CN103675011B

CN103675011B - The industrial melt index soft measurement instrument of optimum support vector machine and method

Info

Publication number: CN103675011B
Application number: CN201310435358.2A
Authority: CN
Inventors: 刘兴高; 张明明; 李见会
Original assignee: Zhejiang University ZJU
Current assignee: Zhejiang University ZJU
Priority date: 2013-09-22
Filing date: 2013-09-22
Publication date: 2015-09-30
Anticipated expiration: 2033-09-22
Also published as: CN103675011A

Abstract

The industrial melt index soft that the invention discloses a kind of optimum support vector machine measures soft measuring instrument and method.This flexible measurement method carries out Fuzzy processing to the output of multiple Weighted Least Squares Support Vector Machines, and adopts particle cluster algorithm to be optimized whole fuzzifying equation system, to obtain optimum hard measurement result.In the present invention, for measure easily survey variable field intelligent instrument, be connected with DCS database for the control station measuring performance variable, hard measurement value display instrument comprises the industrial melt index soft measurement model of optimum support vector machine, DCS database is connected with the input end of soft-sensing model, and the output terminal of the industrial melt index soft measurement model of described optimum support vector machine is connected with melt index flexible measured value display instrument; The present invention has on-line optimization parameter, model upgrades automatically, noise resisting ability strong, promote the good feature of performance.

Description

Industrial melt index soft measuring instrument and method of optimal support vector machine

Technical Field

The invention relates to a soft measuring instrument and a method, in particular to an industrial melt index soft measuring instrument and a method of an optimal support vector machine.

Background

Polypropylene is a semi-crystalline thermoplastic polymerized from propylene, has high impact resistance, strong mechanical properties, resistance to various organic solvents and acid and alkali corrosion, is widely applied in the industry, and is one of the most common high polymer materials. The Melt Index (MI) is one of the important quality indicators in polypropylene production that determines the grade of the final product, and it determines the different uses of the product. The accurate and timely measurement of the melt index plays an important role and a very important guiding significance for production and scientific research. However, the online analysis and measurement of the melt index are still difficult to achieve at present, and the lack of an online analyzer of the melt index is a major problem which limits the quality of polypropylene products. MI can be obtained only by manual sampling and offline assay analysis, and is generally analyzed once every 2-4 hours, so that the time lag is large, and the requirement of real-time production control is difficult to meet.

Most of the recent research on online prediction of MI has focused on artificial neural networks, which has achieved good results. However, artificial neural networks have their own drawbacks, such as overfitting, the number of nodes in the hidden layer, and poor parameter determination. Secondly, noise, manual operation errors and the like of DCS data acquired in an industrial field have certain uncertain errors, so that a forecasting model using an artificial neural network with strong certainty is not strong in popularization capability generally.

Zadeh first proposed the concept of fuzzy aggregation in 1965 by american mathematician l. Fuzzy logic then begins to replace the classical logic that persists in that everything can be represented in terms of binary terms in a way that it more closely resembles the question and semantic statement of everyday people. Fuzzy logic has been successfully applied in various fields of industry, such as home appliances, industrial control, etc. In 2003, Demirci proposed the concept of fuzzy equations by constructing a new input matrix using fuzzy membership matrices and their variants, and then deriving the analytic values as final outputs in local equations by the centroid method in the inverse fuzzy method. For soft measurement of melt index in propylene polymerization production, considering noise effects and operational errors in industrial production, the effect of fuzzy performance of fuzzy logic on overall prediction accuracy can be reduced.

Support vector machines, introduced by Vapnik in 1998, are widely used in pattern recognition, fitting and classification problems due to their good generalization ability. Since the standard support vector machine is sensitive to isolated points and noise points, a weighted least squares support vector machine was proposed later. The weighted least squares support vector machine is better able to process noisy sample data than the standard support vector machine, and is chosen here as the local equation in the fuzzy equation.

Particle Swarm Optimization, namely Particle Swarm Optimization, is a biological intelligent Optimization algorithm proposed by professors Kennedy and Eberhart to seek global optimality by simulating bird flight behavior, called PSO for short. The algorithm reduces the risk that the search algorithm is trapped in the local optimal solution through the mutual influence among particles in the group, and has good global search performance. The particle swarm algorithm is used for searching the optimal parameter combination of the weighted least square support vector machine so as to achieve the aim of optimizing the model.

Disclosure of Invention

In order to overcome the defects of low measurement precision, low noise sensitivity and poor popularization performance in the conventional propylene polymerization production process, the invention provides the industrial melt index soft measurement instrument and method of the optimal support vector machine, which have the advantages of online measurement, high calculation speed, automatic model updating, strong noise resistance and good popularization performance.

An industrial melt index soft measuring instrument of an optimal support vector machine comprises a propylene polymerization production process, a field intelligent instrument for measuring easily-measured variables, a control station for measuring operation variables, a DCS database for storing data and a melt index soft measurement value display instrument, the on-site intelligent instrument and the control station are connected with the propylene polymerization production process, the on-site intelligent instrument and the control station are connected with the DCS database, the soft measuring instrument also comprises an industrial melt index soft measuring model of an optimal support vector machine, the DCS database is connected with the input end of the industrial melt index soft measuring model of the optimal support vector machine, the output end of the industrial melt index soft measurement model of the optimal support vector machine is connected with a melt index soft measurement value display instrument, and the industrial melt index soft measurement model of the optimal support vector machine comprises:

the data preprocessing module is used for preprocessing the model training samples input from the DCS database, centralizing the training samples, namely subtracting the average value of the samples, and then normalizing the training samples:

calculating an average value:

<math> <mrow> <mover> <mi>TX</mi> <mo>&OverBar;</mo> </mover> <mo>=</mo> <mfrac> <mn>1</mn> <mi>N</mi> </mfrac> <munderover> <mi>Σ</mi> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>TX</mi> <mi>i</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </math>

calculating the variance:

<math> <mrow> <msubsup> <mi>σ</mi> <mi>x</mi> <mn>2</mn> </msubsup> <mo>=</mo> <mfrac> <mn>1</mn> <mrow> <mi>N</mi> <mo>-</mo> <mn>1</mn> </mrow> </mfrac> <munderover> <mi>Σ</mi> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <mrow> <mo>(</mo> <msub> <mi>TX</mi> <mi>i</mi> </msub> <mo>-</mo> <mover> <mi>TX</mi> <mo>&OverBar;</mo> </mover> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow> </math>

and (3) standardization:

<math> <mrow> <mi>X</mi> <mo>=</mo> <mfrac> <mrow> <mi>TX</mi> <mo>-</mo> <mover> <mi>TX</mi> <mo>&OverBar;</mo> </mover> </mrow> <msub> <mi>σ</mi> <mi>x</mi> </msub> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>3</mn> <mo>)</mo> </mrow> </mrow> </math>

wherein, TX_iIs the ith training sample, N is the number of training samples,is the mean of the training samples, and X is the normalized training sample. Sigma_xRepresenting the standard deviation, σ, of the training samples² _xRepresenting the variance of the training samples.

And the fuzzy equation module is used for fuzzifying the standardized training sample X transmitted from the data preprocessing module. Let there be c in the system of fuzzy equations^*A fuzzy group, fuzzyThe centers of groups k and j are v_k、v_jThen the normalized ith training sample X_iMembership mu for fuzzy group k_ikComprises the following steps:

in the formula, n is a blocking matrix index required in the fuzzy classification process, and is usually taken as 2, | | · |, as a norm expression.

Using the above membership values or its variants to obtain a new input matrix, for the fuzzy group k, its input matrix variant is:

Φ_ik(X_i,μ_ik)=[1func(μ_ik)X_i] （5）

wherein func (. mu.)_ik) Is a membership value mu_ikDeformation function of, in general, takeexp(μ_ik) Equal, phi_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikThe corresponding new input matrix.

And the weighted least square support vector machine is used as a local equation of the fuzzy equation system, and optimal fitting is carried out on each fuzzy group. Let the ith target output of the model training sample be O_iThe weighted support vector machine equates the fitting problem to the following quadratic programming problem by transformation:

where R (w, xi) is the objective function of the optimization problem, minR (w, xi) is the minimum of the objective function of the optimization problem, N is the number of training samples, xi = { xi =₁,…,ξ_NIs the relaxation variable, ξ_iIs the ith component of the relaxation variable, w is the normal vector to the hyperplane of the support vector machine, b is the corresponding offset, and ω is_iI =1, …, N and γ being the weight and penalty factors, respectively, of a weighted least squares support vector machine,is the ith component xi of the relaxation variable of the weighted least squares support vector machine_iEstimation of the standard deviation, c₁Is a constant, here take 2.5, c₂Is a constant, here taken as 3, the superscript T denotes the transpose of the matrix, μ_ikRepresents the normalized ith training sample X_iMembership, Φ, for fuzzy group k_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikThe corresponding new input matrix. From equations (6), (7) and (8), the output of the fuzzy group k in the training sample i is derived as:

wherein,is the output of the fuzzy group k at the training sample i. K<·>Is the kernel function of a weighted least squares support vector machine, where K<·>Taking a linear kernel function; alpha is alpha_mM =1, …, N is the mth component of the corresponding lagrange multiplier. Mu.s_mkRepresents the m-th training sample X_mMembership, Φ, for fuzzy group k_mk(X_m,μ_mk) Denotes the m-th input variable X_mAnd membership mu of its fuzzy group k_mkThe corresponding new input matrix.

The output of the final fuzzy equation system is obtained by the gravity center method in the anti-fuzzy method:

wherein,is the output of the fuzzy group k at the training sample i.

The particle swarm optimization module is used for optimizing a penalty factor and an error tolerance value of a local equation of a weighted least square support vector machine in a fuzzy equation by adopting a particle swarm optimization, and the specific implementation steps are as follows:

determining optimization parameters of the particle swarm as a penalty factor and an error tolerance value of a local equation of a weighted least square support vector machine, the individual number of the particle swarm popsize, and the maximum cyclic optimization number iter_maxInitial position r of the p-th particle_pInitial velocity v_pLocal optimum value Lbest_pAnd a global optimum Gbest for the entire population of particles.

Setting an optimization objective function, converting the optimization objective function into fitness, and evaluating each local fuzzy equation; calculating a fitness function through a corresponding error function, considering that the fitness of the particle with large error is small, and expressing the fitness function of the particle p as follows:

f_p=1/(E_p+1) （11）

in the formula, E_pIs the error function of the fuzzy equation system, expressed as:

in the formula,is the predicted output of a system of fuzzy equations, O_iIs the target output of the fuzzy equation system;

ninthly, circularly updating the speed and the position of each particle according to the following formula,

v_p(iter+1)=ω×v_p(iter)+m₁a₁(Lbest_p-r_p(iter))+m₂a₂(Gbest-r_p(iter))

（13）

r_p(iter+1)=r_p(iter)+v_p(iter+1) （14）

in the formula, v_pIndicates the velocity, r, of the update particle p_pIndicating the position of the update particle p, Lbest_pRepresenting the individual optimal value of the updated particle p, Gbest being the global optimal value of the whole particle swarm, iter representing the cycle number, omega being the inertial weight in the particle swarm algorithm, m₁、m₂Is the corresponding acceleration factor, a₁、a₂Is [0,1 ]]A random number in between;

and (c) for the particle p, if the new fitness is larger than the original individual optimal value, updating the individual optimal value of the particle:

Lbest_p=f_p （15）

if the individual optimum value Lbest of particle p_pAnd (3) updating the original particle swarm global optimum value Gbest if the particle swarm global optimum value Gbest is larger than the original particle swarm global optimum value Gbest:

Gbest=Lbest_p （16）

judging whether the performance requirements are met, if so, finishing the optimization to obtain a group of optimized local equation parameters of the fuzzy equation; otherwise, returning to the step III, continuing the iteration optimization until reaching the maximum iteration number iter_max。

Preferably, the industrial melt index soft measurement model of the optimal support vector machine further includes: and the model updating module is used for updating the model on line, inputting offline experimental data into a training set regularly and updating the fuzzy equation model.

An industrial melt index soft measurement method of an optimal support vector machine comprises the following specific implementation steps:

1) selecting an operation variable and an easily-measured variable as input of a model for a propylene polymerization production process object according to process analysis and operation analysis, wherein the operation variable and the easily-measured variable are obtained by a DCS (distributed control system) database;

2) preprocessing a model training sample input from a DCS database, centralizing the training sample, namely subtracting the average value of the sample, and then normalizing the training sample so that the average value is 0 and the variance is 1. The processing is accomplished using the following mathematical process:

2.1) calculating the mean value:

2.2) calculating the variance:

2.3) standardization:

3) And fuzzifying the training sample transmitted from the data preprocessing module. Let there be c in the system of fuzzy equations^*A fuzzy group, the centers of the fuzzy groups k and j are v_k、v_jThen the normalized ith training sample X_iMembership mu for fuzzy group k_ikComprises the following steps:

Φ_ik(X_i,μ_ik)=[1func(μ_ik)X_i] （5）

where R (w, ξ) is the objective function of the optimization problem, minR (w, ξ) is the minimum of the objective function of the optimization problem, and N is the trainingNumber of samples, ξ = { ξ =₁,…,ξ_NIs the relaxation variable, ξ_iIs the ith component of the relaxation variable, w is the normal vector to the hyperplane of the support vector machine, b is the corresponding offset, and ω is_iI =1, …, N and γ being the weight and penalty factors, respectively, of a weighted least squares support vector machine,is the ith component xi of the relaxation variable of the weighted least squares support vector machine_iEstimation of the standard deviation, c₁Is a constant, here take 2.5, c₂Is a constant, here taken as 3, the superscript T denotes the transpose of the matrix, μ_ikRepresents the normalized ith training sample X_iMembership, Φ, for fuzzy group k_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikThe corresponding new input matrix. From equations (6), (7) and (8), the output of the fuzzy group k in the training sample i is derived as:

wherein,is the output of the fuzzy group k at the training sample i.

4) Optimizing a penalty factor and an error tolerance value of a local equation of a weighted least square support vector machine in a fuzzy equation by adopting a particle swarm optimization, and specifically realizing the following steps:

f_p=1/(E_p+1) （11）

v_p(iter+1)=ω×v_p(iter)+m₁a₁(Lbest_p-r_p(iter))+m₂a₂(Gbest-r_p(iter))

（13）

r_p(iter+1)=r_p(iter)+v_p(iter+1) （14）

Lbest_p=f_p （15）

Gbest=Lbest_p （16）

As a preferred solution: the soft measurement method further comprises the following steps: 5) and inputting the offline experimental data into a training set regularly, and updating the fuzzy equation model.

The technical conception of the invention is as follows: the melt index of important quality index in propylene polymerization production process is subjected to online soft measurement, the defects of low measurement precision, low noise sensitivity and poor popularization performance of the existing polypropylene melt index measuring instrument are overcome, the particle swarm optimization is introduced to automatically optimize a fuzzy equation model, and the parameters of the local equation of the weighted least square support vector machine in the fuzzy equation are not required to be adjusted for many times through artificial experience. Compared with the existing melt index soft measurement model, the model has the following advantages: (1) the influence of noise and manual operation errors on the model forecasting precision is reduced; (2) the popularization performance of the model is enhanced, and overfitting is effectively inhibited; (3) the parameters of the model are automatically optimized, the stability of the model is improved, and the possibility that the model falls into local optimization is reduced.

The invention has the following beneficial effects: 1. online measurement; 2. automatically optimizing online parameters; 3. the model is automatically updated; 4. the anti-noise interference capability is strong, 5 and the precision is high; 6. the popularization capability is strong.

Drawings

FIG. 1 is a schematic diagram of the basic structure of an industrial melt index soft measuring instrument and method of an optimal support vector machine;

FIG. 2 is a schematic diagram of an industrial melt index soft measurement model structure of an optimal support vector machine.

Detailed Description

The invention is further described below with reference to the accompanying drawings. The examples are intended to illustrate the invention, but not to limit the invention, and any modifications and variations of the invention within the spirit and scope of the claims are intended to fall within the scope of the invention.

Example 1

Referring to fig. 1 and fig. 2, an industrial melt index soft measurement instrument of an optimal support vector machine comprises a propylene polymerization production process 1, an on-site intelligent instrument 2 for measuring easily-measured variables, a control station 3 for measuring operation variables, a DCS database 4 for storing data and a melt index soft measurement value display 6, wherein the on-site intelligent instrument 2 and the control station 3 are connected with the propylene polymerization production process 1, the on-site intelligent instrument 2 and the control station 3 are connected with the DCS database 4, the soft measurement instrument further comprises a soft measurement model 5 of a particle swarm optimization weighted least square support vector machine fuzzy equation, the DCS database 4 is connected with an input end of an industrial melt index soft measurement model 5 of the optimal support vector machine, an output end of the industrial melt index soft measurement model 5 of the optimal support vector machine is connected with the melt index soft measurement value display 6, the industrial melt index soft measurement model of the optimal support vector machine comprises:

calculating an average value:

calculating the variance:

and (3) standardization:

And the fuzzy equation module is used for fuzzifying the standardized training sample X transmitted from the data preprocessing module. Let there be c in the system of fuzzy equations^*A fuzzy group, the centers of the fuzzy groups k and j are v_k、v_jThen the normalized ith training sample X_iMembership mu for fuzzy group k_ikComprises the following steps:

Φ_ik(X_i,μ_ik)=[1func(μ_ik)X_i] （5）

Weighted least squares support vector machine as a fuzzyAnd (4) performing optimization fitting on each fuzzy group by using a local equation of the equation system. Let the ith target output of the model training sample be O_iThe weighted support vector machine equates the fitting problem to the following quadratic programming problem by transformation:

wherein,is the output of the fuzzy group k at the training sample i. K<·>Is the kernel function of a weighted least squares support vector machine, where K<·>Taking a linear kernel function; alpha is alpha_mM =1, …, N being the corresponding lagrange multiplier, μ_mkRepresents the m-th training sample X_mMembership, Φ, for fuzzy group k_mk(X_m,μ_mk) Denotes the m-th input variable X_mAnd membership mu of its fuzzy group k_mkThe corresponding new input matrix.

wherein,is the output of the fuzzy group k at the training sample i.

firstly, determining optimization parameters of a particle swarm as a penalty factor and an error tolerance value of a local equation of a weighted least square support vector machine, the individual number of the particle swarm popsize and the maximum cyclic optimization number iter_maxInitial position r of the p-th particle_pInitial velocity v_pLocal optimum value Lbest_pAnd a global optimum Gbest for the entire population of particles.

Secondly, setting an optimization objective function, converting the optimization objective function into fitness, and evaluating each local fuzzy equation; calculating a fitness function through a corresponding error function, considering that the fitness of the particle with large error is small, and expressing the fitness function of the particle p as follows:

f_p=1/(E_p+1) （11）

thirdly, according to the following formula, the speed and the position of each particle are circularly updated,

v_p(iter+1)=ω×v_p(iter)+m₁a₁(Lbest_p-r_p(iter))+m₂a₂(Gbest-r_p(iter))

（13）

r_p(iter+1)=r_p(iter)+v_p(iter+1) （14）

and fourthly, for the particle p, if the new fitness is larger than the original individual optimal value, updating the individual optimal value of the particle:

Lbest_p=f_p （15）

if the individual optimum value Lbest of the particle p_pGreater than original particle swarm global optimumAnd updating the original particle swarm global optimum value Gbest by the value Gbest:

Gbest=Lbest_p （16）

sixthly, judging whether the performance requirement is met, if so, finishing the optimization to obtain a set of optimized local equation parameters of the fuzzy equation; otherwise, returning to the step III, continuing the iteration optimization until reaching the maximum iteration number iter_max。

Preferably, the industrial melt index soft measurement model of the optimal support vector machine further includes: and the model updating module is used for updating the model on line, inputting offline experimental data into a training set regularly and updating the fuzzy equation system model.

According to the reaction mechanism and the process analysis, in consideration of various factors influencing the melt index in the production process of polypropylene, nine commonly used operation variables and easily-measured variables in the actual production process are taken as modeling variables, including: three propylene feed flow rates, main catalyst flow rate, auxiliary catalyst flow rate, temperature, pressure, liquid level in the kettle, and hydrogen volume concentration in the kettle. Table 1 lists 9 modeling variables input as the soft measurement model 5, which are the temperature in the kettle (T), the pressure in the kettle (p), the liquid level in the kettle (L), and the volume concentration of hydrogen in the kettle (X)_v) 3 propylene feed flow rates (first propylene feed flow rate f1, second propylene feed flow rate f2, third propylene feed flow rate f 3), 2 catalyst feed flow rates (main catalyst flow rate f4, cocatalyst flow rate f 5). The polymerization reaction in the reaction kettle is carried out after reaction materials are repeatedly mixed, so that the process variable of the model input variable related to the materials adopts the average value of a plurality of previous moments. The data in this example were averaged over the previous hour. The melt index off-line assay value is used as an output variable of the soft measurement model 5. The test sample is obtained by manual sampling and offline assay analysis, and is analyzed and collected every 4 hours.

The on-site intelligent instrument 2 and the control station 3 are connected with the propylene polymerization production process 1 and the DCS database 4; the soft measurement model 5 is connected with the DCS database and the soft measurement value display instrument 6. The on-site intelligent instrument 2 measures the easily-measured variable of the propylene polymerization production object and transmits the easily-measured variable to the DCS database 4; the control station 3 controls manipulated variables of the propylene polymerization production target, and transmits the manipulated variables to the DCS database 4. The variable data recorded in the DCS database 4 is used as the input of the industrial melt index soft measurement model 5 of the optimal support vector machine, and the soft measurement value display instrument 6 is used for displaying the output, namely the soft measurement value, of the industrial melt index soft measurement model 5 of the optimal support vector machine.

Table 1: modeling variable required by industrial melt index soft measurement model of optimal support vector machine

Variable sign	Meaning of variables	Variable sign	Meaning of variables
				T	Temperature in the kettle	f1	First propylene feed flow rate
p	Pressure intensity in kettle	f2	Second propylene feed flow rate

L	Liquid level in the kettle	f3	Third propylene feed flow rate
				X_v	Volume concentration of hydrogen in the autoclave	f4	Main catalyst flow rate
		f5	Flow rate of cocatalyst

The industrial melt index soft measurement model 5 of the optimal support vector machine comprises the following 4 parts:

a data preprocessing module 7, configured to preprocess the model training samples input from the DCS database, centralize the training samples, that is, subtract an average value of the samples, and then normalize them:

calculating an average value:

calculating the variance:

and (3) standardization:

wherein, TX_iIs the ith training sample, N is the number of training samples,is the mean of the training samples, and X is the normalized training sample. Sigma_xRepresenting the standard deviation, σ, of the training samples² _xRepresenting training samplesThe variance of (c).

And the fuzzy equation module 8 is used for fuzzifying the standardized training sample X transmitted from the data preprocessing module. Let there be c in the system of fuzzy equations^*A fuzzy group, the centers of the fuzzy groups k and j are v_k、v_jThen the normalized ith training sample X_iMembership mu for fuzzy group k_ikComprises the following steps:

Φ_ik(X_i,μ_ik)=[1func(μ_ik)X_i] （5）

wherein,is the output of the fuzzy group k at the training sample i.

The particle swarm optimization module 9 is configured to optimize a penalty factor and an error tolerance value of a local equation of a weighted least squares support vector machine in a fuzzy equation by using a particle swarm optimization, and specifically includes the following steps:

f_p=1/(E_p+1) （11）

v_p(iter+1)=ω×v_p(iter)+m₁a₁(Lbest_p-r_p(iter))+m₂a₂(Gbest-r_p(iter))

（13）

r_p(iter+1)=r_p(iter)+v_p(iter+1) （14）

Lbest_p=f_p （15）

if the individual optimum value Lbest of the particle p_pAnd (3) updating the original particle swarm global optimum value Gbest if the particle swarm global optimum value Gbest is larger than the original particle swarm global optimum value Gbest:

Gbest=Lbest_p （16）

And the model updating module 10 is used for updating the model on line, inputting offline experimental data into a training set regularly and updating the fuzzy equation model.

Example 2

Referring to fig. 1 and 2, a particle swarm optimization-based method for soft measurement of melt index in industrial polypropylene production by using a fuzzy equation model of a weighted least squares support vector machine is specifically implemented as follows:

2.1) calculating the mean value:

2.2) calculating the variance:

2.3) standardization:

Φ_ik(X_i,μ_ik)=[1func(μ_ik)Xi] （5）

where R (w, ξ) is the purpose of the optimization problemThe objective function, minR (w, xi) is the minimum value of the objective function of the optimization problem, N is the number of training samples, xi = { xi =₁,…,ξ_NIs the relaxation variable, ξ_iIs the ith component of the relaxation variable, w is the normal vector to the hyperplane of the support vector machine, b is the corresponding offset, and ω is_iI =1, …, N and γ being the weight and penalty factors, respectively, of a weighted least squares support vector machine,is the ith component xi of the relaxation variable of the weighted least squares support vector machine_iEstimation of the standard deviation, c₁Is a constant, here take 2.5, c₂Is a constant, here taken as 3, the superscript T denotes the transpose of the matrix, μ_ikRepresents the normalized ith training sample X_iMembership, Φ, for fuzzy group k_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikThe corresponding new input matrix. From equations (6), (7) and (8), the output of the fuzzy group k in the training sample i is derived as:

wherein,is the output of the fuzzy group k at the training sample i.

f_p=1/(E_p+1) （11）

v_p(iter+1)=ω×v_p(iter)+m₁a₁(Lbest_p-r_p(iter))+m₂a₂(Gbest-r_p(iter))

（13）

r_p(iter+1)=r_p(iter)+v_p(iter+1) （14）

Lbest_p=f_p （15）

Gbest=Lbest_p （16）

As a preferred solution: the soft measurement method further comprises the following steps: 4) and inputting the offline experimental data into a training set regularly, and updating the fuzzy equation model.

The method of the embodiment comprises the following specific implementation steps:

step 1: for the propylene polymerization production process object 1, the manipulated variables and easily measurable variables are selected as the inputs of the model according to the process analysis and the operational analysis. The manipulated variables and easily measurable variables are obtained from the DCS database 4.

Step 2: and sample data is preprocessed and completed by a data preprocessing module 7.

And step 3: and establishing an initial fuzzy equation model 8 based on model training sample data. Input data is obtained as described in step 2 and output data is obtained from an off-line assay.

And 4, step 4: the local weighted least squares support vector machine equation parameters of the initial fuzzy equation model 8 are optimized by the particle swarm algorithm 9.

And 5: the model updating module 10 periodically inputs offline experimental data into a training set, updates the fuzzy equation model, and completes the establishment of the soft measurement model 5 of the fuzzy equation model of the particle swarm optimization-based weighted least square support vector machine.

Step 6: the melt index soft measurement value display instrument 6 displays the output of the industrial melt index soft measurement model 5 of the optimal support vector machine, and the melt index soft measurement display of the industrial polypropylene production is completed.

Claims

1. The utility model provides an industry melt index soft measurement instrument of best support vector machine, includes the field intelligent instrument that is used for measuring easy measurability variable, is used for measuring the control station of operating variable, deposits the DCS database of data and the soft measurement value display appearance of melt index, field intelligent instrument, control station and DCS database connection, its characterized in that: the soft measuring instrument further comprises an industrial melt index soft measuring model of an optimal support vector machine, the DCS database is connected with the input end of the industrial melt index soft measuring model of the optimal support vector machine, the output end of the industrial melt index soft measuring model of the optimal support vector machine is connected with a melt index soft measuring value display instrument, and the industrial melt index soft measuring model of the optimal support vector machine comprises:

calculating an average value:

calculating the variance:

and (3) standardization:

wherein, TX_iIs the ith training sample, N is the number of training samples,is the mean value of the training sample, and X is the training sample after standardization; sigma_xRepresenting the standard deviation, σ, of the training samples² _xRepresenting the variance of the training samples;

the fuzzy equation module is used for fuzzifying the standardized training sample X transmitted from the data preprocessing module; let there be c in the system of fuzzy equations^*A fuzzy group, the centers of the fuzzy groups k and j are v_k、v_jThen the normalized ith training sample X_iMembership mu for fuzzy group k_ikComprises the following steps:

in the formula, n is a blocking matrix index required in the fuzzy classification process, and is taken as 2, | | · | | is a norm expression;

Φ_ik(X_i,μ_ik)＝[1 func(μ_ik) X_i] (5)

wherein func (. mu.)_ik) Is a membership value mu_ikIs taken as a deformation function ofexp(μ_ik)，Φ_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikThe corresponding new input matrix;

the weighted least square support vector machine is used as a local equation of a fuzzy equation system, and optimal fitting is carried out on each fuzzy group; let the ith target output of the model training sample be O_iThe weighted support vector machine equates the fitting problem to the following quadratic programming problem by transformation:

where R (w, ξ) is the objective function of the optimization problem, minR (w, ξ) is the minimum of the objective function of the optimization problem, N is the number of training samples, ξ ═ ξ₁,...,ξ_NIs the relaxation variable, ξ_iIs the ith component of the relaxation variable, w is the normal vector to the hyperplane of the support vector machine, b is the corresponding offset, and ω is_iN and y are the weight and penalty factors of a weighted least squares support vector machine, respectively,is the ith component xi of the relaxation variable of the weighted least squares support vector machine_iEstimation of the standard deviation, c₁Is a constant, here take 2.5, c₂Is constant, hereTake 3, superscript T denotes the transpose of the matrix, μ_ikRepresents the normalized ith training sample X_iMembership, Φ, for fuzzy group k_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikThe corresponding new input matrix; from equations (6), (7) and (8), the output of the fuzzy group k in the training sample i is derived as:

wherein,the output of the fuzzy group k in the training sample i; k<·>Is the kernel function of a weighted least squares support vector machine, where K<·>Taking a linear kernel function; alpha is alpha_mM is 1, …, N is the mth component of the corresponding lagrange multiplier; mu.s_mkTo representThe mth training sample X_mMembership, Φ, for fuzzy group k_mk(X_m,μ_mk) Denotes the m-th input variable X_mAnd membership mu of its fuzzy group k_mkThe corresponding new input matrix;

wherein,the output of the fuzzy group k in the training sample i;

firstly, determining optimization parameters of a particle swarm as a penalty factor and an error tolerance value of a local equation of a weighted least square support vector machine, the individual number of the particle swarm popsize and the maximum cyclic optimization number iter_maxInitial position r of the p-th particle_pInitial velocity v_pLocal optimum value Lbest_pAnd the global optimum value Gbest of the whole particle swarm;

f_p＝1/(E_p+1) (11)

v_p(iter+1)＝ω×v_p(iter)+m₁a₁(Lbest_p-r_p(iter))+m₂a₂(Gbest-r_p(iter))

(13)

r_p(iter+1)＝r_p(iter)+v_p(iter+1) (14)

Lbest_p＝f_p (15)

Gbest＝Lbest_p (16)

sixthly, judging whether the performance requirement is met, if so, finishing the optimization to obtain a set of optimized local equation parameters of the fuzzy equation; otherwise, returning to the step III, continuing the iteration optimization until reaching the maximum iteration number iter_max；

The industrial melt index soft measurement model of the optimal support vector machine further comprises:

and the model updating module is used for updating the model on line, inputting offline experimental data into a training set regularly and updating the fuzzy equation model.

2. A soft-sensing method implemented by the industrial melt index soft-sensing instrument of the optimal support vector machine of claim 1, wherein: the soft measurement method comprises the following concrete implementation steps:

2) preprocessing a model training sample input from a DCS database, centralizing the training sample, namely subtracting the average value of the sample, and then standardizing the training sample to ensure that the average value is 0 and the variance is 1; the processing is accomplished using the following mathematical process:

2.1) calculating the mean value:

2.2) calculating the variance:

2.3) standardization:

3) fuzzification is carried out on the training samples transmitted from the data preprocessing module; let there be c in the system of fuzzy equations^*A fuzzy group, the centers of the fuzzy groups k and j are v_k、v_jThen the normalized ith training sample X_iMembership mu for fuzzy group k_ikComprises the following steps:

Φ_ik(X_i,μ_ik)＝[1 func(μ_ik) X_i] (5)

wherein func (. mu.)_ik) Is a membership value mu_ikIs taken as a deformation function ofexp(μ_ik)，Φ_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikCorresponding new input matrix；

where R (w, ξ) is the objective function of the optimization problem, minR (w, ξ) is the minimum of the objective function of the optimization problem, N is the number of training samples, ξ ═ ξ₁,...,ξ_NIs the relaxation variable, ξ_iIs the ith component of the relaxation variable, w is the normal vector to the hyperplane of the support vector machine, b is the corresponding offset, and ω is_iN and y are the weight and penalty factors of a weighted least squares support vector machine, respectively,is the ith component xi of the relaxation variable of the weighted least squares support vector machine_iEstimation of the standard deviation, c₁Is a constant, here take 2.5, c₂Is a constant, here taken as 3, the superscript T denotes the transpose of the matrix, μ_ikRepresents the normalized ith training sample X_iMembership, Φ, for fuzzy group k_ik(X_i,μ_ik) Denotes the ith input variable X_iAnd membership mu of its fuzzy group k_ikThe corresponding new input matrix; from equations (6), (7) and (8), the output of the fuzzy group k in the training sample i is derived as:

wherein,the output of the fuzzy group k in the training sample i; k<·>Is the kernel function of a weighted least squares support vector machine, where K<·>Taking a linear kernel function; alpha is alpha_mM is 1, …, N is the mth component of the corresponding lagrange multiplier; mu.s_mkRepresents the m-th training sample X_mMembership, Φ, for fuzzy group k_mk(X_m,μ_mk) Denotes the m-th input variable X_mAnd membership mu of its fuzzy group k_mkThe corresponding new input matrix;

wherein,the output of the fuzzy group k in the training sample i;

f_p＝1/(E_p+1) (11)

(13)

r_p(iter+1)＝r_p(iter)+v_p(iter+1) (14)

Lbest_p＝f_p (15)

Gbest＝Lbest_p (16)

The soft measurement method further comprises the following steps: 5) and inputting the offline experimental data into a training set regularly, and updating the fuzzy equation model.