (k) in the formula (8) and φ (k) all can obtain.
Model shown in the formula (8) is only applicable near some working points, and when bigger variation took place the running environment of system, parameter identification was difficult to follow the tracks of actual change, therefore can cause model inaccurate.Nonlinear system is launched in a plurality of working points, utilized the local linearization model on a plurality of working points to be similar to nonlinear system, can set up the object dynamic model that more approaches real system.According to this thought, the rare earth extraction process is launched in different working point, and with n input I
1, I
2..., I
nDescriptive system, near local linear model discrete each working point is as follows:
Wherein, θ
l=[A
l, G
l, H
l]
TBe parameter matrix, the matrix that φ (k) is made up of inputoutput data.
According to formula (9),, obtain the relational expression of monitoring point component purity y of place and input quantity u in conjunction with rare-earth cascade extraction process shown in Figure 1:
(10)
Wherein,
Be the parameter vector of l model, φ (k)=[y (k) ,-y (k-1), u
1(k), u
2(k), u
3(k), x
I, F(k)]
T
Said process has been determined the version of submodel, i.e. linear model, and ensuing groundwork is exactly how to determine the working point, just the steady-state operation point in the rare earth extraction process.
The cluster analysis of sample data:
Cluster analysis is not have one the sample set of mark to be divided into some classes by certain criterion, makes similar sample be classified as a class as far as possible, and the sample that does not have a similarity is divided in the different classifications as far as possible.Adopt the subtractive clustering method that sample data is carried out cluster analysis among the present invention, the clusters number n that obtains is the number of submodel in the multi-model, and cluster centre is the working point of determining extraction process.Cluster numbers n gets following target function:
Wherein, N is the sample data number, and n is a clusters number, X
i∈ R
qBe i input sample data, and X in the methods of the invention
i=[u
1i, u
2i, u
3i, x
I, F]
T, u
1i, u
2i, u
3iI flow representing organic solvent, aqueous phase liquid and water cleansing solution respectively, x
I, FThe feed concentration of representing i flow correspondence.As seen the rare earth extraction process is many input processes that 4 inputs are arranged.
Be j cluster centre, promptly
μ
IjBe the degree of membership of i sample data j cluster.The present invention copies the definition of subordinate function in the fuzzy logic with μ
IjBe expressed as follows:
The present invention adopts the subtractive clustering algorithm that sample data is classified, and with target function J
mWeigh the effect of cluster, main calculation procedure is as follows:
1) initial parameter δ is set
a=δ
Min, chosen steps increment ε, and ε>0;
5) repeating step 3), 4) calculate up to
ξ is a less positive constant, obtains j cluster centre
Wherein n is the cluster centre number that obtains, and n<N;
6) calculate the inferior cluster index of k (k>1)
Readjust δ
a=δ
aIf+ε is δ
a∈ [δ
Minδ
Max] repeated execution of steps 2 then)-5);
7) order
(k=1 ..., K), wherein K is the number of times of cluster, gets J
mCorresponding clusters number n and cluster centre
(l=1 ..., n) sample data is classified, according to nearest neighbouring rule, obtain n data set Ω
l
Utilize the subtractive clustering algorithm to obtain cluster centre
As the working point of local linear model, just the point of the steady-state operation in the actual production process is used categorized data set Ω
lReach the initial parameter of least squares identification partial model l, obtain the initial model of the online detection of rare earth extraction process elementary composition content.
The design of control method:
When production index is constant in the rare earth extraction production run, generally not to feed liquid flow value u
2Adjust, the control of process is by adjusting the flow u of extractant and cleansing solution
1, u
3Realize.GPC (Generalized Predictive Control) algorithm adopts the CARIMA model to control, and with the initial model that obtains through the polyteny model that obtains after preceding 100 groups of sample data optimizations is:
Following formula is represented with structure as follows
(13)
In the formula,
φ (k)=[y (k) ,-y (k-1), u
1(k), u
2(k), u
3(k), x
F(k)]
TIn conjunction with CARIMA structure of models form, following formula is transformed to:
Obtain k component concentration y and flow u constantly
iThe pass be:
Compare with the CARIMA model structure, (14), (15) two formulas are subtracted each other obtain
(16)
The factor that influences component concentration in the rare earth extraction process is a lot, and the online detection polyteny of the component concentration of Jian Liing model is component concentration output y and extractant flow u here
1, feed liquid flow u
2, cleansing solution flow u
3And feed concentration x
FBetween relation.When the design forecast Control Algorithm, respectively at different influence factor design control methods, adjust controlled quentity controlled variable, make component concentration output reach desirable desired value.
1, only to extractant u
1Control the design forecast Control Algorithm
To u
1When controlling, at first make the following assumptions: y, u
1Be variable, u
2, u
3, x
FBe constant.So, formula (14) is subtracted each other and can be got with formula (15)
In the formula, Δ u
1(k)=u
1(k)-u
1(k-1).
In conjunction with the polyteny model that obtains, corresponding each submodel design corresponding controller is:
When 2, only cleansing solution being controlled, the design forecast Control Algorithm
In like manner, to u
3When controlling, make the following assumptions: y, u
3Be variable, u
1, u
2, x
FBe constant.So, formula (14) is subtracted each other and can be got with formula (15)
In the formula, Δ u
3(k)=u
3(k)-u
3(k-1).
Utilize the model parameter after optimizing to calculate, obtain each submodel at cleansing solution u
3CARIMA model when controlling is shown below:
3, control u simultaneously
1, u
3, the design forecast Control Algorithm
Control u simultaneously
1, u
3, promptly adjust u simultaneously
1And u
3, make component concentration output valve y follow the tracks of the dreamboat value.Suppose: y, u
1, u
3Be variable, u
2, x
FBe constant.Can get by formula (14), (15)
The control method that in like manner obtains corresponding each submodel is:
In sum, under different assumed condition, foundation meets the CARIMA model of GPC (Generalized Predictive Control) algorithm rule, and design generalized forecast control method, calculate accurate changing value by adjusting different controlled quentity controlled variables, can make do not reach desirable desired value the monitoring point component concentration follow the tracks of desirable desired value fast, and then guaranteed the quality of exit rare-earth products.
The present invention's beneficial effect compared with the prior art is, the Rare Earth Separation process adopts cascade extraction technology mostly, and flow process progression is many, reflection mechanism complexity, and retardation time is long, is difficult to set up the controlling models of extraction process, thereby influences product quality and yield.The technical program is at first set up the soft-sensing model of component concentration, predict the component concentration at check point place in advance, obtain the deviation of component concentration and setting value, utilize the generalized predictive controller amount of being accurately controlled, thereby changed the blindness of regulating by rule of thumb, make component concentration in time reach setting value, solved the large time delay problem, guaranteed the quality of two ends exported products.The technical program is simple and practical, reduces the human resources input, improves enterprises production efficiency, reduces cost, and has realized the robotization control of rare earth extraction process.
The present invention is applicable to the online detection of rare earth extraction detachment process component concentration and the automatic control of product purity.
Description of drawings
Fig. 1 is the multi-model PREDICTIVE CONTROL structural drawing that switches based on model;
Fig. 2 is a l level elementary composition content detection point;
Fig. 3 is actual output and model curve of output in the online detection of rare earth component concentration of polyteny model;
Fig. 4 is a relative error curve in the online detection of rare earth component concentration of polyteny model;
Fig. 5 is for regulating the component concentration aircraft pursuit course of different controlled quentity controlled variables;
Fig. 6 is the control curve of output of Δ u3 correspondence when regulating u3 separately;
Fig. 7 is the control curve of output of u3 correspondence when regulating u3 separately;
Fig. 8 is the control curve of output of Δ u3 semi-invariant correspondence when regulating u3 separately;
Fig. 9 is for regulating u1 simultaneously, the control curve of output of Δ u1 correspondence during u3;
Figure 10 is for regulating u1 simultaneously, the control curve of output of u1 correspondence during u3;
Figure 11 is for regulating u1 simultaneously, the control curve of output of Δ u1 semi-invariant correspondence during u3;
Figure 12 is for regulating u1 simultaneously, the control curve of output of Δ u3 correspondence during u3;
Figure 13 is for regulating u1 simultaneously, the control curve of output of u3 correspondence during u3;
Figure 14 is for regulating u1 simultaneously, the control curve of output of Δ u3 semi-invariant correspondence during u3.
Embodiment
The invention process utilizes the measured data of certain Rare Earth Company extract and separate product yttrium to set up the polyteny model of rare earth component concentration On-line Estimation.Sensitivity level in the rare earth extraction process is set up the monitoring point, go out the elementary composition purity of this monitoring point by the model On-line Estimation, when its purity does not meet the demands, in time regulate the flow of extractant, feed liquid or acid solution, and then the elementary composition content of monitoring point is controlled in the setting range.
The place measures 150 groups of valid data in the monitoring point, preceding 100 groups of numbers are carried out cluster analysis, obtain 4 cluster centres shown in the table 1, be the steady-state operation point of extraction process, with the categorized data set identification of correspondence to local linear model that should operating point, obtain elementary composition content y (k) initial model Model 1~Model 4 shown in (1) formula, wherein, u
1, u
2, u
3The flow of representing extractant, feed liquid and acid solution respectively, x
FThe expression feed concentration.Utilize 50 groups of remaining data to carry out model measurement, obtain the structure shown in accompanying drawing 3 and the accompanying drawing 4.
Table 1 steady-state operation point
Model?1~Model?4:
As can be seen, the model output error is bigger in preceding 100 groups of data from Fig. 3 and Fig. 4, this be since model in constantly optimizing, the final mask that 50 groups of data utilization optimizations afterwards obtain is tested, error is less.So utilize this model more exactly real-time online detect the component concentration at monitoring point place.
Utilize GPC (Generalized Predictive Control) algorithm, carry out conversion in conjunction with the final detection model of determining, controlled model i.e. (CARIMA) model.At a time, obtain the component concentration y (k)=0.6 that the monitoring point is located by model, this moment, the corresponding flow value was respectively u
1=51.42L, u
2=4.5L, u
3=8.11L, feed concentration x
F=0.481.Judge that rule of thumb will make outlet purity satisfy production requirement, the component concentration at this check point place must reach 0.9.According to the deviation between controlling models and component concentration and setting value, regulate u
1Perhaps u
3Flow value, make the component concentration at monitoring point place reach setting value fast, and then guarantee the purity of exported product.
As prediction time domain N=12, during control time domain Nu=7, obtain Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Figure 10, Figure 11, Figure 12, Figure 13, aircraft pursuit course and controlled quentity controlled variable change curve shown in Figure 14.Fig. 5 is for regulating the component concentration aircraft pursuit course of different controlled quentity controlled variables, and as shown in Figure 5, output 1 is to regulate acid solution u separately
3The time aircraft pursuit course that obtains, output 2 is for regulating extractant u simultaneously
1With acid solution u
3The time aircraft pursuit course, relatively draw, regulate u simultaneously
1And u
3Can comparatively fast reach setting value 0.9, but concussion occur during the course, prediction time domain N is provided with and causes more greatly.Fig. 6, Fig. 7, Fig. 8 are the control curve of output of correspondence when regulating u3 separately, as shown in Figure 7, and the controlled quentity controlled variable u of corresponding output 1
3Change curve, place, monitoring point component concentration is lower than setting value 0.9, needs to reduce u
3Value, after reaching 0.9, u
3Flow value be 6.712L, accumulation reduces 1.398L.Fig. 9,, the control curve of output of Figure 10, Figure 11 correspondence when regulating u1 simultaneously; Figure 12, Figure 13, Figure 14 are the control curve of output of correspondence when regulating u3 simultaneously, as Figure 10, shown in Figure 11, and the controlled quentity controlled variable u of corresponding output 2
1, u
3Change curve needs to increase extractant u simultaneously
1, stable back u
1Value be 51.663L, cumulative rises 0.247L as Figure 13, shown in Figure 14, reduces acid solution flow u at the same time
3, accumulation reduces 0.564L, final u
3Be 7.547L.
Claims (4)
1. rare earth extraction process component concentration multi-model forecast Control Algorithm, it is characterized in that, described method is 1 polyteny model at rare earth extraction detachment process component concentration check point place foundation assembly extraction progression, utilize GPC (Generalized Predictive Control) algorithm, with the polyteny model conversion of setting up is controlled autoregressive integration running mean (CARIMA) model, by the deviation of component concentration On-line Estimation value and setting value, calculate the variation of the flow control amount of extractant or acid solution.
2. rare earth extraction process component concentration multi-model forecast Control Algorithm according to claim 1 is characterized in that, said method comprising the steps of:
(1), determines the linear structure of submodel based on the material balance equation of extraction separation process; Secondly the sample data of gathering is carried out cluster analysis, obtain the steady-state operation point and the corresponding data set of extraction process, set up partial model according to the model structure of determining, by the initial parameter of each submodel of least square method of recursion identification; The model of selecting suitable multi-model scheduling mechanism (model switches or the model weighting) to get is to the end once more exported, and can realize the On-line Estimation of rare earth extraction process component concentration.
(2) on the component concentration On-line Estimation model based of above-mentioned foundation, in conjunction with generalized predictive control, the polyteny model is changed, obtain the relational expression of component concentration and extractant or acid solution flow control amount, deviation by On-line Estimation value and setting value, calculate the controlled quentity controlled variable of extractant or acid solution, realize the on-line monitoring and the robotization control of rare earth extraction process component concentration.
3. rare earth extraction process component concentration multi-model forecast Control Algorithm according to claim 2, it is characterized in that, the linear structure of described definite submodel, dynamic perfromance according to the rare earth extraction process is divided into several extraction sections with extraction process, each section is abstracted into a virtual extracting stage, utilizes segmentation to assemble modeling model is simplified; And nonlinear system launched in a plurality of working points, utilize the local linearization model on a plurality of working points to be similar to nonlinear system, can set up the object dynamic model that more approaches real system.
4. rare earth extraction process component concentration multi-model forecast Control Algorithm according to claim 2, it is characterized in that, described sample data to collection is carried out cluster analysis, adopt the subtractive clustering method that sample data is carried out cluster analysis, the clusters number n that obtains is the number of submodel in the multi-model, and cluster centre is the working point of determining extraction process.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201010555634 CN102004444A (en) | 2010-11-23 | 2010-11-23 | Multi-model predictive control method for component content in process of extracting rare earth |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201010555634 CN102004444A (en) | 2010-11-23 | 2010-11-23 | Multi-model predictive control method for component content in process of extracting rare earth |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102004444A true CN102004444A (en) | 2011-04-06 |
Family
ID=43811870
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 201010555634 Pending CN102004444A (en) | 2010-11-23 | 2010-11-23 | Multi-model predictive control method for component content in process of extracting rare earth |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102004444A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102841540A (en) * | 2012-09-10 | 2012-12-26 | 广东电网公司电力科学研究院 | MMPC-based supercritical unit coordination and control method |
CN103744292A (en) * | 2014-01-22 | 2014-04-23 | 东北大学 | Rare earth extraction process dynamic operating control method based on data driving |
CN104090491A (en) * | 2014-07-04 | 2014-10-08 | 东南大学 | Gas steam combined cycle unit multivariable constrained prediction function load control method |
CN105425593A (en) * | 2016-01-22 | 2016-03-23 | 东华理工大学 | Multi-model smooth and stable switching control method for increase and decrease of state variables |
CN108490779A (en) * | 2018-03-20 | 2018-09-04 | 华东交通大学 | A kind of Rare-Earth Extraction Process decoupling control method |
CN109839825A (en) * | 2019-01-28 | 2019-06-04 | 华东交通大学 | A kind of forecast Control Algorithm and system of Rare-Earth Extraction Process constituent content |
CN111624954A (en) * | 2019-03-12 | 2020-09-04 | 妙微(杭州)科技有限公司 | Advanced process control machine learning framework for flow type manufacturing industry |
CN113126501A (en) * | 2021-04-23 | 2021-07-16 | 华东交通大学 | Rare earth extraction process prediction control method and system |
CN113359512A (en) * | 2021-06-26 | 2021-09-07 | 华东交通大学 | Component content digital twinning characteristic analysis method in rare earth extraction separation process |
CN113377072A (en) * | 2021-06-24 | 2021-09-10 | 华东交通大学 | Rare earth extraction process drug dosage optimization setting method based on instant learning |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1670646A (en) * | 2005-04-14 | 2005-09-21 | 东北大学 | Rare earth cascade extraction separation intelligence control system and method |
-
2010
- 2010-11-23 CN CN 201010555634 patent/CN102004444A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1670646A (en) * | 2005-04-14 | 2005-09-21 | 东北大学 | Rare earth cascade extraction separation intelligence control system and method |
Non-Patent Citations (1)
Title |
---|
《Proceedings of the 8th World Congress on Intelligent Control and Automation July 6-9 2010, Jinan, China》 20100709 Hui yang等 The Multiple Models Predictive Control of Component 5836-5841 1-4 , 2 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102841540A (en) * | 2012-09-10 | 2012-12-26 | 广东电网公司电力科学研究院 | MMPC-based supercritical unit coordination and control method |
CN103744292A (en) * | 2014-01-22 | 2014-04-23 | 东北大学 | Rare earth extraction process dynamic operating control method based on data driving |
CN103744292B (en) * | 2014-01-22 | 2016-08-31 | 东北大学 | A kind of Rare-Earth Extraction Process dynamic operation control method based on data-driven |
CN104090491B (en) * | 2014-07-04 | 2017-02-01 | 东南大学 | Gas steam combined cycle unit multivariable constrained prediction function load control method |
CN104090491A (en) * | 2014-07-04 | 2014-10-08 | 东南大学 | Gas steam combined cycle unit multivariable constrained prediction function load control method |
CN105425593B (en) * | 2016-01-22 | 2018-05-08 | 东华理工大学 | The multi-model smooth steady method for handover control of state variable increase and decrease |
CN105425593A (en) * | 2016-01-22 | 2016-03-23 | 东华理工大学 | Multi-model smooth and stable switching control method for increase and decrease of state variables |
CN108490779A (en) * | 2018-03-20 | 2018-09-04 | 华东交通大学 | A kind of Rare-Earth Extraction Process decoupling control method |
CN108490779B (en) * | 2018-03-20 | 2021-01-08 | 华东交通大学 | Rare earth extraction process decoupling control method |
CN109839825A (en) * | 2019-01-28 | 2019-06-04 | 华东交通大学 | A kind of forecast Control Algorithm and system of Rare-Earth Extraction Process constituent content |
CN109839825B (en) * | 2019-01-28 | 2022-04-12 | 华东交通大学 | Method and system for predictive control of component content in rare earth extraction process |
CN111624954A (en) * | 2019-03-12 | 2020-09-04 | 妙微(杭州)科技有限公司 | Advanced process control machine learning framework for flow type manufacturing industry |
CN113126501A (en) * | 2021-04-23 | 2021-07-16 | 华东交通大学 | Rare earth extraction process prediction control method and system |
CN113377072A (en) * | 2021-06-24 | 2021-09-10 | 华东交通大学 | Rare earth extraction process drug dosage optimization setting method based on instant learning |
CN113359512A (en) * | 2021-06-26 | 2021-09-07 | 华东交通大学 | Component content digital twinning characteristic analysis method in rare earth extraction separation process |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102004444A (en) | Multi-model predictive control method for component content in process of extracting rare earth | |
CN102831269B (en) | Method for determining technological parameters in flow industrial process | |
CN101587328B (en) | Device for identifying non-linear process dynamic model | |
CN108920863B (en) | Method for establishing energy consumption estimation model of robot servo system | |
CN108241901B (en) | Transformer early warning evaluation method and device based on prediction data | |
CN102693452A (en) | Multiple-model soft-measuring method based on semi-supervised regression learning | |
CN104914723A (en) | Industrial process soft measurement modeling method based on cooperative training partial least squares model | |
CN105425583B (en) | The control method of penicillin production process based on coorinated training LWPLS | |
CN106778838A (en) | A kind of method for predicting air quality | |
CN107403196A (en) | Instant learning modeling method based on spectral clustering analysis | |
CN105260798A (en) | Big data miner for multi-dimensional load characteristic analysis | |
CN113130014B (en) | Rare earth extraction simulation method and system based on multi-branch neural network | |
CN114169254A (en) | Abnormal energy consumption diagnosis method and system based on short-term building energy consumption prediction model | |
CN102117365A (en) | On-line modeling and optimizing method suitable for recovering coking coarse benzene | |
CN101916394A (en) | Knowledge fusion-based online soft measurement method | |
CN114117852B (en) | Regional heat load rolling prediction method based on finite difference working domain division | |
CN102129242A (en) | Product quality control method during batch processing production process based on two-layer hybrid intelligent optimization | |
CN101446828A (en) | Nonlinear process quality prediction method | |
Ding et al. | Short-term forecasting of building cooling load based on data integrity judgment and feature transfer | |
CN1327376C (en) | Soft measuring meter moduling method based on supporting vector machine | |
CN113064401A (en) | Closed loop system micro fault detection and estimation method based on data driving | |
CN110163437B (en) | Day-ahead photovoltaic power generation power prediction method based on DPK-means | |
CN109210268B (en) | Big data processing method based on ultralow-power electromagnetic valve | |
CN106548199A (en) | A kind of large user's load characteristics clustering method based on demand response | |
CN106548188A (en) | A kind of nickel valency Forecasting Methodology based on sample optimizing support vector machine |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110406 |