CN100470417C - Fault diagnostic system and method for under industrial producing process small sample condition - Google Patents

Abstract

The fault diagnosis system includes on site intelligent instrument connected to objects of industrial process, DCS system, and upper device. The DCS system is composed of data interface, control station, and database. The intelligent instrument is connected to DCS system, and upper device. The upper device includes the standardization process module, functional module of pivot element analysis, functional module for supporting classifier of vector machine, failure predication module, signal acquisition module, module for determining data to be diagnosed, and fault diagnosis module. The invention also discloses a fault diagnosis method. The invention discloses fault diagnosis system and method capable of well treating measured non-linear data obtained from small sample with high capability of generalizing model, as well as capable of obtaining good diagnosis result under condition of small sample in industrial process.

Description

Fault diagnosis system under the industrial producing process small sample condition and method

(1) technical field

The present invention relates to the industrial process fault diagnosis field, especially, relate to fault diagnosis system and method under a kind of industrial producing process small sample condition.

(2) background technology

Because product quality, economic benefit, safety and environmental protection requirement, it is very complicated that industrial process and relevant control system become, and in order to guarantee the normal operation of industrial system, Fault Diagnosis is being played the part of very important role with detection in industrial process.In recent years, statistical study is applied to process monitoring and fault diagnosis has obtained extensive studies.

Utilize industrial measured data, adopt the method for statistics to carry out fault diagnosis, avoided complicated Analysis on Mechanism, find the solution simple and efficient relatively.But present industrial process trouble-shooter, asymptotic statistical method when being based on number of samples mostly and being tending towards infinity, and in practical problems, sample number is limited often, so some very outstanding in theory method for diagnosing faults show often unsatisfactory in actual applications.How in the face of small sample, non-linear and to require the strong industrial processes of model generalization ability to carry out effective fault diagnosis be existing fault diagnostic system institute urgent problem.

(3) summary of the invention

Can not effectively handle small sample, non-linear and require the strong measured data of model generalization ability, be difficult to obtain the deficiency of diagnosis effect preferably for what overcome existing industrial process trouble-shooter, the invention provides a kind of can the processing and handle small sample, non-linear and require the strong measured data of model generalization ability, can access fault diagnosis system and method under the industrial producing process small sample condition of good diagnosis effect preferably.

The technical solution adopted for the present invention to solve the technical problems is:

Fault diagnosis system under a kind of industrial producing process small sample condition comprises the field intelligent instrument, DCS system and the host computer that are connected with the industrial processes object, and described DCS system is made of data-interface, control station, database; Intelligence instrument, DCS system, host computer link to each other successively, and described host computer comprises:

The standardization module, be used for to the database acquisition system just often the key variables data carry out standardization, the average of each variable is 0, variance is 1, obtains input matrix X, adopts following process to finish:

1) computation of mean values: $\stackrel{\‾}{\mathrm{TX}}=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{TX}}_i,---\left(1\right)$

2) calculate variance: ${\mathrm{\σ}}_x^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{TX}}_i-\stackrel{\‾}{\mathrm{TX}}),---\left(2\right)$

3) standardization: $X=\frac{\mathrm{TX}-\stackrel{\‾}{\mathrm{TX}}}{{\mathrm{\σ}}_x},---\left(3\right)$

Wherein, TX is a training sample, and N is a number of training, and TX is the average of training sample;

The pivot analysis functional module is used to carry out pivot analysis and extracts major component, according to the pivot variance extraction ratio that is provided with, adopts the method for covariance svd, adopts following steps to realize:

1. calculate the covariance matrix of X, be designated as ∑ _X

2. to ∑ _XCarry out svd, obtain characteristic root λ ₁, λ ₂..., λ _p, λ wherein ₁〉=λ 2 〉=... 〉=λ p, the characteristic of correspondence vector matrix is U;

3. calculate population variance and each eigenwert corresponding variance contribution rate, adding up from big to small by the variance contribution ratio of each eigenwert reaches set-point up to total variance contribution ratio, and it is k that note is chosen number;

4. the preceding k of selected characteristic vector matrix U is listed as, as transformation matrix T;

5. calculate pivot, calculate pivot F by formula F=T * X;

The support vector machine classifier functional module is used for kernel function and adopts radial basis function K (x _i, x)=exp (‖ x-x _i‖/σ ²), the fiducial probability according to being provided with turns to following quadratic programming with training process and finds the solution problem:

$\mathrm{\ω}\left(\mathrm{\α}\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i-\frac{1}{2}\underset{i,j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_{j}K(x_i,x_j)---\left(4\right)$

Obtain classification function, promptly as the sign function of minor function:

$f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{y}_i{\mathrm{\α}}_{i}K(x_i,x)+b---\left(5\right)$

Wherein, α _i(i=1 ..., N) be the Lagrange multiplier, x _i(i=1 ..., N) be input vector, y is an output variable, ω is the normal vector of support vector machine lineoid, the decision lineoid direction, b is the parameter of decision lineoid position, δ is a nuclear parameter;

Definition is as f (x) 〉=0, data sample is in normal condition; When f (x)＜0, be in abnormality;

Signal acquisition module is used to set the time slot of each sampling, the signal of collection site intelligence instrument;

The diagnostic data determination module, the data that are used for gathering are sent to the DCS real-time data base, from the real-time data base of DCS database, obtain up-to-date variable data as diagnostic data VX at each timing cycle;

Fault diagnosis module, be used for TX that data to be tested VX the time is obtained with training and Carry out standardization, and with the input of the data after the standardization as the pivot analysis module; The transform coefficient matrix T that obtains during with training carries out conversion to input, and matrix is input to the support vector machine classifier module after the conversion; With the discriminant function that input substitution training obtains, the computational discrimination functional value is differentiated status of processes.

As preferred a kind of scheme: described host computer also comprises: the discrimination model update module, be used for regularly adding the normal point of process status to training set VX, output to standardization module, wavelet decomposition module, pivot analysis functional module, wavelet reconstruction functional module, and upgrade the disaggregated model of support vector machine classifier.

As preferred another kind of scheme: described host computer also comprises: display module as a result, be used for fault diagnosis result is passed to the DCS system, and at the control station procedure for displaying state of DCS, and by DCS system and fieldbus process status information is delivered to operator station and shows.

A kind of industrial polypropylene production melting index detection failure diagnostic method, described method for diagnosing faults may further comprise the steps:

(1), determine the key variables that fault diagnosis is used, from the historical data base of DCS database respectively during the normal and fault of acquisition system the data of described variable as training sample TX;

(2), in pivot analysis module and support vector machine classifier module, parameters such as pivot analysis variance extraction ratio, support vector machine nuclear parameter and fiducial probability are set respectively, set the sampling period among the DCS;

(3), training sample TX in host computer, data are carried out standardization, make that the average of each variable is 0, variance is 1, obtains input matrix X, adopts following process to finish:

3.1) computation of mean values: $\stackrel{\‾}{\mathrm{TX}}=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{TX}}_i---\left(1\right)$

3.2) the calculating variance: ${\mathrm{\σ}}_x^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{TX}}_i-\stackrel{\‾}{\mathrm{TX}})---\left(2\right)$

3.3) standardization: $X=\frac{\mathrm{TX}-\stackrel{\‾}{\mathrm{TX}}}{{\mathrm{\σ}}_x}---\left(3\right)$

Wherein, N is a number of training, and N is a number of training, and TX is the average of training sample;

(4), carry out pivot analysis and extract major component, according to the pivot variance extraction ratio that is provided with, adopt the method for covariance svd, adopt following steps to realize:

1. calculate the covariance matrix of X, be designated as ∑ _X

2. to ∑ _XCarry out svd, obtain characteristic root λ ₁, λ ₂..., λ _p, λ wherein ₁〉=λ 2 〉=... 〉=λ p, the characteristic of correspondence vector matrix is U;

3. calculate population variance and each eigenwert corresponding variance contribution rate, adding up from big to small by the variance contribution ratio of each eigenwert reaches set-point up to total variance contribution ratio, and it is k that note is chosen number;

4. the preceding k of selected characteristic vector matrix U is listed as, as transformation matrix T;

5. calculate pivot, calculate pivot F by formula F=T * X;

(5), kernel function adopts radial basis function K (x _i, x)=exp (‖ x-x _i‖/σ ²), the fiducial probability according to being provided with turns to following quadratic programming with training process and finds the solution problem:

$\mathrm{\ω}\left(\mathrm{\α}\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i-\frac{1}{2}\underset{i,j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_{j}K(x_i,x_j)---\left(4\right)$

Obtain classification function, promptly as the sign function of minor function:

$f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{y}_i{\mathrm{\α}}_{i}K(x_i,x)+b---\left(5\right)$

Wherein, α _i(i=1 ..., N) be the Lagrange multiplier, x _i(i=1 ..., N) be input vector, y is an output variable, ω is the normal vector of support vector machine lineoid, the decision lineoid direction, b is the parameter of decision lineoid position, δ is a nuclear parameter;

Definition is as f (x) 〉=0, data sample is in normal condition; When f (x)＜0, be in abnormality;

(6), the data of gathering are sent in the DCS real-time data base, from the real-time data base of DCS database, obtain up-to-date variable data at each timing cycle as diagnostic data VX; The TX that data to be tested VX the time is obtained with training and Carry out standardization, and with the input of the data after the standardization as the pivot analysis module; The transform coefficient matrix T that obtains during with training carries out conversion to input, and matrix is input to the support vector machine classifier module after the conversion; With the discriminant function that input substitution training obtains, the computational discrimination functional value is differentiated status of processes.

As preferred a kind of scheme: described method for diagnosing faults also comprises: (7), regularly the normal point of process status is added among the training set VX, repeat the training process of (3)～(5), so that the disaggregated model of the support vector machine classifier that upgrades in time.

As preferred another kind of scheme: in described (6), the computational discrimination functional value, and on the man-machine interface of host computer the state of procedure for displaying, host computer is passed to the DCS system with fault diagnosis result, and, by DCS system and fieldbus process status information is delivered to operator station simultaneously and shows at the control station procedure for displaying state of DCS.

Good multivariable nonlinearity mapping ability that strong decorrelation sexuality that the pivot analysis functional module has in the host computer and support vector machine classifier functional module have and strong generalization ability.Can be good at handling small sample, non-linear and require the strong measured data of model generalization ability.

Beneficial effect of the present invention mainly shows: the two combines well with pivot analysis functional module and support vector machine classifier functional module, performance advantage separately, make fault diagnosis effectively reliable more, can better instruct production, improve productivity effect.

(4) description of drawings

Fig. 1 is the hardware structure diagram of fault diagnosis system proposed by the invention.

Fig. 2 is a fault diagnosis system functional block diagram proposed by the invention.

Fig. 3 is the theory diagram of host computer of the present invention.

(5) embodiment

Below in conjunction with accompanying drawing the present invention is further described.

Embodiment 1

With reference to Fig. 1, Fig. 2, Fig. 3, fault diagnosis system under a kind of industrial producing process small sample condition, comprise the field intelligent instrument 2, DCS system and the host computer 6 that are connected with industrial processes object 1, described DCS system is made of data-interface 3, control station 4, database 5; Intelligence instrument 2, DCS system, host computer 6 link to each other successively by fieldbus, and described host computer 6 comprises:

Standardization module 7 is used for data are carried out standardization, and the average of each variable is 0, and variance is 1, obtains input matrix X, adopts following process to finish:

1) computation of mean values: $\stackrel{\‾}{\mathrm{TX}}=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{TX}}_i,---\left(1\right)$

2) calculate variance: ${\mathrm{\σ}}_x^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{TX}}_i-\stackrel{\‾}{\mathrm{TX}}),---\left(2\right)$

3) standardization: $X=\frac{\mathrm{TX}-\stackrel{\‾}{\mathrm{TX}}}{{\mathrm{\σ}}_x},---\left(3\right)$

Wherein, TX is a training sample, and N is a number of training, and TX is the average of training sample;

Pivot analysis functional module 8 is used to carry out pivot analysis and extracts major component, according to the pivot variance extraction ratio that is provided with, adopts the method for covariance svd, adopts following steps to realize:

1. calculate the covariance matrix of X, be designated as ∑ _X

2. to ∑ _XCarry out svd, obtain characteristic root λ ₁, λ ₂..., λ _p, λ wherein ₁〉=λ 2 〉=... 〉=λ p, the characteristic of correspondence vector matrix is U;

3. calculate population variance and each eigenwert corresponding variance contribution rate, adding up from big to small by the variance contribution ratio of each eigenwert reaches set-point up to total variance contribution ratio, and it is k that note is chosen number;

4. the preceding k of selected characteristic vector matrix U is listed as, as transformation matrix T;

5. calculate pivot, calculate pivot F by formula F=T * X;

Support vector machine classifier functional module 9 is used for kernel function and adopts radial basis function K (x _i, x)=exp (‖ x-x _i‖/σ ²), the fiducial probability according to being provided with turns to following quadratic programming with training process and finds the solution problem:

$\mathrm{\ω}\left(\mathrm{\α}\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i-\frac{1}{2}\underset{i,j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_{j}K(x_i,x_j)---\left(4\right)$

Obtain classification function, promptly as the sign function of minor function:

$f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{y}_i{\mathrm{\α}}_{i}K(x_i,x)+b---\left(5\right)$

Wherein, α _i(i=1 ..., N) be the Lagrange multiplier, x _i(i=1 ..., N) be input vector, y is an output variable, ω is the normal vector of support vector machine lineoid, the decision lineoid direction, b is the parameter of decision lineoid position, δ is a nuclear parameter;

Definition is as f (x) 〉=0, data sample is in normal condition; When f (x)＜0, be in abnormality;

Signal acquisition module 10 is used to set the time slot of each sampling, the signal of collection site intelligence instrument;

Diagnostic data determination module 11, the data that are used for gathering are sent to the DCS real-time data base, from the real-time data base of DCS database, obtain up-to-date variable data as diagnostic data VX at each timing cycle;

Fault diagnosis module 12 is used for TX and σ x that data to be tested VX the time is obtained with training ²Carry out standardization, and with the input of the data after the standardization as the wavelet decomposition module, parameter identical during with training is carried out wavelet decomposition to the input data, the input of pivot analysis module in the coefficient conduct that obtains; The transform coefficient matrix T that obtains during with training carries out conversion to input, and matrix is input to the wavelet reconstruction module after the conversion; The data addition of correspondence is promptly obtained the major component of former testing data, and resulting major component is input to the support vector machine classifier module; With the discriminant function that input substitution training obtains, the computational discrimination functional value is differentiated status of processes.

Described host computer also comprises: discrimination model update module 13, be used for regularly adding the normal point of process status to training set VX, output to standardization module 7, pivot analysis functional module 8,, and upgrade the disaggregated model of support vector machine classifier module 9.

Described host computer also comprises: display module 14 as a result, are used for fault diagnosis result is passed to DCS, and at the control station procedure for displaying state of DCS, and by DCS system and fieldbus process status information are delivered to operator station and show.

The hardware structure diagram of the industrial process fault diagnosis system of present embodiment as shown in Figure 1, described fault diagnosis system core is made of the host computer 6 that comprises standardization module 7, pivot analysis module 8, support vector machine classifier module 9 three big functional modules and man-machine interface, comprise in addition: field intelligent instrument 2, DCS system and fieldbus.Described DCS system is made of data-interface 3, control station 4, database 5; Industrial process object 1, intelligence instrument 2, DCS system, host computer 6 link to each other successively by fieldbus, realize uploading and assigning of information flow.Fault diagnosis system is moved on host computer 6, can carry out message exchange with first floor system easily, in time the answering system fault.

The functional block diagram of the fault diagnosis system of present embodiment mainly comprises standardization module 7, pivot analysis module 8, support vector machine classifier module 9 etc. as shown in Figure 2.

Described method for diagnosing faults is implemented according to following steps:

1, determine the key variables that fault diagnosis is used, from the historical data base of DCS database 5 acquisition system just often the data of these variablees as training sample TX;

2, in the pivot analysis functional module 8 and support vector machine classifier functional module 9 of host computer 6, parameters such as pivot analysis variance extraction ratio, support vector machine classifier nuclear parameter and fiducial probability are set respectively, set the sampling period among the DCS;

3, training sample TX passes through functional modules such as standardization 7, pivot analysis 8, support vector machine classifier 9 successively in host computer 6, adopts following steps to finish the training of diagnostic system;

1) the standardization functional module 7 of host computer 6 is carried out standardization to data, makes that the average of each variable is 0, and variance is 1, obtains input matrix X.Adopt following process to finish:

1. computation of mean values: $\stackrel{\‾}{\mathrm{TX}}=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{TX}}_i---\left(1\right)$

2. calculate variance: ${\mathrm{\σ}}_x^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{TX}}_i-\stackrel{\‾}{\mathrm{TX}})---\left(2\right)$

3. standardization: $X=\frac{\mathrm{TX}-\stackrel{\‾}{\mathrm{TX}}}{{\mathrm{\σ}}_x}---\left(3\right)$

Wherein N is a number of training, and N is a number of training, and TX is the average of training sample.

The standardization that the standardization functional module 7 of host computer 6 is carried out can eliminate each variable because the influence that the dimension difference causes.

2) the pivot analysis functional module 8 of host computer 6 is carried out pivot analysis.The population variance extraction ratio of pivot analysis functional module 8 adopts following steps to realize greater than 80% in the described host computer 6:

1. calculate the covariance matrix of X, be designated as ∑ _X

2. to ∑ _XCarry out svd, obtain characteristic root λ ₁, λ ₂..., λ _p, λ wherein ₁〉=λ 2 〉=... λ p, the characteristic of correspondence vector matrix is U;

3. calculate population variance and each eigenwert corresponding variance contribution rate, adding up from big to small by the variance contribution ratio of each eigenwert reaches set-point up to total variance contribution ratio, and it is k that note is chosen number;

4. the preceding k of selected characteristic vector matrix U is listed as, as transformation matrix T;

5. calculate pivot, calculate pivot F by formula F=T * X.

Pivot analysis is lost under the minimum principle making every effort to data message, to the variable space dimensionality reduction of higher-dimension.Its essence is a few linear combination of research variable system, and the generalized variable that this several linear combination constituted will keep the information of former variable variation aspect as much as possible.Obviously, analytic system at a lower dimensional space than much easier at a higher dimensional space.

3) disaggregated model of the support vector machine classifier functional module 9 in the training host computer 6.

The kernel function of the support vector machine classifier functional module 9 in the described host computer 6 adopts radial basis function K (x _i, x)=exp (‖ x-x _i‖/σ ²), training process is turned to following quadratic programming finds the solution problem:

$\mathrm{\ω}\left(\mathrm{\α}\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i-\frac{1}{2}\underset{i,j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_{j}K(x_i,x_j)---\left(4\right)$

Thereby obtain classification function, promptly as the sign function of minor function:

$f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{y}_i{\mathrm{\α}}_{i}K(x_i,x)+b---\left(5\right)$

Wherein, α _i(i=1 ..., N) be the Lagrange multiplier, x _i(i=1 ..., N) be input vector, y is an output variable, ω is the normal vector of support vector machine lineoid, the decision lineoid direction, b is the parameter of decision lineoid position, δ is a nuclear parameter;

Definition is as f (x) 〉=0, data sample is in normal condition; When f (x)＜0, be in abnormality.

Support vector machine classifier adopts the structural risk minimization criterion based on Statistical Learning Theory, has solved difficult problems such as small sample, local minimum point, high dimension well, is used for classification problem and can improves nicety of grading.

4, system begins to put into operation:

1) uses timer, set the time interval of each sampling;

2) field intelligent instrument 2 testing process data and being sent in the real-time data base of DCS database 5;

3) host computer 6 from the real-time data base of DCS database 5, obtains up-to-date variable data at each timing cycle, as diagnostic data VX;

4) data to be tested VX, in the standardization functional module 7 of host computer 6, the TX that the time obtains with training and

Carry out standardization, and with the input of the data after the standardization as pivot analysis module 8;

5) the pivot analysis module 8 in the host computer 6, the transformation matrix T that obtains during with training carries out conversion to input, and the matrix after the conversion is input to support vector machine classifier module 9, as the input of support vector machine classifier module 9;

6) the support vector machine classifier module 9 in the host computer 6, the discriminant function that will input data substitution training obtains, the computational discrimination functional value is differentiated and the state of procedure for displaying on the man-machine interface of host computer 6;

7) host computer 6 is passed to DCS with fault diagnosis result, and at the control station 4 procedure for displaying states of DCS, by DCS system and fieldbus process status information is delivered to operator station simultaneously and shows, makes the execute-in-place worker in time to tackle.

5, sorter model upgrades

In system puts into operation process, regularly the normal point of process status is added among the training set TX, the training process of repeating step 3 is so that the disaggregated model in the support vector machine classifier 9 of the host computer 6 that upgrades in time keeps sorter model to have classifying quality preferably.

Embodiment 2

With reference to Fig. 1, Fig. 2, Fig. 3, the method for diagnosing faults under a kind of industrial producing process small sample condition, described method for diagnosing faults may further comprise the steps:

(1), determine the key variables that fault diagnosis is used, from the historical data base of DCS database respectively during the normal and fault of acquisition system the data of described variable as training sample TX;

(2), in pivot analysis module 9 and support vector machine classifier module 11, parameters such as pivot analysis variance extraction ratio, support vector machine nuclear parameter and fiducial probability are set respectively, set the sampling period among the DCS;

(3), training sample TX in host computer 6, data are carried out standardization, make that the average of each variable is 0, variance is 1, obtains input matrix X, adopts following process to finish:

3.1) computation of mean values: $\stackrel{\‾}{\mathrm{TX}}=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{TX}}_i---\left(1\right)$

3.2) the calculating variance: ${\mathrm{\σ}}_x^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{TX}}_i-\stackrel{\‾}{\mathrm{TX}})---\left(2\right)$

3.3) standardization: $X=\frac{\mathrm{TX}-\stackrel{\‾}{\mathrm{TX}}}{{\mathrm{\σ}}_x}---\left(3\right)$

Wherein, N is a number of training, and N is a number of training, and TX is the average of training sample.

(4), carry out pivot analysis and extract major component, adopt the method for covariance svd, adopt following steps to realize:

1. calculate the covariance matrix of X, be designated as ∑ _X

2. to ∑ _XCarry out svd, obtain characteristic root λ ₁, λ ₂..., λ _p, λ wherein ₁〉=λ 2 〉=... 〉=λ p, the characteristic of correspondence vector matrix is U;

3. calculate population variance and each eigenwert corresponding variance contribution rate, adding up from big to small by the variance contribution ratio of each eigenwert reaches set-point up to total variance contribution ratio, and it is k that note is chosen number;

4. the preceding k of selected characteristic vector matrix U is listed as, as transformation matrix T;

5. calculate pivot, calculate pivot F by formula F=T * X;

(5), kernel function adopts radial basis function K (x _i, x)=exp (‖ x-x _i‖/σ ²), training process is turned to following quadratic programming finds the solution problem:

$\mathrm{\ω}\left(\mathrm{\α}\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i-\frac{1}{2}\underset{i,j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_{j}K(x_i,x_j)---\left(4\right)$

Obtain classification function, promptly as the sign function of minor function:

$f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{y}_i{\mathrm{\α}}_{i}K(x_i,x)+b---\left(5\right)$

Wherein, α _i(i=1 ..., N) be the Lagrange multiplier, x _i(i=1 ..., N) be input vector, y is an output variable, ω is the normal vector of support vector machine lineoid, the decision lineoid direction, b is the parameter of decision lineoid position, δ is a nuclear parameter;

Definition is as f (x) 〉=0, data sample is in normal condition; When f (x)＜0, be in abnormality;

(6), the data of gathering are sent in the DCS real-time data base, from the real-time data base of DCS database, obtain up-to-date variable data at each timing cycle as diagnostic data VX; TX and σ x that data to be tested VX the time is obtained with training ²Carry out standardization, and with the input of the data after the standardization as the pivot analysis module; The transformation matrix T that obtains during with training carries out conversion to input, and matrix is input to the support vector machine classifier module after the conversion; With the discriminant function that input substitution training obtains, the computational discrimination functional value is differentiated status of processes.

Described method for diagnosing faults also comprises: (7), regularly the normal point of process status is added among the training set VX, repeat the training process of (3)～(5), so that the disaggregated model of the support vector machine classifier that upgrades in time.

In described (6), the computational discrimination functional value, and on the man-machine interface of host computer the state of procedure for displaying, host computer is passed to the DCS system with fault diagnosis result, and, by DCS system and fieldbus process status information is delivered to operator station simultaneously and shows at the control station procedure for displaying state of DCS.

Fault diagnosis system under the industrial producing process small sample condition proposed by the invention and method, be described by above-mentioned concrete implementation step, person skilled obviously can be in not breaking away from content of the present invention, spirit and scope to device as herein described with method of operating is changed or suitably change and combination, realize the technology of the present invention.Special needs to be pointed out is, the replacement that all are similar and change apparent to one skilled in the artly, they all can be regarded as being included in spirit of the present invention, scope and the content.

Claims (6)

1, the fault diagnosis system under a kind of industrial producing process small sample condition comprises the field intelligent instrument, DCS system and the host computer that are connected with the industrial processes object, and described DCS system is made of data-interface, control station and database; Intelligence instrument, DCS system link to each other successively with host computer, it is characterized in that: described host computer comprises:

The standardization module, be used for to the database acquisition system just often the key variables data carry out standardization, the average of each variable is 0, variance is 1, obtains input matrix X, adopts following process to finish:

1) computation of mean values: $\stackrel{\‾}{\mathrm{TX}}=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{TX}}_i,---\left(1\right)$

2) calculate variance: ${\mathrm{\σ}}_x^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{TX}}_i-\stackrel{\‾}{\mathrm{TX}}),---\left(2\right)$

3) standardization: $X=\frac{\mathrm{TX}-\stackrel{\‾}{\mathrm{TX}}}{{\mathrm{\σ}}_x},---\left(3\right)$

Wherein, TX is a training sample, and N is a number of training, and TX is the average of training sample;

The pivot analysis functional module is used to carry out pivot analysis and extracts major component, according to the pivot variance extraction ratio that is provided with, adopts the method for covariance svd, adopts following steps to realize:

1. calculate the covariance matrix of X, be designated as ∑ _X

2. to ∑ _XCarry out svd, obtain characteristic root λ ₁, λ ₂..., λ _p, λ wherein ₁〉=λ ₂〉=... 〉=λ p, the characteristic of correspondence vector matrix is U;

3. calculate population variance and each eigenwert corresponding variance contribution rate, adding up from big to small by the variance contribution ratio of each eigenwert reaches set-point up to total variance contribution ratio, and it is k that note is chosen number;

4. the preceding k of selected characteristic vector matrix U is listed as, as transformation matrix T;

5. calculate pivot, calculate pivot F by formula F=T * X;

The support vector machine classifier functional module is used for kernel function is made as radial basis function K (x _i, x)=exp (‖ x-x _i‖/σ ²), the fiducial probability according to being provided with turns to following quadratic programming with training process and finds the solution problem:

$\mathrm{\ω}\left(\mathrm{\α}\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i-\frac{1}{2}\underset{i,j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_{j}K(x_i,x_j)---\left(4\right)$

Obtain classification function, promptly as the sign function of minor function:

$f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{y}_i{\mathrm{\α}}_{i}K(x_i,x)+b---\left(5\right)$

Wherein, α _i, α _jBe the Lagrange multiplier, x _i, x _jBe input vector, y _i, y _jBe output variable, i=1 wherein ..., N, j=1 ..., N, ω are the normal vectors of support vector machine lineoid, the direction of decision lineoid, and b is the parameter of decision lineoid position;

Definition is as f (x) 〉=0, data sample is in normal condition; When f (x)＜0, be in abnormality;

Signal acquisition module is used to set the time slot of each sampling, the signal of collection site intelligence instrument;

The diagnostic data determination module, the data that are used for gathering are sent to the DCS real-time data base, from the real-time data base of DCS database, obtain up-to-date variable data as diagnostic data VX at each timing cycle;

Fault diagnosis module, be used for TX that diagnostic data VX the time is obtained with training and

Carry out standardization, and with the input of the data after the standardization as the pivot analysis functional module; The transform coefficient matrix T that obtains during with training carries out conversion to input, and matrix is input to the support vector machine classifier module after the conversion; With the discriminant function that input substitution training obtains, the computational discrimination functional value is differentiated status of processes.

2, the fault diagnosis system under the industrial producing process small sample condition as claimed in claim 1, it is characterized in that: described host computer also comprises:

The discrimination model update module is used for regularly adding the normal point of process status to training set, outputs to standardization module, pivot analysis functional module, and upgrades the disaggregated model of support vector machine classifier.

3, the fault diagnosis system under the industrial producing process small sample condition as claimed in claim 1 or 2, it is characterized in that: described host computer also comprises:

Display module is used for fault diagnosis result is passed to the DCS system as a result, and at the control station procedure for displaying state of DCS, and by DCS system and fieldbus process status information is delivered to operator station and shows.

4, the method for diagnosing faults of the realization of the fault diagnosis system under a kind of usefulness industrial producing process small sample condition as claimed in claim 1, it is characterized in that: described method for diagnosing faults may further comprise the steps:

1), determine the key variables that fault diagnosis is used, from the historical data base of DCS database respectively during the normal and fault of acquisition system the data of described variable as training sample TX;

2), in pivot analysis functional module and support vector machine classifier functional module, pivot analysis variance extraction ratio, support vector machine nuclear parameter and fiducial probability parameter are set respectively, set the sampling period among the DCS;

3), training sample TX in host computer, data are carried out standardization, make that the average of each variable is 0, variance is 1, obtains input matrix X, adopts following process to finish:

3.1) computation of mean values: $\stackrel{\‾}{\mathrm{TX}}=\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{TX}}_i---\left(1\right)$

3.2) the calculating variance: ${\mathrm{\σ}}_x^2=\frac{1}{N-1}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}({\mathrm{TX}}_i-\stackrel{\‾}{\mathrm{TX}})---\left(2\right)$

3.3) standardization: $X=\frac{\mathrm{TX}-\stackrel{\‾}{\mathrm{TX}}}{{\mathrm{\σ}}_x}---\left(3\right)$

Wherein, TX is a training sample, and N is a number of training, and TX is the average of training sample;

4), carry out pivot analysis and extract major component, according to the pivot variance extraction ratio that is provided with, adopt the method for covariance svd, adopt following steps to realize:

1. calculate the covariance matrix of X, be designated as ∑ _X

2. to ∑ _XCarry out svd, obtain characteristic root λ ₁, λ ₂..., λ _p, λ wherein ₁〉=λ ₂〉=... 〉=λ p, the characteristic of correspondence vector matrix is U;

3. calculate population variance and each eigenwert corresponding variance contribution rate, adding up from big to small by the variance contribution ratio of each eigenwert reaches set-point up to total variance contribution ratio, and it is k that note is chosen number;

4. the preceding k of selected characteristic vector matrix U is listed as, as transformation matrix T;

5. calculate pivot, calculate pivot F by formula F=T * X;

5), kernel function adopts radial basis function K (x _i, x)=exp (‖ x-x _i‖/σ ²), the fiducial probability according to being provided with turns to following quadratic programming with training process and finds the solution problem:

$\mathrm{\ω}\left(\mathrm{\α}\right)=\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i-\frac{1}{2}\underset{i,j=1}{\overset{N}{\mathrm{\Σ}}}{\mathrm{\α}}_i{\mathrm{\α}}_j{y}_i{y}_{j}K(x_i,x_j)---\left(4\right)$

Obtain classification function, promptly as the sign function of minor function:

$f\left(x\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{y}_i{\mathrm{\α}}_{i}K(x_i,x)+b---\left(5\right)$

Wherein, α _i, α _jBe the Lagrange multiplier, x _i, x _jBe input vector, y _i, y _jBe output variable, i=1 wherein ..., N, j=1 ..., N, ω are the normal vectors of support vector machine lineoid, the direction of decision lineoid, and b is the parameter of decision lineoid position;

Definition is as f (x) 〉=0, data sample is in normal condition; When f (x)＜0, be in abnormality;

6), the data of gathering are sent in the DCS real-time data base, from the real-time data base of DCS database, obtain up-to-date variable data as diagnostic data VX at each timing cycle; The TX that diagnostic data VX the time is obtained with training and

Carry out standardization, and with the input of the data after the standardization as the pivot analysis functional module; The transform coefficient matrix T that obtains during with training carries out conversion to input, and matrix is input to the support vector machine classifier module after the conversion; With the discriminant function that input substitution training obtains, the computational discrimination functional value is differentiated status of processes.

5, method for diagnosing faults as claimed in claim 4, it is characterized in that: described method for diagnosing faults also comprises: step 7), regularly the normal point of process status is added in the training set, repeating step 3)～5 training process) is so that the disaggregated model of the support vector machine classifier that upgrades in time.

6, as claim 4 or 5 described method for diagnosing faults, it is characterized in that: in described step 6), the computational discrimination functional value, and on the man-machine interface of host computer the state of procedure for displaying, host computer is passed to the DCS system with fault diagnosis result, and, by DCS system and fieldbus process status information is delivered to operator station simultaneously and shows at the control station procedure for displaying state of DCS.

Images