JPH06222831A - Fault diagnostic device based on observation of vibration waveform - Google Patents

Abstract

PURPOSE:To provide a vibration waveform observation type fault diagnosing device which can easily and quickly diagnose the faults. CONSTITUTION:A fault diagnosing device 1 contains the connectors M1 and M2 which can be freely attached to or detached from the input/output terminals of a controlled object and a regulator of control system or orher diagnosed device. A storage 15 stores the constant of each diagnosed device. A characteristic calculator means 20 detects the angle frequency omega of the surge signals that can appear on both connectors M1 and M2 and then calculates the gain and the phase delay of a device at the angle omega by using the stored constant of the device connected between both connectors M1 and M2. A measuring means 21 measures the gain and the phase delay between the signals of both connectors, and a deciding means 22 calculates the ratios between the measured value and the calculated value of the gain and the phase delay, respectively. Then, a fact that the device causes the oscillation is diagnosed when either one of both ratios is larger than the prescribed threshold value. If it is diagnosed that there is no factor of oscillation in the device, it is decided that the parameter value of the regulator causes the oscillation.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormality diagnosing device by observing vibration waveforms, and in particular, when an undesired vibration occurs in a control amount which is feedback-controlled by a controller, it is possible to observe the waveform of the vibrations without adding disturbance. The present invention relates to an abnormality diagnosis device for observing the cause of oscillation by observing a vibration waveform.

[0002]

2. Description of the Related Art In a control system including a controller, vibration called "hunting" may occur in a controlled variable due to malfunction of an instrument or a device. In some cases, the vibration may be left for a while. It was For example, the controller 3 controls the steam flow in the heat recovery equipment for exhaust gas of a thermal power plant shown in FIG.
In the device that adjusts with the valve 7 and the valve 7, since the adjustment of the steam flow is a kind of auxiliary machine, when vibration is slight and there is no direct hindrance to the operation of the main machine such as the boiler or the main generator, it is appropriate. It was sometimes left unattended until the scheduled shutdown. However, the vibration of the control amount in the auxiliary machine can be a factor that hinders the safe operation of the main machine, so immediately investigate the cause, take measures such as updating the malfunctioning equipment and re-adjusting the control system, and eliminate the vibration. There is a need.

Conventionally, since there is no apparatus suitable for investigating the cause of this type of vibration, an experienced engineer visually observes the state of vibration and changes the parameters of the controller or the equipment. It was treated by estimating the abnormal point and making necessary adjustments and replacements.

[0004]

However, the above-mentioned conventional treatment requires an experienced technician,
Repeated trial and error tends to be unavoidable, and it takes a long time to perform the treatment. Since the treatment mainly by an engineer's intuition is strong, the recurrence of vibration due to incorrect estimation tends to occur in a short period of time. There was such a problem.

Therefore, it is an object of the present invention to provide an abnormality diagnosis device by observing a vibration waveform, which requires no skilled person and enables rapid treatment.

[0006]

The inventor has paid attention to the following three points in detecting the cause of oscillation of the controlled variable. That is, (1) The main cause of vibration is the excessive gain and phase delay of the equipment in the control system, and (2) Simple measurement that can measure the gain and phase delay between any two points in the control system. Lack of equipment,
(3) There is no simple measuring device that can directly measure the characteristics of the equipment in operation.

The present invention provides a simple diagnostic device having a connector which can be detachably connected to any two points in a control system.
It succeeds in diagnosing the cause of oscillation by measuring the characteristic values of the controlled object and related equipment during operation, and obtaining the magnitudes of gain and phase delay between two connected points.

Referring to FIGS. 1 and 4, an abnormality diagnosing device 1 for observing a vibration waveform by the present inventor is detachably connected to any two points in a control system including a controller 3 and a controlled object 9. The first connector M1 and the second connector M2, which are the controller 3, the control target 9, and the device 8 (see FIG. 6) in the control system, each of which has an initial constant value and a gain change threshold value ( th _g ) and the storage device 15 for storing the threshold value (th _p ) of the change in the phase delay; the angular frequency (ω) of the vibration waveform existing in the signals on the first and second connectors M1 and M2 is detected. And a characteristic calculating means 20 for calculating a gain and a phase delay at the angular frequency of the device from the initial value of the constant of the device connected between the two connectors and the angular frequency; on the first connector M1 Measuring the gain and phase lag at said angular frequency between the signal and the signal on the second connector M2 Measuring means 21 for calculating the gain and phase lag and the ratio of the calculated gain and phase lag to the measured gain and phase lag, and any one of the ratios (the threshold of the gain change corresponding to the device). th _g )
Alternatively, when it is equal to or larger than the threshold value (th _p ) of the change in the phase delay, it is determined that the device is the cause of the vibration.
Is used.

[0009] Preferably, when none of the controller 3, the controlled object 9 and the device 8 in the control system causes oscillation,
It is assumed that the parameter of the controller 3 is the cause of oscillation. Also,
The constants stored in the storage device 15 include the proportional band of the controller 3, the integration time, the differential time, the gain of the control target 9, the time constant, and the initial value of the dead time (PB ₀ , T _I0 , T _D0 , K ₀ , T ₀ , L ₀ ) are included.

When a device other than the controller 3 and the controlled object 9 is included in the control system, all the devices other than the controller 3 or a plurality of adjacent devices are batched together and a first-order delay with dead time is provided. It can be approximated by an equivalent circuit. That is, a plurality of adjacent devices in the control system, for example, all devices other than the controller 3 are connected between the first and second connectors of the abnormality diagnosing device 1, and the angular frequency of the vibration waveform detected by the characteristic calculating means 20. Using (ω) and the gain at the angular frequency measured by the measuring means 21 and the constant of the controller 3 in the storage device 15, the equivalent gain (K _e ) of the first-order delay equivalent circuit with dead time corresponding to the plurality of devices. , Equivalent time constant (T _e ) and equivalent dead time (L _e ) may be provided. In this case, the both connectors by the characteristic calculating means 20
The gain and phase delay of the equivalent circuit between M1 and M2 can be calculated using the equivalent gain (K _e ), the equivalent time constant (T _e ), and the equivalent dead time (L _e ).

[0011]

The principle of the present invention will be described in the following order. A. Criteria B. Calculated value of gain and phase delay C. Measured values of gain and phase delay D. Operation flow chart

A. [Judgment Criteria] In order to diagnose the cause of the sustained vibration of the controlled variable, on the other hand, after measuring the angular frequency ω of the vibration, the gain or phase delay of the controller 3 and the controlled object 9 at the angular frequency ω is set to the initial value of the related constant. Calculated as a calculated value based on the value, and on the other hand, the gain or phase delay during vibration continuation is obtained as a measured value by measurement, and the measured value deviates significantly from the calculated value, and their ratio exceeds a predetermined threshold value. At this time, it is assumed that the controller or the controlled object has an oscillation cause. When there is no abnormality in the gain or phase delay of the controller and the controlled object, it is assumed that the parameter value of the controller has an oscillation cause.

B. [Calculated Value of Gain and Phase Delay] B-1 [Calculated Value of Controller Gain] The transfer function Gc (s) of the controller 3 in FIG. 9 is given by the following equation.

[0014]

[Equation 1]

Where Kp ₀ initial value of proportional gain T _I0 initial value of integration time T _D0 initial value of differential time η differential gain s Laplace operator Gc (s) transfer function of controller

If s = jω and Eq. (1) is rewritten as a frequency transfer function, Eq. (2) is obtained.

[0017]

[Equation 2]

From equation (2), the gain at angular frequency ω | Gc
(jω) | can be obtained as follows.

[0019]

[Equation 3]

FIG. 11 shows both connectors when the first and second connectors M1 and M2 of the abnormality diagnosing device 1 of the present invention are connected to the input and output ends of the controller 3 during continuous vibration as shown in FIG. The vibration waveform above is shown. The angular frequency ω ₀ of this vibration is obtained as follows from the times T ₁₀ and T ₁₂ of the adjacent peaks of the controlled variable PV shown in FIG.

[0021]

[Equation 4]

The proportional gain initial value Kp ₀ used in the equation (1), the integration time initial value T _I0 , the differential time initial value T _D0 , the differential gain η, and ω ₀ in the above equation are converted into the equation (3). Substituting, angular frequency ω ₀
Calculated value of controller gain under continuous vibration | Gc (j
ω ₀ ) | ₀ is obtained.

B-2 [Calculated value of phase lag of controller] From equation (2), phase lag Φ {Gc at the angular frequency ω of the controller.
(jω)} is calculated as follows.

[0024]

[Equation 5]

The initial value Kp ₀ of the proportional gain used for calculating the calculated value of the controller gain, the initial value T _{I0 of} the integration time, the initial value T _D0 of the differential time, the differential gain η, and ω _{0 of the} equation (4). By substituting into equation (5), the calculated value Φc ₀ of the phase lag of the controller under continuous oscillation of angular frequency ω ₀ can be obtained.

B-3 [Calculated Value of Gain of Controlled Object] The transfer function Gp (s) of the controlled object 9 of FIG. 10 is given by the following equation.

[0027]

[Equation 6]

Where K is the gain of the controlled object T is the time constant of the controlled object L is the dead time of the controlled object s Laplace operator Gp (s) is the transfer function of the controlled object

If the exponential function Ke ^-Ls is expanded by the Maclaurin expansion and rearranged, the equation (6) can be approximately written as follows.

[0030]

[Equation 7]

When s = jω and Eq. (7) is rewritten as a frequency transfer function, Eq. (8) is obtained.

[0032]

[Equation 8]

The calculated value | Gp (jω ₀ ) | ₀ of the gain when the controlled object oscillates at the angular frequency ω ₀ is obtained from the equation (8) as follows.

[0034]

[Equation 9]

By the way, the controlled object 9 is a plurality of individual devices 8
It may consist of a combination of. FIG. 6 (A) shows an initial value K ₀ of a controlled object 9a having two individual devices 8 and an initial value of a time constant.
A method of measuring T ₀ and the initial value L ₀ of the dead time by the abnormality diagnosis device 1 of the present invention connected in the same manner as in the normal control operation will be described. That is, with the valve 7 kept unloaded, as shown in FIG.
As shown in (B), the stepwise output ΔM from the controller 3 is added to the controlled object 9a, and the dead time L, the maximum gain ΔPV1 / ΔMV, and the time constant up to the maximum gain in the output from the controlled object 9a corresponding to this are added. T ₁ is the initial value L ₀ of the dead time of the controlled object 9 a, the initial value K _{0 of the} gain, and the initial value T _{0 of the} time constant.
Can be an approximate value of. In addition, the measurement method of this figure is applicable not only to the plurality of individual devices 8 but also to the controlled object 9 of any configuration.

The dead time initial value L ₀ , the gain initial value K ₀ , the time constant initial value T ₀ , and ω ₀ in the equation (4) obtained in this way are given by (9)
By substituting into the equation, the gain calculation value | Gp (jω ₀ ) | ₀ at the angular frequency ω ₀ of the controlled object 9 or 9a can be calculated.

B-4 [Calculated value of phase delay of controlled object] Calculated value of phase delay Φp of the controlled object 9 at the angular frequency ω _0
0 can be obtained as the sum of the phase delay Φp _T due to the time constant initial value T ₀ and the phase delay Φp _L due to the dead time initial value L ₀ as follows.

[0038]

[Equation 10]

C. [Measurement value of gain and phase delay] C-1 [Measurement value of gain of controller] Connect the first and second connectors M1 and M2 of the abnormality diagnosis device 1 to the input / output terminals of the controller as shown in FIG. In Fig. 11 showing the waveforms at both connectors when the
Since it is inside the housing 3a of the controller 3, the abnormality diagnostic device 1 is directly
Cannot be entered into. However, since the deviation E is the difference between the DC set value SV and the oscillating control amount PV, the change in the deviation E inverts the phase of the oscillating control amount PV or π.
(Radian) Only shifted. Therefore, controller 3
Gain value at angular frequency ω ₀ of | Gc (jω ₀ ) | _m
Can be obtained as follows as a ratio of the manipulated variable change ΔMV of FIG. 11 measured by the connection of FIG. 9A and the control change ΔPV.

[0040]

[Equation 11]

C-2 [Measured value of phase lag of controller] Measured value of phase lag at angular frequency ω ₀ of controller 3 Φ _cm
Is the phase difference angle between the deviation E as an input and the manipulated variable MV as an output. Therefore, the specific peak time t ₁₁ of the deviation E in FIG. 11 measured by the connection in FIG. 9A and the manipulated variable MV corresponding thereto Is obtained as follows as a phase difference from the specific peak time t _{21 of}

[0042]

[Equation 12]

C-3 [Measured value of gain of controlled object] The measured value of gain | Gp (jω ₀ ) | _{m at} the angular frequency ω ₀ of the controlled object 9 of FIG. 10 is measured by the connection of FIG. The ratio of the change amount ΔMV of the manipulated variable as an input to the change amount ΔPV of the control amount as an output can be obtained as follows.

[0044]

[Equation 13]

C-4 [Measured value of phase delay of controlled object] Measured value of phase delay Φp of controlled object 9 at angular frequency ω _{0
Since m} is the phase difference angle between the manipulated variable MV as the input and the controlled variable PV as the output, the specific peak time t ₂₀ of the manipulated variable MV in FIG. 12 measured by the connection in FIG. 10 and the controlled variable PV corresponding thereto. The phase difference between the specific peak time t ₁₁ and the specific peak time t ₁₁ is calculated as follows.

[0046]

[Equation 14]

D. [Flowchart of Operation] The abnormality diagnosis operation of the embodiment of the present invention will be described with reference to the flowcharts of FIGS. Step 134: Calculation of the calculated gain | Gc (jω ₀ ) | ₀ of the controller 3 by the equations (3) and (4) and calculation of the calculated phase delay Φc ₀ of the controller 3 by the equation (5) Step 135 : Measured value of gain of controller 3 | Gc (jω ₀ ) | _m by equation (11) and measured value of phase delay of controller 3 by equation (12) Φc _m Step 136: Equation (9) Calculation of the gain of the controlled object 9 by | Gp (jω ₀ ) | ₀ and calculation of the phase delay calculated value Φp ₀ of the controlled object 9 by the equation (10) Step 137: Gain of the controlled object 9 by the equation (13) Of the measured value | Gp (jω ₀ ) | _m and the measured value of the phase delay Φp _m of the controlled object 9 according to Eq. (14).

First, it is determined whether or not the controller 3 is the cause of oscillation. That is, the ratio between the calculated value of the gain and the measured value of the following equations (15) and (16) and the ratio of the calculated value of the phase delay and the measured value are respectively the threshold value thc _g of the predetermined gain change and It determines at step 138 to 141 depending whether the threshold value thc _p or more phase delay change.

[0049]

[Equation 15]

[0050]

[Equation 16]

Then, it is determined whether or not the controlled object 9 is the cause of oscillation by the ratio of the calculated value of the gain and the measured value of the following equations (17) and (18) and the calculated value of the phase delay and the measured value. Is determined in steps 144 to 147 based on whether or not the ratio is within a predetermined gain change threshold thp _g and phase delay change threshold thp _p .

[0052]

[Equation 17]

[0053]

[Equation 18]

Finally, when it is determined that the cause of oscillation is neither in the controller nor in the controlled object, it is assumed that the parameter set values of the controller have an oscillation cause. In this case, the parameter is examined by a method other than the present invention, and the set value is corrected if necessary.

[0055] The present inventors, in the case of pressure control in FIG. 4, the threshold value thc _g in abnormality diagnosis apparatus 1 of the present invention, thc _p, thp _g
And thp _p are set to 1.4, 1.4, 1.3, and 1.3, respectively, to quickly detect abnormal operation in the electropneumatic converter 5 and the cause of oscillation in the movable part of the positioner 6 without interruption, leading to a serious accident. We succeeded in efficiently detecting the anomaly obtained.

Therefore, the object of the present invention is to provide the "abnormality diagnostic apparatus by vibration waveform observation that enables rapid treatment without requiring an expert".

[0057]

EXAMPLE FIG. 18 shows a flow chart of an example including optimization of the proportional band PB. Here, “optimization” means a combination of the two individual devices 8 in FIG. 6 or a linear element 8a and a non-linear element 8b in FIG.
When the comprehensive equivalent gain K _e , the equivalent time constant T _e , and the equivalent dead time L _e for the complex controlled object 9a including the combination with the above are known, the set value S for the controlled variable PV is known.
Controller for PID control so that a control response that satisfies the "required evaluation criteria" can be obtained when V is changed stepwise 3
Proportional gain of Kp (= (y / PB), y is 1 if standardized by a constant,
(PB is a proportional band), the integration time T _I , and the derivative time T _D are set. Although various optimization methods are known, according to the so-called Chien, Hrones, Reswick (Chien, etc.) method,
"Required evaluation criteria" is "No overshoot, minimum settling time"
In the case of PID control as, the optimized proportional gain Kp _op , the optimized proportional band PB _op , the optimized integration time T _Iop , and the optimized differential time T _Dop are as follows (Kazuo Hiroi, “Practical digital control technology” , "Page 57," Optimum adjustment of PID parameters and complex loop control ", Kogyo Gijutsu, October 1992).

[0058]

[Formula 19]

The optimizing means 24 of FIG. 1 calculates the optimized value of the PID parameter by the above equation, for example. The inventor has found that the optimized proportional band PB _op thus obtained is
When the controller 3 before the start of measurement is not within a certain range compared to the proportional band PB ₀ of the initial value set in the storage device 15,
It was experimentally found that the controlled object 9 has a cause of oscillation.
In the illustrated example, the constant range is 0.5PB ₀ <PB _op <2PB ₀
Is. When the controlled object 9 has a cause of oscillation, the optimized proportional band PB _op shows an abnormal value outside the fixed range because the saturation characteristic becomes large or the phase delay becomes large.

The equivalent gain K _e during vibration, the equivalent time constant T _e , and the equivalent dead time L _e in the equation (19) can be determined by numerical calculation as follows. The open loop transfer function G ₀ (s) of the system in which the controller 3 having the transfer function of the equation (1) and the controlled object 9 having the transfer function of the equation (6) are connected can be written as follows.

[0061]

[Equation 20]

The e- ^Les term in the above equation can be approximated by the Maclaurin series as in the case of ^deriving the equation (6) from the equation (6). In the approximation formula in which the power term is replaced by a series, the constants Kp ₀ , T _I0 , T _D0 , and η of the controller 3 are known, so there are only three unknown coefficients K _e , T _e , and L _e. . On the other hand, as boundary conditions, (i) vibrating at a measurable frequency ω, (ii) performing continuous vibration, and (iii) damping vibration And control amount PV
The unknown coefficient K _e , T _e , and L _e can be determined by numerical calculation, because it can be determined from the measured value of the change in the peak value of, and (iv) the gain can be measured by the same method as in Eq. (13). You can ask.

The present inventor substitutes the equivalent gain K _e , the equivalent time constant T _e , and the equivalent dead time L _e at the time of vibration obtained as described above into the equation (19) to optimize the proportional band PB _op. Figure 18
It was experimentally confirmed that the existence of the oscillation cause in the controlled object 9 can be effectively diagnosed by making the determinations in steps 149 and 150.

The above method determines the equivalent gain, equivalent time constant, and equivalent dead time of the system during normal operation corresponding to the equivalent gain K _e during vibration, the equivalent time constant T _e , and the equivalent dead time L _e. to enable. The present inventor has found that the optimization of the PID control by the above equation (19) and the like is extremely effective for stabilizing the flow rate control, reducing the driving power, and the like, based on the thus obtained equivalent constants of the control system.

The operation of one embodiment of the present invention will be described with reference to FIGS.
It will be described according to the flowchart of FIG. The method of using the abnormality diagnosis device 1 is generally shown in FIG. Steps to perform diagnostics
Initial data needs to be created by the step response test shown in 100. After the controlled object 9 starts to operate in step 101, when the sustained vibration is observed in the controlled variable PV in step 102, the waveform of the sustained vibration of the controlled variable PV is observed by the abnormality diagnosis device 1 in step 103, and the step If the cause of oscillation is diagnosed in 104 and the processing based on the diagnosis result is performed in step 105,
The operation of step 106 can be restarted.

To explain the operation after FIG. 14, step 11
At 0, a number is set for the device on the controlled side and the initial values of those characteristics are stored, and at step 111, the controller 3
The set values SV, PID values, constants such as the overshoot amount, the number of measurements u, the measurement start time, and the like that are being executed are set in the storage device 15 of the abnormality diagnosis device 1. In step 112, it is determined whether or not the diagnosis of the individual device is necessary. When proceeding to the measurement of the individual device, the number of times of measurement n corresponding to the number of the individual devices 8 to be measured or the like is updated for each measurement in step 113 of FIG. 19, and it is confirmed in step 114 that the measurement is not in progress. Step 115
At, the measured angular frequency ω ₀ and gain initial value (K ₀ ), time constant initial value (T ₀ ), and dead time initial value (L ₀ ) are substituted into equations (9) and (10), and The calculated value | Gp (jω ₀ ) | ₀ of the vibration gain of the device 8 and the calculated value Φp ₀ of the vibration phase delay are obtained. The gain initial value (K ₀ ), the time constant initial value (T ₀ ), and the dead time initial value (L ₀ ) used here are, for example, as shown in FIG. To the reference output device 4, first, by the reference output device 4, a reference input indicated by ΔIn in FIG. 3 (B) is added to the individual device 8, and the gain measurement value (K ₀ ) and the time constant of FIG. 3 (B) are set. Measured value (T ₀ ) and dead time measured value (L ₀ )
Ask by. The reference output device 4 can be replaced by the manual MV value of the controller 3. When the individual device 8 is the controller 3, the initial value of the characteristic may be obtained by the circuit of FIG. However, the original purpose of the circuit of FIG. 2 is to check the function.

In the subsequent step 116, the vibration waveform input MV, which can be regarded as a sine wave having an angular frequency ω as shown in FIG. 12, is connected to the first connector M1 of the abnormality diagnosis device 1 and the input of the individual device under test 8. It is applied to the point, observed by the second connector M2 output abnormality diagnosis apparatus waveforms PV occurring in 1 of the measured individual device 8 accordingly, the both waveforms amplitude Pv, .DELTA.MV and time difference corresponding peak (t ₁₁ - t ₂₀ ), and substitute it into the equations (13) and (14) to measure the vibration gain of the individual device 8 | Gp (jω ₀ ) | _m
And the measured value of the phase delay Φ p _m .

For the characteristic measurement of the individual device, the connection and signal of FIG. 2 may be used in the case of the controller 3, and the connection of FIG. 4 and the signal of FIG. 5 may be used in the case of the positioner 6. When investigating the overall characteristics of a plurality of individual devices, the connections and signals of FIG. 6 can be used. FIG. 7 shows the controller 3 with a single linear element 8a.
FIG. 8 shows a control system connected to a combination of the linear element 8a and the non-linear element 8b.

In step 117, the calculated value | Gp (jω ₀ ) | ₀ and the measured value | Gp (j
The ratio with ω ₀ ) | _m is determined, and if the ratio is equal to or greater than the predetermined threshold value thp _{g in} step 118, it is determined that the device 8 causes oscillation, and the process proceeds to step 119 to take appropriate measures.
In step 120, the ratio between the calculated value Φ p _{0 of the} phase delay during vibration of the individual device 8 and the measured value Φ p _m thereof is obtained, and in step 121, if the ratio is equal to or greater than a predetermined threshold thp _p , the device concerned. Is determined to be the cause of oscillation, and the process proceeds to step 119 and appropriate measures are taken. If it is not determined that there is an oscillation cause in steps 118 and 121, step 122
Then, it is determined that the individual device 8 is not the cause of oscillation, and it is determined in step 123 whether or not the total number of measurements has been completed. If not, the number of measurements n is incremented by 1 in step 124, and then FIG.
Return to the above step 113 of. If it is determined in step 123 that all measurements have been completed, it is determined in step 125 of FIG. 21 whether or not the next measurement should be continued, and in the case of continuation, the step 110 in FIG. Return to the front.

In the case of optimum adjustment, the method of determining whether or not the cause of oscillation is the same as in the case of individual elements, but in order to optimize the proportional gain, integral time and derivative time of the controller 3, the control system The difference is that a total equivalent gain K _e , an equivalent time constant T _e , and an equivalent dead time L _e in consideration of all devices other than the controller 3 therein are used. In step 130, the number of times u of calculation of the optimum adjustment value is recorded, and in step 131
After confirming that the abnormality diagnosing device 1 is not under measurement in step 132, in step 132, using the above equation (20), the equivalent gain of the comprehensive controlled object 9a (Figs. 6 and 8) including all devices other than the controller 3 is calculated. K _e , the equivalent time constant T _e , and the equivalent dead time L _e are obtained.
In step 133, the optimized proportional band PB _op , the optimized integration time T _Iop , and the optimized differential time T _Dop are determined by the above equation (19) based on these equivalent constants. Hereinafter, in steps 134 to 150, the cause of oscillation is diagnosed so as to be divided into the controller 3, the controlled object 9, and the parameters of the controller 3 as described in the section of the operation.

Step 151 is a case where the cause of the undesired vibration in the controlled variable PV of the controlled object 9 is the parameter of the controller 3 as a result of the diagnosis by the abnormality diagnosis device 1 by the vibration waveform observation according to the present invention. It shows the treatment of. That is, the parameters are corrected so as to suppress the oscillation of the controlled variable PV. After the treatment is completed, it is judged in step 152 whether or not the predetermined number of times u of calculation of the optimized value has ended. If not, the number u of optimum adjustment is increased by 1 in step 153 and then the process returns to step 130. Then, the above operation is repeated. When the predetermined number of times of calculation u has been completed, the process proceeds to step 125 to determine whether or not to end the measurement, for example, from the manual input device 16 in FIG. To continue the measurement, the operation returns to before step 110.

In the embodiment of FIG. 1, each of the connectors M1 and M2 of the abnormality diagnosing device 1 has an air connection end P1 and an electrical connection end, respectively.
With E1, the pneumatic signal at the air connection P1 is converted into an electrical signal by the pressure transmitter 10. Static switch
14 selectively A / D the input signal from the connector E1 or P1
In addition to the converter 11, the digitized signal here reaches the central control unit 13 via the noise elimination filter 12 and is processed there. The storage device 15 stores predetermined items used in the processing described above with reference to the flowchart. The output device 17 is, for example, a printer, and selectively outputs the results of the various calculations described above. The display device 18 is used for input from the manual input device 16 and for displaying the contents of the storage device 15, and the pen recorder 19 records the waveform and the like at the first connector M1 and the second connector M2 by writing with a pen. It is a thing.

[0073]

As described in detail above, the abnormality diagnosis apparatus for observing the vibration waveform according to the present invention connects the two connectors to any two points in the control system and diagnoses the cause of oscillation only by observing the waveform. Therefore, the following remarkable effects are obtained.

(1) Oscillations of the control system can be quickly eliminated and the control can be stabilized. Oscillations (hunting) in auxiliary equipment that accompanies the main engine of a thermal power generator or the like were not a problem as long as the operation of the main engine was not hindered, but by easily detecting the cause of oscillation and eliminating the oscillation, the main engine or plant The overall operating efficiency of can be improved. For example,
Although the thermal power plant is recovered thermal energy of the exhaust gas from the boiler to the chimney using a preheating of air to the boiler, the temperature rise 1 ^o C of exhaust gas temperature in the chimney outlet, generator main engine output 250MW Fuel oil consumption is about 0.4
Corresponding to an increase of 4 KL / day, the hunting with the amplitude of 3-4 ^o C recorded previously is reported to be a loss of several hundred thousand yen per generator per year. Although hunting in the flow rate of the preheated steam in FIG. 4 with respect to the air sucked into the exhaust gas heat recovery apparatus is a vibration of the control amount of the auxiliary machine, it is possible to considerably save energy by removing it.

(2) Since diagnosis is performed without applying disturbance such as inspection signals, measurement / diagnosis can be performed while continuing normal operation without affecting control.

(3) Since the connector can be selectively coupled to each part of the control circuit by a clip or the like, the cause of oscillation can be quickly detected by observing the waveform of the connector. As an example, the cause of oscillation that could not be detected by the conventional method even after several days of inspection could be found in about 2 hours by using the abnormality diagnosis device of the present invention.

(4) The detection of the cause of hunting can be easily performed by an unskilled person with a simple operation.

(5) Appropriate measures can be displayed for each of the different causes of oscillation to provide a so-called expert system.

(6) The structure is simple, and the miniaturization and portability are easy.

(7) It is possible to easily incorporate a means for optimizing the constant of PID control, prevent overshoot of the control amount and shorten the settling time easily, and improve the quality of control. Even if the control of the auxiliary machine is optimized, the overall operation efficiency of the main machine or the plant can be further improved as described above.

[Brief description of drawings]

FIG. 1 is a block diagram of an abnormality diagnosis device of the present invention.

FIG. 2 is an explanatory diagram of a characteristic test circuit of a controller.

FIG. 3 is an explanatory diagram of a characteristic test circuit of an individual device.

FIG. 4 is an explanatory diagram of a positioner characteristic test circuit.

FIG. 5 is an explanatory diagram of signal waveforms in the characteristic test circuit of the positioner.

FIG. 6 is an explanatory diagram of a characteristic test circuit of a combination of a plurality of individual devices.

FIG. 7 is an explanatory diagram of a system including a controller and a linear element.

FIG. 8 is an explanatory diagram of a system including a controller and linear and nonlinear elements.

FIG. 9 is an explanatory diagram of connection at the time of abnormality diagnosis of the controller.

FIG. 10 is an explanatory diagram of connection at the time of abnormality diagnosis of a control target.

FIG. 11 is a graph showing a signal waveform at the time of abnormality diagnosis of the controller.

FIG. 12 is a graph showing a signal waveform at the time of abnormality diagnosis of a control target.

FIG. 13 is a general flow chart of the operation of the abnormality diagnosis device of the present invention.

FIG. 14 is a first partial view of an operation flow chart of the abnormality diagnosis device of the present invention.

FIG. 15 is a second partial view of the operation flow chart of the abnormality diagnosis device of the present invention.

FIG. 16 is a third partial view of the operation flow chart of the abnormality diagnosis device of the present invention.

FIG. 17 is a fourth partial view of the operation flow chart of the abnormality diagnosis device of the present invention.

FIG. 18 is a fifth partial view of the operation flow chart of the abnormality diagnosis device of the present invention.

FIG. 19 is a sixth partial view of the operation flowchart of the abnormality diagnosis device of the present invention.

FIG. 20 is a seventh partial view of the operation flow chart of the abnormality diagnosis device of the present invention.

FIG. 21 is an eighth partial view of an operation flowchart of the abnormality diagnosis device of the present invention.

FIG. 22 is an explanatory diagram of attenuation of response characteristic of control amount.

[Explanation of symbols]

1 Abnormality diagnosis device 3 Controller 4 Reference output device 5 Electro-pneumatic converter 6 Positioner 7 Valve 8 Individual device 9 Control target 10 Pressure transmitter 11 A / D converter 12 Noise elimination filter 13 Central processing unit 14 Pneumatic switching switch 15 Storage device 16 Manual input device 17 Output device 18 Display device 19 Pen recorder 20 Characteristic calculating means 21 Measuring means 22 Judging means 23 Equivalent constant calculating means, 24 Optimizing means.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kiyoshi Shimomura 1-1 Uhei Kawara, Kawara-ki, Hachinohe City, Aomori Prefecture Tohoku Electric Power Co., Inc. Hachinohe Thermal Power Station (72) Nobuhisa Akabane Narita Nishi 3-chome, Suginami-ku, Tokyo No. 8 Okura Electric Co., Ltd. (72) Inventor Jinichi Kikuchi 3-20-8 Narita Nishi, Suginami-ku, Tokyo Okura Electric Co., Ltd. (72) Masaru Niwata 2-31 Ishido, Hachinohe-shi, Aomori Prefecture No. Northern Japan Instrumentation Control Co., Ltd.

Claims (9)

[Claims] 1. A first connector and a second connector which are detachably connected to any two points in a control system including a controller and a controlled object; each of the controller, the controlled object, and equipment in the controlled system. A storage device for storing an initial value of a constant, a gain change threshold value (th _g ) and a phase delay change threshold value (th _p );
A device for detecting the angular frequency (ω) of the waveform of vibration existing in the signals on the first and second connectors and from the initial value of the constant of the device connected between the connectors and the angular frequency. Characteristic calculating means for calculating a gain and a phase delay at the angular frequency; a measuring means for measuring a gain and a phase delay at the angular frequency between a signal on the first connector and a signal on the second connector; And a ratio between the calculated gain and phase delay and the measured gain and phase delay, respectively, and any one of these ratios corresponds to the threshold value of change in the gain of the device (th _g ) or An abnormality diagnosis apparatus for observing a vibration waveform, comprising a determination means for determining that the device is the cause of oscillation of the vibration when the phase delay variation threshold (th _p ) or more. 2. The abnormality waveform diagnosing apparatus according to claim 1, wherein when the controller, the controlled object, and the device in the control system are not the cause of oscillation, the parameter of the controller is the cause of oscillation. Abnormality diagnosis device. 3. The abnormality diagnosing device according to claim 1, wherein the constants stored in the storage device include a proportional band of the controller, an integration time, a differential time, and an initial value of a gain, a time constant, and a dead time of the controlled object. (PB ₀ , T _I0 , T _D0 , K ₀ , T ₀ , L ₀ ). 4. The abnormality diagnosing device according to claim 1, 2 or 3, wherein when two or more adjacent devices in the control system are connected between the first and second connectors, the characteristic calculating means. When the angular frequency of the vibration waveform detected by the and the gain at the angular frequency measured by the measuring means and the constant of the controller in the storage device are used to approximate the plurality of devices by a first-order lag equivalent circuit with dead time Equivalent gain (K _e ), equivalent time constant (T _e ), and equivalent dead time (L _e ) of the equivalent circuit are provided with equivalent constant calculation means, and the characteristic calculation means is used to equalize between the two connectors. An abnormality diagnosing device by observing a vibration waveform, wherein the gain and phase delay of a circuit are calculated using the equivalent gain (K _e ), the equivalent time constant (T _e ), and the equivalent dead time (L _e ). 5. The abnormality diagnostic device according to claim 1, wherein an optimized value (Kp _op ) of the controller proportional gain (Kp) that satisfies a predetermined evaluation criterion for the response characteristic of the controller is set to the gain of the control target. An abnormality diagnosing apparatus for observing a vibration waveform, which is provided with an optimizing means for calculating from an initial value (K ₀ ), an initial value of a time constant (T ₀ ) and an initial value of a dead time (L ₀ ). 6. The abnormality diagnosing device according to claim 4, wherein the optimized value (PB _op ) of the controller proportional band (PB) that satisfies a predetermined evaluation standard for the response characteristic of the controller is set for all of the controllers. An optimization means for calculating the equivalent gain (K _e ), the equivalent time constant (T _e ), and the equivalent dead time (L _e ) with respect to the circuit including the device is provided, and the proportional band optimization value (PB) is provided by the determination means. _op ) and the proportional band initial value (PB ₀ ) of the controller are not within the specified range, it is considered that there is an oscillation cause in the circuit consisting of all devices other than the controller. apparatus. 7. The abnormality diagnosing device according to claim 5 or 6, wherein the predetermined evaluation criterion for the response characteristic of the controller is an overshoot in a change in a controlled variable of a controlled object in response to a stepwise change in a controller input signal. Abnormality diagnosis device by observing vibration waveforms with 20% or less and minimum settling time. 8. An arbitrary device in a control system including a controller and an object to be controlled is connected between two connectors of the abnormality diagnosis device according to claim 1, and there is vibration of an angular frequency ω on the two connectors. The calculated values of the gain and phase delay of the device at the angular frequency ω at the time are obtained by the characteristic calculation means from the initial values of the constants of the device in the storage device, and calculated by the characteristic calculation means by the determination means. The ratio between the gain and the phase delay and the gain and the phase delay measured by the measuring means are respectively obtained, and one of the ratios is the threshold (th _g ) or the phase delay of the gain change corresponding to the device. Abnormality diagnosis method by observing the vibration waveform that causes the device to oscillate when it is above the change threshold (th _p ). 9. A plurality of adjacent plural devices in a control system including a controller and a controlled object are connected between two connectors of the abnormality diagnosis device according to claim 4, and a total gain of the plural devices is stored in the storage device. Change threshold (th _g ) or phase delay change threshold (th _p ) is stored, and the plurality of devices at the angular frequency ω when there is vibration of the angular frequency ω on the two connectors. By the equivalent gain (K _e ), the equivalent time constant (T _e ), and the equivalent dead time (L _e ) of the plurality of devices calculated by the equivalent constant calculation means by the equivalent constant calculation means. Then, the determining means determines the ratio between the gain and the phase delay calculated by the characteristic calculating means and the gain and the phase delay measured by the measuring means, respectively, and any one of the ratios is the total gain of the plurality of devices. Change threshold (th _g ) Or a phase delay change threshold value (th _p ) or more, an abnormality diagnosis method by observing vibration waveforms that causes multiple devices to oscillate.