JPH03277939A - Abnormality diagnostic method for plant - Google Patents

Abstract

PURPOSE:To improve real time property and man-machine interface for an operator by introducing the operating data for a plant and abnormality inspecting data, forming a diagnosis matrix, inputting the matrix into a diagnosing and processing part, and diagnosing the abnormality. CONSTITUTION:A diagnosis-matrix processing part 22 has an abnormality symtom matrix indicating the cause of abnormality and the strength of linking with the phenomenon (basic matrix) and a magnification matrix indicating the relationship with the abnormal level for every judged item of the cause of the abnormality and every item of the cause. When it is judged that a shaft vibration signal which is inputted into an abnormality diagnosing device 20 is abnormal in an abnormality detecting part 14, a diagnosis matrix is formed by using the basic matrix and the magnification matrix based on the abnormality detecting data table for a plant transmitted from the detecting part 14 and the operation data table of the plant transmitted from a state judging part 21. A diagnosing part 23 receives the diagnosis matrix and computes the possibility of the cause of each abnormality. The result of the diagnosis processed by an adding method and the like for every item is displayed 17, and the data are stored 18.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Application Field) The present invention provides a method for diagnosing abnormalities in operating equipment, such as shaft vibration of large rotating machinery in a power generation plant and performance deterioration of equipment. Regarding.

(Prior technology) Thermal power plants are equipped with numerous rotating machines, including a main turbine/generator with a rotor weighing several hundred tons that rotates at high speed. A large number of heat exchangers and various hydraulic equipment and control equipment are installed.

In such plants, especially in high-speed rotating machinery, many of the abnormalities that occur during operation often appear as changes in vibration phenomena, and there is a risk that they may develop rapidly and develop into serious troubles. It is equipped with detailed operation monitoring equipment and protection and control equipment in the event of an abnormality.

Particularly in recent years, the diversification of electricity demand has required frequent operations such as frequent startups and shutdowns and rapid load fluctuations. There is an increasing demand for the ability to quickly repair the situation.

For this reason, various plant abnormality diagnosis devices and equipment diagnosis devices have been proposed. For example, regarding shaft vibration diagnosis of rotating machines, there is ``Abnormality Diagnosis Device for Rotating Body'' in Japanese Patent Application Laid-Open No. 61-128128.

FIG. 6 shows an example in which a conventional vibration diagnosis device is applied to a main turbine/generator unit of a thermal power plant.

In the figure, a high/intermediate pressure turbine 2, low pressure turbines 3a, 3b, and a generator 4 of a main turbine/generator unit 1 are shown.
Vibration detectors 6a to 6n for detecting shaft vibration are attached to bearings 58 to 5n that support each rotor. Also included in the unit 1 are a tachometer 7 that detects the rotational speed of the rotor, a phase angle reference pulse transmitter 8 that transmits reference pulses for detecting the vibration phase angle at each bearing 58 to 5n, and a power generator. Various sensor groups 9 are installed to detect plant operating conditions such as machine output (load), steam temperature, and bearing temperature.

Signals from the tachometer 7 and various sensor groups 9 are sent to a control computer 10 that monitors and controls the operation of the plant, and signals from the vibration detectors 6a to 6n and the phase angle reference pulse transmitter 8 are used for vibration monitoring. The information is sent to the control computer 10 via the device 11 to support operation control by the operator.

That is, the vibration monitoring device 11 determines the vibration state in each of the bearings 5a to 5n, and if it determines that the vibration is at a predetermined signal level or higher, transmits an alarm signal or a turbine trip signal to the control computer 10. In addition, the vibration detector 6a~
The vibration signal from 6n and the signals from the various sensor groups 9 are also sent to the analog recorder 12 and are continuously recorded.

The plant abnormality diagnosis device M13 includes an abnormality detection processing section 14,
It is composed of a state judgment processing section 15, a diagnosis processing section 16, a display processing section 17, and a data storage processing section 18.
A signal from the vibration monitoring device 11 is guided to an abnormality detection processing section 14.

The abnormality detection processing unit 14 detects signs of an abnormal phenomenon from trends such as vibration amplitude values, amplitude increase rates, frequency components, or phase changes, and starts diagnosis of the cause when it is determined to be an abnormal sign. , transmits the detection information to the diagnostic processing section 16.

On the other hand, operating state information other than vibration information, such as information such as the rotation speed and load factor of the main turbine/generator, is provided to the control computer 10.
The information is manually inputted to the state judgment processing section 15 via. The output of this state judgment processing section 15 is sent to the abnormality detection processing section 14,
This information is used as information when an abnormality is detected, is sent to the diagnostic processing section 16 and used as plant information when diagnosing the cause, and is sent to the display processing section 17 to display the monitoring status and diagnosis results.

The diagnostic processing unit 16 contains a symptom matrix.

There are many types and contents of this symptom matrix, but for example, Fig. 7 shows a part of the symptom matrix showing the causal relationship between vibration causes and phenomena in a rotating body, and Fig. 8 shows performance performance. 4 illustrates a symptom matrix for deterioration diagnosis;

The diagnostic processing unit 16 uses these symptom matrices to estimate the cause of the abnormality based on the abnormality detection information and the plant information.

The output of the diagnostic processing section 16 is sent to the display processing section 17 and is used as direct support and guide information for operators and operators. Furthermore, data that needs to be saved is led to the data storage processing section 18 and saved as historical data. These historical data include reference values indicating normal conditions and records of past abnormalities, and are processed by the abnormality detection processing unit 1 as necessary.
4 and the diagnostic processing unit 16 for use.

(Problem to be solved by the invention) Real-time performance is required to perform diagnosis using online plant signals and data, such as diagnosing shaft vibration of rotating machines and plant performance deterioration. Regarding shaft vibration phenomena that have the possibility of progressing, it is necessary to make judgments according to the operating status and abnormal conditions of the plant.

However, in the conventional abnormality diagnosis device, even if the abnormality level is recognized at the stage of the monitoring function by the abnormality detection processing unit 14, at the stage of the symptom matrix for cause diagnosis, the symptom stage where the cause of the abnormality symptom is missing is determined. It serves as a guideline.

For this reason, even when the abnormality sign is at a low level or when the phenomenon has progressed, the diagnostic processing criteria remain unchanged, and there is a risk that, for example, a serious abnormality sign may be mistaken for a minor vibration cause.

In order to avoid such risks, it is possible to use multiple symptom matrices, but this would complicate the processing such as selection of criteria, risk impairing the real-time nature of diagnosis, and the diagnostic equipment There is a need to significantly increase the processing power of computer systems.

The present invention was made in order to solve the above-mentioned drawbacks of the conventional technology, and by using a diagnostic algorithm suitable for time-series changes in abnormal phenomena, it provides excellent real-time performance and man-machine operation for operators. The purpose is to provide a method for diagnosing plant abnormalities.

[Structure of the invention]

(Means for Solving the Problems) The abnormality diagnosis method of the present invention detects an abnormality in a plant, analyzes the cause, and diagnoses the abnormality to support the decision-making work of a plant operator. The diagnosis matrix processing unit is equipped with an anomaly symptom matrix that indicates the strength of the connection between the event and the phenomenon, and a magnification matrix that indicates the relationship between the abnormality cause judgment items and the abnormality level for each cause item. The present invention is characterized in that information is introduced to create a diagnostic matrix, and this diagnostic matrix is input to a diagnostic processing section to diagnose abnormalities in the plant.

(Function) In the plant abnormality diagnosis method of the present invention configured as described above, for example, taking vibration diagnosis of rotating machinery as an example,
First, the shaft vibration signal input to the abnormality diagnosis device is judged by the abnormality detection processing section to see if there is an abnormality sign, and if it is determined that there is an abnormality sign, the item and abnormality level are created as an abnormality detection information table. .

Next, among the operating status values of the plant, information that changes the setting values for abnormality detection is sent from the status judgment processing unit to the abnormality detection processing unit, and information directly used for diagnosis is processed as an operating information table in the diagnostic matrix. sent to the department.

The diagnosis matrix processing unit has an abnormality symptom matrix (basic matrix) that shows the strength of the connection between the cause of the abnormality and the phenomenon, and a magnification matrix that shows the relationship between the judgment items of the abnormality cause and the abnormality level for each cause item. ing.

When an abnormality is detected, a diagnosis matrix is created using the above basic matrix and magnification matrix from the abnormality detection information table sent from the abnormality detection processing section and the operation information table sent from the status judgment processing section. create. This diagnostic matrix is displayed as a multi-layered weighting coefficient, which is obtained by adding weights based on the abnormality level and cause item to the weighting coefficients for each item included in the basic matrix.

On the other hand, the diagnostic processing unit is the part that calculates the possibility of each abnormality cause using this diagnostic matrix, and the diagnostic results processed by the addition method for each item are sent to the operator in the display processing unit and data storage processing unit. The display of results and data saving operations are performed as support information.

Here, in the algorithm of the present invention, plant information and abnormality levels are each stored as table information, so there is no need to divide the judgment items of the abnormality symptom matrix (basic matrix) into small items and create a large matrix. In addition to being able to describe the characteristics relatively easily, it is also effective in reducing verification work, and is effective not only during processing but also when creating diagnostic data.

(Example) Hereinafter, an example in which the present invention is applied to vibration diagnosis of a rotating machine will be described. In the following description, the same parts as in FIG. 6 are denoted by the same reference numerals, and the description thereof will be omitted unless necessary.

In FIG. 1, an abnormality diagnosis device 20 includes an abnormality detection processing section 1.
4, a state judgment processing section 21, a diagnosis matrix processing section 22, a diagnosis processing section 23, a display processing section 17, and a data storage processing section 18.

The abnormality detection processing unit 14 analyzes the vibration signal (analog signal) sent via the vibration monitoring device 11 and detects signs of abnormal phenomena. The state judgment processing section 21 analyzes plant information (digital signals) from the control computer 10 and the like, and outputs the results to the abnormality detection processing section 14, the display processing section 17, and the diagnosis matrix processing section 22.

The diagnostic matrix processing unit 22 creates a diagnostic matrix that takes into account the abnormality level from the abnormality detection information and the plant information. Further, the diagnostic processing unit 23 determines the cause of vibration based on the diagnostic matrix, and the output thereof is sent to the display processing unit 1.
7 and is used as direct support and guide information for operators and operators. Furthermore, data that needs to be saved is led to the data storage processing section 18 and saved as historical data.

FIG. 2 shows the flow of signals introduced into the abnormality diagnosis device, in which plant data is led to the status judgment processing section 21,
Region determination 21a is performed for each rotation speed and operating load, and threshold setting 21b for abnormality determination is performed. Further, plant information necessary for diagnosis is also sent to the diagnosis matrix processing section 22.

The abnormality detection processing unit 14 determines vibration data. When making this determination, not only the amplitude and increase rate of the vibration data but also processing such as frequency analysis and vector analysis are used to detect abnormalities.
a, and detects the presence or absence of an abnormality 14b based on the threshold value determined in the previous operating state. In this case, historical data stored in the data storage processing section 18 is used as past data to be compared with the reference data.

If no signs of abnormality are detected by the above processing <N
O) returns to the original 21g and periodically analyzes the data. If abnormal symptoms are detected during that time (YES),
The data is sent to the next diagnostic matrix processing section 22.

The diagnostic matrix processing unit 22 first creates a detection information table 22a based on the sent data. This includes an operation information table 22b and an abnormality detection information table 22c as shown in FIGS. 3A and 3B.

Next, based on the abnormality symptom matrix 22d, it is determined whether the conditions given in the driving information table 22b and the abnormality detection information table 22c are met, and a basic matrix 22e is created. Figure 4 shows symptom matrices A and B showing the causal relationship between vibration causes, driving range, detection conditions, etc.
, C, . . .

Subsequently, the basic matrix 22e is weighted based on the abnormality detection information table 22c and the magnification matrix 22f as illustrated in FIG. 3C to create a diagnostic matrix 22g. An example of the format of this diagnostic matrix is shown in FIG.

The diagnosis processing unit 23 determines the cause based on the diagnosis matrix 22g. In this case, for example, diagnostic matrix 2
2g by adding the points of each element of the matrix for each cause,
The probability and high probability are determined based on the score for each cause as a percentage of the total score.

In the above, the magnification matrix 22f works particularly effectively when trying to determine items that are in a subordinate relationship among the determination items of abnormality causes.

In other words, for example, when expressing the causal relationship with the rubbing phenomenon using the vibration amplitude value and the amount of vector change, the basic symptom matrix will give a strong causal relationship to both the change in the amplitude value and the amount of vector change. This condition should naturally appear in the amount of vector change when the amplitude value is large,
The magnification matrix is determined so that abnormality detection of vector change amount becomes large proportionately depending on abnormal level of amplitude value.

Conversely, when the amplitude value is large, if no abnormality is detected in the amount of vector change, it is possible to more strongly reduce the possibility of a rubbing phenomenon and increase the possibility of other cause items.

As described above, according to the embodiment of the present invention, it is possible to obtain a diagnostic matrix having a weight according to the level of the abnormality from the abnormality detection information.
The condition was difficult to distinguish from other cause items, but
When the level of abnormality increases, it is possible to obtain that information, clearly differentiate it from other causes, and obtain a diagnostic result with a strong possibility.

〔Effect of the invention〕

As described above, according to the plant abnormality diagnosis method of the present invention, the weight of each element of the diagnosis matrix is determined based not only on the abnormality detection conditions, but also on the level of abnormality, the operating status, and the conditions between each cause. can be changed dynamically.

Therefore, when the driving situation or abnormality level changes, it is possible to easily perform diagnosis according to the situation. In this case, the basic symptom matrix of causal relationships can be summarized into a relatively simple small matrix, which not only improves the judgment speed of the computer but also makes it possible to easily construct a knowledge base for diagnosis. There are secondary effects.

Furthermore, according to the present invention, abnormalities can be detected online and in real time at an early stage, diagnosis can be performed according to the progress of the abnormality, and by providing timely information to plant operators, plant This can contribute to improved safe driving and reliability.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an example in which the present invention is applied to vibration diagnosis of a rotating machine, FIG. 2 is a flowchart explaining the operation of the embodiment of the present invention, and FIG. 3 (A). (B) and (C) are explanatory diagrams illustrating the driving information table, abnormality detection information table, and magnification matrix in the embodiment of the present invention, FIG. 4 is an explanatory diagram illustrating the symptom matrix in the embodiment of the present invention, FIG. 5 is an explanatory diagram illustrating a diagnostic matrix in an embodiment of the present invention,
Fig. 6 is a block diagram illustrating a conventional plant abnormality diagnosis device, Fig. 7 is an explanatory diagram illustrating a part of a symptom matrix showing the causal relationship between vibration causes and phenomena in a rotating body, and Fig. 8 is a performance deterioration diagnosis. It is an explanatory diagram illustrating a symptom matrix regarding. 1... Main turbine/generator unit, 2... High/medium pressure turbine, 3a, 3b... Low pressure turbine, 4...
Generator, 5a to 5n... Bearing section, 6a to 6n... Vibration detector, 7... Tachometer, 8... Phase angle reference pulse transmitter, 9... Various sensor groups, 10. ... Control computer, 11... Vibration monitoring device, 12... Analog recorder, 13... Abnormality diagnosis device, 14... Abnormality detection processing section, 15... State judgment processing section, 16 ...Diagnosis processing section, 17... Display processing section, 18... Data storage processing section, 20... Abnormality diagnosis device, 21... State judgment processing section, 22... Diagnosis matrix processing section 23...
Diagnostic processing unit.

Claims (1)

[Claims] In an anomaly diagnosis method that supports plant operators' judgment work by detecting plant anomalies, analyzing and diagnosing their causes, an anomaly symptom matrix that shows the strength of the connection between the cause of an anomaly and a phenomenon, and an anomaly cause are used. A diagnostic matrix is created by introducing plant operation information and abnormality detection information into the diagnostic matrix processing unit, which has a magnification matrix that shows the relationship with the abnormality level for each judgment item and cause item. 1. A method for diagnosing an abnormality in a plant, the method comprising diagnosing an abnormality in a plant by inputting data into a section.