Online quantitative evaluation method for viscosity of intelligent control valve

Abstract

The invention discloses an online quantitative evaluation method for viscosity of an intelligent control valve, and belongs to the field of performance evaluation of control valves. The method comprises the following steps: segmenting a multi-cycle superposed OP-MV map into a plurality of OP-MV sub-maps formed by single cycles; four end points in each OP-MV subgraph are found by solving a nonlinear optimization problem with constraints; judging the position of phase trajectory end-to-end disconnection in each OP-MV subgraph; the head-tail disconnection position is avoided, and the inflection point of the time continuous period is searched by using the angle change; calculating the sizes of viscosity characteristic parameters S and J corresponding to each OP-MV subgraph by using the end point coordinates and the inflection point coordinates; and a dynamic change curve of the two viscosity characteristic parameters is obtained. From the perspective of online quantitative evaluation, interference of non-stationary factors can be effectively eliminated, meanwhile, viscosity characteristic parameters can be quantified online, time-varying characteristics of viscosity faults can be captured, and a method basis is provided for online monitoring and performance evaluation of the intelligent control valve.

Description

Technical Field

[0001] The present invention belongs to the field of control valve performance evaluation, and in particular relates to an online quantitative evaluation method for viscosity of an intelligent control valve. Background Art

[0002] Control valve sticking is one of the most significant valve failure modes, severely degrading control loop performance. The 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. Consequently, valve fault detection and diagnosis have garnered significant attention from researchers and industry professionals over the past two decades. To date, research on valve sticking detection and diagnosis has yielded fruitful results. Common methods for quantifying control valve sticking include the apparent sticking method (Choudhury, 2006), the Hammerstein-based method (Srinivasan, 2005), the nonlinear process model-based method (Wang, 2009), and hybrid methods (Zabiri, 2009). However, these methods assume that the valve position signal is unknown and that the valve sticking characteristic parameters remain constant over time during monitoring, thus failing to capture the time-varying characteristics of valve sticking. However, with the increasing prevalence of intelligent control valves, valve position signals can now be acquired by intelligent positioners. Therefore, most reported methods for valve sticking quantification may not be suitable for handling online quantitative evaluation of intelligent control valves. Summary of the Invention

[0003] The purpose of the present invention is to solve the problems existing in the prior art and provide an online quantitative evaluation method for viscosity of an intelligent control valve.

[0004] The specific technical solutions adopted in the present invention are as follows:

[0005] In a first aspect, the present invention provides an online quantitative evaluation method for viscosity of an intelligent control valve, as follows:

[0006] S1: Obtain the control signal OP, valve position signal MV and sampling frequency f of the target intelligent control valve s , a double-constrained period segmentation method is used to segment the multi-period superimposed OP-MV graph into multiple OP-MV subgraphs formed by single periods;

[0007] S2: By solving two constrained nonlinear optimization problems, the coordinates of the four endpoints, namely, the lower left corner endpoint M, the upper right corner endpoint G, the lower right corner endpoint B, and the upper left corner endpoint H, in each of the OP-MV subgraphs in S1 are obtained;

[0008] S3: Based on the result of S2, determine the position where the phase trajectory of each OP-MV subgraph is disconnected;

[0009] S4: Based on the results of S3, avoid the disconnection position of the phase trajectory and use the angle change to obtain the coordinates of the two inflection points, namely the lower right inflection point E and the upper left inflection point F, of the time continuous segment;

[0010] S5: Using the coordinates of the four endpoints obtained in S2 and the coordinates of the two inflection points obtained in S4, calculate the magnitudes of the two viscosity characteristic parameters, the sticky band and the dead zone S and the slip jump J, corresponding to each of the OP-MV subgraphs;

[0011] S6: Taking the middle moment of the time interval corresponding to each of the OP-MV subgraphs as the abscissa and the corresponding S and J as the ordinates, a dynamic change curve of the two viscosity characteristic parameters is obtained.

[0012] Preferably, in S1, a dual-constraint period segmentation method is specifically as follows:

[0013] S11: Construct a three-dimensional observation vector P(k) = [t(k), OP(k), MV(k)], k = 1, 2, 3...N; where t(k) is the sampling time, OP(k) is the control signal, MV(k) is the valve position signal, and N is the data length;

[0014] S12: Set the current data pointer k=1, the segment start pointer k start =1, first-category segment counter j=1, second-category segment counter i=1;

[0015] S13: If kk start ≤N min , then jump to S16; otherwise, execute S14;

[0016] S14: Calculate the cumulative phase θ acc ; If |θ acc |>2π, then the first type of segment Seg1(i)=[P(k start ); P(k start +1);…;P(k)], update i=i+1,k start =k+1, jump to S16; otherwise, execute S15;

[0017] S15: If kk start >N min , then generate the second type of segment Seg2(j)=[P(k start ); P(k start +1); ...; P(k)], update j = j + 1, k start =k+1, execute S16; otherwise, execute S16 directly;

[0018] S16: The pointer is incremented by k=k+1; if k>N, the program ends and outputs the segment set {Seg1, Seg2}; otherwise, jump to S13.

[0019] Furthermore, the N min and N max The calculation formula is as follows:

[0020]

[0021]

[0022] in, For rounding up operation, T mean is the average period maximum value of MV signal and OP signal, σ T is the maximum value of the period standard deviation of the MV signal and the OP signal, α is 0.5, and β is 1.5.

[0023] Furthermore, the average period and standard deviation of the MV signal and the OP signal are obtained through the autocorrelation function.

[0024] Furthermore, the accumulated phase θ acc The calculation formula is as follows:

[0025]

[0026] Among them, Δθ(n)=mod(θ(n)-θ(n-1)+π,2π)-π,

[0027] Preferably, in S2, the four endpoints are obtained by the following method:

[0028] S21: Assume that the observation data set is Among them, P(i)=[t(i),OP(i),MV(i)]∈R 3 ; Construct binary decision variables (i, j)∈Z 2 , establish a nonlinear optimization problem with constraints:

[0029]

[0030]

[0031] sign(OP(j)-OP(i))=1} Constraint 2

[0032]

[0033]

[0034] Among them, a=MV(m)-MV(g), b=OP(g)-OP(m), c=OP(m)MV(g)-OP(g)MV(m); m,g,h,b∈{k start ,k start +1,…,k};

[0035] S22: Among the point pairs that satisfy constraints 1 and 2, select the pair with the largest distance between the two points to obtain points M and G; where the i corresponding to the maximum distance between the two points is denoted as m, and the j corresponding to the maximum distance between the two points is denoted as g, then P(m) = [t(m), OP(m), MV(m)], P(g) = [t(g), OP(g), MV(g)];

[0036] S23: Among the point pairs that satisfy constraint 1, select the combination with the largest sum of the straight-line distances from the two points to MG to obtain points H and B; among them, the corresponding i when the sum of the distances is the largest is denoted as b, and the corresponding j is denoted as h, then P(h) = [t(h), OP(h), MV(h)], P(b) = [t(b), OP(b), MV(b)], thus obtaining the coordinates of the four endpoints.

[0037] Preferably, in S3, the judgment logic of the disconnection position of the phase track is as follows:

[0038] The four endpoints M, B, G and H correspond to the times t(m), t(b), t(g) and t(h) respectively. The disconnection area is located between the two points with the maximum and minimum time among the four endpoints, thus defining the extreme point

[0039]

[0040] Among them, argmax is the operator that determines the maximum value position, argmin is the operator that determines the minimum value position, and the disconnection area is located at p max With p min The index between the two points.

[0041] Preferably, in S4, the logic for determining two inflection points of a continuous time segment is as follows:

[0042] When the continuous segment is BG, the starting point is point B, P(b) = [t(b), OP(b), MV(b)], and the angle formed by the current point and the next point is calculated starting from point B, that is, Calculate the angle difference in sequence, that is, Δθ(i) = |θ(i+1)-θ(i)|; if Δθ(i)>π / 36, stop the calculation, let e=i+1, output the coordinates of point E, P(e)=[t(e),OP(e),MV(e)]; otherwise (that is, if Δθ(i)≤π / 36), i=i+1, recalculate θ(i) and Δθ(i), until Δθ(i)>π / 36 and stop the calculation.

[0043] When the continuous segment is HM, the starting point is point H, P(h) = [t(h), OP(h), MV(h)], and the angle formed by the current point and the next point is calculated in sequence from point H, that is, Calculate the angle difference in sequence, that is, Δθ(i) = |θ(i+1)-θ(i)|; if Δθ(i)>π / 36, stop the calculation, let f=i+1, output the coordinates of point F, P(f)=[t(f),OP(f),MV(f)]; otherwise (that is, if Δθ(i)≤π / 36), i=i+1, recalculate θ(i) and Δθ(i), until Δθ(i)>π / 36 and stop the calculation.

[0044] Preferably, in S5, sticking refers to the increase in the control signal from the time the valve stops moving to the time it starts moving due to the influence of static friction; dead zone refers to the change in the control signal when the valve position remains unchanged when the valve receives a change in the control signal; and sliding jump refers to the jump amount of the valve position when the valve overcomes static friction and moves suddenly.

[0045] Preferably, in S5, the calculation steps of the two viscosity characteristic parameters S and J are as follows:

[0046] S51: If p min =h and p max =g, then S=OP(b)-OP(m), J=[MV(f)-MV(h)+MV(e)-MV(b)] / 2,

[0047] Otherwise, execute S2;

[0048] S52: If p min =b and p max =m, then S=OP(g)-OP(h), J=[MV(f)-MV(h)+MV(e)-MV(b)] / 2,

[0049] Otherwise, execute S3;

[0050] S53: If p min =m and p max=h, then S=[OP(b)-OP(m)+OP(g)-OP(h)] / 2, J=MV(e)-MV(b), otherwise S=[OP(b)-OP(m)+OP(g)-OP(h)] / 2, J=MV(f)-MV(h).

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] From the perspective of online quantitative evaluation, the present invention can effectively eliminate the interference of non-stationary factors, and at the same time can quantify the viscosity characteristic parameters online and capture the time-varying characteristics of viscosity faults, providing a methodological basis for online monitoring and performance evaluation of intelligent control valves. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Schematic diagram of the process of the dual-constraint period segmentation method of the present invention;

[0054] Figure 2 are the four endpoints in the OP-MV graph obtained by the present invention through the constrained nonlinear optimization problem;

[0055] Figure 3 These are the four cases where the phase trajectory ends are disconnected in the OP-MV subgraph;

[0056] Figure 4 is an online quantization curve of the viscosity parameter S in an embodiment of the present invention;

[0057] Figure 5 is an online quantization curve of the viscosity parameter J in an embodiment of the present invention;

[0058] Figure 6 This is the input-output diagram (OP-MV diagram) of a typical control valve with sticking fault. DETAILED DESCRIPTION

[0059] 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.

[0060] In a preferred implementation of the present invention, the above-mentioned online quantitative evaluation method for viscosity of an intelligent control valve includes the following steps S1 to S6, the specific implementation processes of which are described in detail below.

[0061] S1: Obtain the control signal OP, valve position signal MV and sampling frequency f of the target intelligent control valve to be detected s , a double-constrained period segmentation method is used to segment the multi-period superimposed OP-MV graph into multiple OP-MV subgraphs formed by single periods.

[0062] As a preferred embodiment of the present invention, in this step, Figure 1 As shown, a dual-constraint period segmentation method has the following specific steps:

[0063] S11: Construct a three-dimensional observation vector P(k) = [t(k), OP(k), MV(k)], k = 1, 2, 3…N; where t(k) is the sampling time, OP(k) is the control signal, MV(k) is the valve position signal, and N is the data length.

[0064] S12: Initialize parameters: set the current data pointer k=1, the segment start pointer k start =1, first category segment counter j=1, second category segment counter i=1.

[0065] S13: Minimum length N min Verification: If kk start ≤N min , then jump to S16; otherwise, execute S14.

[0066] S14: Phase constraint judgment: calculate the cumulative phase θ acc ; If |θ acc |>2π, then the first type of segment Seg1(i)=[P(k start ); P(k start +1);…;P(k)], update i=i+1,k start =k+1, jump to S16; otherwise, execute S15.

[0067] S15: Maximum length N max Forced split: If kk start >N min , then generate the second type of segment Seg2(j)=[P(k start ); P(k start +1); ...; P(k)], update j = j + 1, k start =k+1, execute S16; otherwise, execute S16 directly.

[0068] S16: The pointer is incremented by k=k+1; if k>N, the program ends and outputs the segment set {Seg1, Seg2}; otherwise, jump to S13.

[0069] In the above step S1 of this embodiment, N min and Nmax The calculation formula is as follows:

[0070]

[0071]

[0072] in, For rounding up operation, T mean is the average period maximum value of MV signal and OP signal, σ T is the maximum value of the period standard deviation of the MV signal and the OP signal, α is 0.5, and β is 1.5.

[0073] In actual use, the average period and standard deviation of the MV signal and the OP signal can be solved by the autocorrelation function. The specific solution process is based on the literature "Thornhill NF, Huang B, Zhang H. Detection of multiple oscillations in control loops [J]. Journal of Process Control, 2003, 13 (1): 91-100.", which will not be repeated here.

[0074] In actual use, the cumulative phase θ acc The calculation formula is as follows:

[0075]

[0076] Among them, Δθ(n)=mod(θ(n)-θ(n-1)+π,2π)-π,

[0077] S2: By solving two constrained nonlinear optimization problems, the coordinates of the four endpoints in each OP-MV subgraph in S1, namely the lower left corner endpoint M, the upper right corner endpoint G, the lower right corner endpoint B and the upper left corner endpoint H, are found.

[0078] As a preferred embodiment of the present invention, in this step, Figure 2 As shown, the specific method of finding the four endpoints is as follows:

[0079] S21: Assume that the observation data set is Among them, P(i)=[t(i),OP(i),MV(i)]∈R 3 ; Construct binary decision variables (i, j)∈Z 2 , establish a nonlinear optimization problem with constraints:

[0080]

[0081]

[0082] sign(OP(j)-OP(i))=1} Constraint 2

[0083]

[0084]

[0085] Among them, a=MV(m)-MV(g), b=OP(g)-OP(m), c=OP(m)MV(g)-OP(g)MV(m); m,g,h,b∈{k start ,k start +1,…,k};

[0086] S22: Among the point pairs that satisfy constraints 1 and 2, select the pair with the largest distance between the two points to obtain points M and G; where the i corresponding to the maximum distance between the two points is denoted as m, and the j corresponding to the maximum distance between the two points is denoted as g, then P(m) = [t(m), OP(m), MV(m)], P(g) = [t(g), OP(g), MV(g)];

[0087] S23: Among the point pairs that satisfy constraint 1, select the combination with the largest sum of the straight-line distances from the two points to MG to obtain points H and B; among them, the corresponding i when the sum of the distances is the largest is denoted as b, and the corresponding j is denoted as h, then P(h) = [t(h), OP(h), MV(h)], P(b) = [t(b), OP(b), MV(b)], thus obtaining the coordinates of the four endpoints.

[0088] S3: Based on the results of S2, determine the location where the phase trajectory is disconnected at the beginning and end of each OP-MV subgraph, such as Figure 3 shown.

[0089] As a preferred embodiment of the present invention, in this step, the judgment logic of the disconnection position of the phase track is as follows:

[0090] The four endpoints M, B, G and H correspond to the times t(m), t(b), t(g) and t(h) respectively. The disconnection area is located between the two points with the maximum and minimum time among the four endpoints, thus defining the extreme point

[0091]

[0092] Among them, argmax is the operator that determines the maximum value position, argmin is the operator that determines the minimum value position, and the disconnection area is located at p max With p min The index between the two points.

[0093] S4: Based on the results of S3, avoid the disconnection positions at the beginning and end of the phase trajectory, and use the angle change to obtain the coordinates of the two inflection points, namely the lower right inflection point E and the upper left inflection point F, of the time continuous segment.

[0094] As a preferred embodiment of the present invention, in this step, the logic for finding the inflection point of the time continuous segment is as follows:

[0095] When the continuous segment is BG, the starting point is point B, P(b) = [t(b), OP(b), MV(b)], and the angle formed by the current point and the next point is calculated starting from point B, that is, Calculate the angle difference in sequence, that is, Δθ(i) = |θ(i+1)-θ(i)|; if Δθ(i)>π / 36, stop the calculation, let e=i+1, output the coordinates of point E, P(e)=[t(e),OP(e),MV(e)]; otherwise, i=i+1, recalculate θ(i) and Δθ(i), until Δθ(i)>π / 36 and stop the calculation.

[0096] When the continuous segment is HM, the starting point is point H, P(h) = [t(h), OP(h), MV(h)], and the angle formed by the current point and the next point is calculated in sequence from point H, that is, Calculate the angle difference in sequence, that is, Δθ(i) = |θ(i+1)-θ(i)|; if Δθ(i)>π / 36, stop the calculation, let f=i+1, output the coordinates of point F, P(f)=[t(f),OP(f),MV(f)]; otherwise, i=i+1, recalculate θ(i) and Δθ(i), until Δθ(i)>π / 36 and stop the calculation.

[0097] S5: Using the coordinates of the four endpoints obtained in S2 and the coordinates of the two inflection points obtained in S4, calculate the sizes of the two viscosity characteristic parameters, the sticky band and the dead zone S and the sliding jump J, corresponding to each OP-MV subgraph.

[0098] As a preferred embodiment of the present invention, in this step, sticking refers to the increase in the control signal from the time the valve stops moving to the time it starts moving due to the influence of static friction; dead zone refers to the change in the control signal when the valve receives a change in the control signal but its valve position remains unchanged; and sliding jump refers to the jump in the valve position when the valve overcomes static friction and suddenly moves. Figure 6 As shown, S=stickband+deadband (ie S=Stickband+Deadband), J=slip jump (ie J=Slip jump).

[0099] The calculation steps of the two viscous characteristic parameters S and J are as follows:

[0100] S51: If p min =h and p max=g, then S=OP(b)-OP(m), J=[MV(f)-MV(h)+MV(e)-MV(b)] / 2,

[0101] Otherwise, execute S2;

[0102] S52: If p min =b and p max =m, then S=OP(g)-OP(h), J=[MV(f)-MV(h)+MV(e)-MV(b)] / 2,

[0103] Otherwise, execute S3;

[0104] S53: If p min =m and p max =h, then S=[OP(b)-OP(m)+OP(g)-OP(h)] / 2, J=MV(e)-MV(b), otherwise S=[OP(b)-OP(m)+OP(g)-OP(h)] / 2, J=MV(f)-MV(h).

[0105] S6: The dynamic change curves of the two viscosity characteristic parameters are obtained by taking the middle moment of the time interval corresponding to each OP-MV subgraph as the horizontal coordinate and the corresponding S and J as the vertical coordinates.

[0106] In order to better demonstrate the specific implementation and technical effects of the present invention, an online quantitative evaluation method for viscosity of an intelligent control valve shown in steps S1 to S6 in the above preferred implementation is applied to a specific example.

[0107] Example

[0108] The specific implementation process of the online quantitative evaluation method for viscosity of an intelligent control valve used in this embodiment is as described above and will not be repeated here.

[0109] The following example uses a typical single-input single-output (SISO) system. This system is used to generate simulation data in feedback control and is a recognized benchmark data set in viscosity research. The controlled process transfer function is as follows:

[0110]

[0111] The process is operated under the regulatory control of a PI controller, G c(s) = 0.15 + 0.15 / s. The valve hysteresis model uses the Choudhury model. The controlled process transfer function, PID parameters, and valve hysteresis model of this system are derived from the literature MAASChoudhury, SLShah, NFThornhill, Diagnosis of poor control-loop performance using higher-order statistics, Automatica 40(10)(2004) 1719–1728.

[0112] In this embodiment, the system is set to run for 2000s, and the signal sampling frequency is 1Hz. From 0 to 2000s, S increases uniformly from 1 to 2, and J increases uniformly from 0.2 to 0.5. In the system set up above, the S and J sizes are quantified online by the collected OP and MV signals, as shown in Figure 2. Figure 4 and 5 shown. Figure 4 The horizontal axis is time, the vertical axis is S, the solid line in the figure is the estimated value of S, and the dotted line is the true value of S. Figure 5 The horizontal axis is time, the vertical axis is J, the solid line is the estimated value of J, and the dotted line is the true value of J. Figure 4 and Figure 5 As can be seen from the above, this method can effectively capture the time-varying characteristics of the true viscosity parameters S and J and realize the function of online quantification. This embodiment demonstrates the advancement and effectiveness of the method of the present invention.

[0113] From the perspective of online quantitative evaluation, the present invention can effectively eliminate the interference of non-stationary factors, and at the same time can quantify the viscosity characteristic parameters online and capture the time-varying characteristics of viscosity faults, providing a methodological basis for online monitoring and performance evaluation of intelligent control valves.

[0114] 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. An online quantitative evaluation method for viscosity of intelligent control valves, characterized in that: The details are as follows: S1: Obtain the control signal OP, valve position signal MV and sampling frequency f of the target intelligent control valve s , a double-constrained period segmentation method is used to segment the multi-period superimposed OP-MV graph into multiple OP-MV subgraphs formed by single periods; S2: By solving two constrained nonlinear optimization problems, the coordinates of the four endpoints, namely, the lower left corner endpoint M, the upper right corner endpoint G, the lower right corner endpoint B, and the upper left corner endpoint H, in each of the OP-MV subgraphs in S1 are obtained; S3: Based on the result of S2, determine the position where the phase trajectory of each OP-MV subgraph is disconnected; S4: Based on the results of S3, avoid the disconnection position of the phase trajectory and use the angle change to obtain the coordinates of the two inflection points, namely the lower right inflection point E and the upper left inflection point F, of the time continuous segment; S5: Using the coordinates of the four endpoints obtained in S2 and the coordinates of the two inflection points obtained in S4, calculate the magnitudes of the two viscosity characteristic parameters, the sticky band and the dead zone S and the slip jump J, corresponding to each of the OP-MV subgraphs; S6: Taking the middle moment of the time interval corresponding to each of the OP-MV subgraphs as the abscissa and the corresponding S and J as the ordinates, a dynamic change curve of the two viscosity characteristic parameters is obtained.

2. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 1 is characterized in that: In S1, a dual-constraint period segmentation method is specifically as follows: S11: Construct a three-dimensional observation vector P(k) = [t(k), OP(k), MV(k)], k = 1, 2, 3...N; where t(k) is the sampling time, OP(k) is the control signal, MV(k) is the valve position signal, and N is the data length; S12: Set the current data pointer k=1, the segment start pointer k start =1, first-category segment counter j=1, second-category segment counter i=1; S13: If kk start ≤N min , then jump to S16; otherwise, execute S14; S14: Calculate the cumulative phase θ acc ; If |θ acc |>2π, then the first type of segment Seg1(i)=[P(k start ); P(k start +1);…;P(k)], update i=i+1,k start =k+1, jump to S16; otherwise, execute S15; S15: If kk start >N min , then generate the second type of segment Seg2(j)=[P(k start ); P(k start +1); ...; P(k)], update j = j + 1, k start =k+1, execute S16; otherwise, execute S16 directly; S16: The pointer is incremented by k=k+1; if k>N, the program ends and outputs the segment set {Seg1, Seg2}; otherwise, jump to S13.

3. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 2 is characterized in that: The N min and N max The calculation formula is as follows: in, For rounding up operation, T mean is the average period maximum value of MV signal and OP signal, σ T is the maximum value of the period standard deviation of the MV signal and the OP signal, α is 0.5, and β is 1.

5.

4. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 3 is characterized in that: The average period and standard deviation of the MV signal and OP signal are obtained through the autocorrelation function.

5. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 2, characterized in that: The accumulated phase θ acc The calculation formula is as follows: Where, Δθ(n) = mod(θ(n)-θ(n-1)+π,2π)-π, 6. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 1 is characterized in that: In S2, the four endpoints are obtained by the following method: S21: Assume that the observation data set is Among them, P(i)=[t(i),OP(i),MV(i)]∈R 3 ; Construct binary decision variables (i, j)∈Z 2 , establish a nonlinear optimization problem with constraints: sign(OP(j)-OP(i))=1} Constraint 2 where a = MV(m) - MV(g), b = OP(g) - OP(m), c = OP(m)MV(g) - OP(g)MV(m); m, g, h, b ∈ {k start , k start + 1, …, k}; S22: Among the point pairs that satisfy constraints 1 and 2, select the pair with the largest distance between the two points to obtain points M and G; where the i corresponding to the maximum distance between the two points is denoted as m, and the j corresponding to the maximum distance between the two points is denoted as g, then P(m) = [t(m), OP(m), MV(m)], P(g) = [t(g), OP(g), MV(g)]; S23: Among the point pairs that satisfy constraint 1, select the combination with the largest sum of the straight-line distances from the two points to MG to obtain points H and B; among them, the corresponding i when the sum of the distances is the largest is denoted as b, and the corresponding j is denoted as h, then P(h) = [t(h), OP(h), MV(h)], P(b) = [t(b), OP(b), MV(b)], thus obtaining the coordinates of the four endpoints.

7. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 1 is characterized in that: In S3, the judgment logic of the disconnection position of the phase track is as follows: The four endpoints M, B, G and H correspond to the times t(m), t(b), t(g) and t(h) respectively. The disconnection area is located between the two points with the maximum and minimum time among the four endpoints, thus defining the extreme point Among them, argmax is the operator that determines the maximum value position, argmin is the operator that determines the minimum value position, and the disconnection area is located at p max With p min The index between the two points.

8. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 1 is characterized in that: In S4, the logic for determining the two inflection points of a continuous time segment is as follows: When the continuous segment is BG, the starting point is point B, P(b) = [t(b), OP(b), MV(b)], and the angle formed by the current point and the next point is calculated starting from point B, that is, Calculate the angle difference in sequence, i.e., Δθ(i) = |θ(i+1)-θ(i)|; if Δθ(i) > π / 36, stop the calculation, set e = i+1, and output the coordinates of point E, P(e) = [t(e), OP(e), MV(e)]; Otherwise, i=i+1, recalculate θ(i) and Δθ(i) until Δθ(i)>π / 36 and stop the calculation; When the continuous segment is HM, the starting point is point H, P(h) = [t(h), OP(h), MV(h)], and the angle formed by the current point and the next point is calculated in sequence from point H, that is, Calculate the angle difference in sequence, that is, Δθ(i) = |θ(i+1)-θ(i)|; if Δθ(i)>π / 36, stop the calculation, let f=i+1, output the coordinates of point F, P(f)=[t(f),OP(f),MV(f)]; otherwise, i=i+1, recalculate θ(i) and Δθ(i), until Δθ(i)>π / 36 and stop the calculation.

9. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 1, characterized in that: In S5, sticking refers to the increase in the control signal from the time the valve stops moving to the time it starts moving due to the influence of static friction; dead zone refers to the change in the control signal when the valve position remains unchanged when the valve receives a change in the control signal; and slip refers to the jump in the valve position when the valve suddenly moves to overcome static friction.

10. The online quantitative evaluation method for viscosity of an intelligent control valve according to claim 1, characterized in that: In S5, the calculation steps of the two viscosity characteristic parameters S and J are as follows: S51: If p min =h and p max =g, then S=OP(b)-OP(m), J=[MV(f)-MV(h)+MV(e)-MV(b)] / 2, Otherwise, execute S2; S52: If p min =b and p max =m, then S=OP(g)-OP(h), J=[MV(f)-MV(h)+MV(e)-MV(b)] / 2, Otherwise, execute S3; S53: If p min = m and p max = h, then S = [OP(b) - OP(m) + OP(g) - OP(h)] / 2, J = MV(e) - MV(b), otherwise S=[OP(b)-OP(m)+OP(g)-OP(h)] / 2, J=MV(f)-MV(h).

Images