JPH06307921A - Diagnostic monitoring system for rotating machine - Google Patents

Abstract

PURPOSE:To assist instantaneous decision required for an operator or a maintenance crew by allowing real time discrimination between a quickly advancing abnormality and a slowly advancing abnormality. CONSTITUTION:Data of a rotating machine is fed in real time to a data processing means 3 where the data is processed. An abnormality detecting means 5 monitors the processed data in real time and compares the processed data with a threshold data thus detecting an abnormality. An abnormality advancing speed decision means 8 compares a newest set data with a data set in past and calculates the advancing speed of abnormality. A diagnostic result decision means 9 diagnoses a quickly advancing abnormality and a slowly advancing abnormality while discriminating from each other based on abnormality detection results from the abnormality detecting means 5 and an abnormality advancing speed from the abnormality advancing speed decision means 8.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a rotating machine monitoring and diagnosing device for monitoring an abnormality of a rotating machine installed in a power plant, a steelmaking plant, a chemical plant, etc. and diagnosing the cause of the abnormality when the abnormality occurs. .

[0002]

2. Description of the Related Art Rotating machines such as power plants tend to be continuously operated for a long period of time and their operating states are diversified, so that long life and highly reliable operation are required. The development and practical application of a monitoring and diagnostic device for rotating machinery are in progress.

This diagnostic device can be divided into two types, one of which is a type that detects abnormalities that are rapidly progressing by monitoring current vibration data and the like, and the other is a type that diagnoses the cause. There is a type of diagnosing the cause by analyzing the data for a long period of time, for example, from a periodic inspection to the next periodic inspection, grasping the tendency, detecting a slow progressing abnormality such as secular change.

First, as the former type of detecting a rapidly advancing abnormality, a steam turbine / generator vibration diagnosis device currently installed in a thermal power plant is disclosed in, for example, Hisa et al., "Large steam turbine shaft vibration diagnosis system", Thermal power,
Vol. 38, No. 12 (published in December 1987), pages 1387 to 1398, etc. have been put into practical use.

This type of diagnostic device monitors vibration data or the like in a short cycle of, for example, 2 seconds, and compares it with a preset threshold value to monitor abnormality. Furthermore, in order to narrow down the possibility of the cause of abnormality, the vibration frequency is used to largely classify the vibration frequency into unbalanced vibration, unstable vibration / subharmonic vibration, harmonic vibration, etc. In order to identify, we use the survey results such as the correlation between the process output and the related process data such as generator output, condenser vacuum, bearing metal temperature, etc.

However, in order to subdivide the cause to the level at which the operator or maintenance staff can decide the countermeasure,
In addition to the present data, it is necessary to have expertise and advanced analysis technology for specializing the transition or vibration from the past data, and it is difficult to incorporate these into a computer. For this reason, at present, the level of grouping the causes of abnormality by vibration classification is at a practical level.

Next, as the latter type of detecting a slow progressing abnormality, a system currently developed and introduced for a thermal power plant, "operating state tendency management system", electric field technology, June 1992. Issue, pages 7 to 13 are at the stage of practical application.

This type of diagnostic device uses vibration data and the like for a long period (for example, one month) for a long period (for example, two years).
Collect over. Further, these data are subjected to statistical processing such as linear regression, exponential regression, multiple regression, etc. off-line in response to the request of the operator or maintenance staff. By processing these data and displaying the data, it is possible to detect abnormal signs such as slow deterioration over time, and support judgment such as "how long can the operation be continued?"

[0009]

A steam turbine / generator vibration diagnostic device installed in a thermal power plant belongs to a type that detects a rapidly progressing abnormality and diagnoses the cause.
The feature is that an abnormality is detected early by catching the current change in vibration.

For example, if an abnormal value of vibration of a certain bearing exceeds a threshold value, an abnormality is detected, and if the result of frequency analysis indicates that the rotation synchronization component is significantly larger than the other frequency components, It is determined to be unbalanced vibration. At this time, if an instantaneous change such as a sudden change phenomenon is observed in the change in vibration, it can be determined that the cause of the abnormality is likely to be a defect in the rotating part (dropping of balance weight, blade scattering, etc.).

However, even if the cause of the same unbalanced vibration is abnormal progress such as rotor bending (change over several hours) due to thermal imbalance at turbine startup, and rotor aging (change over half a year). There are some, and it is difficult to distinguish them from the data in the immediate vicinity.

Therefore, in order to distinguish and detect such slow progressing abnormalities, a diagnostic device such as an operating state tendency management system is separately required as a diagnostic device of a type that detects slow progressing abnormalities.

For example, FIG. 13 shows a case where vibration data during rated load operation is collected once a month and displayed in time series. Here, in the periods of time A to B and time C to D, the vibration tends to increase due to the bending of the rotor over time. On the other hand, in the period from time B to C, the imbalance amount is decreased due to the loss of the rotating part and the like, and the vibration is decreased. In such a case, when the time to reach the control value is predicted using the data of times A to D at the current time D, it becomes time F, and it is determined that the equipment should be adjusted during the periodic inspection.

However, since an abnormality such as a loss of a rotating portion actually occurs between times B and C, it is necessary to make a prediction using the data from times C to D. In such a case, , The control value will be reached at time E. In this case, it means that the equipment needs to be adjusted before the periodic inspection.

Therefore, in the conventional trend management method,
Since short-term changes are easily ignored, there is a possibility that wrong trend prediction may be performed, and it may cause the operator or maintenance staff to be alert.

The present invention has been made in consideration of the above-mentioned circumstances, and makes it possible to distinguish between a fast-moving abnormality and a slow-moving abnormality in real time, and to support the immediate judgment required by an operator or maintenance personnel. It is an object of the present invention to provide a monitoring / diagnosing device for a rotating machine, which is effective for

[0017]

In order to solve the above-mentioned problems, according to claim 1 of the present invention, vibration data of a rotating machine and data relating to an operating state are input in real time, and frequency analysis and change rate calculation are performed. , Data processing means for processing data such as a difference from the reference value data, abnormality detection means for monitoring this processing data in real time and comparing it with threshold data to detect an abnormality, and the processing data is external A data settling determination unit that determines whether or not the settling state is unaffected, and an abnormal progression rate determination unit that compares the latest settled data with the past settled data to calculate an abnormal progression rate. , A diagnostic result determination method for distinguishing between a fast-moving abnormality and a slow-moving abnormality based on the abnormality detection result of the abnormality detecting means and the abnormal progress speed of the abnormal progress speed determining means. Characterized by comprising and.

According to a second aspect of the present invention, in addition to the first aspect of the invention, a reference value data storage section for storing data during normal operation of the rotating machine as reference value data for each operating state,
Threshold data storage unit that stores the comparison value to determine whether the processing data is abnormal as threshold data for each operating condition, and the settling history that stores the processing data in the settling state for a long time A data storage unit, a diagnosis history storage unit that stores the abnormality detection result of the abnormal progress speed determination unit and the diagnosis result of the diagnosis result determination unit, and a display process that displays the diagnosis result, the diagnosis history, and the settling history data in an overlapping manner. And means are provided.

[0019]

According to the present invention having the above-mentioned structure, in a monitoring and diagnosing device for a rotating machine, an abnormality of the rotating machine is detected in real time, and a cause of the abnormality is diagnosed by distinguishing between a fast-moving abnormality and a slow-moving abnormality. Therefore, it is possible to identify abnormalities with extremely rapid progress, abnormalities with gradual progress, and abnormalities with extremely slow progress, which were difficult to identify in real time by the conventional diagnostic device. Become.

[0020]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 2 shows an example in which the monitoring / diagnosing device for a rotating machine according to the present invention is installed in a power plant as an embodiment of the present invention.

The rotary machine monitoring and diagnosing device of this embodiment is
A steam turbine / power generation unit 1 including a high / middle pressure turbine 14, low pressure turbines 15a and 15b, and a generator 16.
18i, vibration detectors 18a, ... 18i at bearings 17a, ... 17i, tachometer 19, phase reference pulse oscillator 20, and various sensor groups 21 are attached to the steam turbine / power generation unit 13.

Detection signals from the vibration detectors 18a, ... 18i and various sensor groups 21 are output to the central operation panel 22. The central control panel 22 includes an operation monitoring device 23, a vibration monitoring device 24, an alarm device 25, and a continuous recorder 26. The operation monitoring device 23 and the vibration monitoring device 24 are connected to the monitoring and diagnosing device 1 of the rotating machine, and the data relating to the operating state and the overall,
Vibration data such as a vibration vector and a raw waveform signal are output to the rotary machine monitoring and diagnostic device 1 in real time.

In addition, the operation monitoring device 23 constantly checks the operating condition data, and the vibration monitoring device 24 constantly checks the vibration data (overall). When the data exceeds a predetermined limit value, a command signal is sent to the alarm device 25. It is emitted and an alarm and an automatic trip signal are output. The output signals of the vibration detectors 18a, ... 18i, the tachometer 19, the phase reference pulse oscillator 20, and the various sensor groups 21 are
It is also sent to and recorded in the analog recorder 26. On the other hand, the rotating machine monitoring and diagnosing device 1 makes a diagnosis based on the sent data, and displays the result on the man-machine interface 12.

FIG. 1 shows a concrete configuration of a monitoring / diagnosing device 1 for a rotary machine according to this embodiment.

The monitoring / diagnosing device 1 includes a reference value data storage unit 2 for storing data during normal operation of the rotating machine as reference value data for each operating state, vibration data of the rotating machine, and operating state of the rotating machine. Data processing means 3 for inputting in real time data related to, reference value data, etc., and performing data processing such as frequency analysis, change rate calculation, difference with reference value data, and whether or not this processed data is abnormal A threshold value data storage unit 4 that stores a comparison value as a threshold value data for each operating state, and an abnormality that detects the abnormality by monitoring the processing data in real time and comparing it with the threshold value. The detection means 5 is provided.

Further, the monitoring and diagnosing device 1 determines that the above-mentioned machining data is in a settling state (a state in which the variation of the value is within a normal variation range) without being affected by external influences such as a change in the state of the plant. By comparing the latest settling data and the settling data in the past with the settling history data storage section 7 for storing the settling processing data for a long period of time, the slow progressing abnormality is detected. Abnormal progress speed determination means 8 for detecting and calculating the progress speed of the abnormality
It has and.

Further, the monitoring and diagnosing device 1 includes a diagnostic result judging means 9 for distinguishing between a fast-moving abnormality and a slow-moving abnormality based on the abnormality detection results of the abnormality detecting means 5 and the abnormal progress speed determining means 8. A diagnosis history storage unit 10 for storing the abnormality detection result of the abnormal progress speed determination unit 8 and the diagnosis result of the diagnosis result determination unit 9, and a display processing unit for displaying the diagnosis result, the diagnosis history, and the settling time history data in an overlapping manner. 11 and 11.

Next, the operation of the rotary machine monitoring and diagnosing device 1 in this embodiment will be described.

First, the vibration data and the data relating to the operating state (for example, the number of revolutions, the load, the metal temperature, etc.) input to the monitoring and diagnosing device 1 for the rotating machine are periodically sent to the data processing means 3 in real time. Sent. In this data processing means 3, frequency analysis,
The change rate is calculated, the difference from the reference value data is processed, and the data is processed into a form that can be compared with the threshold value for abnormality detection.

For example, FIG. 3 shows data processing for calculating the amount of change in the vibration vector. In this data processing, when input data (vibration vector) V is given, reference value data (vibration vector) V ₀ corresponding to the operating state such as the rotational speed or load at that time is referred from the reference value data storage unit 2,
This is a process of calculating the variation amount | V−V ₀ | of the vibration vector. Note that V _t is threshold data.

FIG. 4 shows an example of data processing in which a specific frequency component is calculated by frequency-analyzing the input vibration data. In FIG. 4, ωn is the rotation period,
ω ₁ , ω ₂ , ω ₃ are eigenvalues (rotor natural frequencies).

Next, the abnormality detecting means 5 judges the abnormality detection from the data (first column in FIG. 5) input from the data processing means 3 using the abnormality detecting logic shown in FIG. 5 (2 in FIG. 5).
The third column) is performed to obtain the abnormality detection result (third column in FIG. 5).

Specifically, when the data is first input, the threshold value data storage unit 4 refers to the threshold value data corresponding to the operating state such as the rotational speed or the load at that time. Next, the input data and the threshold data are compared in the form as shown in FIGS. As a result, as shown in FIG. 5, the abnormality detection result is obtained depending on which abnormality detection criterion is satisfied.

For example, when the criterion of abnormality detection "more than the amplitude increase rate c" is satisfied, the abnormality detection result is "large amplitude". Further, when the criterion of abnormality detection “rotation synchronization component predominance” is satisfied, the abnormality detection result becomes “unbalanced vibration”.

On the other hand, a part of the data output from the data processing means 3 is sent to the data settling determination means 6 in response to a request. The data settling determination means 6 inputs data under specific conditions and further determines whether the input data is in a settling state. First, as the input data, data whose value increases or decreases according to the progress of the abnormality of the rotating machine over time is selected.

For example, if the purpose is to detect rotor bending over time, the rotation synchronization component is selected, and if the purpose is to detect rotor cracks, the double synchronization component is selected. Next, as a condition for inputting these data, sampling is periodically performed under the same operating condition.

For example, in a nuclear power plant or the like, most of the operation is at a rated load. Therefore, the operating condition is checked every 6 hours, and if operating at the rated load, sampling is performed 30 times in a cycle of 2 seconds. As it happens, sampling is not performed and the next opportunity is waited for unless it is in rated load operation.

Next, it is judged from the sampled data whether or not the data is settled (the variation of the value is within the range of normal variation), and if the data is settled, a representative value is calculated. If not settled, sample again or wait for the next sampling opportunity. If the unstabilized state continues for a long time, no data is sent to the abnormal progress speed determination means 8 during that time, and therefore a slow-moving abnormality is not detected. However, since the data is sent from the abnormality detection means 5 to the diagnosis result determination means 9, the diagnosis is performed.

Here, as a method for judging the settling,
There is a method of determining whether or not the standard deviation of the sampled data or the difference between the maximum value and the minimum value is within the range of normal values (determined and set based on the operating results after periodic inspections). Further, since the representative value is a value representative of the sampled data, it is calculated by the method of taking the average value of the sampled data. The representative value is stored in the settling history data storage unit 7, and at the same time, the abnormal traveling speed determination unit 8 is provided.
Also sent to.

In the settling time history data storage unit 7, as shown in FIG. 6, the following three types of data are stored in time series for each diagnostic result item. First, the item in the first column of FIG. 6 is the date and time when the representative value was calculated, the item in the second column is the value at the time of settling (representative value) calculated by the data settling determination means 6, and the item in the third column is The present invention relates to the detection result of the slow-moving abnormality performed by the abnormal-movement speed determination means 8.

The abnormality detection result of the third column is stored as abnormal if it is determined that an abnormality of slow progress has occurred at the date and time of the first column, and normal otherwise. However,
Adjustment of rotating machinery is performed by periodic inspection, repair, etc.
When the normal state is restored, the point in time can be defined as “reset of abnormality detection” and the already saved information (normal) can be updated. This "reset anomaly detection"
Is saved by inputting the date and time at that time from the man-machine interface 12 when the operator or maintenance person sees the value at the time of settling and determines that the normal state has been restored.

The abnormal progress speed determination means 8 performs the following two processes in order to detect a slow-moving abnormality and to calculate the abnormal progress speed. That is, in the process 1, whether or not the latest settling value (representative value) is abnormal is determined using the past normal settling value. Further, in the process 2, in the case of an abnormality, the progress speed and the amount of change of the abnormality are calculated from the date and time of the latest value and the date and time of the past value used for the abnormality determination, and the result is stored.

Prior to the description of these processing procedures, the principle of abnormality detection will be described below with reference to FIG.

[0045] Figure 7 is obtained by displaying the settling time value (representative value) in time series, the value D _A is the most recent value at time T _A. However, it is assumed that the data before time T _B does not include a value determined to be abnormal. Where the latest value D _A
The previous value D _B is used to determine whether or not is abnormal. The difference | D _A −D _B | of this value is the threshold value ε for abnormality detection.
Since it is smaller, no abnormality is detected at this point.

[0046] Similarly, when the carry on determined using sequential historical data, at time T _C, the difference between the values | D _A -D _C
Since | exceeds the threshold value ε, an abnormality is detected at this point. The difference between these values is the amount of change caused by the abnormality. Further, the time required to change by the threshold value ε (about T _A −T _C ) is the abnormal progress speed.

FIG. 8 shows a processing procedure of the abnormal traveling speed determination means 8. This process is performed for each item of the diagnosis result. Figure 8
In, the processing steps S1 to S6 are processing relating to abnormality detection. Further, processing steps S7 to S9 are processing relating to the calculation of the abnormal progress speed. Here, as an example of the item of the diagnosis result, a case of determining an abnormality of “unbalanced vibration (# 1 bearing)” will be described.

First, in processing step S1, the latest data (hereinafter, referred to as
The latest data. ) Is entered. This latest data consists of the value at the time of settling of the rotation synchronization component (# 1 bearing) and the date and time when the value was calculated.

In processing step S2, it is judged whether or not the latest data is abnormal by using the data output from the data settling judgment means 6 one cycle before the latest data (hereinafter referred to as the previous input data). The determination process of this process step S2 is shown in FIG.

First, the latest data is sent to the processing step S2. This latest data is stored in the history data storage section 7 during settling.
Will also be saved. Here, for the sake of explanation, if the latest data value is D _i , the previous input data value is D _i−1 , ..., and the oldest input data value is D ₁ , these data are as shown in FIG. The time history data is stored in the settling time history data storage unit 7.

Since the abnormality detection column stores the abnormality detection result by the abnormal progress speed determination means 8, the latest data is blank at this point.

Further, in the processing step S2, the latest data and the previous input data are input together. Next, if the previous input data is determined to be abnormal, the process proceeds to processing step S3. If not, the abnormality is detected by the following equation.

[0053]

## EQU1 ## | D _i −D _i−1 |> ε If this expression is satisfied, it is determined that the latest data is abnormal, and the process proceeds to step S6. If this equation is not satisfied, it is determined whether or not the abnormality detection information of the previous input data is abnormality detection reset. As a result, if the abnormality detection reset, it is determined that the latest data is normal,
Proceed to processing step S6. If not, it is assumed that no abnormality is detected here and the process proceeds to step S3.

In this processing step S3, similarly to the processing step S2, it is determined whether or not the latest data is abnormal by using the input data two times before. Here again, if normality or abnormality cannot be determined and no abnormality is detected, further determination is made with the input data before the input data two times before.

After that, similarly, whether or not the latest data is abnormal is sequentially determined by using the past input data.
However, if it is not determined whether it is normal or abnormal even if the oldest input data (last data) is determined in processing step S4, it is determined to be normal in processing step S5.

In process step S6, it is determined whether the latest data is normal or abnormal, and the next process is determined. If it is normal, the process proceeds to step S5, and finally it is determined to be normal. If abnormal, the process proceeds to step S7. In this processing step S7, the difference between the date and time of the latest data stored in the settling history data storage unit 7 and the date and time of the data used when the abnormality is determined is calculated and defined as the speed of progress of the abnormality. The longer this time, the slower the progress speed of the abnormality.

However, when the data collection period is long (for example, the progress of one month is slow), the detection of the abnormality may be delayed a little. Therefore, the progress rate of the abnormality is calculated by the following equation.

[0058]

V _i = (T _i −T _n ) ε / (D _i −D _n ) In this equation, V _i is the speed of progress of abnormality, T _i is the date and time of the latest data, and T _n is used for the determination. Date and time of data, D _i
Is the value of the latest data, D _n is the value of the data used for the determination (where 1 ≦ n <i), and ε is the threshold value for abnormality detection.

Further, in processing step S8, the difference between the value of the latest data and the value of the data used for the determination is calculated by the following equation, and defined as the amount of change in the data. The amount of change in this data is expressed as D _i −D _n . A processing step S9 stores the result of abnormality detection (normal or abnormal) in the settling history data storage unit 7 (third column in FIG. 6). The latest data value at the time of abnormality detection, the speed of abnormality progress, and the amount of data change are stored in the diagnostic history storage unit 10.

Next, the diagnosis history storage unit 10 is shown in FIG.
It will be described based on.

The diagnosis history storage unit 10 stores the abnormality detection result and the diagnosis result for each item of the diagnosis result. FIG. 10 shows a case where the item of the diagnosis result is unbalanced vibration (# 1 bearing). First, as the date and time in the first column, the date and time when the abnormality is detected by the abnormality detecting means 5 or the abnormal progress speed determining means 8 is recorded. The calculation results of the abnormal progress speed determination means 8 are recorded as the amounts of change in the abnormal progress speed in the second row and the rotation synchronization component in the third row.

Therefore, when the abnormality detecting means 5 detects an abnormality, the second and third columns are blank. The value of the rotation synchronization component when the abnormality is detected is recorded in the fourth column. In the final fifth column, the diagnostic result obtained from the diagnostic result judging means 9 is recorded.

Further, the diagnosis result judging means 9 is shown in FIG.
It will be described based on.

The diagnosis result judging means 9 inputs the abnormality detection results of the abnormality detecting means 5 and the abnormal traveling speed judging means 8,
The diagnostic result is obtained using diagnostic logic such as a causal matrix table. FIG. 11 shows a causal matrix table when the item of the diagnosis result is unbalanced vibration (# 1 bearing). This causal matrix table is prepared for each diagnostic result item.

First, the abnormality detection result on the horizontal axis will be described. The first abnormal progress speed is determined by the level of the abnormal progress speed input from the abnormal progress speed determination means 8. For example, if the rate of progress of the abnormality is less than 3 hours, "no change", if 3 hours or more and less than 6 months, "3
"Time" is decided.

Items after the second amplitude increase rate are determined by inputting the abnormality detection result from the abnormality detecting means 5. When the abnormality detection result is determined in this way, the diagnosis result is evaluated according to the symbols ◯, Δ, and × in the table. For example, if the speed of progress of the abnormality is "6 months", the diagnostic result is "the possibility of aged rotor bending (slow progressing abnormality) is large (marked with a circle)".

If the rate of increase in amplitude is "large", the diagnostic result is that there is a high possibility that there is a defect in the rotating portion (abnormality of rapid progress) (marked by "○"), bending of the thermal rotor, etc. (abnormality of slow progress). There is almost no possibility of () (X mark). "

As other symbols, “Δ” is a little possible and “−” is unknown. By such a method, whether the abnormality is a fast-moving abnormality or a slow-moving abnormality is distinguished, and the most probable one is obtained as a final diagnosis result. This diagnosis result is stored in the diagnosis storage unit 1 of FIG.
Stored in 0.

As the final processing, the display processing means 11 will be described with reference to FIG. FIG. 12 is an example of a screen displayed on the man-machine interface 12, and the display processing means 11 performs processing for display.

First, the graph shown in FIG. 12 (A) shows the data at the time of settling (the value of the rotation synchronization component) in time series, and is distinguished by three kinds of symbols. The symbol “◯” indicates normal data in the diagnosis history storage unit 10 (FIG. 6). Similarly, the symbol “▲” indicates abnormality detection reset data in the diagnostic history storage unit 10. The symbols “× 1” and “× 2” represent data of values in the diagnosis history storage unit 10 (FIG. 10) at the time of abnormality.

The diagram shown in FIG. 12B is created by referring to the date and time, the speed of progress of the abnormality, and the diagnostic result from the diagnostic history storage unit 10 and used as a diagnostic history. Further, the diagnosis history can be scrolled and displayed by an instruction from the man-machine interface 12. With such a display, the relationship between the diagnosis result and the data change can be easily grasped.

As described above, according to this embodiment, it is difficult to discriminate in real time by the conventional diagnostic apparatus, that is, a very fast progressing abnormality such as a defect in the rotating portion.
It is possible to identify a gradual anomaly such as a bending of the thermal rotor and an anomaly of the aging rotor that is very slow in progress.

Further, since the speed of progress of the abnormality can be grasped together, the operator or the maintenance staff can judge whether the operation is required immediately or the repair is required after the periodic inspection. It will be easier.

[0074]

As described above, according to the monitoring and diagnosing device for a rotating machine of the present invention, the vibration data of the rotating machine and the data relating to the operating state are inputted in real time, and the frequency analysis, the change rate calculation, A data processing means for processing data such as a difference from the reference value data, an anomaly detecting means for monitoring the processed data in real time and detecting an anomaly by comparing the processed data with the threshold data, and the processed data having an external influence. Data settling determination means for determining whether or not it is in a settling state without receiving, an abnormal progression speed determination means for comparing the latest settling data and the data settling in the past to calculate an abnormal progression speed, A diagnostic result judging means for distinguishing between a fast-moving abnormality and a slow-moving abnormality based on the abnormality detection result of the abnormality detecting means and the abnormal progress speed of the abnormal progress speed determining means. By having the door, it is possible to identify the slow and fast progress abnormalities progress abnormalities in real time.

When a slow-moving abnormality is detected,
Since the abnormal progress speed determination means calculates the abnormal progress speed, it is necessary for the operator or maintenance staff to immediately perform an operation operation, within a few hours of operation, or to perform repairs after a periodic inspection. It becomes easy to judge whether there is any.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a monitoring and diagnosing device for a rotary machine according to the present invention.

FIG. 2 is a schematic configuration diagram showing an example in which the embodiment of FIG. 1 is installed in a power plant.

FIG. 3 is an explanatory view showing an example of data processing of an abnormality detecting means.

FIG. 4 is a graph showing an example of data processing by the abnormality detecting means.

FIG. 5 is an explanatory diagram showing an abnormality detection logic of abnormality detection means.

FIG. 6 is an explanatory diagram showing an example of data stored in a settling history data storage unit.

FIG. 7 is a diagram in which values at the time of settling are displayed in time series by the abnormal traveling speed determination means.

FIG. 8 is a flowchart showing a processing procedure of an abnormal traveling speed determination means.

9 is a flowchart showing a determination processing procedure of processing step S2 in FIG.

FIG. 10 is an explanatory diagram showing an example of data stored in a diagnosis history storage unit.

FIG. 11 is an explanatory diagram showing an example of a diagnostic logic used in a diagnostic result determination means.

12A and 12B are diagrams showing a display example of a screen created by the display processing means.

FIG. 13 is an explanatory diagram showing a problem of conventional trend management.

[Explanation of symbols]

 1 Monitoring and Diagnostic Device for Rotating Machinery 2 Reference Value Data Storage Section 3 Data Processing Section 4 Threshold Data Storage Section 5 Abnormality Detection Section 6 Data Settling Determination Section 7 Settling History Data Storage Section 8 Abnormal Progression Speed Determination Section 9 Diagnostic Result Determination Means 10 Diagnostic history storage unit 11 Display processing means 12 Man-machine interface

Claims (2)

[Claims] 1. Real-time input of vibration data of a rotating machine and data relating to an operating state, frequency analysis,
Data processing means for performing data processing such as change rate calculation and difference with reference value data, anomaly detecting means for monitoring the processed data in real time and detecting anomaly by comparing with the threshold data, and the processing data Data settling judgment means to judge whether or not is in a settling state without being affected by the outside, and the latest settling data and the past settling data are compared to calculate the abnormal progressing speed. And a diagnosis result determining means for distinguishing between a fast progressing abnormality and a slow progressing abnormality based on the abnormality detection result of the abnormality detecting means and the abnormal progressing speed of the abnormal progressing speed determining means. A monitoring and diagnostic device for rotating machinery. 2. A reference value data storage unit for storing data during normal operation of a rotating machine as reference value data for each operating state, and a comparison value for determining whether or not machining data is abnormal. Threshold data storage for each operation state as value data, settling history data storage for long-term settling processing data, abnormality detection result and diagnosis result of abnormal progress speed determination means The rotary machine according to claim 1, further comprising: a diagnosis history storage unit that stores the diagnosis result of the determination unit, and a display processing unit that displays the diagnosis result, the diagnosis history, and the settling time history data in an overlapping manner. Monitoring and diagnostic equipment.