A dynamic response time-based modeling method for stick-slip fault of pneumatic control valve

Abstract

The application discloses a kind of dynamic response time-based pneumatic regulating valve stick fault modeling method, it is related to valve fault analysis field.The stick fault modeling method of the application is five-parameter data-driven model, parameters include: dimensionless maximum static friction fs, dimensionless dynamic friction fd, response coefficient RF, inflation delay coefficient Da, deflation delay coefficient Db.This model considers the influence of the relationship between actual control signal transmission frequency and valve dynamic response time on valve stick behavior, and uses delay coefficient to simulate the law of control signal OP and gas film pressure when lower gas source pressure or poor performance of positioner, more comprehensively and accurately realizes the stick fault modeling of pneumatic regulating valve, makes up the technical deficiency of existing stick model based on data-driven, provides model basis for subsequent valve stick fault diagnosis, analysis and compensation.

Description

Technical Field

[0001] This invention belongs to the field of valve fault analysis, and specifically relates to a method for modeling viscous faults in pneumatic control valves based on dynamic response time. Background Technology

[0002] Sticky failures in pneumatic control valves are common in process control systems, typically leading to control loop oscillations, impacting product quality, increasing energy consumption, and accelerating equipment wear. Therefore, studying the mechanism of sticky failures and establishing reliable and accurate sticky models is crucial. Current research largely utilizes data-driven methods, requiring only simple parameters to describe sticky behavior. However, some models are either incomplete or difficult to understand. Most current models do not consider the impact of the valve's dynamic response time on the sticky behavior of pneumatic control valves. Whether static friction occurs in each sampling interval is a gray area in sticky models, and most models ignore the delay effect between the control signal and film pressure, assuming a simple linear relationship. However, nonlinearity often occurs when the positioner performance is poor or the air source pressure is low. Therefore, how to simply and effectively simulate the nonlinearity of the control signal and film pressure, and consider the impact of the valve's dynamic response time on the valve's sticky behavior for sticky failure modeling, is a problem that needs to be solved. Summary of the Invention

[0003] The purpose of this invention is to overcome the defects in the prior art and to provide a method for modeling viscous faults in pneumatic control valves based on dynamic response time.

[0004] The specific technical solution adopted in this invention is as follows:

[0005] This invention provides a method for modeling viscous faults in pneumatic control valves based on dynamic response time, specifically including the following steps:

[0006] Step 1: Determine the control signal input to the valve positioner;

[0007] Step 2: Set the magnitude of the viscous fault, the dynamic response time of the valve, and the degree of delay when the control signal is converted into film pressure during charging and discharging;

[0008] Step 3: Based on the delay in the conversion of the control signal into air film pressure during inflation and deflation in Step 2, simulate the process of the control signal being converted into air film pressure through the positioner in Step 1, and output the air film pressure and the corresponding time.

[0009] Step 4: Input the gas film pressure at different times obtained from Step 3, and obtain the valve position at the corresponding time according to the magnitude of the viscous fault and the dynamic response time in Step 2.

[0010] Preferably, to facilitate data calculation, the control signal, air film pressure, and valve position should all be normalized to 0-100%.

[0011] Preferably, in step 1, the control signal input from the controller to the valve positioner needs to have its transmission interval ts(s) or transmission frequency f(Hz) and time sequence u(k) determined.

[0012] ts=1 / f (1)

[0013] {u(1), u(2), …, u(k), …, u(kmax)},k=1,2,3…kmax (2)

[0014] Where kmax is the length of a set of time series data; k is a positive integer representing the number of sampling points in the time direction of the time series u(k), and the sampling interval is ts.

[0015] Preferably, in step 2, the magnitude of the viscous fault is represented by dimensionless static friction force fs and dynamic friction force fd, and has the following relationship with the viscosity S and the jump variable J:

[0016]

[0017]

[0018] Preferably, in step 3, since the air film pressure cannot guarantee a good linear relationship with the control signal when the air source pressure is low or the positioner performance is poor, the air film pressure will have a certain delay in reaching the set signal. The control signal is transformed into air film pressure by using inflation delay coefficient Da and deflation delay coefficient Db. The air film pressure is represented by a time series U(n), and T(n) is the cumulative delay corresponding to the air film pressure U(n). The specific steps are as follows:

[0019] Step 301: Begin by inputting a positive integer k = 1;

[0020] Step 302: Determine if the positive integer k is 1;

[0021] If k equals 1, then initial values ​​are assigned to each parameter. The initial value of the delay coefficient D(k) corresponding to the time sequence of the control signal is u(k) is determined by judging the control signal. If the control signal is on an upward trend, then let D(1) = Da, and then proceed to step 304. If the control signal is on a downward trend, then let D(1) = Db, and then proceed to step 304. If the control signal remains unchanged, then continue to compare the magnitudes of the subsequent control signals.

[0022] If k is not equal to 1, assign a value to the delay coefficient D(k) at this moment according to the following rules: if the control signal is on an upward trend, let D(k) = Da; if the control signal is on a downward trend, let D(k) = Db; if the control signal remains unchanged, let D(k) = D(k-1).

[0023] Step 303: Determine whether the film pressure remains constant at this moment. If the difference between the time accumulation k-1 and the delay accumulation n1·Da+n2·Db of the control signal is less than the current delay coefficient D(k), then the film pressure remains constant, and proceed to step 304. Otherwise, the film pressure jumps to the set signal, and the film pressure is recorded as delayed once. If D(k) = Da, it means that the film pressure is delayed once during the inflation process, and the number of delays during the inflation process is n1 = n1 + 1. Otherwise, it means that the film pressure is delayed once during the deflation process, and the number of delays during the deflation process is n2 = n2 + 1. Only the film pressure after the jump is recorded, denoted as U(n1+n2+1), whose value corresponds to the magnitude of the set signal at time (n1·Da+n2·Db)·ts, which is approximated by the linear difference between the control signals u(k) and u(k-1). The corresponding cumulative delay is denoted as T(n1+n2+1).

[0024] U(n)=u(k-1)+(n1·Da+n2·Db-k+2)·[u(k)-u(k-1)] (4)

[0025] T(n)=n1·Da+n2·Db (5)

[0026] n = n1 + n2 + 1 (6)

[0027] Step 304: Let k = k + 1, and compare it with the length of the time series data corresponding to the control signal, kma. If k > kma, the loop ends; otherwise, repeat steps 302 and 303.

[0028] Preferably, in step 4, the valve motion state is determined by comparing the current delay coefficient D(n) with the response coefficient RF, the valve status indicator stop, and the cumulative force cum_U(n) with the frictional force. The specific steps are as follows:

[0029] Step 401: Start inputting the air film pressure and the corresponding cumulative delay: [T(n), U(n)];

[0030] Step 402: Calculate the cumulative force cum_U(n), the motion direction indicator d(n), and the delay coefficient D(n):

[0031] cum_U(n)=cum_U(n-1)+U(n)-U(n-1) (7)

[0032] d(n)=sgn(cum_U(n)) (8)

[0033] D(n)=T(n)-T(n-1) (9)

[0034] Where n represents the number of sampling points in the time direction of the time series U(n), the sampling interval for the inflation process is Da·ts, and the sampling interval for the deflation process is Db·ts.

[0035] Step 403: Determine the magnitudes of the delay coefficient D(n) and the response coefficient RF. The magnitude of the response coefficient RF is the ratio of the valve's dynamic response time Td to the transmission interval ts of the control signal. If D(n) > RF, it means that the valve stops moving before the next gas film pressure signal arrives, and the valve's movement state can only switch between state 1 (sticky to sliding) and state 2 (remaining sticky). Conversely, it means that the valve moves slowly and is still moving before the next gas film pressure signal arrives, and the valve's movement state switches between state 1 (sticky to sliding), state 2 (remaining sticky), state 3 (sliding to sticky), and state 4 (remaining sliding).

[0036] Step 404: When switching from state 1 or state 2 to state 3 or state 4, and when switching from state 3 or state 4 to state 1 or state 2, the valve state indicator stop needs to be used to determine whether the valve stopped at the previous moment; if stop = 1, then the valve stopped at the previous moment, and the valve state is state 1 or state 2; otherwise, the valve moved at the previous moment, and the valve state is state 3 or state 4.

[0037] Step 405: The switching between state 1 and state 2 is determined by comparing the absolute value of the cumulative force cum_U(n) with the magnitude of the dimensionless static friction force fs. If cum_U(n) > fs, the valve is in state 1; otherwise, it is in state 2. The switching between state 3 and state 4 is determined by comparing the sign of fd, the absolute value of the cumulative force cum_U(n) with the magnitude of fd, and the direction of motion indicator d(n). For the valve to be in state 3, one of the following three conditions must be met:

[0038] (1) When fd>0 and |cum_U(n)|>fd, it indicates that there is undercompensated (S>J) frictional force, and the absolute value of the cumulative force cum_U(n) is greater than the static friction at this moment;

[0039] (2) When fd = 0 and d(n)d(n-1) > 0, it indicates that there is uncompensated friction (S = J) and the direction of motion remains unchanged;

[0040] (3) When fd<0 and |cum_U(n)|<-fd, it indicates that there is an over-compensated friction force (S<J), and the absolute value of the cumulative acting force cum_U(n) is less than the absolute value of the static friction force;

[0041] Step 406: Calculate the displacement of each state and related intermediate quantities:

[0042] The state 1 is specifically as follows:

[0043]

[0044] The state 2 and state 3 are specifically as follows:

[0045]

[0046] The state 4 is specifically as follows:

[0047]

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention considers the influence of the relative magnitude between the actual control signal transmission frequency and the valve dynamic response time on the valve stiction behavior, and uses a delay coefficient to simulate the nonlinearity between the control signal and the air film pressure when the air source pressure is low or the positioner performance is poor, so as to more comprehensively and accurately realize the stiction fault modeling of the pneumatic control valve, make up for the technical deficiencies of the existing data-driven stiction model, and provide a model basis for subsequent diagnosis, analysis and compensation of valve stiction faults. Brief Description of the Drawings

[0050] Figure 1 is a flowchart of the conversion algorithm between the control signal and the air film pressure;

[0051] Figure 2 is a flowchart of the stiction model based on the valve dynamic response time;

[0052] Figure 3 is the input-output behavior of a typical valve with stiction fault;

[0053] Figure 4 is a set triangular wave signal;

[0054] Figure 5 is a partial enlargement of the set signal, OP signal and air film pressure signal in region I;

[0055] Figure 6 is a partial enlargement of the set signal, OP signal and air film pressure signal in region II;

[0056] Figure 7 is a displacement-time curve;

[0057] Figure 8 This is the displacement-pressure signal curve. Detailed Implementation

[0058] The present invention will be further described and illustrated below with reference to the accompanying drawings and specific embodiments. The technical features of each embodiment of the present invention can be combined accordingly, provided that there is no mutual conflict.

[0059] like Figure 1 and 2 As shown, this invention provides a method for modeling viscous faults in pneumatic control valves based on dynamic response time. The method specifically includes the following steps:

[0060] Step 1: Determine the control signal input to the valve positioner.

[0061] Step 2: Set the magnitude of the viscous fault, the dynamic response time of the valve, and the delay in the conversion of the control signal into gas film pressure during charging and discharging.

[0062] Step 3: Simulate the process of the control signal being converted into air film pressure through the positioner based on the delay when the control signal is converted into air film pressure during inflation and deflation, and output the air film pressure and the corresponding time.

[0063] Step 4: Input the gas film pressure at different times obtained from Step 3, and obtain the valve position at the corresponding time according to the magnitude of the viscous fault and the dynamic response time.

[0064] In practical applications, to facilitate data calculation, the control signal, air film pressure, and valve position all need to be normalized to 0-100%.

[0065] In practical applications, in step 1, the control signal input from the controller to the valve positioner needs to have its transmission interval ts(s) or transmission frequency f(Hz) and time sequence u(k) determined:

[0066] ts=1 / f (1)

[0067] {u(1), u(2), …, u(k), …, u(kmax)},k=1,2,3…kmax (2)

[0068] Where kmax is the length of a set of time series data; k is a positive integer representing the number of sampling points in the time direction of the time series u(k), and the sampling interval is ts.

[0069] In practical applications, in step 2, the magnitude of the viscous fault is represented by the dimensionless static friction force fs and kinetic friction force fd, and has the following relationship with the viscosity S and the jump variable J: Figure 3 As shown:

[0070]

[0071]

[0072] In practical applications, in step 3, due to low air source pressure or poor positioner performance, the film pressure cannot guarantee a good linear relationship with the control signal. The film pressure will arrive at the set signal with a certain delay. The conversion of the control signal into film pressure is simulated using inflation delay coefficients Da and Db. The film pressure is represented by a time series U(n), where T(n) is the cumulative delay corresponding to the film pressure U(n). The specific steps are as follows: Figure 1 As shown:

[0073] Step 301: Start by inputting a positive integer k = 1.

[0074] Step 302: Determine if the positive integer k is 1. If it is, assign initial values ​​to each parameter. The initial value of the delay coefficient D(k) corresponding to the control signal u(k) needs to be determined by whether the control signal is rising or falling. If it is rising, then D(1) = Da, and jump to step 304; if it is falling, then D(1) = Db, and jump to step 304; if it remains unchanged, continue to compare the magnitude of the subsequent control signals. If the positive integer k is not equal to 1, assign a value to the delay coefficient D(k) at this moment. If the control signal is rising, then D(k) = Da; if it is falling, then D(k) = Db; if it remains unchanged, then D(k) = D(k-1).

[0075] Step 303: Determine whether the film pressure remains constant at this moment. If the difference between the time accumulation k-1 and the delay accumulation n1·Da+n2·Db of the control signal is less than the current delay coefficient D(k), then the film pressure remains constant, and proceed to step 304; otherwise, the film pressure jumps to the set signal, and the film pressure is recorded as delayed once. If D(k) = Da, it means that the film pressure is delayed once during the inflation process, and the number of delays during the inflation process is n1 = n1 + 1; otherwise, it means that the film pressure is delayed once during the deflation process, and the number of delays during the deflation process is n2 = n2 + 1. Only the film pressure after the jump is recorded, denoted as U(n1+n2+1), whose value corresponds to the magnitude of the set signal at time (n1·Da+n2·Db)·ts, which can be approximated by the linear difference between the control signals u(k) and u(k-1). The corresponding cumulative delay is denoted as T(n1+n2+1).

[0076] U(n)=u(k-1)+(n1·Da+n2·Db-k+2)·[u(k)-u(k-1)] (4)

[0077] T(n)=n1·Da+n2·Db (5)

[0078] n = n1 + n2 + 1 (6)

[0079] Step 304: k = k + 1, and compare it with the length of the time series data corresponding to the control signal, kma. If k > kma, the loop ends; otherwise, repeat steps 302 and 303.

[0080] In practical applications, step 4 determines the valve's motion state by comparing the current delay coefficient D(n) with the response coefficient RF, the valve status indicator stop, and the cumulative force cum_U(n) with the frictional force. The specific steps are as follows: Figure 4 As shown:

[0081] Step 401: Start inputting the air film pressure and the corresponding cumulative delay: [T(n), U(n)].

[0082] Step 402: Calculate the cumulative force cum_U(n), the motion direction indicator d(n), and the delay coefficient D(n):

[0083] cum_U(n)=cum_U(n-1)+U(n)-U(n-1) (7)

[0084] d(n)=sgn(cum_U(n)) (8)

[0085] D(n)=T(n)-T(n-1) (9)

[0086] Step 403: Determine the magnitudes of the delay coefficient D(n) and the response coefficient RF. The magnitude of the response coefficient RF is the ratio of the valve's dynamic response time Td to the transmission interval ts of the control signal. If D(n) > RF, it means that the valve stops moving before the next gas film pressure signal arrives, and the valve's movement state can only switch between state 1 (sticky to sliding) and state 2 (remaining sticky). Conversely, it means that the valve moves slowly and is still moving before the next gas film pressure signal arrives, and the valve's movement state switches between state 1 (sticky to sliding), state 2 (remaining sticky), state 3 (sliding to sticky), and state 4 (remaining sliding).

[0087] Step 404: Switching from state 1 or state 2 to state 3 or state 4 (switching from state 3 or state 4 to state 1 or state 2) requires using the valve state indicator stop to determine whether the valve stopped at the previous moment. If stop = 1, then the valve stopped at the previous moment, and the valve state is state 1 or state 2; otherwise, the valve was moving at the previous moment, and the valve state is state 3 or state 4.

[0088] Step 405: The switching between state 1 and state 2 is performed by judging the magnitude of the absolute value of the cumulative acting force cum_U(n) and the dimensionless static friction force fs. If cum_U(n) > fs, the valve state is state 1, otherwise it is state 2; the switching between state 3 and state 4 is performed by judging the sign of fd, the magnitude of the absolute value of the cumulative acting force cum_U(n) and fd, and the motion direction identifier d(n). The valve state is state 3, which satisfies any one of the following three conditions:

[0089] (1) When fd>0 and |cum_U(n)|>fd. It indicates that there is an under-compensated (S>J) friction force, and the absolute value of the cumulative acting force cum_U(n) is greater than the static friction at this moment.

[0090] (2) When fd=0 and d(n)d(n-1)>0. It indicates that there is a non-compensated (S=J) friction force, and the motion direction remains unchanged.

[0091] (3) When fd<0 and |cum_U(n)|<-fd. It indicates that there is an over-compensated (S<J) friction force, and the absolute value of the cumulative acting force cum_U(n) is less than the absolute value of the static friction.

[0092] Step 406: Calculate the displacement of each state and related intermediate quantities

[0093] State 1:

[0094] State 2, state 3:

[0095] State 4:

[0096] The above method will be applied to a specific embodiment below to illustrate the specific implementation manner and effects of the present invention.

[0097] Embodiment

[0098] This embodiment provides a method for modeling stiction fault of a pneumatic control valve based on dynamic response time, which is specifically as follows:

[0099] In this embodiment, a triangular wave with a period of 200s is used as the set signal, as Figure 4 shown, the transmission interval ts of the control signal is 0.3s, the inflation delay coefficient Da is 2.3, the deflation delay coefficient Db is 1.3, the dimensionless static friction force fs is 8, the dynamic friction force fd is 4, and the dynamic response time Td is 0.5s.

[0100] Step 1: Generate a triangular wave with a period of 200s, a transmission interval of 0.3s, and a range of 0-100% as the control signal time series.

[0101] Step Two: Through, for example Figure 1 The flowchart of the control signal-film pressure conversion algorithm shown yields [T(n)U(n)]. Plotting T(n)·ts on the x-axis and U(n) on the y-axis, the curve of the film pressure signal changing over time can be obtained. This realizes the time delay effect of the film pressure relative to the control signal during the two different stages of inflation and deflation. The interval of the film pressure during inflation is 2.3 times the control signal transmission interval, and the interval of the film pressure during deflation is 1.3 times the control signal transmission interval. Specifically... Figure 5 and 6 As shown.

[0102] Step 3: Through, for example Figure 2 The flowchart shown is based on the viscous model of the valve's dynamic response time. The displacement under viscous conditions is obtained. Plotting T(n)·ts on the x-axis and y(n) on the y-axis, the displacement versus time curve is obtained. During the inflation process, the interval between changes in the gas film pressure is greater than the valve's dynamic response time, resulting in multiple instances of significant jamming. However, during the deflation process, the interval between changes in the gas film pressure is less than the valve's dynamic response time, causing the valve to continue moving with each change in gas film pressure, exhibiting continuous sliding. This curve reflects the different valve motion characteristics during the inflation and deflation processes under viscous faults. Figure 7 As shown.

[0103] Step 4: By plotting U(n) on the x-axis and y(n) on the y-axis, the displacement versus gas film pressure curve can be obtained. This reveals the differences in viscous behavior during the valve's charging and discharging process, and intuitively demonstrates the influence of the valve's dynamic response time on its viscous characteristics. Figure 8 As shown.

[0104] This invention considers the influence of the relative magnitude of the actual control signal transmission frequency and the valve dynamic response time on the valve's sticky behavior, and uses a delay coefficient to simulate the nonlinearity of the simulated control signal and air film pressure when the air source pressure is low or the positioner performance is poor. This invention achieves a more comprehensive and accurate modeling of sticky faults in pneumatic control valves, making up for the technical deficiencies of existing data-driven sticky models, and providing a model foundation for the subsequent diagnosis, analysis and compensation of valve sticky faults.

[0105] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, all technical solutions obtained through equivalent substitution or transformation fall within the protection scope of the present invention.

Claims

1. A method for modeling viscous faults in pneumatic control valves based on dynamic response time, characterized in that, Specifically, the following steps are included: Step 1: Determine the control signal input to the valve positioner; Step 2: Set the magnitude of the viscous fault, the dynamic response time of the valve, and the degree of delay when the control signal is converted into film pressure during charging and discharging; The magnitude of the viscous fault is represented by the dimensionless static friction force fs and dynamic friction force fd, and has the following relationship with the viscosity S and the jump variable J: （3）； （4）； The degree of delay is defined as the inflation delay coefficient Da and the deflation delay coefficient Db. Step 3: Based on the delay in the conversion of the control signal into air film pressure during inflation and deflation in Step 2, simulate the process of the control signal being converted into air film pressure through the positioner in Step 1, and output the air film pressure and the corresponding time. The air film pressure is represented by a time series U(n), where T(n) is the cumulative delay corresponding to the air film pressure U(n). Step 4: Input the gas film pressure at different times obtained from Step 3, and obtain the valve position at the corresponding time according to the magnitude of the viscous fault and the dynamic response time in Step 2; Specifically, the valve's motion state is determined by comparing the current delay coefficient D(n) with the response coefficient RF, the valve status indicator stop, and the cumulative force cum_U(n) with the frictional force. The specific steps are as follows: Step 401: Start inputting the air film pressure and the corresponding cumulative delay: [T(n), U(n)]; Step 402: Calculate the cumulative force cum_U(n), the motion direction indicator d(n), and the delay coefficient D(n): cum_U(n)=cum_U(n-1)+U(n)-U(n-1) (7); d(n)=sgn(cum_U(n)) (8; D(n) = T(n) - T(n-1) (9; Where n represents the number of sampling points in the time direction of the time series U(n), the sampling interval for the inflation process is Da·ts, and the sampling interval for the deflation process is Db·ts. Step 403: Determine the magnitudes of the delay coefficient D(n) and the response coefficient RF. The response coefficient RF is the ratio of the valve's dynamic response time Td to the transmission interval ts of the control signal. If D(n) > RF, it means that the valve stops moving before the next film pressure signal arrives, and the valve's movement state can only switch between state 1 and state 2. Conversely, it means that the valve moves slowly and is still moving before the next film pressure signal arrives, and the valve's movement state switches between state 1, state 2, state 3, and state 4. State 1 represents the transition from viscous to sliding state, state 2 represents the viscous state, state 3 represents the transition from sliding to viscous state, and state 4 represents the sliding state. Step 404: When switching from state 1 or state 2 to state 3 or state 4, and when switching from state 3 or state 4 to state 1 or state 2, the valve status indicator stop needs to be used to determine whether the valve stopped at the previous moment; if stop=1, the valve stopped at the previous moment, and the valve status is state 1 or state 2; otherwise, the valve moved at the previous moment, and the valve status is state 3 or state 4. Step 405: The switching between state 1 and state 2 is performed by judging the magnitude relationship between the absolute value of the cumulative acting force cum_U(n) and the dimensionless static friction force fs. If cum_U(n) > fs, the valve state is state 1; otherwise, the valve state is state 2. The switching between state 3 and state 4 is confirmed by judging the sign of fd, the magnitude relationship between the absolute value of the cumulative acting force cum_U(n) and fd, and the motion direction identifier d(n). The valve state is state 3 when any one of the following three conditions is satisfied: (1) When fd > 0 and |cum_U(n)| > fd; it indicates that there is under-compensated friction force, that is, S > J, and the absolute value of the cumulative acting force cum_U(n) is greater than the current static friction force; (2) When fd = 0 and d(n)d(n-1) > 0; it indicates that there is uncompensated friction force, that is, S = J, and the motion direction remains unchanged; (3) When fd < 0 and |cum_U(n)| < -fd; it indicates that there is over-compensated friction force, that is, S < J, and the absolute value of the cumulative acting force cum_U(n) is less than the absolute value of the static friction force; Step 406: Calculate the displacement y(n) of each state and related intermediate quantities: The state 1 is specifically as follows: cum_U(n)=d(n)·fd y(n)=U(n)-cum_U(n) Stop=0 (10); The state 2 and state 3 are specifically as follows: y(n)=y(n-1) Stop=1 (11); The state 4 is specifically as follows: cum_U(n)=d(n)·|fd| y(n)=U(n)-cum_U(n) Stop=0 (12).

2. The method for modeling viscous faults in pneumatic control valves based on dynamic response time according to claim 1, characterized in that, The control signal, air film pressure and valve position are all normalized to 0~100%.

3. The method for modeling viscous faults in pneumatic control valves based on dynamic response time according to claim 1, characterized in that, In the step 1, for the control signal input from the controller to the valve positioner, it is necessary to determine the transmission interval ts or transmission frequency f of the control signal and the time sequence u(k) of the control signal: ts=1 / f (1); {u(1), u(2), …, u(k), …, u(kmax)},k=1,2,3…kmax (2); Wherein, kmax is the length of a set of time sequence data, the unit of the transmission interval ts is s, and the unit of the transmission frequency f is HZ; k is a positive integer, representing the sampling point number in the time direction of the time sequence u(k), and the sampling interval is ts.

4. The method for modeling viscous faults in pneumatic control valves based on dynamic response time according to claim 3, characterized in that, The specific steps of the step 3 are as follows: Step 301: Start to input a positive integer k=1; Step 302: Judge whether the positive integer k is 1; If k is equal to 1, initial values are assigned to each parameter, wherein the initial value of the delay coefficient D(k) corresponding to the time sequence u(k) of the control signal is determined by judging the control signal; if the control signal is in an upward trend, set D(1)=Da, and then proceed to step 304; if the control signal is in a downward trend, set D(1)=Db, and then proceed to step 304; if the control signal remains unchanged, continue to compare the magnitude of the subsequent control signals; If k is not equal to 1, assign a value to the delay coefficient D(k) at this moment according to the following rules: if the control signal is on an upward trend, let D(k) = Da; if the control signal is on a downward trend, let D(k) = Db; if the control signal remains unchanged, let D(k) = D(k-1). Step 303: Determine whether the film pressure remains constant at this moment. If the difference between the time accumulation k-1 and the delay accumulation n1·Da+n2·Db of the control signal is less than the current delay coefficient D(k), then the film pressure remains constant, and proceed to step 304. Otherwise, the film pressure jumps to the set signal, and the film pressure is recorded as delayed once. If D(k) = Da, it means that the film pressure is delayed once during the inflation process, and the number of delays during the inflation process is n1 = n1 + 1. Otherwise, it means that the film pressure is delayed once during the deflation process, and the number of delays during the deflation process is n2 = n2 + 1. Only the film pressure after the jump is recorded, denoted as U(n1+n2+1), whose value corresponds to the magnitude of the set signal at time (n1·Da+n2·Db)·ts, which is approximated by the linear difference between the control signals u(k) and u(k-1). The corresponding cumulative delay is denoted as T(n1+n2+1). U(n)=u(k-1)+(n1·Da+n2·Db-k+2)·[u(k)-u(k-1)] (4); T(n) = n1·Da + n2·Db (5; n = n1 + n2 + 1 (6); Step 304: Let k = k + 1, and compare it with the length of the time series data corresponding to the control signal, kma. If k > kma, the loop ends; otherwise, repeat steps 302 and 303.

Images