Preventing control-induced oscillations of the position of a valve member in a valve with pneumatic actuator

Abstract

A method for preventing control-induced oscillations in a valve with a pneumatic actuator and position control with an integrating component, including the following steps: Checking whether oscillations of the valve member occur by counting the zero crossings or extreme values of the control difference. If oscillations were detected, it is checked whether they result from oscillations of the set point. If not, the dead zone is increased and/or the gain parameter is decreased. If no oscillations were detected, it is checked whether wear in the drive has exceeded a predetermined measure. If so, the dead zone is decreased and/or the gain parameter is increased. In this way, oscillations caused by the I-component of the control can be detected and stopped. Further changes to the parameters are only made when friction is expected to have decreased due to wear.

Claims

1. A method for preventing control-induced oscillations of a position of a valve member in a valve with pneumatic actuator and position control of the valve member, wherein the position control of the valve member has an integrating component; wherein the integrating component of the position control has a gain parameter and a dead zone; the method comprising the following steps: verifying the presence of oscillations of the position of the valve member; in the event that oscillations were detected: checking whether the detected oscillations result from oscillations of a set point of the position of the valve member; in case the oscillations do not result from oscillations of the set point of the position of the valve member, increasing the dead zone and/or decreasing the gain parameter; in the event that no oscillations were detected: checking whether wear in the pneumatic actuator of the valve member has exceeded a predetermined measure; if wear in the pneumatic actuator has exceeded the predetermined measure, decreasing the dead zone and/or increasing the gain parameter.

2. The method according to claim 1, further comprising: checking whether the detected oscillations result from oscillations of the setpoint value of the position of the valve member, and determining a period duration of the setpoint value and a period duration of the control difference, and identifying the detected oscillations as oscillations of the setpoint value if the period duration of the setpoint value is not greater than the period duration of the control difference.

3. The method according to claim 1, wherein in the case where oscillations have been detected which do not result from oscillations of the setpoint of the position of the valve member, increasing the dead band and/or decreasing the gain parameter take place only if the control difference is below a first predetermined threshold and a predetermined time period has elapsed.

4. The method according to claim 1, further comprising: changing the dead zone in each individual case by a maximum of 0.1%; and/or changing the gain parameter in each individual case by a maximum of 33%.

5. The method according to claim 1, further comprising: changing the dead zone in each individual case by a maximum of 0.1%; and/or changing the gain parameter in each individual case by a maximum of 25%.

6. The method according to claim 1, further comprising: changing the dead zone in each individual case by a maximum of 0.1%; and/or changing the gain parameter in each individual case by a maximum of 10%.

7. The method according to claim 1, further comprising: using a second threshold for the total travelled stroke and a third threshold for the number of directional changes of the stroke of the valve member as a predetermined measure for the wear in the pneumatic actuator of the valve member.

8. The method according to claim 1, wherein the pneumatic actuator and/or the position control of the valve member has at least one pressure sensor to determine the pressure in the actuator, the method further comprising: generating a stroke-pressure curve based on the pressure determined by the pressure sensor; and using a fourth threshold for the width of a hysteresis in the stroke-pressure curve as a predetermined measure for the wear in the pneumatic actuator of the valve member.

9. The method according to any claim 1, wherein decreasing the dead zone and/or increasing the gain parameter only if in addition the setpoint value of the position of the valve member changes more slowly than a fifth predetermined threshold, and/or there is a sign change of the control difference, and/or the actual value of the position of the valve member is stationary within the dead zone.

10. The method according to claim 1, further comprising: postponing the decreasing the dead zone and/or the increasing of the gain parameter until the actual value of the position of the valve member has left the dead zone.

11. The method according to claim 7, further comprising: when the dead zone is decreased and/or the gain parameter is increased, setting stored values for the total travelled stroke and the number of directional changes of the stroke of the valve member to zero.

12. Method according to claim 8, further comprising: when the dead zone is decreased and/or the gain parameter is increased, redefining the fourth threshold.

13. A positioner for a valve with a pneumatic actuator, wherein a control of the positioner has an integrating component, the positioner configured to carry out the steps of the method according to claim 1.

14. A valve having a positioner according to claim 13.

15. A process plant with a valve according to claim 14.

16. A non-transitory computer-readable medium having stored thereon program instructions that upon execution by a processing unit, control electronics, a digital signal processor (DSP), a microcontroller, a computer, or a plurality thereof in a network which cause the positioner according to claim 14 to perform a set of method steps according to method claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The embodiment are shown schematically in the figures. Identical reference numerals in the individual figures designate identical or functionally identical elements or elements that correspond to one another in terms of their functions. In detail:

(2) FIG. 1 shows a flowchart of the described method;

(3) FIG. 2 shows a partial flow chart for the section of an embodiment of the described method in which the dead zone is increased and/or the gain parameter is decreased;

(4) FIG. 3 shows a partial flow chart for the section of an embodiment of the described method in which the dead zone is decreased and/or the gain parameter is increased.

DETAILED DESCRIPTION

(5) In the flow chart shown in FIG. 1, the method runs in a loop. Such an embodiment is particularly useful, for example, in a program-controlled form of the method. The loop may be repeated continuously, for example, or it runs at fixed, periodic intervals, e.g. once per hour. After the start of the method, it is first queried whether there is an oscillation of the valve member, i.e. an oscillation of the actual value x. If this is the case, it is determined whether there is an oscillation of the valve member. If this is the case, it is determined whether the oscillation originates from an oscillation of the setpoint value w. Such oscillations should not be prevented by the described method and are therefore not processed. If the oscillation originates from an oscillation of the setpoint value, the loop is therefore terminated and the method returns to the starting point.

(6) However, if the oscillation is not specified by an oscillation of the setpoint value, the necessary steps are taken to increase the dead zone and/or decrease the gain parameter K.sub.i. In the embodiment of the method shown in FIG. 1, both of the above changes are provided, but it is of course possible to adjust only one of these parameters and leave the other constant. After these adjustments, the loop is terminated and the method returns to the starting point.

(7) If no oscillation is detected during the first query, the described method checks whether the wear conditions applicable to the respective valve are fulfilled. This can occur, for example, by the value of a stroke counter exceeding a second threshold and a likewise counted number of directional changes also exceeding a third threshold, or, for example, by the width of the friction-induced hysteresis in a measured stroke-pressure curve falling below a provided fourth threshold. If the wear conditions are not fulfilled, no change of the parameters of the control of the valve is required, the loop is thus terminated. If the wear conditions are fulfilled, the necessary steps are carried out to decrease the dead zone and/or increase the gain parameter K.sub.i. In the embodiment of the method shown in FIG. 1, both of the above changes are provided, but it is of course possible to adjust only one of these parameters and leave the other constant. After these adjustments, the loop is terminated and the method returns to the starting point.

(8) Oscillation detection is typically performed by counting zero crossings of the control deviation e=wx. Alternatively, the extreme values of the control difference can also be counted. A further, particularly robust alternative is counting of extreme values of the actual value x. In order to determine whether a detected oscillation is specified by the setpoint value (and thus wanted), the period durations of the control difference e and the setpoint value w are preferably compared. A check is made whether the period duration of the setpoint value w is greater by a certain factor, e.g. 3 or 4, than the period duration of the control difference e. If this is the case, the control difference oscillates significantly faster than the setpoint value, thus it is ensured that the observed oscillation is not being specified by the setpoint value.

(9) Adjustments to the control parameters, i.e. in particular to the width of the dead zone and to the gain parameter K.sub.i, are preferably made in a very sensitive manner. Particularly preferably, the dead zone is changed by a maximum of 0.1%, while larger changes of up to 33%, preferably up to 25%, particularly preferably up to 10% to the gain parameter K.sub.i are possible.

(10) A possible flow chart for the steps necessary to increase the dead zone and/or decrease the gain parameter K.sub.i is shown in FIG. 2. Immediately after the start, an additional check can be provided to determine whether the setpoint value changes slowly, i.e. whether the magnitude of the first time derivative of the setpoint value is below a specified threshold and does not oscillate (not shown). However, this check is optional and not absolutely necessary, since it has usually already been carried out in the superordinate part of the described process, which is shown in FIG. 1.

(11) It is first determined whether the magnitude of the control difference is small, i.e. below a first predetermined threshold. Typically, this first threshold is selected such that it is ensured that the actual value x already approaches the setpoint value wbut may still lie outside the dead zone. If this is not fulfilled, this procedure is terminated or aborted, since a change of the control parameters would not make sense. However, if the control difference is small enough, a predetermined time period is waited for, which should be adapted to the normal settling behavior of the control. After that, one can be sure that the settling of the controlled system is completed. Only then are the control parameters changed, i.e. preferably both the dead zone is increased and the gain parameter K.sub.i is decreased. Of course, it is also possible to adjust only one of these parameters and leave the other constant.

(12) A possible flowchart for the steps necessary to decrease the dead zone and/or increase the gain parameter K.sub.i is shown in FIG. 3. First, further prerequisites are checked: Does the setpoint value w change only slowly, i.e. below a fifth predetermined threshold? Is there a sign change of the control difference? Is the setpoint value stationary in the dead zone? In the embodiment of the method shown in FIG. 3, these conditions are checked successively. However, other, less preferred embodiments are also possible, in which only two or only one of these prerequisites are checked. If they are not fulfilled, this procedure is terminated or aborted, since it would not make sense to change the control parameters.

(13) However, if the above conditions are met, the procedure remembers the decision that the dead zone should be decreased and/or the gain parameter K.sub.i should be increased. This can occur in different ways, depending on exactly how the method is controlled and on what kind of device it runs. For example, a bit provided for this purpose may be set in an electronic buffer, e.g. in the main memory of a computing device, or a value is assigned to a variable, for example.

(14) Then it is waited until the actual value x has left the dead zone again. This means that the valve member must move before the control parameters are changed to a more aggressive control. This ensures that if the parameters are changed too far, an oscillation can be detected again very quickly, whereupon the parameters would again be adjusted towards a less aggressive control. In the sequence shown in FIG. 3, this waiting is achieved by a sub-loop which preferably runs periodically: If the actual value has not left the dead zone, the decision just described is further memorized.

(15) If the actual value has left the dead zone, however, preferably both the dead zone is reduced and the gain parameter K.sub.i is increased. It is of course also possible to adjust only one of these parameters and leave the other constant. Once this change has been made, the wear indicators are re-initialized, since the control has just been adapted to the reduced friction due to wear. If they are used, the stroke and direction change counters are therefore set to zero. If instead the width of the hysteresis of the stroke-pressure curve serves as wear indicator, the fourth threshold for the width of this hysteresis is re-determined.

GLOSSARY

(16) Control Difference, Control Deviation

(17) The control difference or control deviation e is understood as the difference between setpoint value w and actual value x: e=wx.

(18) Control with Integrating Component

(19) Integrating controllers are used to completely compensate control deviations at every operating point. As long as the control deviation is not equal to zero, the amount of the manipulated variable changes. Only when reference and controlled variable, i.e. setpoint w and actual value x, are equal, at the latest, however, when the manipulated variable reaches its system-dependent limit value (e.g. maximum voltage), the control is steady-state. The mathematical formulation of this integral behavior is: The value of the manipulated variable y is proportional to the time integral of the control difference e:
y=K.sub.ie dt
Here, the gain parameter K.sub.i is usually defined as the reciprocal of the integration time.
Stuffing Box

(20) The stuffing box or gland or packing, formerly also called a cloth gland because felt was used as a sealing material, is a sealing element in mechanical engineering. It seals a rotating shaft or a reciprocating rod from a housing against the pressure of a liquid or vapor as well as against penetrating dirt or escaping lubricant.

(21) A stuffing box consists of the stuffing box packing (the actual seal) and a gland (a flange-like sleeve) with which the stuffing box packing is axially compressed by means of bolts or springs. In the case of elastic sealing material, the axial compression also achieves radial compression of the stuffing box packing on the shaft. In this way, the sealing gap can be adjusted to a minimum suitable for the operating conditions.

(22) Disadvantages: A small amount of leakage cannot be completely ruled out. Due to the large contact area, combined with high pressure, which is required for low-leakage sealing, stuffing box packings cause relatively high friction.

(23) Advantages: Since the sealing pressure is applied externally via the packing gland, many suitable materials are available (e.g. fiber materials, graphite). Some of these can also be used at high temperatures and with aggressive media, for which the elastomer seals predominantly used today (e.g. O-rings, radial shaft seals, etc.) are not suitable. Another advantage is that by retightening the packing gland, leakages caused by wear of the sealing material can be reduced again.

(24) Dead Zone (Also Dead Band)

(25) A dead zone is a range of input values of a control or signal processing system for which the output value is zero. In control systems, such a dead zone is generally used to suppress undesirably frequent cycles of switching operations. In control systems, the dead zone is typically a tolerance range for the actual value x around the setpoint value w, orequivalentlya tolerance range for the control difference e around zero. Typically, a dead zone is connected upstream of the actual controller or only of the integrating component of the controller.