A method for surveying the condition of a control valve, and a valve apparatus

METHOD OF CHECKING THE CONDITION OF THE CONTROL VALVE

Engineering

The present invention relates to a method for monitoring the condition of a control valve system, the position of which is controlled by an electro-pneumatic positioner-controlled actuator operated by a pressure medium, the operation of which is monitored by sensors which read readings at least from the vent10 control signal and the position of the valve, and the microprocessor locates the fault causing the abnormal reading using the readings reported by the sensors and the inference rules stored in the microprocessor.

State of the art

The control valve and its operation are known and need not be described in detail in this context. The valve may be a quarter-rotary or linear valve, 20 of which designations describe the direction of movement of the valve closing member in the control situation. The quarter rotary valve may be^: e.g., a ball valve or a butterfly valve. Examples of a ball valve are described, e.g., in U.S. Pat

747 578. The valve is operated by an actuator which rotates · '25 the axis of rotation of the closing member between the closed and open position.

The actuator can be operated by a cylinder-piston device, which in turn is controlled by a control valve. This control valve is located in the position of the control valve. A positioner is a device that. ·, Confirms the control signal by using a pneumatic actuator • · · ••• J 30 pressure. In the electro-pneumatic positioner, the electric • · '···' signal is amplified as pneumatic operating pressure. The position; ··· witch also positions the valve to match the control signal by means of a feedback member.

Diagnostics of control valves with a pneumatic actuator are generally based on tests on control valves, such as step response and hysteria tests. To perform the tests, the process must be interrupted, in which case the tests will be performed at the valve. Test equipment often also includes complex sensing, which can make it difficult to perform tests ken10 i I ii

<img file="FI104129B_D0001.tif" />

with this (U.S. Patent No. 5,197,328). This is the so-called off-line diagnostics.

Diagnostics are also performed on-line, in which case the positioner of the control valve monitors, for example, the position message of the valve and alerts if the position message deviates too much from the value required for control. On-line diagnostics also include valve operation counters that give an alarm when the number of operations exceeds a set limit. However, this diagnosis does not include reasoning that would perform fault location.

U.S. Patent Nos. 5,329,465 and 4,694,390 both describe valves whose control system uses separate measuring sensors. The inventions described in the publications are directed to valve open-close control, not continuous control.

The system of U.S. Patent No. 5,329,465 uses sensors to monitor valve condition and collect historical data for a database. The fault analysis is performed by comparing the measured values ​​with the values ​​of a previously collected database. The position of the valve is not constantly adjusted, but the valve only moves when it is closed and opened. This may occur a few times a day or the adjustment interval may be up to several months. The sensors only monitor the values ​​required for fault analysis during this opening and closing movement at intervals.

Description of the invention

The features of the method according to the invention are set out in claim 1.

The invention relates to an electrically controlled pneumatic system in which the sensors used for the control event are also used for fault analysis. The values ​​are thus continuously collected from the valve by means of sensors. The positioner unit includes an internal self-diagnostic part which, in bus-structured control, enables the immediate transfer of information to the control unit or control room.

The invention is based on the logic programmed for the digital positioner, on the basis of which the positioner processor performs a continuous (on-line) conditional check of the control valve ·· ♦ • · ·

• · lua. The quantities monitored by the positioner are the control flow signal, the difference between the inlet and outlet pressures of the actuator, the position of the valve and the supply pressure of the positioner. By constantly monitoring these variables and comparing their changes to the rules defined in the logic, the diagnostics seeks to locate the point of failure of the control valve. When locating the fault target, the logic is used to determine whether the fault is in the positioner (slider), the pneumatic actuator or the valve. In addition, the electronic board of the digital positioner has its own internal diagnostics, which allows the positioner to report faults on the circuit board. However, the diagnostics of this electronic card are not within the scope of this invention.

Logic rules consist of rules, all of which must be fulfilled for diagnostics to give a fault message.

As a result, the diagnostics indicate a possible fault location.

The invention relates to a digital position of a control valve¹ Logic for diagnostics that can be programmed into the instrument. Logic 20 consists of a reasoning mechanism formed from Boolean rules.

During the process, the positioner monitors e.g. control valve position information, actuator inlet and outlet pressure difference, control signal and positioner supply pressure level. These 25 values ​​and their changes are compared with the inference rules formed from the Boolean • rules, on the basis of which the condition of the control valve is determined and, if necessary, alarms are given to locate any faulty components (positioner, actuator, valve).

The reasoning rules are divided into five different alarms or faults¹ the notifying part.

Alarm 1 is issued when a fault is detected on the positioner's electronics board (current message, angle sensor, pilot valves, pressure sensors, EEPROM and ROM).

This diagnostics is taken care of by the card's own diagnostics, the more detailed content of which is not the subject of the present invention.

Alarm 2 is issued when the valve control signal changes, but the difference between the inlet and outlet pressure of the actuator and the valve4 * 4 &lt;4 • 4 · • · 44

J ίο

I «

» 4

I • ·· 4 • ♦ ♦

Iin status does not change. In this case, a fault message indicates that the fault is in the positioner slider.

Alarm 3 is issued when the difference between the position of the valve corresponding to the control signal and the measured position of the valve exceeds the set limit continuously for a certain time. The error message then indicates the most probable error destinations. In the event of a friction problem, alarm 4 and / or alarm 5 may also occur.

Alarm 4 is issued when the control signal changes and the difference between the inlet and outlet pressures of the actuator increases, but the position of the valve does not change. This is a friction problem.

Alarm 5 is issued when the load factor trend exceeds the set limit. This is a friction problem.

When the fault diagnosis rules and the reasoning mechanism are in the positioner's memory, information is obtained about the possible indeterminate operation of the control valve continuously during the process. Thus, the reasoning is not based on separate tests, in which case the process would have to be interrupted, but the monitoring of the valve condition and alarms are performed on-line during the process. This is a major innovation in the condition monitoring of pneumatically operated control valves.

Information about alarms can be sent automatically, for example, via the HART communication bus, for example to a control room. Therefore, there is no need to retrieve data separately for each valve, which can sometimes be difficult to access. When it is not necessary to go to the valve to retrieve diagnostic information, then there is also no need for a separate device to read the data from the valve in the field. Nor does it take time to retrieve information.

To assess the condition of the valve, it is not necessary to disconnect the valve from the piping, which is a time consuming and laborious operation. If the diagnostics give alarms, test runs can then be performed on the valve, if necessary, so that possible fault locations can be located even more precisely. Performing these test runs requires stopping the process.

Obtaining continuous on-line information on the condition of the control valves is important for preventive maintenance. The valve can operate indefinitely for a long time, interfering with the operation of the process before it is serviced. In the worst case, the valve can break, causing costly production downtime. These cases are avoided by continuously receiving information about the condition of the valve during the process.

Traditionally, valve maintenance has been a maintenance activity performed during factory outages, in which case the valves in the most critical locations are serviced without prior knowledge of their condition. In this case, it is possible to service valves that have not yet required service, or to omit ίο valves that would have required service. All valves are not serviced during downtime. Once the on-line information on the condition of the valves is available, the valves in need of maintenance can be located and more accurately diagnosed and tested to determine their actual need for maintenance.

The ultimate benefit of on-line diagnostics is increased process safety and reduced costs as process availability increases and production losses decrease. Energy is saved, losses are reduced by 20 and unexpected production downtime is avoided. The size of spare parts stocks can also be reduced.

Brief description of the drawings

The invention and its details will be described in more detail below with reference to the accompanying drawings, in which Fig. 1 schematically shows a control valve apparatus according to the invention, Fig. 2 shows conditions and conclusions of alarm 2, Fig. 3 shows exact values ​​for alarm 2, Fig. 4 shows conditions and conclusions , Fig. 5 shows the conditions and conclusions of alarm 4, Fig. 6 shows the conditions and conclusions of alarm 5, Fig. 7 shows the quantities 35 to be monitored for the alarm 5 and Fig. 8 shows a trend curve of the load factor.

Embodiments of the invention

Figure 1 shows an apparatus comprising a control valve 101, an actuator operated by a cylinder-piston device 102

103 and a positioner 104. The actuator is controlled by a control valve 105. Positioner messages are processed by a circuit board 106 in which control information is stored. Sensor 107 measures the position g of the valve closing member 108, the sensor 109 measures the supply pressure p of the positioner._s and sensor 110 measures the difference between the inlet and outlet pressures of the actuator Ap. These sensors are in the positioner. The diagram shows the flow of various electronic messages and measurements. A control signal i, is applied to the system.

Alarm 1 indicates a positioner circuit board problem, as noted above, but is not within the scope of this invention.

Alarm 2 indicates a control valve slip problem.

Figure 2 shows in the diagram four different conditions by which the reasoning for alarm 2 is made. Figure 3 shows the operation for alarm 2, i.e. the quantities to be monitored as a function of time and their parameters. Point A on the time axis is the position where the values ​​of the control signal i, the difference between the inlet and outlet pressure Ap of the device and the valve, the position s are monitored during repeated read cycles have been within the transmitted limit values ​​for the required time. The values ​​of the quantities of this equilibrium state A, called the initial state, are stored in the registers. If there is a change in the control signal of&gt; 5%, the time counter starts. If the 'signal change returns 5% below the limit within the • selected time period, the time counter will be reset and no changes will be made in the other registers unless the control signal X changes •.

·. there has been no&gt; 1% change in the position of the valve in relation to the value of the initial state A. In this case, a new initial position A must be set. If both control signal i and valve i

If the position s exceeds the values ​​of the equilibrium conditions of the initial situation A, a new initial situation is retrieved.

Condition 1: A change of more than 5% in the control signal X takes more than 35 120 s, i.e. from point B to point D in Fig. 3.

Condition 2: The change in the difference A of the inlet and outlet pressures of the actuator is within the selected limits, in Fig. 3 the change is &lt;0.2 bar in the interval A -&gt; B and from B to C.

<img file="FI104129B_D0002.tif" />

Condition 3: The change in valve position s is within the selected limits, in Figure 3 the change is &lt;1% in the time interval A -&gt; B and from B to C.

Condition 4: No Alarm 1.

On the basis of this information, it is concluded 5:

the slide of the control valve is stuck, whereby the fault location 6 is in the positioner.

The field of view is 10 - 90% of the control valve movement10

<img file="FI104129B_D0003.tif" />

from the area, i.e., does not look at the very beginning or the very end of the opening movement.

In the example of Fig. 3, the control signal i. Change s is 2% and the duration 10 s is selected as the parameters describing the equilibrium state; actuator inlet and outlet pressure difference Δρ change s 0,2 bar and duration 10 s; and control valve position s change s 1% and duration 10 s.

The corresponding parameters of the alarm state are: change of control signal i&gt; 5% and time 120 s, i.e. on the time axis from B to D; the difference between inlet and outlet pressure Δρ change s 0,2 bar and time 30 s, ie on the time axis from B to C; and the change in position s is 1% and the time is 120 s, i.e. from B to D. Of course, these parameter values ​​can be selected as needed. The first and last of the alarms are stored, their times and the total number of alarms. The reading cycle is 0.1 s.

Alarm 3 indicates a large difference between the control signal and the valve position.

Figure 4 shows the condition and conclusion of alarm 3.

Condition 7: The difference between the position of the valve corresponding to the signal and the measured position of the valve is&gt; 5% continuously for more than 10 minutes.

Conclusion 8 is that this means either that the positioner front choke, which reduces the supply air pressure to the electro-pneumatic positioner to suit electrical control, is blocked or the control valve slide unit level seal is leaking or the actuator cylinder has internal or external leakage or the actuator or valve.

Fault location 9 is in the positioner, actuator or valve.

The field of view is 10 - 90% of the valve's operating range.

The above-mentioned values ​​of condition 7, i.e. a difference value&gt; 5% and a duration of 10 minutes, have been chosen as parameters. The first and last of the alarms are stored, their times and the total number of 5 alarms. The reading cycle is 0.1 s.

Alarm 4 indicates a friction problem.

Figure 5 shows the four conditions and conclusions of alarm 4.

io Condition 10: A change of more than 5% of the control signal i lasts more than

120 s, in Figure 3 from point B to point D.

Condition 11: The change in inlet and outlet pressure difference Δρ of the actuator is more than 0,2 bar within 30 s, in Figure 3 from point B to point C.

Condition 12: The change in valve position s is within the selected limits, in Figure 3 the change is &lt;1% between A and B and from B to D.

Condition 13: No alarm 1.

When these conditions are met, the conclusion 14 is that there is a friction problem in question, the location 15 of which is either in the valve or in the actuator.

The field of view is 10 - 90% of the valve's operating range.

The initial situation A and its stored quantities are the same as in Alarm 2 (Figure 3). The operation is the same as in Alarm 2, except that here the difference between inlet and outlet pressure must change, while in Alarm 2 this difference must not change.

The above-mentioned values ​​of conditions 10 to 12, ie the change of the control signal ϊ&gt; 5% and 30 time 120 s, have been selected as the parameters of the alarm state; change in actuator inlet and outlet pressure difference Δρ&gt; 0,2 bar and time 30 s; and control valve position s change £ 1% and time 120 s.

The first and last of the alarms are stored, their times and the total number of alarms. The reading cycle is 35 0.1 s.

Alarm 5 is obtained by following the trend curve of the load factor.

Figure 6 shows condition 16 of alarm 5: The trend of the load factor exceeds the set limit. The conclusion 17 is that there is a friction problem, the location 18 of which is in either the valve or the actuator.

The field of view is 10 - 90% of the valve's operating range.

Figure 7 shows the operation for alarm 5, i.e. the changes in the difference between the inlet and outlet pressure Δρ of the actuator and the position s of the valve as a function of time. The initial situation A and the quantities to be stored are determined as in Alarm 2. The change in valve position s is then considered. When the change s exceeds the value stored in the initial position memory by 1%, this situation is defined as the valve start time E, where the difference between the inlet and outlet pressure Δρ of the actuator and the supply pressure p of the positioner is stored._s. From these values, the load factor k is calculated. The next equilibrium position A 'is chosen as the new initial situation. If the difference between the valve positions in the initial situation A and the new initial situation A 'is greater than 2%, the calculated load factor k is saved to modify the trend curve.

An alarm is issued when five consecutive measurements exceed the set limit.

The actuator load factor k is given by the following formula:

Δρ, k = · 100%

Ps • «· • ·· • · ·· • ·» • · • «• · · where

Ap_L = actuator inlet and outlet pressure difference at valve actuation positioner supply pressure

Ps =

The parameters of alarm 5 are position s change&gt; 1% relative to the stored initial state A value, load factor k, position s change&gt; 2% relative to the stored initial state A value, load factor trend alarm limit and the number of limit crossings leading to the alarm. The reading cycle is 0.1 s.

Figure 8 shows the trend curve calculated from the averages of several load factors as a function of time and the set alarm limit a.