A method for detecting sticking of a control valve in a control loop based on phase space reconstruction

Abstract

The application discloses a kind of based on phase space reconstruction's control loop regulating valve stickiness detection method, it is related to valve fault diagnosis field.The method of the present application is the nonlinear signal analysis method of unsupervised, the method combines phase space reconstruction, recursive analysis and power spectrum analysis and other technologies, proposes stickiness statistical characteristic index and stickiness distribution characteristic index, to characterize the nonlinear and periodicity law brought by stickiness, only by loop parameter realizes the unsupervised regulating valve stickiness fault detection.The method proposed effectively extracts the stickiness characteristics in recurrence plot, not only simple and convenient to calculate, but also has high interpretability and accuracy, shows the great potential of recursive analysis in valve stickiness detection, provides a method basis for the online monitoring and performance evaluation of industrial loop.

Description

Technical Field

[0001] The present invention belongs to the field of valve fault diagnosis, and in particular relates to a method for detecting viscosity of a regulating valve in a control loop based on phase space reconstruction. Background Art

[0002] As one of the most critical components in a control loop, control valves ensure product quality and personnel safety. During extended operation, nonlinear motion caused by valve sticking often leads to control loop oscillation, increasing economic and safety risks in the production process. Among the various factors contributing to control loop oscillation, valve sticking accounts for 20-30% of the cause. Therefore, valve sticking detection remains a crucial issue in control loop performance evaluation. In recent years, nonlinear dynamics theory has provided new insights into signal analysis methods. Recursive analysis based on phase space reconstruction is an analytical approach for dealing with complex nonlinear systems. It is considered a powerful tool for detecting subtle changes in signals and systems. However, most methods have not extensively explored how to effectively extract features from recursive graphs for valve sticking detection. A highly interpretable and accurate valve sticking detection method has yet to be established. Existing neural network methods are mostly black-box models, lacking interpretability and requiring large amounts of labeled data, which poses challenges for both practical application and model optimization. Summary of the Invention

[0003] The purpose of the present invention is to solve the problems existing in the prior art and provide a method for detecting viscosity of a regulating valve in a control loop based on phase space reconstruction.

[0004] In order to achieve the above-mentioned object of the invention, the present invention specifically adopts the following technical solutions:

[0005] A method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction comprises the following steps:

[0006] S1: Acquire the signals in the detected control loop, where the signals in the control loop include a setting signal, a control signal, and a process variable signal, use the difference between the process variable signal and the setting signal as an error signal, and perform normalization processing on the error signal and the control signal;

[0007] S2: Reconstruct the phase space of the normalized control signal to obtain a high-dimensional first time series after signal reconstruction, and reconstruct the phase space of the normalized error signal to obtain a high-dimensional second time series after signal reconstruction;

[0008] S3: constructing a first recursive matrix from the high-dimensional first time series, and visually drawing a first recursive graph of the control signal using the first recursive matrix; constructing a second recursive matrix from the high-dimensional second time series, and visually drawing a second recursive graph of the error signal using the second recursive matrix;

[0009] S4: calculating a first viscosity statistical characteristic index based on the shape of the geometric pattern on the first recursive graph, calculating a second viscosity statistical characteristic index based on the shape of the geometric pattern on the second recursive graph, and calculating a viscosity distribution characteristic index based on the distribution of the geometric pattern on the second recursive graph;

[0010] S5: The difference between the first viscosity statistical characteristic index and the second viscosity statistical characteristic index is used as the viscosity statistical characteristic index increment. If the viscosity statistical characteristic index increment is greater than a preset viscosity statistical threshold and the viscosity distribution characteristic index is greater than a preset viscosity distribution threshold, it indicates that a viscosity fault occurs in the control valve in the control loop.

[0011] Based on the above solution, each step can be implemented in the following preferred specific manner.

[0012] Preferably, in step S1, the standardization process adopts a Z-score standardization method.

[0013] Preferably, in step S2, the phase space reconstruction uses a coordinate delay method based on Takens embedding theorem to calculate the i-th high-dimensional vector X in the high-dimensional first time series or the high-dimensional second time series. i :

[0014] X i ={x(i),x(i+τ),…x(i+(m-1)τ)}i=1,2,…N-(m-1)τ

[0015] Where m represents the embedding dimension; τ represents the delay time; N represents the length of the normalized control signal or the normalized error signal; x(i), x(i+τ), … x(i+(m-1)τ) represent the sample values ​​of the normalized control signal or the normalized error signal at the i-th time point, respectively.

[0016] Preferably, the mutual information method is used to calculate the delay time reference value of the error signal and the delay time reference value of the control signal respectively, and the maximum value of the delay time reference value of the error signal and the delay time reference value of the control signal is used as the final reference value of the delay time for phase space reconstruction.

[0017] Preferably, the false neighbor method is used to calculate the embedding dimension reference value of the error signal and the embedding dimension reference value of the control signal respectively, and the maximum value of the embedding dimension reference value of the error signal and the embedding dimension reference value of the control signal is used as the final reference value of the embedding dimension for phase space reconstruction.

[0018] Preferably, in step S3, the element R in the i-th row and j-th column of the first recursive matrix or the second recursive matrix is i,j The calculation is as follows:

[0019] R i,j =Θ(ε-S i,j )

[0020]

[0021] Among them, S i,j Indicates the calculation of the i-th high-dimensional vector X i and the j-th high-dimensional vector X j θ represents the Heaviside function; ε represents the threshold distance.

[0022] As a preference, in step S3, when R i,j = 1, the elements on the i-th row and j-th column of the first recursive matrix or the second recursive matrix are set to black dots. i,j When =0, the elements in the i-th row and j-th column of the first recursive matrix or the second recursive matrix are set as white dots to obtain the first recursive graph and the second recursive graph.

[0023] As a preference, in step S4, the first viscosity statistical characteristic index SFI_ OP And the second viscosity statistical characteristic index SFI_ PV-SP The specific steps are as follows:

[0024] S411: Obtain positions of black dots in the first recursive graph and the second recursive graph;

[0025] S412: Traverse the black dots in the first recursive graph and score each black dot. If the scored black dot has adjacent black dots in the horizontal or vertical direction, preset scores are given to the adjacent black dots. The scores of all adjacent black dots are added together as the score of the scored black dot. The scores of all scored black dots in the first recursive graph are added together to obtain a first total score score1.

[0026] S413: Traverse the black dots in the second recursive graph and score each black dot. If the scored black dot has adjacent black dots in the horizontal or vertical direction, preset scores are given to the adjacent black dots. The scores of all adjacent black dots are added together as the score of the scored black dot. The scores of all scored black dots in the second recursive graph are added together to obtain a second total score, score2.

[0027] S414: Traverse the black points in the first recursive graph and score each black point. If the scored black point has an adjacent black point on the diagonal line, preset a new score for the adjacent black point on the diagonal line, and the new score is twice the preset score. The scores of all the adjacent black points on the diagonal line are added together as the score of the scored black point. The scores of all the scored black points in the first recursive graph are added together to obtain a third total score score3.

[0028] S415: Traverse the black points in the second recursive graph and score each black point. If the scored black point has adjacent black points on the diagonal line, preset new scores for the adjacent black points on the diagonal line, and add up the scores of all the adjacent black points on the diagonal line as the score of the scored black point. Add up the scores of all the scored black points in the second recursive graph to obtain a fourth total score score4;

[0029] S416: The ratio of the first total score score1 to the third total score score3 is used as the first viscosity statistical characteristic index SFI_ OP :

[0030] SFI_ OP =score1 / score3

[0031] The ratio of the second total score score2 to the fourth total score score4 is used as the second viscosity statistical characteristic index SFI_PV -SP :

[0032] SFI_ PV-SP =score2 / score4.

[0033] As a preference, in step S4, the viscosity distribution characteristic index DFI_ PV-SP The specific steps are as follows:

[0034] S421: taking the number of black dots in each row of the second recursive graph as an element in the time series, and traversing the second recursive graph to obtain the time series;

[0035] S422: Calculate the power spectrum density of the time series, and use the power spectrum density within a preset frequency range as the final power spectrum density;

[0036] S423: Calculate the sparsity of the final power spectrum density to characterize the viscosity distribution characteristic index DFI_ PV-SP :

[0037]

[0038] Among them, N f is the total number of frequencies within the preset frequency range; px x(f) is the power spectral density of the time series L; |·| 2 represents the square of the modulus value; f represents the frequency.

[0039] Preferably, in step S5, the reference value of the viscosity statistical threshold is 0, and the reference value of the viscosity distribution threshold is 0.58.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] This unsupervised nonlinear signal analysis method combines multiple methods, including phase space reconstruction, recursion analysis, and power spectrum analysis. It proposes the viscosity statistical feature index (SFI) and the viscosity distribution feature index (DFI) to characterize the nonlinear and periodic characteristics caused by viscosity, enabling unsupervised control valve viscosity fault detection using only loop parameters. The proposed method effectively extracts viscosity features from the recursion graph, is computationally simple and convenient, and is highly interpretable and accurate, providing a methodological foundation for online monitoring and performance evaluation of industrial circuits. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Schematic diagram of the process of the present invention;

[0043] Figure 2 This is a schematic diagram of the cross-shaped basic structure and score distribution in the recursive graph of the present invention;

[0044] Figure 3 Schematic diagram of the basic structure of the diagonal line and the score distribution in the recursive graph of the present invention;

[0045] Figure 4 Schematic diagram showing the principle of time series in the recursive graph of the present invention;

[0046] Figure 5 Schematic diagram comparing different sticking detection methods in the international sticking fault database according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the above-mentioned objects, features and advantages of the present invention more clearly understood, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below. The technical features in the various embodiments of the present invention can be combined accordingly without conflicting with each other.

[0048] In the description of the present invention, it should be understood that the terms "first" and "second" are used solely for descriptive purposes and are not to be construed as indicating or implying relative importance or implicitly specifying the number of technical features being described. Therefore, features defined as "first" or "second" may explicitly or implicitly include at least one of such features.

[0049] like Figure 1 As shown, in a preferred implementation of the present invention, the above-mentioned method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction includes the following steps S1 to S5. The specific implementation process is described in detail below.

[0050] S1: Acquire the signals in the detected control loop. The signals in the control loop include the set signal SP, the control signal OP, and the process variable signals PV such as flow and pressure. The difference between the process variable signal PV and the set signal SP is used as the error signal (PV-SP). The error signal and the control signal OP are standardized.

[0051] It should be noted that in step S1 of the present invention, the signal sampling interval in the control loop is t s .

[0052] It should also be noted that in step S1 of the present invention, the standardization process adopts the Z-score standardization method, and the above standardization process can be expressed as:

[0053]

[0054] where x and x′ represent the unscaled and scaled variables, respectively; represents the mean value of the variable; σ x Represents the standard deviation of the variable.

[0055] S2: Reconstruct the phase space of the normalized control signal to obtain a high-dimensional first time series after signal reconstruction, and reconstruct the phase space of the normalized error signal to obtain a high-dimensional second time series after signal reconstruction.

[0056] It should be noted that in step S2 of the present invention, the signals for phase space reconstruction are the normalized error signal and control signal. Phase space reconstruction uses the coordinate delay method based on Takens embedding theorem to calculate the i-th high-dimensional vector X in the high-dimensional first time series or the high-dimensional second time series. i :

[0057] X i ={x(i),x(i+τ),…x(i+(m-1)τ)}i=1,2,…N-(m-1)τ

[0058] Where m represents the embedding dimension; τ represents the delay time; N represents the length of the normalized control signal or the normalized error signal; x(i), x(i+τ), … x(i+(m-1)τ) represent the sample values ​​of the normalized control signal or the normalized error signal at the i-th time point, respectively.

[0059] Furthermore, the present invention uses the Average Mutual Information (AMI) method to calculate the delay time reference value τ1 of the error signal and the delay time reference value τ2 of the control signal, respectively, and uses the maximum value of the delay time reference value of the error signal and the delay time reference value of the control signal as the final reference value of the delay time τ for phase space reconstruction:

[0060] τ=max(τ1,τ2)

[0061] Furthermore, the present invention uses the False Nearest Neighbor (FNN) method to calculate the embedding dimension reference value m1 of the error signal and the embedding dimension reference value m2 of the control signal, respectively. The maximum value of the embedding dimension reference value of the error signal and the embedding dimension reference value of the control signal is used as the final reference value m of the embedding dimension for phase space reconstruction:

[0062] m=max(m1,m2)

[0063] S3: constructing a first recursive matrix from the high-dimensional first time series, and visually drawing a first recursive graph of the control signal using the first recursive matrix; constructing a second recursive matrix from the high-dimensional second time series, and visually drawing a second recursive graph of the error signal using the second recursive matrix.

[0064] It should be noted that in step S3, the recursive matrix is ​​constructed by comparing the angular distance between two high-dimensional vectors in the phase space with the threshold. Specifically, the element R in the i-th row and j-th column of the first recursive matrix or the second recursive matrix is i,j The calculation is as follows:

[0065] R i,j =Θ(ε-S i,j )

[0066]

[0067] Among them, S i,j Indicates the calculation of the i-th high-dimensional vector X i and the j-th high-dimensional vector X j θ represents the Heaviside function; ε represents the threshold distance.

[0068] In this embodiment, when calculating the first recursive matrix, Xi 、X j Substitute the i-th and j-th high-dimensional vectors in the high-dimensional first time series; when calculating the second recursive matrix, X i 、X j The corresponding high-dimensional vectors in the second high-dimensional time series are substituted. In addition, the value of the threshold distance can be selected by those skilled in the art according to actual conditions. In this embodiment, it is recommended to be 10% of the maximum phase space diameter.

[0069] It should be noted that, in step S3, when R i,j = 1, the elements on the i-th row and j-th column of the first recursive matrix or the second recursive matrix are set to black dots. i,j When =0, the elements in the i-th row and j-th column of the first recursive matrix or the second recursive matrix are set as white dots, thereby drawing the first recursive graph and the second recursive graph.

[0070] S4: Calculate the first viscosity statistical feature index SFI_ by the shape of the geometric pattern on the first recursive graph OP , calculate the second viscosity statistical feature index SFI_ through the shape of the geometric pattern on the second recursive graph PV-SP , the viscosity distribution characteristic index DFI_ is calculated by the distribution of geometric patterns on the second recursive graph PV-SP .

[0071] It should be noted that in step S4 of the present invention, the viscosity statistical characteristic index is related to the cross-shaped basic structure and the diagonal basic structure in the recursive graph. Figure 2 As shown in , the cross-shaped basic structure is composed of five adjacent black dots in the middle, left, right, top and bottom on the recursive graph. Figure 3 As shown in Figure 2, the diagonal basic structure is composed of three adjacent black dots in the middle, upper left, and lower right of the recursive graph. The proportion of the two basic structures in the first and second recursive graphs is evaluated by a statistical method based on structural similarity to characterize the first and second viscosity statistical characteristic indicators.

[0072] In the present invention, the specific calculation steps of the first and second viscosity statistical characteristic indicators are as follows:

[0073] S411: Obtain the positions of the black dots in the first recursive graph and the second recursive graph, that is, the positions of the black dots in the first recursive matrix or the second recursive matrix that satisfy R i,j =1 position coordinates.

[0074] S412: Traverse the black dots in the first recursive graph and score each black dot. If the scored black dot has adjacent black dots in the horizontal or vertical direction, preset scores are given to the adjacent black dots. The scores of all adjacent black dots are added together as the score of the scored black dot. The scores of all scored black dots in the first recursive graph are added together to obtain a first total score, score1.

[0075] S413: Traverse the black dots in the second recursive graph and score each black dot. If the scored black dot has adjacent black dots in the horizontal or vertical direction, preset scores are given to the adjacent black dots. The scores of all adjacent black dots are added together as the score of the scored black dot. The scores of all scored black dots in the second recursive graph are added together to obtain a second total score, score2.

[0076] It should be noted that in S412-S413 of this embodiment, the preset score for adjacent black dots in the horizontal or vertical direction is 25 points. If a scored black dot has adjacent black dots in both the horizontal and vertical directions, the score of the scored black dot is 25*4=100 points. In other words, if there are adjacent black dots in both the horizontal and vertical directions, 25 points are reduced from 100 points for each missing adjacent black dot, which is used as the score of the scored black dot.

[0077] S414: Traverse the black dots in the first recursive graph and score each black dot. If the scored black dot has adjacent black dots on the diagonal, preset new scores for the adjacent black dots on the diagonal, and satisfy the requirement that the new scores are twice the preset scores. The scores of all adjacent black dots on the diagonal are added together as the score of the scored black dot. The scores of all scored black dots in the first recursive graph are added together to obtain a third total score, score3.

[0078] S415: Traverse the black dots in the second recursive graph and score each black dot. If the scored black dot has adjacent black dots on the diagonal line, preset new scores for the adjacent black dots on the diagonal line, and add up the scores of all the adjacent black dots on the diagonal line as the score of the scored black dot. Add up the scores of all the scored black dots in the second recursive graph to obtain a fourth total score score4.

[0079] It should be noted that in S414-S415 of this embodiment, the preset score for adjacent black dots on the diagonal line is 25*2=50 points. If the scored black dot also has adjacent black dots on the diagonal line, the score for the scored black dot is 50*2=100 points. In other words, if there are adjacent black dots on the diagonal line, each missing adjacent black dot will be reduced by 50 points from the 100-point score to serve as the score for the scored black dot.

[0080] S416: The ratio of the first total score score1 to the third total score score3 is used as the first viscosity statistical characteristic index SFI_ OP :

[0081] SFI_ OP =score1 / score3

[0082] The ratio of the second total score score2 to the fourth total score score4 is used as the second viscosity statistical characteristic index SFI_ PV-SP :

[0083] SFI_ PV-SP =score2 / score4

[0084] It should be noted that, in step S4 of the present invention, Figure 4 As shown in Figure 1, the viscosity distribution characteristic index DFI performs power spectrum analysis by counting the number of black points in each row on the recurrence graph, and uses the sparsity of the power spectrum density to characterize the viscosity distribution characteristic index.

[0085] In the present invention, the specific calculation steps of the viscosity distribution characteristic index are as follows:

[0086] S421: Taking the number of black dots in each row of the second recursive graph as an element in the time series, traversing the second recursive graph to obtain the time series L.

[0087] It should be noted that the sampling interval of the time series is the same as the signal sampling interval in the control loop, which is t s .

[0088] S422: Calculate the power spectrum density of the time series L, and use the power spectrum density within a preset frequency range as the final power spectrum density.

[0089] In the present invention, the calculation method of power spectral density belongs to the prior art. In this embodiment, the range of [0.002, 1] Hz is used as the preset frequency range, and on this basis, the power spectral density within this frequency range is extracted. The specific process of calculating the power spectral density of the time series L is as follows:

[0090]

[0091] Among them, p xx (f) is the Fourier transform of the autocorrelation coefficient, that is, the power spectrum density of the time series L; r xx (k) is the autocorrelation coefficient; k is the lag number; f is the frequency; j′ is the imaginary unit; L(i) and L(i+k) represent the i-th and i+k-th elements in the time series, respectively; is the mean of the time series.

[0092] S423: Calculate the sparsity of the final power spectrum density to characterize the viscosity distribution characteristic index DFI_ PV-SP :

[0093]

[0094] Among them, N fis the total number of frequencies within the preset frequency range.

[0095] S5: The difference between the first viscosity statistical characteristic index and the second viscosity statistical characteristic index is used as the viscosity statistical characteristic index increment ΔSFI. If the viscosity statistical characteristic index increment ΔSFI is greater than the preset viscosity statistical threshold e1 and the viscosity distribution characteristic index DFI_ PV-SP If it is greater than the preset viscosity distribution threshold e2, it means that the control valve in the control loop has a viscosity fault:

[0096] ΔSFI=SFI_ PV-SP -SFI_ OP >e1

[0097] DFI_ PV-SP >e2

[0098] It should be noted that, in step S5 of the present invention, the reference value of the viscosity statistical threshold e1 is 0, and the reference value of the viscosity distribution threshold e2 is 0.58.

[0099] In order to better demonstrate the specific implementation and technical effects of the present invention, the method for detecting viscosity of a regulating valve in a control loop based on phase space reconstruction shown in steps S1 to S5 in the above preferred implementation is applied to a specific example.

[0100] Example

[0101] The specific implementation process of the method for detecting viscosity of a regulating valve in a control loop based on phase space reconstruction adopted in this embodiment is as described above and will not be repeated here.

[0102] This example uses the International Stiction Database (ISDB) to validate the method. The database, available at https: / / sites.ualberta.ca / ~bhuang / Stiction-Book.htm, is a comprehensive process control dataset that includes both self-regulating and integrated control loops. Most of these loops are flow, temperature, level, and pressure loops, most of which are affected by stiction. The dataset includes 108 industrial data types from six sectors, including construction, chemicals, papermaking, power generation, mining, and metallurgy, with 78 data points available for analysis.

[0103] The OP, SP, and PV signals of 78 different industrial circuits in the data set are calculated and judged according to steps S1 to S5 to obtain the corresponding diagnostic results. The method of the present invention is compared with the existing viscosity detection algorithm in the world. The diagnostic results of each method are as follows: Figure 5 shown.

[0104] Diagnostic results show that the method of the present invention correctly diagnosed 60 out of 78 industrial circuits, ranking among the top existing international viscosity detection methods. Currently, algorithms with high accuracy rates are all supervised neural network models, which require high resource costs, have poor interpretability, and are limited in practical application. This example demonstrates the advanced nature and effectiveness of the method of the present invention.

[0105] The embodiment described above is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Persons skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, any technical solution obtained by equivalent substitution or equivalent transformation falls within the scope of protection of the present invention.

Claims

1. A method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction, characterized in that: The following steps are involved: S1: Acquire the signals in the detected control loop, where the signals in the control loop include a setting signal, a control signal, and a process variable signal, use the difference between the process variable signal and the setting signal as an error signal, and perform normalization processing on the error signal and the control signal; S2: Reconstruct the phase space of the normalized control signal to obtain a high-dimensional first time series after signal reconstruction, and reconstruct the phase space of the normalized error signal to obtain a high-dimensional second time series after signal reconstruction; S3: constructing a first recursive matrix from the high-dimensional first time series, and visually drawing a first recursive graph of the control signal using the first recursive matrix; constructing a second recursive matrix from the high-dimensional second time series, and visually drawing a second recursive graph of the error signal using the second recursive matrix; S4: calculating a first viscosity statistical characteristic index based on the shape of the geometric pattern on the first recursive graph, calculating a second viscosity statistical characteristic index based on the shape of the geometric pattern on the second recursive graph, and calculating a viscosity distribution characteristic index based on the distribution of the geometric pattern on the second recursive graph; S5: taking the difference between the first viscosity statistical characteristic index and the second viscosity statistical characteristic index as a viscosity statistical characteristic index increment; if the viscosity statistical characteristic index increment is greater than a preset viscosity statistical threshold and the viscosity distribution characteristic index is greater than a preset viscosity distribution threshold, it indicates that a viscosity fault occurs in the control valve in the control loop; In step S4, the first viscosity statistical characteristic index SFI_ OP And the second viscosity statistical characteristic index SFI_ PV-SP The specific steps are as follows: S411: Obtain positions of black dots in the first recursive graph and the second recursive graph; S412: Traverse the black dots in the first recursive graph and score each black dot. If the scored black dot has adjacent black dots in the horizontal or vertical direction, preset scores are given to the adjacent black dots. The scores of all adjacent black dots are added together as the score of the scored black dot. The scores of all scored black dots in the first recursive graph are added together to obtain a first total score score1. S413: Traverse the black dots in the second recursive graph and score each black dot. If the scored black dot has adjacent black dots in the horizontal or vertical direction, preset scores are given to the adjacent black dots. The scores of all adjacent black dots are added together as the score of the scored black dot. The scores of all scored black dots in the second recursive graph are added together to obtain a second total score, score2. S414: Traverse the black points in the first recursive graph and score each black point. If the scored black point has an adjacent black point on the diagonal line, preset a new score for the adjacent black point on the diagonal line, and the new score is twice the preset score. The scores of all the adjacent black points on the diagonal line are added together as the score of the scored black point. The scores of all the scored black points in the first recursive graph are added together to obtain a third total score score3. S415: Traverse the black points in the second recursive graph and score each black point. If the scored black point has adjacent black points on the diagonal line, preset new scores for the adjacent black points on the diagonal line, and add up the scores of all the adjacent black points on the diagonal line as the score of the scored black point. Add up the scores of all the scored black points in the second recursive graph to obtain a fourth total score score4; S416: The ratio of the first total score score1 to the third total score score3 is used as the first viscosity statistical characteristic index SFI_ OP : SFI_ OP =score1 / score3 The ratio of the second total score score2 to the fourth total score score4 is used as the second viscosity statistical characteristic index SFI_ PV-SP : SFI_ PV-SP =score2 / score4 In step S4, the viscosity distribution characteristic index DFI_ PV-SP The specific steps are as follows: S421: taking the number of black dots in each row of the second recursive graph as an element in the time series, and traversing the second recursive graph to obtain the time series; S422: Calculate the power spectrum density of the time series, and use the power spectrum density within a preset frequency range as the final power spectrum density; S423: Calculate the sparsity of the final power spectrum density to characterize the viscosity distribution characteristic index DFI_ PV-SP : ； in, is the total number of frequencies within the preset frequency range; For time series The power spectral density of Represents the square of the modulus value; Indicates frequency.

2. The method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction according to claim 1, characterized in that: In step S1, the normalization process adopts the Z-score normalization method.

3. The method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction according to claim 1, characterized in that: In step S2, the phase space reconstruction uses the coordinate delay method based on Takens embedding theorem to calculate the i-th high-dimensional vector in the high-dimensional first time series or the high-dimensional second time series. : ； Where m represents the embedding dimension; τ represents the delay time; Indicates the length of the normalized control signal or the normalized error signal; represent the sampling value of the normalized control signal or the normalized error signal at the i-th time point.

4. The method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction according to claim 3, characterized in that: The mutual information method is used to calculate the delay time reference value of the error signal and the delay time reference value of the control signal respectively, and the maximum value of the delay time reference value of the error signal and the delay time reference value of the control signal is used as the final reference value of the delay time for phase space reconstruction.

5. The method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction according to claim 3, characterized in that: The false neighbor method is used to calculate the embedding dimension reference value of the error signal and the embedding dimension reference value of the control signal respectively, and the maximum value of the embedding dimension reference value of the error signal and the embedding dimension reference value of the control signal is used as the final reference value of the embedding dimension for phase space reconstruction.

6. The method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction according to claim 3, characterized in that: In step S3, the element in the i-th row and j-th column of the first recursive matrix or the second recursive matrix The calculation is as follows: ； ； ； in, Indicates the calculation of the i-th high-dimensional vector and the jth high-dimensional vector angular distance; Represents the Heaviside function; Indicates the threshold distance.

7. The method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction according to claim 6, characterized in that: In step S3, when When , the elements on the i-th row and j-th column of the first recursive matrix or the second recursive matrix are set to black dots. When , the elements on the i-th row and j-th column of the first recursive matrix or the second recursive matrix are set to white dots to obtain the first recursive graph and the second recursive graph.

8. The method for detecting sticking of a regulating valve in a control loop based on phase space reconstruction according to claim 1, characterized in that: In step S5, the reference value of the viscosity statistical threshold is 0, and the reference value of the viscosity distribution threshold is 0.58.

Images