KR20220102364A - System for Predicting Flaw of Facility Using Vibration Sensor - Google Patents

Abstract

A facility predictive maintenance system through a vibration sensor may collect vibration data by measuring the vibration of a rotating body by the vibration sensor, analyze and visualize the data to monitor the current state of the facility in real time, and perform failure diagnosis for each part of the rotating body by detecting the failure of the facility in advance. The present invention is utilized for failure diagnosis of a device including the rotating body, and has the effect of detecting the abnormal state of a product in advance and transmitting corresponding information to the server or cloud to cope with the accident.

Description

System for Predicting Flaw of Facility Using Vibration Sensor

The present invention relates to a facility predictive maintenance system, and more particularly, collects, analyzes and visualizes vibration data by measuring the vibration of a rotating body by a vibration sensor, monitors the current state of the facility in real time, and prevents the failure of the facility in advance. It relates to a facility predictive maintenance system through a vibration sensor that detects and performs fault diagnosis for each part of a rotating body.

With the recent development of industrial technology, as rotating machines have become larger and more precise, if a failure occurs in these facilities, it can cause enormous economic loss or damage to human life, so fault monitoring of rotating machines has become very important.

In general, in a rotating machine, vibration characteristics occurring in a rotating shaft and an increase in the amount of vibration in mechanical elements such as bearings, gearboxes, and fans also occur, and accordingly, changes in characteristics of vibration occur.

As the operating time of mechanical equipment increases, the signal output from the machine deviates from the normal signal value, and this signal value gradually changes with time and becomes the same as the failure signal that occurs when the machine breaks down, causing a machine failure.

Vibration analysis is used to analyze the cause of failure of mechanical equipment. However, it is very difficult to analyze the vibration frequency and accurately analyze the cause of the failure, even for experts who are professionally trained and experienced in the field.

Therefore, in the conventional vibration analysis, it is difficult to precisely detect which part of the failure is a problem only with the vibration information of the mechanical equipment. There is a problem that can occur or cause a major accident.

Korean Patent Publication No. 10-1046748

In order to solve this problem, the present invention collects, analyzes and visualizes the vibration data by measuring the vibration of a rotating body by a vibration sensor, monitors the current state of the facility in real time, and detects a failure of the facility in advance to detect the vibration of the rotating body. The purpose of this is to provide a predictive maintenance system for equipment through a vibration sensor that performs fault diagnosis for each component.

A facility predictive maintenance system through a vibration sensor according to a feature of the present invention for achieving the above object,

a sensor device including a vibration sensor installed for each rotating body that is a mechanical facility; and

A first data collection unit for receiving a first vibration signal of the rotating body, which is normal data for each part in the rotating body, from the vibration sensor, or receiving a second vibration signal of the rotating body, which is input, and the first vibration of the rotating body A preprocessor that extracts characteristic information including a specific frequency and amplitude value from a signal, applies a fast Fourier transform to the extracted characteristic information to convert it into frequency domain data, and sets a frequency range detected for each part in the rotating body and a threshold value of the set frequency range is set based on the first vibration signal of the normal data, and when the second vibration signal is greater than or equal to the threshold value, the vibration comprising a detailed diagnosis unit that determines the occurrence of an abnormality in the component analysis section.

The vibration analyzer generates a Mahalanobis space using the first vibration signal, which is the normal data, and quantitatively determines whether the vibration signal is abnormal by comparing how far the second vibration signal, which is the input data, is away from the Mahalanobis space. It may further include a Mahalanobis distance classification unit that determines as .

The detailed diagnosis unit collects the first vibration signal of the rotating body, which is normal data for each component in the rotating body, as many times as a predetermined number of measurements, and sets the average value of the collected first vibration signal as normal data corresponding to each component, can be saved as

According to the above-described configuration, the present invention has an effect that can be used for fault diagnosis of a device including a rotating body, detect an abnormal state of a product in advance, and transmit the corresponding information to a server or cloud to cope with an accident.

The present invention has the effect of increasing the accuracy of diagnosis because the fault diagnosis is performed by sensing the vibration and sound of the rotating body.

The present invention has an effect that is expected to be highly useful in various small vibrating devices and automobiles due to simple installation and miniaturization of products.

1 is a view showing the configuration of a facility predictive maintenance system through a vibration sensor according to an embodiment of the present invention.
2 is a diagram schematically showing the internal configuration of the equipment predictive maintenance device according to an embodiment of the present invention.
3 is a diagram illustrating an embodiment in which the simple diagnosis unit determines whether an abnormality is based on an absolute criterion according to an embodiment of the present invention.
4 is a diagram illustrating an embodiment in which the simple diagnosis unit determines whether there is an abnormality based on the relative determination criterion according to an embodiment of the present invention.
5 is a diagram illustrating a process of determining the occurrence of an abnormality in the detailed diagnosis unit according to an embodiment of the present invention.
6 is a view showing an example of a specific frequency sensed for each part in the rotating body according to an embodiment of the present invention.
7 is a view showing an example of monitoring whether there is an abnormality for each part in the rotating body according to an embodiment of the present invention.
8 is a diagram illustrating an example of monitoring whether a part is abnormal by Mahalanobis distance analysis in the Mahalanobis distance classification unit according to an embodiment of the present invention.
9 is a diagram for explaining the meaning of a Mahalanobis space according to an embodiment of the present invention.

Throughout the specification, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated.

The present invention provides a predictive maintenance system for equipment capable of detecting a signal output from a machine part and comparing it with a normal value to predict whether an abnormality (failure) of a machine and a machine part is present.

Signals output from mechanical parts may appear as vibrations or sounds. When the mechanical part is in a normal state, the parts in the machine output a signal corresponding to a normal signal value within a certain range.

1 is a view showing the configuration of a system for predictive maintenance of equipment through a vibration sensor according to an embodiment of the present invention, and FIG. 2 is a diagram schematically showing the internal configuration of a system for predictive maintenance of equipment according to an embodiment of the present invention.

The equipment predictive maintenance system 100 using a vibration sensor according to an embodiment of the present invention includes a mechanical equipment 10 , a sensor device 110 installed for each mechanical equipment 10 , and an equipment predictive maintenance equipment 120 .

The mechanical equipment 10 represents a rotating body such as a motor, a pump, and a heater. Here, the rotating body is connected to an object rotating to supply rotational force, such as a pump, a motor, etc., as well as an object rotating in order to transmit rotational force to an arbitrary place by being connected to such a rotating object (a rotating shaft connected to a rotating shaft of a pump, a motor, etc.) Gears, impellers, propellers, etc.) includes all objects connected to and rotating on the shaft, including the shaft.

The sensor device 110 includes a vibration sensor 111 and an acoustic sensor 112 installed on the rotating body 10 .

The vibration sensor 111 may be a sensor that measures a vibration signal including vibration displacement information and acceleration information of a rotating body.

The vibration sensor 111 is mounted on a rotating body such as an elevator, a generator, a pump, and a fan to output a vibration signal due to an imbalance symptom in which the center of gravity and the center of mass do not match.

The sound sensor 112 measures the sound generated from the parts in the rotating body 10 and outputs a corresponding sound signal. The acoustic sensor 112 may be a microphone that measures the sound generated by the mechanical part.

The facility predictive maintenance device 120 includes a first data collection unit 121 , a second data collection unit 122 , a second A/D converter 123 , and a vibration analysis unit 130 , and a vibration analysis unit ( 130 includes a preprocessor 131 , a first A/D converter 132 , a Mahalanobis distance classification unit 133 , and a diagnosis unit 134 . The diagnosis unit 134 includes a simple diagnosis unit 135 and a detailed diagnosis unit 136 .

The big data management unit 140 includes a normal data storage unit 141 , an abnormal data storage unit 142 , and an acoustic data storage unit 143 .

The present invention may provide feedback to the production facility process of the corresponding part by utilizing the big data management unit 140 .

The first data collection unit 121 receives a vibration signal of the rotating body from the vibration sensor 111 installed on the rotating body.

The first data collection unit 121 may receive a vibration signal of a rotating body, which is normal data, from the vibration sensor 111 , or may receive a vibration signal of a rotating body in which an abnormality has occurred.

In other words, the first data collection unit 121 receives the first vibration signal of the rotating body, which is normal data for each component in the rotating body 10 , from the vibration sensor 111 , or receives the inputted second vibration signal of the rotating body 10 . Receive a vibration signal. Here, the first vibration signal is a vibration signal generated from the rotating body 10 in a normal state, and the second vibration signal is a vibration signal generated from the rotating body 10 as input data for determining whether there is an abnormality through vibration analysis. to be.

The first A/D converter 132 receives the vibration signal, which is an analog signal, received from the first data collection unit 121 and converts it into a digital signal.

The simple diagnosis unit 135 receives the vibration signal converted by the first A/D converter 132 and calculates a root mean square (RMS) of the vibration speed from the input vibration signal.

The simple diagnosis unit 135 classifies the calculated effective value of the vibration speed into four areas A, B, C, and D as a vibration analysis judgment criterion based on ISO10813.

The simple diagnosis unit 135 may diagnose whether the rotating body is normal by using the effective value of the vibration speed.

3 is a diagram illustrating an embodiment in which the simple diagnosis unit determines whether an abnormality is based on an absolute criterion in the simple diagnosis unit according to an embodiment of the present invention, and FIG. It is a diagram showing an embodiment of determining whether there is an abnormality.

As shown in FIG. 3 , the effective value belonging to the area A is the area to which the vibration of the newly installed rotating body belongs, and the effective value belonging to the area B is generally regarded as allowing long-term operation without limitation, and belongs to the C area. The rms value is usually considered to be unsuitable for long-term continuous operation without limitation, and in general, the rotating body can be operated for a limited period in this state until there is a suitable opportunity to take corrective action, and the rms value belonging to the D area is It is considered to be very severe enough to cause damage to the rotating body.

In this case, regions A, B, C, and D may represent Good, Satisfactory, Unsatisfactory, and Unacceptable states, respectively.

And the A, B, C, and D regions represent the effective value (RMS) of the vibration speed due to the imbalance symptom in which the center of mass and the center of the axis of the rotating body do not coincide.

When the effective value of the vibration speed belongs to the C region or the D region, the simple diagnosis unit 135 determines that an abnormality has occurred in the rotating body (absolute judgment criterion).

As another embodiment, as shown in FIG. 4 , when the effective value of the vibration speed belongs to the first vibration value to the second vibration value, the simple diagnosis unit 135 is in the normal range, and the second vibration value to the third vibration value are in the normal range. If it belongs to the vibration value, it is a warning range, and if it is higher than the third vibration value, it can be determined as an abnormal range (relative judgment standard).

The simple diagnosis unit 135 may output the absolute determination and the relative determination in the form of a user interface through the display unit 137 ( FIGS. 3 and 4 ).

5 is a view showing a process of determining the occurrence of an abnormality in the detailed diagnosis unit according to an embodiment of the present invention, and FIG. 6 is a view showing an example of a specific frequency detected by each part in the rotating body according to the embodiment of the present invention. , FIG. 7 is a view showing an example of monitoring for abnormalities by parts in a rotating body according to an embodiment of the present invention.

The preprocessor 131 removes noise from the vibration signal of the rotating body received from the first data collection unit 121 and extracts feature information from the vibration signal from which the noise has been removed.

The preprocessor 131 extracts characteristic information including a specific frequency and amplitude value from the first vibration signal of the rotating body 10, and applies a fast Fourier transform to the extracted characteristic information to convert it into frequency domain data.

Here, the characteristic information may include the shape, size (amplitude), phase, etc. of a waveform extracted from a waveform representing a change in vibration displacement or acceleration with respect to the rotating body 10 over time.

The preprocessor 131 applies FFT (Fast Fourier Transform) or STFT (Short Time Fourier Transform) to the extracted feature information to convert the noise-removed vibration signal into frequency domain data or time-frequency domain data. have.

The preprocessor 131 extracts a specific frequency and amplitude value through FFT or STFT.

The preprocessor 131 transmits characteristic information including a specific frequency and amplitude value to the detailed diagnosis unit 136 .

As shown in FIG. 5 , the detailed diagnosis unit 136 sets a frequency range detected for each part in the rotating body 10 ( S10 ), and the threshold of the frequency range set based on the first vibration signal of normal data A value is set ( S11 ), and when the second vibration signal is equal to or greater than a threshold value, it is determined that an abnormality occurs in the component.

The detailed diagnosis unit 136 may output the frequency range setting, threshold value setting, occurrence of component abnormality, etc. in the form of a user interface through the display unit 137 ( FIG. 5 ).

The detailed diagnosis unit 136 collects the first vibration signal of the rotating body, which is normal data for each part in the rotating body, as many times as a preset number of measurements, and sets the average value of the collected first vibration signal as normal data corresponding to each component. to store each part (S12).

The big data management unit 140 may include a normal data storage unit 141 that stores normal data corresponding to each part from the detailed diagnosis unit 136 .

As shown in FIG. 6 , for example, one rotating body 10 may be detected at 100 Hz, 200 Hz, and 300 Hz for each component.

Referring to FIG. 7 , for example, when the measured values of the components (inner ring damage, outer ring damage, misalignment, etc.) of the rotating body 10 exceed a threshold value, it may be determined as a failure of the corresponding components. The detailed diagnosis unit 136 may output a threshold value setting for each part of the rotating body 10, normal data, etc. in the form of a user interface through the display unit 137, as shown in FIGS. 6 and 7 ( FIGS. 6 and 7 ). 7).

Vibration analyzer 120 detects a specific frequency to detect the parts of each rotating body 10, and when the specific frequency generated in the sensed parts exceeds a threshold value, abnormality in the parts of each rotating body 10 judge the occurrence.

The vibration analyzer 120 receives a first vibration signal for each part of the rotating body, which is normal data, and stores a specific frequency of each received first vibration signal for each part to detect the parts of each rotating body 10 . can do.

When the second vibration signal is within 50% of the threshold value, the detailed diagnosis unit 136 may determine that the component in the rotating body 10 is in a 'good' state.

The detailed diagnosis unit 136 is a 'primary warning' that, when the second vibration signal belongs to 50% to 70% of the threshold, a failure is predicted at the initial stage of damage to the parts in the rotating body 10, and a precise diagnosis is performed. can be judged as being.

When the second vibration signal falls within 70% to 90% of the threshold value, the detailed diagnosis unit 136 determines that the component in the rotating body 10 is in a 'secondary warning' state in which the damage severity increases and repair is performed. can do.

The detailed diagnosis unit 136 may generate diagnostic status information of parts in the rotating body 10 and transmit it to an external central server or cloud. Due to this, the facility predictive maintenance system 100 of the present invention can be a useful element technology of a smart factory by detecting an abnormal state of the rotating body 10 in advance and notifying it to the outside.

8 is a view showing an example of monitoring whether a part is abnormal by Mahalanobis distance analysis in the Mahalanobis distance classification unit according to an embodiment of the present invention, and FIG. 9 is Mahalanobis according to an embodiment of the present invention It is a drawing for explaining the meaning of space.

The Mahalanobis distance classification unit 133 receives the normal data from the first data collection unit 121, and the state of the rotating body (normal state or abnormal state) based on the Mahalanobis space constructed from the received normal data. can be determined based on the Mahalanobis distance.

Mahalanobis distance analysis is a method of determining whether a part is normal or abnormal based on the Mahalanobis distance. It creates a reference space using everyday data, and compares how far the target data is from the reference space with the Mahalanobis distance. Thus, the abnormality is quantitatively judged.

As shown in FIG. 8 , it is possible to designate a reference space for each frequency region, and calculate an average value in the reference space for each region to determine a distance from the average value.

The Mahalanobis distance classifier 133 is a distance measure used in cluster analysis, and when there is a correlation between variables in multivariate data, the Mahalanobis distance is measured through data standardization using a covariance matrix.

In other words, the Mahalanobis distance classification unit 133 configures a Mahalanobis space using normal data, measures the distance to the space for arbitrary data, and determines whether the data is the same class as the normal data according to the distance do.

The Mahalanobis distance classification unit 133 configures the Mahalanobis space by Equation 1 below by using the normal data of the vibration signal to detect an abnormal state.

은 i번째 데이터의 j번째의 통계적 지표값,

은 통계적 지표들의 평균,

은 표준편차이다.where n is the steady-state data of the vibration signal, k is the number of statistical indicators,

is the j-th statistical index value of the i-th data,

is the average of statistical indicators,

is the standard deviation.

The Mahalanobis distance classifying unit 133 may calculate the Mahalanobis distance by the following Equation (2) as a ratio of the covariance to the degree of the distance of the input data from the population mean in the multivariate normal distribution.

Here, D ² is the Mahalanobis distance, x is the input data, m is the average value, and C is the covariance matrix.

The Mahalanobis distance classification unit 133 generates a Mahalanobis space using the vibration signal, which is normal data, and compares how far the input vibration signal is from the Mahalanobis space to determine whether the vibration signal is abnormal. .

The Mahalanobis distance classification unit 133 generates a Mahalanobis space using the first vibration signal, which is normal data, and compares how far the second vibration signal, which is the input data, is away from the Mahalanobis space to generate the vibration signal. Abnormality can be quantitatively determined.

As shown in FIG. 9 , the circle means the Mahalanobis space of the normal vibration signal, the center point of the circle is the average value of the data in the Mahalanobis space, the distance located inside the circle is normal data, and the distance located outside the circle is abnormal is data.

The Mahalanobis distance classifying unit 133 may determine the normal data and the abnormal data of the vibration signal to know how much the difference is from the normal value average value.

The Mahalanobis distance classification unit 133 may output the Mahalanobis space, the Mahalanobis distance, the average value, each frequency region, etc. in the form of a user interface through the display unit 137 as shown in FIGS. 8 and 9 .

The second data collection unit 122 receives the sound signal of the rotating body from the acoustic sensor 112 installed in the rotating body 10 for each component in the rotating body 10 .

The second data collection unit 122 receives an acoustic signal of the rotating body 10 that is normal data from the acoustic sensor 112 or receives an acoustic signal of the rotating body in which an abnormality occurs.

The second A/D converter 123 receives the sound signal, which is an analog signal, from the sound sensor 112 , converts it into a digital signal, and transmits it to the detailed diagnosis unit 136 .

The detailed diagnosis unit 136 receives the first acoustic signal generated from the first vibration signal of the component in the rotating body 10 or receives the second acoustic signal generated from the second vibration signal of the component, and the second vibration When the signal is equal to or greater than the threshold value, the second vibration signal equal to or greater than the threshold value and the second sound signal corresponding thereto are set as abnormal data and transmitted to the big data management unit 140 .

The present invention can also be used as a vibration noise measuring device for automobiles, aircraft, engines, etc. using the vibration sensor 111 and the acoustic sensor 112 .

The big data management unit 140 receives abnormality data obtained by matching the vibration signal and the acoustic signal from the detailed diagnosis unit 136 and stores the data in the acoustic data storage unit 143 .

When the second vibration signal for each part in the rotating body 10 is equal to or greater than a threshold value, the big data management unit 140 determines the occurrence of an abnormality in the corresponding part, sets it as the first abnormal data corresponding to each part, and stores it for each part. can

The big data management unit 140 receives the result of determining the abnormal occurrence of the vibration signal by comparing how far the second vibration signal is from the Mahalanobis space in the Mahalanobis distance classification unit 133, and corresponds to each part. It may further include an abnormality data storage unit 142 that sets the second abnormality data to be stored for each part.

The vibration analysis unit 130 performs a detailed diagnosis of determining the occurrence of an abnormality in the component in the detailed diagnosis unit 136 on the second vibration signal, which is the input data, and at the same time, an abnormality occurs in the component in the Mahalanobis distance classification unit 133 . Perform Mahalanobis distance analysis to determine .

The vibration analysis unit 130 selects one diagnostic result as anomaly data when an abnormality occurs in a part in detailed diagnosis or Mahalanobis distance analysis, and when both abnormality occurs in a part in detailed diagnosis and Mahalanobis distance analysis , the detailed diagnosis and the diagnosis result of Mahalanobis distance analysis can be selected as abnormal data.

The present invention can be utilized for fault diagnosis of a device including a rotating body used in a smart factory.

In the above, the embodiment of the present invention is not implemented only through an apparatus and/or method, and may be implemented through a program for realizing a function corresponding to the configuration of the embodiment of the present invention, a recording medium in which the program is recorded, etc. And, such an implementation can be easily implemented by an expert in the technical field to which the present invention belongs from the description of the above-described embodiment.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto. is within the scope of the right.

10: rotating body 100: equipment predictive maintenance system
110: sensor device 111: vibration sensor
112: acoustic sensor 120: equipment predictive maintenance device
121: first data collection unit 122: second data collection unit
123: second A/D converter 130: vibration analysis unit
131: preprocessor 132: first A/D converter
133: Mahalanobis Street Classification Unit 134: Diagnosis Unit
135: simple diagnosis unit 136: detailed diagnosis unit
140: big data processing unit 141: normal data storage unit
142: abnormal data storage unit 143: sound data storage unit

Claims (12)

a sensor device including a vibration sensor installed for each rotating body that is a mechanical facility; and
A first data collection unit for receiving a first vibration signal of the rotating body, which is normal data for each part in the rotating body, from the vibration sensor, or receiving a second vibration signal of the rotating body, which is input, and the first vibration of the rotating body A preprocessor that extracts characteristic information including a specific frequency and amplitude value from a signal, applies a fast Fourier transform to the extracted characteristic information to convert it into frequency domain data, and sets a frequency range detected for each part in the rotating body and a threshold value of the set frequency range is set based on the first vibration signal of the normal data, and when the second vibration signal is greater than or equal to the threshold value, the vibration comprising a detailed diagnosis unit that determines the occurrence of an abnormality in the component Predictive equipment maintenance system through vibration sensor including analysis unit. The method according to claim 1,
The vibration analyzer generates a Mahalanobis space using the first vibration signal, which is the normal data, and compares how far the second vibration signal, which is the input data, is away from the Mahalanobis space to determine whether the vibration signal is abnormal. Predictive maintenance system for facilities through a vibration sensor further comprising a Mahalanobis distance classification unit that quantitatively determines. 3. The method according to claim 2,
The detailed diagnostic unit collects the first vibration signal of the rotating body, which is normal data for each part in the rotating body, as many times as a preset number of measurements, and sets the average value of the collected first vibration signal as normal data corresponding to each component. Predictive maintenance system for equipment through vibration sensor that stores each part. 4. The method of claim 3,
Further comprising a big data management unit including a low-temperature data storage unit for storing normal data corresponding to each part from the detailed diagnosis unit,
When the second vibration signal for each part in the rotating body is equal to or greater than the threshold value, the big data management unit determines the occurrence of an abnormality in the corresponding part, sets it as the first abnormal data corresponding to each part, and stores it for each part, ,
The Mahalanobis distance classification unit compares how far the second vibration signal is from the Mahalanobis space and receives the result of determining the occurrence of an abnormality in the vibration signal, and as second anomaly data corresponding to each part. Equipment predictive maintenance system through a vibration sensor further comprising an abnormal data storage unit to set and store for each part. The method according to claim 1,
The vibration analyzer includes: a first A/D converter that receives a first vibration signal, which is an analog signal, received from the first data collection unit and converts it into a digital signal; and
Receives the first vibration signal converted by the first A/D converter, calculates a root mean square of the vibration speed from the input first vibration signal, and sets the calculated effective value of the vibration speed to ISO10813 It is classified into four areas A, B, C, and D as a vibration analysis judgment criterion based on Predictive maintenance system for equipment through vibration sensor including simple diagnosis unit. The method according to claim 1,
The sensor device further comprises a sound sensor that measures the sound generated from the parts in the rotating body and outputs a corresponding sound signal,
The vibration analyzer comprises: a second data collection unit configured to receive an acoustic signal of the rotating body from the acoustic sensor for each component in the rotating body; and
and a second A/D converter that receives an acoustic signal, which is an analog signal, received from the second data collection unit, converts it into a digital signal, and transmits it to the vibration analysis unit. 7. The method of claim 6,
The detailed diagnosis unit receives a first acoustic signal generated from a first vibration signal of the component in the rotating body, or receives a second acoustic signal generated from a second vibration signal of the component, and the second vibration signal is When the threshold value is greater than the threshold value, the equipment predictive maintenance system through the vibration sensor for setting the second vibration signal equal to or greater than the threshold value and the second sound signal corresponding thereto as abnormal data and transmitting it to the big data management unit. The method according to claim 1,
When the second vibration signal is within 50% of the threshold value, the detailed diagnosis unit is a facility predictive maintenance system using a vibration sensor to determine that the component in the rotating body is in a 'good' state. The method according to claim 1,
When the second vibration signal belongs to 50% to 70% of the threshold value, the detailed diagnosis unit is in a 'primary warning' state in which a failure is predicted at the initial stage of damage to the parts in the rotating body and precise diagnosis is performed. Predictive maintenance system for equipment through vibration sensor that judges. The method according to claim 1,
When the second vibration signal falls within 70% to 90% of the threshold value, the detailed diagnosis unit is a vibration sensor for determining a state of a 'secondary warning' in which the damage severity of the parts in the rotating body increases and repair is performed Equipment predictive maintenance system through. The method according to claim 1,
The vibration analyzer detects the specific frequency to detect the parts of each rotating body, and when the specific frequency generated in the sensed parts exceeds the threshold value, it is determined that an abnormality occurs in the parts of each rotating body, and , A facility through a vibration sensor that receives a first vibration signal for each part of the rotating body, which is the normal data, and stores a specific frequency of each of the received first vibration signals for each part to detect the parts of each rotating body Predictive Conservation System. 3. The method according to claim 2,
The vibration analysis unit performs a detailed diagnosis of determining the occurrence of an abnormality in the component in the detailed diagnosis unit on the second vibration signal, which is the input data, and at the same time, the Mahalanobis distance classification unit determines the occurrence of an abnormality in the component. After performing the Lanovis distance analysis, if an abnormality occurs in the part in the detailed diagnosis or the Mahalanobis distance analysis, one diagnosis result is selected as anomaly data, and in the detailed diagnosis and the Mahalanobis distance analysis Predictive maintenance system through a vibration sensor that selects the detailed diagnosis and the diagnosis result of the Mahalanobis distance analysis as abnormal data when all abnormalities occur in the parts.

Images