CN113237619A - Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration - Google Patents
Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration Download PDFInfo
- Publication number
- CN113237619A CN113237619A CN202110411554.0A CN202110411554A CN113237619A CN 113237619 A CN113237619 A CN 113237619A CN 202110411554 A CN202110411554 A CN 202110411554A CN 113237619 A CN113237619 A CN 113237619A
- Authority
- CN
- China
- Prior art keywords
- series
- under
- vibration
- frequency multiplication
- frequency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M7/00—Vibration-testing of structures; Shock-testing of structures
- G01M7/02—Vibration-testing by means of a shake table
- G01M7/025—Measuring arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a fault early warning method for variable-speed rotating machinery vibration in the field of mechanical fault detection, and aims to solve the technical problem that vibration fault detection and early warning are not accurate enough by only depending on vibration amplitude as a vibration analysis basis. It includes: acquiring historical frequency vectors and online frequency vectors under each series of characteristic frequency multiplication; constructing a vibration target map under frequency multiplication of each series of characteristics; establishing a normal behavior model under each series of characteristic frequency multiplication; calculating model residual errors corresponding to historical frequency vectors under each series of characteristic frequency multiplication on the vibration target graph; obtaining residual distribution under each series of characteristic frequency multiplication; determining abnormal alarm threshold values under frequency multiplication of various series of characteristics; calculating model residual errors corresponding to the online frequency vectors under each series of characteristic frequency multiplication on the vibration target graph; and comparing the online frequency vector with an abnormal alarm threshold value, thereby realizing fault early warning under frequency multiplication of each series of characteristics. The invention can realize real-time early warning of each vibration characteristic frequency of the equipment.
Description
Technical Field
The invention relates to a fault early warning method, a fault early warning device, fault early warning equipment and a storage medium for variable-speed rotating machinery vibration, and belongs to the technical field of mechanical fault detection.
Background
For rotating equipment, most faults are vibration-related. Therefore, vibration monitoring is a common technical means in the monitoring of the state of a rotating machine, and can realize early fault early warning and fault diagnosis. The vibration monitoring mainly comprises amplitude, frequency and phase, and the amplitude of vibration is usually only regarded as important in practical application, namely the size of vibration is taken as the basis of vibration monitoring and equipment state analysis.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, provides a fault early warning method, a fault early warning device, fault early warning equipment and a storage medium for variable-speed rotating machinery vibration, and solves the technical problem that vibration early warning is inaccurate because vibration phenomena are covered under some special conditions by only depending on vibration amplitude as a basis for vibration analysis in the prior art.
In order to achieve the purpose, the invention is realized by adopting the following technical scheme:
in a first aspect, the invention provides a fault early warning method for vibration of a variable-speed rotating machine, which comprises the following steps:
acquiring historical frequency vectors and online frequency vectors under frequency multiplication of various series of characteristics according to historical data and online data of normal operation vibration of equipment;
constructing a vibration target map under each series of characteristic frequency multiplication by taking the historical frequency vector and the online frequency vector as polar coordinates;
establishing a normal behavior model under each series of characteristic frequency multiplication based on the historical frequency vector;
based on the trained normal behavior model, calculating model residual errors corresponding to historical frequency vectors under each series of characteristic frequency multiplication on the vibration target map;
obtaining residual error distribution under frequency multiplication of each series of characteristics based on a model residual error sequence consisting of model residual errors corresponding to historical frequency vectors;
determining abnormal alarm threshold values under frequency multiplication of various series of characteristics based on residual distribution;
based on the trained normal behavior model, calculating model residual errors corresponding to the on-line frequency vectors under each series of characteristic frequency multiplication on the vibration target graph;
and comparing the online frequency vector with an abnormal alarm threshold value, thereby realizing fault early warning under frequency multiplication of each series of characteristics.
As an alternative embodiment: the step of acquiring historical frequency vectors and online frequency vectors under frequency multiplication of each series of characteristics according to the historical data and the online data of normal operation vibration of the equipment comprises the following steps:
decomposing the historical data and the online data of the normal operation vibration of the equipment into historical vibration components and online vibration components under frequency multiplication of various series of characteristics by adopting fast Fourier transform;
and acquiring the phase and amplitude of each vibration component according to each vibration component under each series of characteristic frequency multiplication, and taking the phase and amplitude as a historical frequency vector and an online frequency vector under each series of characteristic frequency multiplication.
As an alternative embodiment: the normal behavior model further comprises a statistical index of a vibration waveform and a working condition parameter, wherein the statistical index of the vibration waveform comprises a peak value, a kurtosis and a peak factor of a vibration signal of historical data; the operating condition parameters of the equipment comprise the operating rotating speed of the equipment.
As an alternative embodiment: the normal behavior model is a three-layer BP neural network.
As an alternative embodiment: the method for calculating the model residual error corresponding to the historical frequency vector under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model comprises the following steps:
training a normal behavior model based on the historical frequency vector;
inputting the historical frequency vector serving as a test sample into a trained normal behavior model to obtain an estimation sample;
the test sample is compared with the estimated sample,obtaining corresponding series characteristic frequency multiplication lower model residual error ri：
Wherein ri represents the model residual of the ith test sample at the series of characteristic multiples,
representing the phase and amplitude of the ith test sample at the frequency multiplication of the series of features,
representing the phase and amplitude of the i-th estimated sample at the characteristic frequency multiplication of the series.
As an alternative embodiment: the obtaining of the residual error distribution under the frequency multiplication of each series of characteristics based on the model residual error sequence composed of the model residual errors corresponding to the historical frequency vectors comprises:
obtaining a model residual error sequence R corresponding to the verification sample set under each series of characteristic frequency multiplication according to the model residual error under each series of characteristic frequency multiplicationk：
Rk＝[r1,r2,r3,…,rN]
Wherein N represents the total number of test samples in the test sample set; rkRepresenting a model residual sequence of the kth series characteristic frequency multiplication;
obtaining the residual distribution under each series of characteristic frequency multiplication by fitting the residual sequence with normal distribution comprises the following steps:
wherein the content of the first and second substances,
denotes the residual distribution, μ, at the frequency multiplication of the kth series of featureskAnd σkRespectively represents the frequency doubling lower model of the kth series of characteristicsMean and standard deviation of the residuals.
As an alternative embodiment: the determining the abnormal alarm threshold value under each series of characteristic frequency multiplication based on the residual distribution comprises:
based on the 3 sigma principle, the three-level abnormal alarm threshold under the frequency multiplication of the kth series of characteristics is as follows:
wherein, alarm _1 is a first-level abnormal alarm threshold, alarm _2 is a second-level abnormal alarm threshold, and alarm _3 is a third-level abnormal alarm threshold.
In a second aspect, the present invention provides a fault warning device for vibration of a variable-speed rotating machine, including:
the vibration data preprocessing module: the system is used for acquiring historical frequency vectors and online frequency vectors under frequency multiplication of various series of characteristics according to historical data and online data of normal operation vibration of equipment;
the vibration target map generation module: the system is used for constructing a vibration target map under each series of characteristic frequency multiplication by taking the historical frequency vector and the online frequency vector as polar coordinates;
a model construction module: the system is used for establishing a normal behavior model under each series of characteristic frequency multiplication based on historical frequency vectors;
a historical model residual error acquisition module: the method is used for calculating model residual errors corresponding to historical frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model;
a residual distribution module: the model residual error sequence is composed of model residual errors corresponding to historical frequency vectors, and residual error distribution under frequency multiplication of each series of characteristics is obtained;
a threshold determination module: the method is used for determining abnormal alarm threshold values under frequency multiplication of various series of characteristics based on residual distribution;
an online model residual error acquisition module: the method is used for calculating model residual errors corresponding to online frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model;
a fault early warning module: the method is used for realizing fault early warning under frequency multiplication of various series of characteristics by comparing the online frequency vector with the abnormal warning threshold value.
In a third aspect, the invention provides a fault early warning device for vibration of a variable-speed rotating machine, which is characterized by comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any of the above.
In a fourth aspect, the invention provides a computer-readable storage medium, on which a computer program is stored, characterized in that the program, when executed by a processor, performs the steps of any of the methods described above.
Compared with the prior art, the invention has the following beneficial effects:
the invention provides a fault early warning method, a device, equipment and a storage medium for variable-speed rotating mechanical vibration.A data source is divided into off-line data and on-line data, wherein the off-line data is used for training a normal behavior model of vibration characteristic frequency of the equipment and residual distribution of actual measured values of the characteristic frequency and estimated values of the model; and the online data is used for giving out residual errors between actual measured values of all characteristic frequency vectors and estimated values of the models in real time along with monitoring intervals according to the trained normal behavior model, so that real-time monitoring and early warning of all vibration characteristic frequencies of the equipment are realized. The normal behavior model also combines the kurtosis, peak value factors and other statistical indexes of the original vibration signal and working condition parameters of equipment such as the running speed and the load, and further improves the accuracy of early warning.
Drawings
FIG. 1 is a schematic diagram of a fault warning method for vibration of a variable-speed rotating machine according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a fault warning method for variable-speed rotating machinery vibration according to an embodiment of the present invention;
FIG. 3 is a diagram of a vibration target of a training sample 1X provided by an embodiment of the present invention;
FIG. 4 is a schematic diagram of a normal behavior model provided by an embodiment of the invention;
FIG. 5 is a diagram illustrating the actual value and the estimated value of a sample 1X according to an embodiment of the present invention;
fig. 6 is a table of statistical distribution of residual errors of frequency multiplication of various series of features provided in the embodiment of the present invention;
fig. 7 is an alarm threshold table for frequency multiplication of various series of features according to an embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The first embodiment is as follows:
as shown in fig. 1-2, the invention provides a fault early warning method for vibration of a variable-speed rotating machine, which comprises the following steps:
step 1, acquiring historical frequency vectors and online frequency vectors under frequency multiplication of various series of characteristics according to historical data and online data of normal operation vibration of equipment;
1.1, decomposing historical data and online data of normal operation vibration of equipment into historical vibration components and online vibration components under frequency multiplication of various series of characteristics by adopting fast Fourier transform;
and 1.2, acquiring the phase and amplitude of each vibration component according to each vibration component under each series of characteristic frequency multiplication, and taking the phase and amplitude as a historical frequency vector and an online frequency vector under each series of characteristic frequency multiplication.
Step 2, constructing a vibration target map under each series of characteristic frequency multiplication by taking the historical frequency vector and the online frequency vector as polar coordinates; polar coordinates
Wherein
Representing phase, A representing amplitudeIn the vibration target map, the amplitude increases radially from the center, and the phase angle increases in the circumferential direction.
Step 3, establishing a normal behavior model under each series of characteristic frequency multiplication based on the historical frequency vector; the normal behavior model is preferably a three-layer BP neural network, and further combines the statistical indexes of the vibration waveform and the working condition parameters, wherein the statistical indexes of the vibration waveform comprise the peak value, kurtosis and peak value factors of the vibration signal of historical data; the operating condition parameters of the equipment comprise the operating rotating speed of the equipment.
Step 4, calculating model residual errors corresponding to historical frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model;
step 4.1, training a normal behavior model based on the historical frequency vector;
step 4.2, inputting the historical frequency vector serving as a test sample into the trained normal behavior model to obtain an estimation sample;
step 4.3, comparing the test sample with the estimated sample to obtain corresponding series characteristic frequency multiplication lower model residual errors ri：
Wherein ri represents the model residual of the ith test sample at the series of characteristic multiples,
representing the phase and amplitude of the ith test sample at the frequency multiplication of the series of features,
representing the phase and amplitude of the i-th estimated sample at the characteristic frequency multiplication of the series.
Step 5, obtaining residual error distribution under frequency multiplication of each series of characteristics based on a model residual error sequence consisting of model residual errors corresponding to historical frequency vectors;
step 5.1, obtaining each series of model residual errors under frequency multiplication according to each series of characteristicsModel residual sequence R corresponding to verification sample set under characteristic frequency multiplicationk：
Rk＝[r1,r2,r3,…,rN]
Wherein N represents the total number of test samples in the test sample set; rkRepresenting a model residual sequence of the kth series characteristic frequency multiplication;
step 5.2, obtaining the residual distribution under each series of characteristic frequency multiplication by fitting the residual sequence through normal distribution comprises the following steps:
wherein the content of the first and second substances,
denotes the residual distribution, μ, at the frequency multiplication of the kth series of featureskAnd σkRespectively representing the mean value and the standard deviation of model residuals under the frequency multiplication of the kth series of characteristics.
Step 6, based on residual distribution, determining abnormal alarm threshold values under frequency multiplication of various series of characteristics;
based on the 3 sigma principle, the three-level abnormal alarm threshold under the frequency multiplication of the kth series of characteristics is as follows:
wherein, alarm _1 is a first-level abnormal alarm threshold, alarm _2 is a second-level abnormal alarm threshold, and alarm _3 is a third-level abnormal alarm threshold.
Step 7, calculating model residual errors corresponding to the on-line frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model; the method of this step is the same as steps 4.2 and 4.3, and the data source for the input is different.
And 8, comparing the online frequency vector with an abnormal alarm threshold value, thereby realizing fault early warning under frequency multiplication of each series of characteristics.
When the model residual error of the online frequency vector is larger than alarm _1, triggering a primary alarm; if the alarm is smaller than alarm _1 and larger than alarm _2, triggering a secondary alarm; and when the alarm level is smaller than the alarm _2 and larger than the alarm _3, triggering three-level alarm, and gradually reducing the alarm level.
In this embodiment, the obtained vibration data in the two speed-up and speed-down processes during normal operation of a certain rotor device respectively include 114 and 66, each piece of vibration data includes 512 sampling points, the vibration acquisition mode is synchronous whole-period sampling, and the sampling frequency is 64 points/rotation. Of these, 114 pieces of data were used to train the normal behavior model of the plant vibration, and another 66 pieces of data were used to test the normal behavior model.
(1) The 114 pieces of normal operation vibration data in the training sample set are preprocessed. Calculating peak value, kurtosis and peak factor; performing fast fourier transform on each piece of vibration data to obtain corresponding characteristic frequency vectors, wherein 0.5X, 1X, 2X, 3X … NX (X is frequency multiplication) is selected in the embodiment; the raw vibration data can be converted into a 12-feature data set with a size of 114 x 12 two-dimensional array in combination with the rotational speed information of the equipment operation. Each eigenfrequency vector is described by a vibration target map, and a 1X distribution scattergram of 114 training samples is shown in fig. 3, taking 1X as an example.
(2) And establishing a normal operation model of the equipment by adopting a three-layer BP neural network. The number of neurons in the corresponding input layer is 12, and the number of the neurons in the corresponding input layer is 12, wherein each vibration characteristic vector (rotating speed, peak value, kurtosis, peak value factor, 0.5X amplitude, 0.5X phase, 1X amplitude, 1X phase, 2X amplitude, 2X phase, 3X amplitude and 3X phase) corresponds to the in-out layer; the model output layer corresponds to each characteristic frequency vector (0.5X amplitude, 0.5X phase, 1X amplitude, 1X phase, 2X amplitude, 2X phase, 3X amplitude and 3X phase), the number of corresponding input layer neurons is 12, and the number of corresponding output layer neurons is 8; the model structure is shown in fig. 4.
(3) And (3) training the normal behavior model established in the step (2) by using the training sample set in the step (1). And carrying out normalization processing on the input layer, wherein the network activation function is a sigmoid function, and the number of neurons in the hidden layer of the model is obtained by adopting an empirical formula and is set to be 10.
(4) And (4) taking 66 normal operation sample sets which are not subjected to parameter and model training as a verification set, and substituting the verification set into the model trained in the step (3) to obtain an estimation value corresponding to each characteristic frequency vector. Taking a sample 1X as an example, the actual and estimated values of the vibration vector are shown in fig. 5. And calculating the distance between the actual value and the estimated value in the polar coordinates to obtain a model residual sequence of each vibration frequency, and fitting by using normal distribution. In this embodiment, the model residual distribution parameters of each characteristic frequency are shown in fig. 6.
(5) Three levels of alarm thresholds for each series of characteristic multiples can be obtained as shown in fig. 7.
Example two:
the invention provides a fault early warning device for variable-speed rotating machinery vibration, which comprises:
the vibration data preprocessing module: the system is used for acquiring historical frequency vectors and online frequency vectors under frequency multiplication of various series of characteristics according to historical data and online data of normal operation vibration of equipment;
the vibration target map generation module: the system is used for constructing a vibration target map under each series of characteristic frequency multiplication by taking the historical frequency vector and the online frequency vector as polar coordinates;
a model construction module: the system is used for establishing a normal behavior model under each series of characteristic frequency multiplication based on historical frequency vectors;
a historical model residual error acquisition module: the method is used for calculating model residual errors corresponding to historical frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model;
a residual distribution module: the model residual error sequence is composed of model residual errors corresponding to historical frequency vectors, and residual error distribution under frequency multiplication of each series of characteristics is obtained;
a threshold determination module: the method is used for determining abnormal alarm threshold values under frequency multiplication of various series of characteristics based on residual distribution;
an online model residual error acquisition module: the method is used for calculating model residual errors corresponding to online frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model;
a fault early warning module: the method is used for realizing fault early warning under frequency multiplication of various series of characteristics by comparing the online frequency vector with the abnormal warning threshold value.
Example three:
the invention provides a fault early warning device for variable-speed rotating machinery vibration, which comprises a processor and a storage medium, wherein the processor is used for processing the fault early warning device;
a storage medium to store instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any one of the first embodiment.
Example four:
the invention provides a computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, is adapted to carry out the steps of the method of any of the embodiments.
The invention relates to a fault early warning method, a fault early warning device and a fault early warning storage medium for variable-speed rotating machinery vibration. A normal behavior model of the vibration signal characteristic frequency is constructed by adopting machine learning algorithms such as a neural network and the like, and the relation among the series of characteristics is considered to be stable when the equipment operates normally; and when the amplitude and the phase of the current vibration characteristic frequency of the equipment deviate from the estimated value of the model, the current state of the equipment is considered to be abnormal. In the method, the amplitude and the phase of the vibration characteristic frequency are combined to construct a characteristic frequency vibration target graph, and each point in the graph represents a vector formed by the current phase and the amplitude of the characteristic frequency; and calculating residual errors between actual measured values of the characteristic frequencies and estimated values of the models in a target graph based on the established normal behavior model, namely, calculating a criterion for judging whether the equipment is abnormal. Therefore, the method can realize vibration monitoring and early fault early warning of more information by only using vibration original data and combining a time-frequency conversion technology and a machine learning technology, and provides reliable equipment state information for equipment predictive maintenance.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A fault early warning method for vibration of variable-speed rotating machinery is characterized by comprising the following steps:
acquiring historical frequency vectors and online frequency vectors under frequency multiplication of various series of characteristics according to historical data and online data of normal operation vibration of equipment;
constructing a vibration target map under each series of characteristic frequency multiplication by taking the historical frequency vector and the online frequency vector as polar coordinates;
establishing a normal behavior model under each series of characteristic frequency multiplication based on the historical frequency vector;
based on the trained normal behavior model, calculating model residual errors corresponding to historical frequency vectors under each series of characteristic frequency multiplication on the vibration target map;
obtaining residual error distribution under frequency multiplication of each series of characteristics based on a model residual error sequence consisting of model residual errors corresponding to historical frequency vectors;
determining abnormal alarm threshold values under frequency multiplication of various series of characteristics based on residual distribution;
based on the trained normal behavior model, calculating model residual errors corresponding to the on-line frequency vectors under each series of characteristic frequency multiplication on the vibration target graph;
and comparing the online frequency vector with an abnormal alarm threshold value, thereby realizing fault early warning under frequency multiplication of each series of characteristics.
2. The method of claim 1, wherein the obtaining of the historical frequency vector and the online frequency vector under the frequency multiplication of each series of characteristics according to the historical data and the online data of the normal operation vibration of the equipment comprises:
decomposing the historical data and the online data of the normal operation vibration of the equipment into historical vibration components and online vibration components under frequency multiplication of various series of characteristics by adopting fast Fourier transform;
and acquiring the phase and amplitude of each vibration component according to each vibration component under each series of characteristic frequency multiplication, and taking the phase and amplitude as a historical frequency vector and an online frequency vector under each series of characteristic frequency multiplication.
3. The method as claimed in claim 2, wherein the normal behavior model further comprises statistical indicators of vibration waveform including peak-to-peak value, kurtosis, and peak factor of vibration signal of historical data, and operating condition parameters; the operating condition parameters of the equipment comprise the operating rotating speed of the equipment.
4. The method of claim 1, wherein the normal behavior model is a three-layer BP neural network.
5. The method of claim 2, wherein the step of calculating model residuals corresponding to the historical frequency vectors under each series of characteristic frequency doubling on the vibration target map based on the trained normal behavior model comprises:
training a normal behavior model based on the historical frequency vector;
inputting the historical frequency vector serving as a test sample into a trained normal behavior model to obtain an estimation sample;
comparing the test sample with the estimated sample to obtain corresponding series characteristic frequency multiplication lower model residual error ri：
Wherein ri represents the seriesThe model residuals for the ith test sample at the column characteristic frequency doubling,
representing the phase and amplitude of the ith test sample at the frequency multiplication of the series of features,
representing the phase and amplitude of the i-th estimated sample at the characteristic frequency multiplication of the series.
6. The method according to claim 5, wherein obtaining the residual error distribution under the frequency multiplication of each series of characteristics based on the model residual error sequence composed of the model residual errors corresponding to the historical frequency vectors comprises:
obtaining a model residual error sequence R corresponding to the verification sample set under each series of characteristic frequency multiplication according to the model residual error under each series of characteristic frequency multiplicationk：
Rk＝[r1,r2,r3,…,rN]
Wherein N represents the total number of test samples in the test sample set; rkRepresenting a model residual sequence of the kth series characteristic frequency multiplication;
obtaining the residual distribution under each series of characteristic frequency multiplication by fitting the residual sequence with normal distribution comprises the following steps:
wherein the content of the first and second substances,
denotes the residual distribution, μ, at the frequency multiplication of the kth series of featureskAnd σkRespectively representing the mean value and the standard deviation of model residuals under the frequency multiplication of the kth series of characteristics.
7. The method of claim 6, wherein the determining the abnormal alarm threshold under the frequency multiplication of each series of characteristics based on the residual distribution comprises:
based on the 3 sigma principle, the three-level abnormal alarm threshold under the frequency multiplication of the kth series of characteristics is as follows:
wherein, alarm _1 is a first-level abnormal alarm threshold, alarm _2 is a second-level abnormal alarm threshold, and alarm _3 is a third-level abnormal alarm threshold.
8. The utility model provides a trouble early warning device of variable rotational speed rotating machine vibration which characterized in that, trouble early warning device includes:
the vibration data preprocessing module: the system is used for acquiring historical frequency vectors and online frequency vectors under frequency multiplication of various series of characteristics according to historical data and online data of normal operation vibration of equipment;
the vibration target map generation module: the system is used for constructing a vibration target map under each series of characteristic frequency multiplication by taking the historical frequency vector and the online frequency vector as polar coordinates;
a model construction module: the system is used for establishing a normal behavior model under each series of characteristic frequency multiplication based on historical frequency vectors;
a historical model residual error acquisition module: the method is used for calculating model residual errors corresponding to historical frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model;
a residual distribution module: the model residual error sequence is composed of model residual errors corresponding to historical frequency vectors, and residual error distribution under frequency multiplication of each series of characteristics is obtained;
a threshold determination module: the method is used for determining abnormal alarm threshold values under frequency multiplication of various series of characteristics based on residual distribution;
an online model residual error acquisition module: the method is used for calculating model residual errors corresponding to online frequency vectors under each series of characteristic frequency multiplication on the vibration target map based on the trained normal behavior model;
a fault early warning module: the method is used for realizing fault early warning under frequency multiplication of various series of characteristics by comparing the online frequency vector with the abnormal warning threshold value.
9. The fault early warning equipment for the vibration of the variable-speed rotating machinery is characterized by comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any one of claims 1 to 7.
10. Computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110411554.0A CN113237619B (en) | 2021-04-16 | 2021-04-16 | Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110411554.0A CN113237619B (en) | 2021-04-16 | 2021-04-16 | Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113237619A true CN113237619A (en) | 2021-08-10 |
| CN113237619B CN113237619B (en) | 2022-11-15 |
Family
ID=77128321
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110411554.0A Active CN113237619B (en) | 2021-04-16 | 2021-04-16 | Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113237619B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114048261A (en) * | 2021-09-22 | 2022-02-15 | 国家能源集团乐东发电有限公司 | Vibration vector early warning value determination method and system based on data description |
| CN115327373A (en) * | 2022-04-20 | 2022-11-11 | 岱特智能科技（上海）有限公司 | Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310589A (en) * | 2020-01-20 | 2020-06-19 | 河北科技大学 | Fault diagnosis method and device for mechanical system and terminal |
| CN111504647A (en) * | 2020-06-09 | 2020-08-07 | 华东交通大学 | AR-MSET-based performance degradation evaluation method for rolling bearing |
| CN112036240A (en) * | 2020-07-27 | 2020-12-04 | 国家能源集团新疆吉林台水电开发有限公司 | Method and system for diagnosing lower rack loosening fault based on frequency division algorithm |
| CN112082720A (en) * | 2020-09-04 | 2020-12-15 | 江苏方天电力技术有限公司 | Method for determining early warning value of vibration fault of fixed-speed rotating machine |
-
2021
- 2021-04-16 CN CN202110411554.0A patent/CN113237619B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310589A (en) * | 2020-01-20 | 2020-06-19 | 河北科技大学 | Fault diagnosis method and device for mechanical system and terminal |
| CN111504647A (en) * | 2020-06-09 | 2020-08-07 | 华东交通大学 | AR-MSET-based performance degradation evaluation method for rolling bearing |
| CN112036240A (en) * | 2020-07-27 | 2020-12-04 | 国家能源集团新疆吉林台水电开发有限公司 | Method and system for diagnosing lower rack loosening fault based on frequency division algorithm |
| CN112082720A (en) * | 2020-09-04 | 2020-12-15 | 江苏方天电力技术有限公司 | Method for determining early warning value of vibration fault of fixed-speed rotating machine |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114048261A (en) * | 2021-09-22 | 2022-02-15 | 国家能源集团乐东发电有限公司 | Vibration vector early warning value determination method and system based on data description |
| CN115327373A (en) * | 2022-04-20 | 2022-11-11 | 岱特智能科技（上海）有限公司 | Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113237619B (en) | 2022-11-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113237619B (en) | Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration | |
| CN110994604B (en) | Power system transient stability assessment method based on LSTM-DNN model | |
| CN112418277A (en) | Method, system, medium, and apparatus for predicting remaining life of rotating machine component | |
| CN107944721B (en) | Universal machine learning method, device and system based on data mining | |
| Hwang et al. | SVM-RBM based predictive maintenance scheme for IoT-enabled smart factory | |
| Simani et al. | Hybrid model–based fault detection of wind turbine sensors | |
| CN113408068A (en) | Random forest classification machine pump fault diagnosis method and device | |
| CN111504647A (en) | AR-MSET-based performance degradation evaluation method for rolling bearing | |
| CN110825074A (en) | Phase modulator fault diagnosis system and working method thereof | |
| CN115793590A (en) | Data processing method and platform suitable for system safety operation and maintenance | |
| CN113742932A (en) | Method for predicting and diagnosing faults of gearbox of wind turbine generator | |
| CN110007660B (en) | Online soft measurement method for transient equivalent thermal stress of steam turbine set of thermal power plant | |
| CN106596110B (en) | The automatic analyzing and diagnosing method of turbine-generator units waterpower imbalance fault based on online data | |
| CN113792940A (en) | Hydro-turbo generator set runout data prediction method and device based on deep learning | |
| CN108306997B (en) | Domain name resolution monitoring method and device | |
| CN109635430A (en) | Grid power transmission route transient signal monitoring method and system | |
| CN112082720A (en) | Method for determining early warning value of vibration fault of fixed-speed rotating machine | |
| CN112380782A (en) | Rotating equipment fault prediction method based on mixed indexes and neural network | |
| CN116956215A (en) | Fault diagnosis method and system for transmission system | |
| CN116625683A (en) | Wind turbine generator system bearing fault identification method, system and device and electronic equipment | |
| CN111874813A (en) | Shore bridge crane reducer state detection method based on improved decision tree algorithm | |
| CN115128345B (en) | Power grid safety early warning method and system based on harmonic monitoring | |
| CN115685045A (en) | Online evaluation method for voltage transformer | |
| CN115859077A (en) | Multi-feature fusion motor small sample fault diagnosis method under variable working conditions | |
| CN116026588A (en) | Bearing fault diagnosis and early warning method based on ensemble learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |