JP2601896B2 - Plant abnormality diagnosis system - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a plant abnormality diagnosis system applied to failure and abnormality diagnosis of a large-scale plant such as a chemical plant, and more specifically, to a symptom used for diagnosis. Regarding improvement of the creation.

[Conventional technology]

2. Description of the Related Art Conventionally, in a chemical plant, etc., various instruments such as a pressure gauge and a hydrodynamic gauge are attached to specific locations of the plant, and the operation state of the plant is monitored based on measurement data, which is a process value obtained from these instruments. A system that issues an alarm when necessary is known.

It is generally known that a condition for generating an alarm or the like in such a system is such that a threshold value for a process value is fixedly set and a case where the process value is detected to deviate from the threshold value is set. Have been.

[Problems to be solved by the invention]

However, according to such a method, an abnormality occurs in the process, and as the disturbance of the process due to the influence of the abnormality spreads, the number of alarms generated also increases, what is occurring in the process, and There is a problem that it is not always easy for an operator to know what the cause is.

The present inventors have assiduously studied and have proposed a plant abnormality diagnosis apparatus capable of easily presenting the type and cause of an abnormality to the operator when an abnormality occurs in the plant (Japanese Utility Model Application No. 63-96202). This diagnostic apparatus is provided with a knowledge memory in which a correspondence between a hypothesis that a failure or the like has occurred in a plant and a symptom that appears when the hypothesis has occurred is classified for each symptom. A diagnostic unit compares the correspondence between the contents and the symptoms and the like that are actually occurring in the plant, diagnoses whether or not the hypothesis has occurred, and displays the result.

According to such a diagnostic apparatus, the abnormality and its cause considered to be naturally caused by a physical change or the abnormality and its cause empirically recognized are stored in the knowledge memory, so that the abnormality in each part of the plant can be performed. The advantage is that the diagnosis can be made individually and clearly.

In addition, the present inventors determine the reference value as an average value in accordance with the operating state of the plant that changes from moment to moment, and by taking measures such as statistically obtaining the probability of change, etc. Has already proposed a device that determines the allowable range based on data and detects signs of plant abnormalities.
63-250596).

According to this, as a reference value for determining whether a physical quantity at a specific point of a plant is within a predetermined allowable range, an average value obtained at predetermined time intervals is used, and this average value is set as a reference. It is possible to judge whether or not the value of the physical quantity belongs to an allowable range, and it is possible to detect an abnormal sign corresponding to a long-term variation of the process.

However, it must be noted that simply making the correspondence between the hypothesis stored in the knowledge memory and the symptom is not always enough to perform the abnormality diagnosis of the plant with high accuracy. In other words, it is necessary to take into consideration that the process value indicating a symptom such as an abnormality input to the diagnosis unit, that is, the measurement data itself must have high reliability.

For example, the value of a physical quantity at a specific point of a plant is often allowed to fluctuate with a certain allowable width due to the nature of the specific point. Such an allowable range is determined based on a certain reference value. However, if this reference value is fixedly set as the above-mentioned threshold value, a change in plant operating conditions, a change in outside temperature, etc. In some cases, fluctuations that should be originally permitted due to the above are output to the diagnosis unit as symptoms such as abnormalities, and erroneous diagnosis may be executed.

Therefore, the present invention can dramatically improve the reliability of symptoms for performing abnormality diagnosis if the advantages of the above-described proposed plant abnormality diagnosis device and the advantages of the abnormality sign detection device can be combined. The purpose of the purpose is to obtain a diagnostic symptom indicating a symptom such as a plant abnormality in accordance with the operating conditions of the plant and output this, and based on this, It is an object of the present invention to provide a plant abnormality diagnosis system that can obtain high reliability of a diagnosis result by performing abnormality diagnosis and can avoid unnecessary diagnosis.

[Means for solving the problem]

In order to achieve the above object, a plant abnormality diagnosis system according to the present invention includes a measurement unit that measures a physical quantity at a specific point of a plant, and a physical quantity measured during a time interval before that at predetermined time intervals. The average value calculator for calculating the average value, the physical quantity and the average value indicating a range in which the physical quantity can be considered normal based on the existence range of the difference between the physical quantity and the average value and statistically obtained. An allowable range calculator that determines an allowable range of the difference between the physical quantity at the present time and the average value of the physical quantities measured during the previous time section during the time section. Abnormality sign detection means including comparison means for outputting as a symptom signal, and a diagnosis unit for diagnosing a plant abnormality or the like by inputting detection information such as a predetermined abnormality together with the diagnosis symptom signal. And it is configured as a feature that there was e.

[Action]

According to the present invention, the physical quantity at a specific point of the plant is measured by the measuring means. The measured values are sequentially output to an average calculator, and the average calculator calculates an average value in a predetermined cycle based on the input measured values. The measured value and the average value are input to the comparing means to determine a difference between the measured value and the average value, and it is determined whether or not the difference is within an allowable range determined by an allowable range calculator, and the result is determined. The comparison means outputs the symptom signal for diagnosis to the diagnosis unit. The new diagnostic symptom input to the diagnostic unit is compared with information for detecting abnormality or the like, and the presence / absence of abnormality is diagnosed.

〔Example〕

 Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows a block configuration of a plant abnormality diagnosis system according to the present invention. In this figure, an abnormal sign detecting means 1 includes a pressure gauge for measuring a physical quantity at a specific point of a plant, a sensor 2 of a flow meter, an average value calculator 3 for receiving an output of the sensor 2 to obtain an average value of the physical quantity,
A comparison means 4 is provided which receives the output of the average calculator 3 and the output of the sensor 2 as inputs, performs predetermined processing on these, and outputs a diagnostic symptom signal.

The average calculator 3 is configured to calculate an average value of physical quantities in the immediately preceding time interval, for example, at predetermined intervals. Further, the comparing means 4 includes a subtracter 5 for calculating a difference between the physical quantity at the plant specific point from the sensor 2 and the average value from the average value calculator 3, and sets the difference obtained by the subtractor 5 to an allowable range. And a comparator 7 for judging whether or not it belongs to a preset allowable range via a detector 6 and outputting the result as a diagnostic symptom signal to a diagnostic section 8 at the next stage. At this time, the average value calculated by the average value calculator 3 is regarded as a reference value of the physical quantity.

Note that the average value calculator 3 is provided so that a reset signal is supplied as necessary, so that it can cope with a change in a physical quantity that occurs when the operating condition of the plant or the like is changed. ing.

The diagnostic unit 8 receives the diagnostic symptom signal output from the comparator 7 and compares and compares the diagnostic symptom information with information for detecting abnormality or the like stored in the knowledge memory 9. It is displayed on the display unit 10.

The information for detecting an abnormality or the like is a correspondence between a hypothesis that a failure or an abnormal state has occurred in a plant and a symptom that appears when this hypothesis has occurred. Are classified according to the strength and the symptoms that do not appear. Therefore, the diagnostic unit 8
In the above, a diagnostic symptom signal indicating a symptom output from the abnormal sign detection means 1 is input, and information on the occurrence form of the input symptom, information on the symptom not occurring, and the knowledge memory 9
By comparing the relationship between the hypothesis and the symptom occurrence mode and the relationship with the symptom that does not occur, it is possible to diagnose whether each hypothesis actually occurs.

The term “appearance or detectability” in the classification criterion in the knowledge memory 9 refers to whether the disturbance of the process value due to the occurrence of the abnormality is large enough to detect it, or large from the occurrence of the abnormality to the appearance of the symptoms. This is an index indicating whether it is appropriate to use the symptom for diagnosis without time delay, etc.For example, symptoms that are classified as appearing or more likely to be detected are not detected. "A corresponding hypothesis has occurred"
Is not detected, and conversely, the fact that no corresponding hypothesis has occurred is linked to the determination by the diagnostic unit 8 that no corresponding hypothesis has occurred.

Symptoms that have been classified as having a lower appearance or lower detectability lead to a determination in the diagnosis unit 8 that “the corresponding hypothesis has occurred” when they are detected, but are not detected because they were not detected. Therefore, this does not lead to the determination in the diagnosis unit 8 that “the corresponding hypothesis has not occurred”.

Further, the “non-appearing symptom” in the classification criteria in the knowledge memory 9 indicates that the occurrence of a symptom under the influence of a hypothesis is not considered from the physical properties of the plant, and is classified as a non-appearing symptom. The detected symptom leads to a determination in the diagnosis unit 8 that “the corresponding hypothesis has not occurred” by detecting the symptom.

In this way, the correspondence between the hypothesis that a failure or abnormal state has occurred in the plant and the symptom that appears when this hypothesis has occurred indicates the strength of the symptom or the detectability of each hypothesis and does not appear. The symptoms are classified for each symptom and stored in the knowledge memory 9, and the symptom presently occurring in the plant is input to the diagnosis section 8, and the knowledge memory 9 is stored.
By diagnosing whether each hypothesis has occurred or not by referring to the description contents, it is possible to present the type and cause of the abnormality occurrence to the operator. It is possible to make an appropriate determination according to the classification.

Next, the embodiment will be described in more detail based on specific application examples.

FIG. 2 shows a liquid level control system of a tank used in a chemical plant. In this figure, an inflow flow meter 11 for measuring a flow rate of a liquid flowing in the inflow pipe 10 and an inflow control valve 12 for adjusting the flow rate are arranged in the middle of the inflow pipe 10. The liquid level of the liquid that has flowed into the tank 13 via the inflow control valve 12 is adjusted by a liquid level controller 14, and the liquid level controller 14 controls the liquid level so that the liquid level is maintained at a set value. It is configured to output a control signal to the control valve 12. An outflow pipe 15 extends from the bottom side of the tank 13 and has an outflow control valve for adjusting the flow rate of the liquid.
An outflow flow meter 17 for measuring the flow rate of the outflow pipe 16 and the outflow pipe 15 and outputting a control signal to the outflow control valve 16 so that the flow rate is maintained at a set value is provided. Incidentally, in FIG.
Reference numeral 9 denotes a setting device for each of the liquid level controller 14 and the outflow flow meter 17, and each of the meters including the outflow flow meter 11 constitutes an abnormal sign detecting means 1 shown in FIG. This corresponds to 2.

With such a configuration, the following examples can be given as symptoms of the plant.

I. The indicated value of the inflow flow meter 11 is high.

II. The indicated value of the inflow flow meter 11 is low.

III. The liquid level indicator of the liquid level controller 14 is high.

IV. The liquid level reading of the liquid level controller 14 is low.

V. The control signal of the liquid level controller 14 is high (in this case, the inflow control valve
12 is operated to the open side).

VI. The control signal of the liquid level controller 14 is low (in this case, the inflow control valve 12 is operated to the closed side).

VII. The value obtained by subtracting the indicated value of the liquid level controller 14 from the estimated liquid level is high.

Here, the `` estimated liquid level value '' is a time-integrated calculation of the difference between the indicated value of the inflow flow meter 11 and the indicated value of the outflow flow meter 17 to estimate the liquid holding amount in the tank 13, and The liquid level in the tank 13 corresponding to the liquid holding amount is estimated from the 13 geometric shapes. At this time, the estimated liquid level is
This is obtained by interposing a subtractor for subtracting the indicated value of 11 and the indicated value of the outflow flow meter 17, an integrator for integrating the value of the subtractor, and an arithmetic unit for estimating the liquid level. The output can be an input value of the sensor 2.
The estimated liquid level is represented by the following equation.

Here, Lg (t): estimated liquid level, t: time, FI (t):
Indicated value of inflow flow meter 11, FC (t): indicated value of outflow flow meter 17, M _O : value of liquid holding amount in tank 13 at time zero,
T: current time, LM: a function (monotonically increasing function) for obtaining the liquid level in the tank 13 from the liquid holding amount in the tank 13 and its geometric shape.

VIII. The value obtained by subtracting the indicated value of the liquid level controller 14 from the estimated liquid level is low.

IX. The value obtained by subtracting the indicated value of the inflow flow meter 11 from the estimated inflow amount is high.

Here, the `` estimated inflow value '' is a value estimated based on a physical quantity relationship between the opening degree of the inflow control valve 12 and the flow rate of the liquid flowing through the inflow pipe 10, and is the same as in the case of obtaining the estimated liquid level,
It can be obtained by an arithmetic unit or the like provided before the sensor 2. Further, the opening degree of the inflow control valve 12 is replaced by an operation signal of the liquid level controller 14. The estimated value of inflow is given by the following equation.

Here, FIg: estimated value of inflow amount, Mi: operation signal for operating the inflow control valve 12, Cvi (Mi): flow coefficient of the inflow control valve 12 (the flow coefficient is a function of the operation signal and is a monotonically increasing function. ), Dpi: pressure difference of the inflow control valve 12 (assumed constant here).

X. The value obtained by subtracting the indicated value of the inflow flow meter 11 from the estimated inflow amount is low.

The symptom items I to X exemplified above are merely examples, and the items taken up as symptoms are not limited, and various symptom items may be selected as necessary according to a knowledge description method described later. Of course.

In the present embodiment, a hypothesis of an abnormal cause to be diagnosed is assumed as follows.

Incorrect indication of inflow flowmeter 11 (indicated more than actual flow through inflow pipe 10) Incorrect indication of inflow flowmeter 11 (indicated lower than actual flow through inflow pipe 10) Blockage of the inflow pipe 10 High error indication of the liquid level controller 14 (indicated higher than the actual liquid level) Low error instruction of the liquid level controller 14 (indicated lower than the actual liquid level) Leakage of the tank 13 Assuming the classification of the symptom according to the form as shown in FIG. 3, the knowledge of the correspondence relation of the symptom occurrence associated with each abnormality occurrence in the knowledge memory 9 and the basis thereof will be described for the above-mentioned hypothesis. In the following description, each hypothesis will be described in order using the classification symbol and the symptom number shown in FIG. 3, but the classification symbol A,
All except B and C. In addition, it goes without saying that various methods of classifying the minimum necessary symptoms into D for discrimination from other hypotheses can be considered. FIG. 4 shows knowledge obtained as a conclusion, and FIGS. 5 to 9 show examples of knowledge of each hypothesis.

Therefore, according to such an embodiment, when the diagnostic pattern input to the diagnostic unit 8 is obtained, the processing by the abnormal sign detecting means 1 uses only a fixed value, that is, a deviation from a stable state. Since the allowable range can be set without setting the allowable range, it is possible to set the allowable range narrow, and it is possible to shorten the time required for the fluctuation of the physical quantity caused by the abnormality to deviate from the allowable range. Further, since the average value of the physical quantity in the stable state as a reference value can correspond to a long-term process variation, there is an effect that the process variation can be flexibly dealt with.

As described above, it is possible to easily and accurately grasp the diagnostic symptom signal obtained by the abnormal sign detecting means 1 and to use the diagnostic symptom signal for the diagnosis by the diagnosis unit 8, so that the diagnosis result has high reliability. Can be placed,
Monitoring of the operation state of the plant can be performed more effectively, and the necessary inspection and maintenance work that the operator should take in response to the result of the abnormality diagnosis can be performed accurately and quickly.

Furthermore, since a reset signal can be input to the average value calculator 3 and the average value of the physical quantity is recalculated from the time when the reset signal is input, the average value calculator 3 may be changed due to an artificial factor such as a change in plant operating conditions. Even when the stable value of the physical quantity is displaced, it is possible to cope with this.

Note that the abnormal sign detection means in the above embodiment is not necessarily limited to the configuration example in FIG.
The configuration shown in FIG. 13 to FIG. 13 may be used.

The abnormality sign detecting means 1A in FIG. 10 is different from the above embodiment in that the hold circuit 20 is provided in the average value calculator 3. When the difference between the measured value and the average value deviates from the allowable range, the hold circuit 20 functions so as not to update the average value of the physical quantities in the immediately preceding section.

Therefore, when such an abnormal sign detection means 1A is used, an average value can be obtained based only on the stable value of the physical quantity, and a more reliable diagnostic symptom signal can be obtained. An effect can be added.

In addition, the abnormal sign detection means 1B in FIG. 11 statistically calculates the difference between the physical quantity measured by the sensor 2 and the average value of the physical quantity calculated by the average value calculator 3, that is, the range in which the deviation exists. It has features. Therefore, an allowable range calculator 21 is provided between the average calculator 3 and the comparator 7.

In this modified example, assuming that the variation of the change follows a normal distribution, the existence range of the deviation can be statistically determined from the following equation.

Here, X (i): the measured value of the physical quantity at the time point i: N: the number of samples of the physical quantity to be processed m: average value of the physical quantity σ: standard deviation of X A: allowable range of deviation.

Assuming that the distribution of X is a normal distribution, if A = ± 3σ, there is a deviation in this range with a probability of 99% or more. Therefore, if the value of A is used as an allowable range, abnormalities can be more easily detected. Can be detected.

In addition, the allowable range of the deviation can be automatically determined, and the determination of the allowable range of the variation from the value at the time of the normal operation can be automatically and extremely easily performed. Much time and effort can be dramatically reduced.

Further, the abnormality sign detecting means 1C shown in FIG. 12 considers this as an abnormality on condition that the deviation from the average value of the physical quantity continuously deviates from the allowable range for a predetermined time. The timer is built in the comparing means 4A.

The timer 22 counts the timing when a signal indicating the deviation from the allowable range is input from the comparator 7, and outputs a diagnostic symptom signal indicating that an abnormality is recognized when the deviation continues for a predetermined time. It is configured to output from.

According to such a modification, even when the measured value change of the physical quantity is caused by disturbance such as accidental noise,
This can be clearly identified, and the effect that unnecessary confusion does not occur can be further added.

Further, the abnormal sign detection means 1D shown in FIG. 13 is configured so that reset signals can be artificially input to the average value calculator 3 and further to the allowable range calculator 21. The feature is that the average value and the allowable range are calculated using the physical quantity and updated.

Therefore, according to such a modified example, it is possible to newly calculate the average value and the deviation of the physical quantity at the artificially changed stable point, and to automatically determine the existence range of the deviation according to the operating conditions. And an abnormality detection having a wide applicability in various plants can be performed.

In the description of the embodiment and the modification, the sensor 1 is exemplified as a flow meter, a pressure gauge, and the like.
The present invention is applicable to various instrumentation widely used in plants.

In addition, the abnormal sign detection means may appropriately use the above-described device according to the characteristics of the plant specific point to be measured or the like,
Alternatively, it may be configured to be used in combination. Further, the knowledge memory 9 may use a relationship in which a hypothesis and a symptom appearing when the hypothesis occurs are classified and stored for each plant condition. Reference numeral 8 may be such that the correspondence that best matches the current condition is selected from the description contents of the knowledge memory to perform the abnormality diagnosis. In such a case, the application range of the abnormality diagnosis can be further expanded. More elaborate diagnosis can be realized.

〔The invention's effect〕

As described above, according to the present invention, a diagnostic symptom indicating a symptom such as an abnormality of a plant is obtained in accordance with the operating condition of the plant and the like, and is output, and abnormality diagnosis is performed based on the symptom. Thus, there is an effect that a highly reliable diagnosis result can be obtained, and a plant abnormality diagnosis system that can avoid unnecessary diagnosis can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the overall configuration of a plant abnormality diagnosis system according to one embodiment of the present invention, FIG. 2 is a diagram showing a specific example in a liquid level control system of a plant to which the embodiment is applied, and FIG. FIG. 4 is a diagram showing classification criteria of the knowledge memory, FIG. 4 is a diagram showing an example of a correspondence relationship between a hypothesis stored in the knowledge memory and a symptom occurrence mode, and FIGS. 5 to 9 are knowledge examples of each hypothesis. And FIG. 10 to FIG. 13 are block diagrams showing different modifications of the abnormal sign detecting means. 1 ... abnormal sign detection means, 2 ... sensors as measurement means, 3 ... average value calculator, 4, 4A ... comparison means, 21 ...
Acceptable range calculator, 8 Diagnostic unit, 9 Knowledge memory, 10
... Output display section.

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Ryohei Funakoshi 2-5-1 Marunouchi, Chiyoda-ku, Tokyo Inside Mitsubishi Heavy Industries, Ltd. (72) Inventor Shigeru Kasada 11-11 Minamihama-cho, Chita City, Aichi Prefecture In-house (72) Inventor Koichi Takahashi 11 Minamihama-cho, Chita-shi, Aichi Idemitsu Kosan Co., Ltd. (56) References JP-A-60-108740 (JP, A) JP-A-56-17404 (JP, A) JP-A-59-77312 (JP, A) JP-A-S60-11657 (JP, A) JP-A-57-65497 (JP, U)

Claims (6)

(57) [Claims] A measuring means for measuring a physical quantity at a specific point of a plant; an average calculator for obtaining an average value of the physical quantities measured during a time section before every predetermined time; An allowable range calculator that determines an allowable range of the difference between the physical quantity and the average value indicating a range in which the physical quantity can be regarded as normal based on the existence range of the difference from the average value statistically determined based on the existing range; A comparison means for outputting as a diagnostic symptom signal whether or not a difference between the physical quantity at the present time and an average value of the physical quantities measured during the previous time section has exceeded the allowable range. A plant abnormality diagnosis system, comprising: a detection unit; and a diagnosis unit that inputs preset detection information such as abnormality together with the diagnosis symptom signal and diagnoses a plant abnormality or the like. 2. The plant abnormality diagnosis system according to claim 1, wherein the allowable range calculator determines a standard deviation by regarding a variation state of the existence range as a normal distribution, and calculates the standard range based on the standard deviation. A plant abnormality diagnosis system characterized by determining. 3. The system according to claim 1, wherein the average value calculator outputs a diagnostic symptom signal when a difference between the physical quantity and the average value exceeds an allowable range. A system for diagnosing plant abnormalities, wherein an average value is not updated during a period. 4. The plant abnormality diagnosis system according to claim 1, wherein said comparing means is provided on condition that a difference between said average value and said physical quantity deviates from an allowable range for a predetermined time continuously. And outputting a diagnostic symptom signal. 5. The plant abnormality diagnosis system according to claim 1, wherein a reset signal is supplied to said average value calculator and said allowable range computer, and said reset signal is supplied. A plant abnormality diagnosis system, wherein the average value and the allowable range are recalculated based on the physical quantity from the point of time. 6. The plant abnormality diagnosis system according to claim 1, wherein said abnormality detection information is classified for each plant operating condition. system.