CN115235762A - Method for acquiring and evaluating local damage vibration envelope signal of metallurgical transmission mechanism - Google Patents
Method for acquiring and evaluating local damage vibration envelope signal of metallurgical transmission mechanism Download PDFInfo
- Publication number
- CN115235762A CN115235762A CN202211169217.6A CN202211169217A CN115235762A CN 115235762 A CN115235762 A CN 115235762A CN 202211169217 A CN202211169217 A CN 202211169217A CN 115235762 A CN115235762 A CN 115235762A
- Authority
- CN
- China
- Prior art keywords
- vibration
- sequence
- gearbox
- envelope
- information
- 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.)
- Pending
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
- G01M13/00—Testing of machine parts
- G01M13/02—Gearings; Transmission mechanisms
- G01M13/021—Gearings
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/02—Gearings; Transmission mechanisms
- G01M13/028—Acoustic or vibration analysis
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention relates to the technical field of vibration testing and evaluation of structural components, in particular to a method for acquiring and evaluating a local damage vibration envelope signal of a metallurgical transmission mechanism. The method comprises the steps of obtaining envelope information of a vibration signal of the gearbox, analyzing fluctuation characteristics of the envelope information, audio information and temperature information of the gearbox in time sequence, and constructing a characteristic matrix. The gearboxes are grouped based on the characteristic matrix, and a plurality of gearbox groups are obtained. And training a corresponding CEEMD-GRU information prediction network according to the vibration characteristic change sequence in each gearbox group, and further obtaining a future vibration characteristic change sequence of the gearbox to be evaluated. And evaluating the abnormality of the gearbox to be evaluated according to elements in the future vibration characteristic change sequence. According to the method, the abnormal gearbox is timely evaluated through the acquisition and analysis of the vibration data of the gearbox of the transmission mechanism and the prediction means, so that the evaluation of the local damage of the transmission structure is realized.
Description
Technical Field
The invention relates to the technical field of vibration testing and evaluation of structural components, in particular to a method for acquiring and evaluating local damage vibration envelope signals of a metallurgical transmission mechanism.
Background
The equipment for completing a series of continuous casting processes in the metallurgical field is called continuous casting complete equipment, and the core part equipment of the continuous casting complete equipment is formed by mechanical-electrical-hydraulic integration of steel casting equipment, continuous casting machine body equipment, cutting area equipment and dummy bar collecting and conveying equipment. The continuous casting complete equipment is combined by various equipment, so that a large-scale transmission mechanism is relatively large in quantity in the continuous casting process. The transmission mechanism is composed of a plurality of mechanical structures, and a large number of mechanical structures exist, so that mechanical faults are easily caused by mechanical vibration in the operation process along with the operation of the transmission mechanism, especially the mechanical vibration of various gear boxes in the transmission mechanism is difficult to find manually, and if vibration signals are not obtained and evaluated in time, the problem caused by vibration is ignored, the gears in the gear boxes are subjected to irreversible abrasion, namely, the transmission mechanism is subjected to local damage, and the whole process is further influenced.
Because the metallurgical transmission mechanism has a large scale, the gear boxes have multiple types, the vibration signals of each gear box have different variation trends, and the vibration signals of the same gear box under different working conditions have different variation trends, if the real-time vibration signals of each gear box are directly subjected to envelope analysis, abnormal gear boxes cannot be evaluated in time, and if the abnormal gear boxes are not maintained and replaced in time, the process cost is increased, and the process efficiency is influenced.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a method for acquiring and evaluating local damage vibration envelope signals of a metallurgical transmission mechanism, which adopts the following technical scheme:
the invention provides a method for acquiring and evaluating a local damage vibration envelope signal of a metallurgical transmission mechanism, which comprises the following steps:
collecting vibration signals, audio information and temperature information of each gear box of the metallurgical transmission mechanism; carrying out envelope processing on the vibration signal to obtain various kinds of envelope information;
acquiring an envelope information sequence of each kind of envelope information according to a preset sampling frequency; obtaining a fluctuation frequency factor according to the difference between adjacent elements in the envelope information sequence and the element mean value; obtaining a fluctuation amplitude factor according to the element range and the element mean value in the envelope information sequence; acquiring an audio information sequence and a temperature information sequence according to the sampling frequency, acquiring a first volatility factor according to the fluctuation amplitude and the fluctuation frequency of the audio information sequence, and acquiring a second volatility factor of the temperature information sequence;
combining a plurality of fluctuation frequency factors, the first volatility factor and the second volatility factor in the same sampling time period to obtain a feature vector; the feature vectors in the continuous sampling time period form a feature matrix;
grouping the plurality of gear boxes according to the characteristic matrix to obtain a plurality of gear box groups; extracting the vibration characteristics of the characteristic matrix to obtain a vibration characteristic change sequence on each gear box time sequence in the gear box group; taking the vibration characteristic change sequence in each gearbox group as training data, and obtaining a CEEMD-GRU information prediction network according to the training data;
and predicting a future vibration characteristic change sequence of the gearbox to be evaluated by the network according to the corresponding CEEMD-GRU information, and judging whether the gearbox to be evaluated is abnormal or not according to the sizes and element positions of elements in the future vibration characteristic change sequence.
Further, the envelope processing of the vibration signal to obtain various kinds of envelope information includes:
and performing analog-to-digital conversion, frequency spectrum conversion, band-pass filtering and acceleration envelope processing on the vibration signal to obtain envelope information, wherein the envelope information comprises vibration acceleration, vibration speed and acceleration envelope value.
Further, the obtaining a fluctuation frequency factor according to the difference between adjacent elements in the envelope information sequence and the element mean value includes:
obtaining the fluctuation frequency factor according to a fluctuation frequency factor formula, wherein the fluctuation frequency factor formula comprises:
wherein, the first and the second end of the pipe are connected with each other,
for the purpose of said wave frequency factor,
is the mean value of the elements in the envelope information sequence,
for the first in the envelope information sequence
The value of each of the elements is,
for the number of elements in the envelope information sequence,
as a function of absolute value.
Further, the obtaining a fluctuation amplitude factor according to the element range and the element mean in the envelope information sequence includes:
and obtaining the element value difference between the maximum element value and the element average value in the envelope information sequence, and taking the ratio of the element range of the envelope information sequence to the element value difference as the fluctuation amplitude factor.
Further, the obtaining a first volatility factor according to the fluctuation amplitude and the fluctuation frequency of the audio information sequence includes:
obtaining the first volatility factor according to a first volatility factor formula, the first volatility factor formula comprising:
wherein the content of the first and second substances,
for the purpose of the first volatility factor,
is the average of the elements of the sequence of audio information,
is the variance of an element of the sequence of audio information,
for the number of elements in the sequence of audio information,
for the first in the audio information sequence
And (4) each element.
Further, the grouping the plurality of gearboxes according to the feature matrix, and obtaining a plurality of gearbox groups comprises:
and clustering the research samples by using the characteristic matrix as a research sample by using a quantum clustering algorithm to obtain a plurality of clustering clusters, wherein each clustering cluster corresponds to one gear box group.
Further, the extracting the vibration characteristics of the characteristic matrix and obtaining a vibration characteristic variation sequence on a time sequence of each gearbox in the gearbox group includes:
dividing the characteristic matrix into a plurality of determinants with the same size according to time sequence; and obtaining the row and column values of the determinant, taking the absolute values of the row and column values as the vibration characteristics, and arranging the vibration characteristics according to a time sequence to obtain the vibration characteristic change sequence.
Further, the predicting future vibration characteristic change sequence of the gearbox to be evaluated according to the corresponding CEEMD-GRU information prediction network comprises the following steps:
obtaining a real-time characteristic matrix and a real-time vibration characteristic change sequence of the gearbox to be evaluated, matching the real-time characteristic matrix with the characteristic matrix of the central sample of each gearbox group, and obtaining a best-matched gearbox group corresponding to the gearbox to be evaluated; and inputting the real-time vibration characteristic change sequence of the gearbox to be evaluated into the CEEMD-GRU information prediction network corresponding to the mostly matched gearbox group to obtain the future vibration characteristic change sequence.
Further, the judging whether the gearbox to be evaluated is abnormal or not according to the sizes and the element positions of the elements in the future vibration characteristic change sequence comprises the following steps:
if a certain element exists in the future vibration characteristic change sequence and is larger than a preset characteristic threshold value, obtaining a time sequence position of the element, obtaining an abnormal time point according to the time sequence position, and if a time period from the abnormal time point to a real-time point is smaller than a preset time length threshold value, determining that the gear box to be evaluated is abnormal.
The invention has the following beneficial effects:
the embodiment of the invention carries out envelope processing on the vibration signal of the gearbox to obtain various envelope information, and is used for representing the current state of the gearbox influenced by vibration according to the audio information and the temperature information. By processing the information data in time sequence, the characteristic of each gearbox along with time change is obtained, a characteristic matrix with strong referential is constructed, the gearboxes are grouped, and a plurality of gearbox groups are obtained. The evaluation efficiency of the gear boxes can be improved by grouping and analyzing the gear boxes, and the CEEMD-GRU information prediction network is trained by utilizing the vibration characteristic change information in each gear box group, so that the richness of training data is ensured. The CEEMD-GRU information prediction network can predict a future vibration characteristic change sequence of the gearbox to be evaluated at a future moment, and then whether the gearbox to be evaluated is abnormal or not is evaluated according to element values and time sequence positions in the future vibration characteristic change sequence. According to the embodiment of the invention, the abnormal gearbox can be judged in time according to the trend of the vibration characteristic change by an information prediction means, so that the maintenance and the replacement of the abnormal gearbox are facilitated, and the detection, the maintenance and the early warning of the local damage in the transmission mechanism are realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a method for acquiring and evaluating a local damage vibration envelope signal of a metallurgical transmission mechanism according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description, the structure, the features and the effects of the method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to the present invention are provided with reference to the accompanying drawings and the preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of the method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism provided by the invention in detail by combining with the accompanying drawings.
Referring to fig. 1, a flowchart of a method for acquiring and evaluating a local damage vibration envelope signal of a metallurgical transmission mechanism according to an embodiment of the present invention is shown, where the method includes:
step S1: collecting vibration signals, audio information and temperature information of each gear box of the metallurgical transmission mechanism; and carrying out envelope processing on the vibration signal to obtain various kinds of envelope information.
The vibration signal can directly feed back the vibration information of the current gearbox. The audio information and the temperature information can reflect the state of the influence of the current vibration on the gearbox. The stronger the abnormal vibration is, the more irregular and violent the distribution of the vibration signal is; the more intense the abnormal vibration, the more abnormal audio information and abnormal temperature information are generated. Further, envelope processing is carried out on the vibration signals, and the generated envelope information can more intuitively express the current vibration information. The specific acquisition method of the envelope information comprises the following steps:
and carrying out analog-to-digital conversion, frequency spectrum conversion, band-pass filtering and acceleration envelope processing on the vibration signals to obtain envelope information, wherein the envelope information comprises vibration acceleration, vibration speed and acceleration envelope values.
It should be noted that the acceleration envelope technique is a numerical technique for those skilled in the art, and only the processing procedure thereof is briefly described here:
the acceleration envelope technology is mainly used for extracting small-amplitude and repetitive impact signals in vibration signals, and the small-amplitude repetitive impact models are often covered by high-energy low-frequency signals in the measurement process. The acceleration enveloping method adopts a band-pass filter to filter out low-frequency synchronous vibration signals, the selection of the filter depends on the actual situation, repeated shock signals are separated from an intricate and complex vibration system, after the original vibration signals pass through the filter, some high-frequency shock signals are left, the high-frequency shock signals are subjected to enveloping processing, the enveloping processing is realized through a circuit, the function of the circuit is changed into that the input signals are squared, according to the product of trigonometric functions, after one vibration signal passes through the filter, only the frequency which is 50 times higher than the defect frequency can pass through the vibration signal, therefore, after a series of frequency multiplication and self multiplication, the frequencies of the addition part and the subtraction part can be obtained according to an integration and difference formula in the multiplication process, and because the addition part exceeds the analysis measurement range, only the analysis subtraction part further repeats the calculation process, and the acceleration enveloping value spectrum diagram can be obtained. And obtaining the information of the vibration acceleration, the vibration speed and the acceleration envelope value according to the envelope value spectrogram.
The method for acquiring each kind of information collected in the embodiment of the present invention is specifically described as follows:
(1) For the vibration signal: in the embodiment of the invention, a piezoelectric vibration acceleration sensor is arranged in the radial direction of a bearing seat of a rotating shaft in a gearbox, the sensor outputs a vibration signal of an analog signal type, the vibration signal is transmitted to an envelope processing unit through a double-shielded signal cable for envelope processing, and then various envelope information is obtained.
(2) For audio information and temperature information: because the vibration sources of the gearbox are diverse, the vibration information to which the vibration signal is reflected may be a composite signal of multiple vibration sources of the plant, which composite signal includes at least three components: 1. peripheral vibration conduction; 2. vibration caused by unbalanced structural relationship, misalignment, looseness and the like of equipment; 3. vibration caused by defects of transmission parts such as bearings, gears and the like. Therefore, in order to improve the accuracy and the reference of the collected data, the audio information and the temperature information are further collected to represent the current vibration-affected state of the gearbox. The audio sensor is placed at the bottom of the bearing seat, and the sound frequency inside the gear box can be collected according to the audio sensor. Place temperature sensor in the gear box, because gear self defect can lead to gear engagement to go wrong, produce unusual friction, and then lead to the temperature anomaly in the gear box, and then obtain the temperature information of gear box according to temperature sensor.
It should be noted that, because all the information acquisition modes are in a relatively complex environment, the acquired information may include a certain noise, which affects the result of the subsequent determination, and therefore, the acquired information needs to be subjected to noise reduction and smoothing processing.
Step S2: acquiring an envelope information sequence of each kind of envelope information according to a preset sampling frequency; obtaining a fluctuation frequency factor according to the difference between adjacent elements in the envelope information sequence and the element mean value; obtaining a fluctuation amplitude factor according to the element range and the element mean value in the envelope information sequence; and acquiring an audio information sequence and a temperature information sequence according to the sampling frequency, acquiring a first volatility factor according to the fluctuation amplitude and the fluctuation frequency of the audio information sequence, and acquiring a second volatility factor of the temperature information sequence.
In the embodiment of the present invention, when the envelope information sequence, the audio information sequence, and the temperature information sequence are integrated, the sampling frequency is set to once every 5 minutes, and the length of the data sequence is set to 12, that is, one sequence contains 12 elements.
The envelope information sequence is a sequence in a time sequence, so that the data change of the sequence can reflect the vibration change condition of the gearbox, and the fluctuation frequency factor is obtained according to the difference between adjacent elements in the envelope information sequence and the element mean value, and the method specifically comprises the following steps:
obtaining a fluctuation frequency factor according to a fluctuation frequency factor formula, wherein the fluctuation frequency factor formula comprises:
wherein the content of the first and second substances,
in order to fluctuate the frequency factor, the frequency of the wave,
is the average value of the elements in the envelope information sequence,
for the first in the envelope information sequence
The value of each of the elements is,
for the number of elements in the envelope information sequence,
as a function of absolute value.
The influence of abnormal vibration is represented as a numerical gain on the value of the envelope information, and therefore fluctuates in the frequency factor formula ifThe influence of abnormal vibration appears at the next moment of a certain moment, then
Is relatively large. Therefore, when the accumulated value is larger, the fluctuation in the sequence is more frequently shown, namely, the fluctuation frequency factor is larger.
It should be noted that, the envelope information sequence of each kind of envelope information needs to obtain a corresponding fluctuation frequency factor, the envelope information of the embodiment of the present invention includes a vibration acceleration, a vibration velocity, and an acceleration envelope value, and the corresponding band frequency factors are respectively:
、
and
。
further obtaining a fluctuation amplitude factor according to the element range and the element mean value in the envelope information sequence, specifically comprising: obtaining the element value difference between the maximum element value and the element average value in the envelope information sequence, and taking the ratio of the element range and the element value difference of the envelope information sequence as a fluctuation amplitude factor, namely the expression of the fluctuation amplitude factor is as follows:
wherein, the first and the second end of the pipe are connected with each other,
in order to be a wave amplitude factor,
for the largest element value in the sequence of envelope information,
for enveloping a sequence of informationThe value of the smallest element in the set of elements,
is the mean value of the elements in the envelope information sequence.
I.e., the greater the range of the sequence, the more dramatic the fluctuations in the sequence are; the smaller the difference between the element values of the maximum element value and the element average value is, the larger the distribution of the element values with higher numerical values exists in the sequence, and the more severe the wave band in the sequence is. The larger the range, the smaller the element value difference, and the larger the band amplitude of the description sequence, the larger the fluctuation amplitude factor.
It should be noted that, the envelope information sequence of each kind of envelope information needs to obtain a corresponding fluctuation amplitude factor, the envelope information of the embodiment of the present invention includes a vibration acceleration, a vibration velocity, and an acceleration envelope value, and the corresponding band amplitude factors are respectively:
、
and
。
further analyzing the band characteristics of the audio information sequence and the temperature information sequence, and obtaining a first fluctuation factor according to the band amplitude and the fluctuation frequency of the audio information sequence, wherein the method specifically comprises the following steps:
obtaining a first volatility factor according to a first volatility factor formula, the first volatility factor formula comprising:
wherein, the first and the second end of the pipe are connected with each other,
in order to be the first volatility factor,
is the average of the elements of the sequence of audio information,
is the variance of an element of a sequence of audio information,
for the number of elements in the sequence of audio information,
for the first in the audio information sequence
And (4) each element.
In the first volatility factor formula, the product of the element mean and the element variance represents the fluctuation range, that is, the larger the element mean, the larger the element variance, and the larger the fluctuation range of the sequence;
b denotes the fluctuation frequency. The larger the band amplitude, the larger the fluctuation frequency, and the larger the first volatility factor.
According to the same calculation formula as the first volatility factor, the second volatility factor of the temperature information sequence can be obtained
。
And step S3: combining a plurality of fluctuation frequency factors, the first fluctuation factor and the second fluctuation factor in the same sampling time period to obtain a feature vector; the feature vectors within successive sampling time periods constitute a feature matrix.
Through the processing of the information data in the step S2, in the embodiment of the present invention, eight fluctuation features can be obtained within one sampling time period, so that a feature vector of each sampling time period is obtained by combining multiple fluctuation frequency factors, a first fluctuation factor, and a second fluctuation factor, that is, the length of the feature vector is 8.
Further, in order to analyze the variation trend of the vibration caused on the time sequence, the characteristics in the continuous sampling time periods are adjacent to form a characteristic matrix. In the embodiment of the invention, each feature is adjacent to serve as one column of the feature matrix, and the feature vectors are arranged according to the time sequence information to obtain the feature matrix.
The feature matrix comprises vibration information, audio information and temperature information, and the information is richer and has stronger representativeness.
And step S4: grouping the plurality of gear boxes according to the characteristic matrix to obtain a plurality of gear box groups; extracting the vibration characteristics of the characteristic matrix to obtain a vibration characteristic change sequence on the time sequence of each gearbox in the gearbox group; and taking the vibration characteristic change sequence in each gearbox group as training data, and obtaining a CEEMD-GRU information prediction network according to the training data.
Because the transmission structure comprises a large number of gearboxes of various kinds, the gearboxes of different kinds have different vibration characteristics, and the gearboxes of the same kind under different working conditions can also generate different vibration characteristics. Therefore, in order to facilitate analysis of vibration information and expansion of a data set of subsequent network training, all gearboxes in the transmission structure need to be classified, and because the feature matrix has abundant vibration features, the gearboxes can be grouped according to the feature matrix to obtain a plurality of gearbox groups, and the gearboxes in each gearbox group have the same vibration features and vibration change features in time sequence.
Preferably, the feature matrix is used as a research sample, the research sample is clustered by using a quantum clustering algorithm, a plurality of cluster clusters are obtained, and each cluster corresponds to one gear box group. The quantum clustering algorithm is a clustering algorithm based on division and is an unsupervised clustering algorithm, the algorithm does not need to preset clustering centers and clustering numbers, and the basic idea of the algorithm is as follows: based on quantum theory, researching the distribution rule of samples, solving the potential energy of each research sample based on an iterative gradient descent algorithm and a Schrodinger equation without time, wherein other research samples are distributed around the research sample with zero or minimum potential energy, so that the research sample with zero or minimum potential energy is the cluster center of the class, and the other research samples distributed around the cluster center are classified into the class. It should be noted that the quantum clustering algorithm is a technical means well known to those skilled in the art, and details are not repeated here, and only the schrodinger equation adopted in the embodiment of the present invention is briefly described:
schrodinger's equation is used to solve a wave function with potential field constraints, which is used to describe the quantum states used to describe the microscopic particles. In the embodiment of the invention, a Schrodinger equation without time display is adopted, namely the Schrodinger equation without time display is assumed to not change along with the change of time:
wherein the content of the first and second substances,
the expression Hamilton operator is used to represent,
the function of the wave is represented by,
the function of the potential energy is expressed,
representing the energy eigenvalue of the Hamilton operator,
the shape splitting operator is represented as a shape splitting operator,
width adjustment parameters representing wave functions, which will be described in embodiments of the present invention
Is arranged as
。
The gear box in every gear box group all corresponds a characteristic matrix, for the trend that the vibration of the gear box of every kind of vibration characteristics changes in the time sequence of convenient analysis, extracts the vibration characteristic of characteristic matrix, obtains the vibration characteristic change sequence on every gear box time sequence in the gear box group, specifically includes:
the feature matrix is divided into a plurality of determinants of the same size according to time sequence. And obtaining row and column values of the determinant, taking the row and column values as vibration characteristics, and arranging the row and column values according to a time sequence to obtain a vibration characteristic change sequence.
In the embodiment of the present invention, considering that the length of the feature vector is 8, the number of columns of each determinant is also 8, that is, the size of each determinant is 8 × 8, that is, each determinant is the variation of each vibration feature factor of the corresponding gearbox at 8 sampling frequencies.
It should be noted that calculating the column and row values of the determinant is a well-known technical means for those skilled in the art, and the brief process is summarized as follows: the rank value is the result of the sum of the right slant products minus the sum of the left slant products in the rank.
Because the vibration characteristics and the vibration change characteristics in each gearbox group are similar, the vibration characteristic change sequence in each gearbox group can be used as training data of a network, all sequence data in the group are integrated, a training data set of the network is expanded, the accuracy of the trained network is higher, and the CEEMD-GRU information prediction network corresponding to each gearbox group is obtained through training of the training data set.
The CEEMD-GRU information prediction network is a prediction model based on a combination of CEEMD (complementary empirical mode decomposition) and GRU (gated cyclic unit). In the network, firstly, an input vibration characteristic change sequence is decomposed into a plurality of intrinsic mode function components and a residual component based on a CEEMD algorithm, fluctuation information implied by data is deeply mined, then training is completed under the network structure of a GRU, and a future vibration characteristic change sequence at a future moment is output. It should be noted that the network structure and the specific training method of the CEEMD-GRU are well known to those skilled in the art, and only a brief training process in the embodiment of the present invention is briefly described here:
(1) A plurality of vibration characteristic change sequences are selected as a training data set, and for training, normalization operation is firstly carried out after training data are input into a network.
(2) And carrying out stabilization processing on the training data set through a CEEMD algorithm in the network, and deeply mining fluctuation information implied by the data to obtain a plurality of intrinsic mode function components and residual components.
(3) Recombining a plurality of subcomponents into data input by the GRU neural network, and dividing the data into a training set and a test set, wherein the ratio of the training set to the test set is 20.
(4) And (3) taking the recombined data as GRU neural network input, taking the vibration characteristic change sequence as GRU neural network output, and carrying out multi-input single-output GRU neural network training.
(5) And when the network meets the convergence condition, obtaining the CEEMD-GRU information prediction network after training.
Step S5: and predicting a future vibration characteristic change sequence of the gearbox to be evaluated by the network according to the corresponding CEEMD-GRU information prediction network, and judging whether the gearbox to be evaluated is abnormal or not according to the sizes and the element positions of elements in the future vibration characteristic change sequence.
The method comprises the steps of storing vibration test and data acquisition of a transmission mechanism structural component into a historical database, obtaining a CEEMD-GRU information prediction network capable of predicting vibration characteristic change trend based on historical data in the historical database, collecting data of a gear box to be evaluated when the gear box to be evaluated is actually required to be evaluated, obtaining a real-time vibration characteristic change sequence, and predicting a future vibration characteristic change sequence of the gear box to be evaluated according to the corresponding CEEMD-GRU information prediction network, wherein the method specifically comprises the following steps:
and obtaining a real-time characteristic matrix and a real-time vibration characteristic change sequence of the gearbox to be evaluated, matching the real-time characteristic matrix with the characteristic matrix of the central sample of each gearbox group, and obtaining the most matched gearbox group corresponding to the gearbox to be evaluated. And inputting the real-time vibration characteristic change sequence of the gearbox to be evaluated into a CEEMD-GRU information prediction network corresponding to the mostly-matched gearbox group to obtain a future vibration characteristic change sequence.
It should be noted that the matching method is a prior art known to those skilled in the art, and various matching methods such as distance matching and similarity matching may be selected in specific implementation, and may be specifically set according to specific implementation conditions, which is not limited herein.
The future vibration characteristic change sequence reflects the change trend of the vibration characteristics of the gearbox to be evaluated at the future moment, and abnormal vibration can cause the obtained numerical value of the vibration characteristics to be increased, so that whether the gearbox to be evaluated is abnormal or not can be judged according to the size and the element position of the elements in the future vibration characteristic change sequence, and the method specifically comprises the following steps:
if a certain element exists in the future vibration characteristic change sequence and is larger than a preset characteristic threshold value, a time sequence position of the element is obtained, an abnormal time point is obtained according to the time sequence position, and if the time period from the abnormal time point to the real-time point is smaller than a preset time length threshold value, the fact that abnormal vibration of the gearbox to be evaluated, which is generated by various factors, is about to cause large influence on the gearbox is indicated, namely the fact that the gearbox to be evaluated is abnormal is considered. Because the abnormal condition is judged through the prediction data, the working personnel have enough time to replace or maintain the abnormal gear box, and the influence on the efficiency of the continuous casting process caused by the abnormal gear box which is not found in time to cause the abnormal operation of the transmission mechanism is avoided. It should be noted that the characteristic threshold may be specifically set according to the requirements of the transmission mechanism on various gearboxes in a specific real-time scenario, and is not limited herein.
In summary, the embodiment of the invention acquires the envelope information of the vibration signal of the gearbox, analyzes the fluctuation characteristics of the envelope information, the audio information and the temperature information of the gearbox in time sequence, and constructs the characteristic matrix. The gearboxes are grouped based on the characteristic matrix, and a plurality of gearbox groups are obtained. And extracting the vibration characteristics in the characteristic matrix to obtain a vibration characteristic change sequence, and training a corresponding CEEMD-GRU information prediction network according to the vibration characteristic change sequence in each gearbox group. And obtaining a future vibration characteristic change sequence of the gearbox to be evaluated through a CEEMD-GRU information prediction network. And evaluating the abnormity of the gearbox to be evaluated according to the elements in the future vibration characteristic change sequence. According to the embodiment of the invention, the abnormal gear box is timely evaluated by a prediction means through acquiring and analyzing the vibration data of the gear box of the transmission mechanism, so that the efficiency of the metallurgical process is ensured.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. The processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (9)
1. A method for acquiring and evaluating local damage vibration envelope signals of a metallurgical transmission mechanism is characterized by comprising the following steps:
collecting vibration signals, audio information and temperature information of each gear box of the metallurgical transmission mechanism; carrying out envelope processing on the vibration signal to obtain various kinds of envelope information;
acquiring an envelope information sequence of each kind of envelope information according to a preset sampling frequency; obtaining a fluctuation frequency factor according to the difference between adjacent elements in the envelope information sequence and the element mean value; obtaining a fluctuation amplitude factor according to the element range and the element mean value in the envelope information sequence; obtaining an audio information sequence and a temperature information sequence according to the sampling frequency, obtaining a first volatility factor according to the fluctuation amplitude and the fluctuation frequency of the audio information sequence, and obtaining a second volatility factor of the temperature information sequence;
combining a plurality of fluctuation frequency factors, the first volatility factor and the second volatility factor in the same sampling time period to obtain a feature vector; the feature vectors in the continuous sampling time period form a feature matrix;
grouping the plurality of gear boxes according to the characteristic matrix to obtain a plurality of gear box groups; extracting the vibration characteristics of the characteristic matrix to obtain a vibration characteristic change sequence on the time sequence of each gearbox in the gearbox group; using the vibration characteristic change sequence in each gearbox group as training data, and obtaining a CEEMD-GRU information prediction network according to the training data;
and predicting a future vibration characteristic change sequence of the gearbox to be evaluated by the network according to the corresponding CEEMD-GRU information, and judging whether the gearbox to be evaluated is abnormal or not according to the sizes and element positions of elements in the future vibration characteristic change sequence.
2. The method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the step of performing envelope processing on the vibration signal to acquire a plurality of kinds of envelope information comprises the steps of:
and performing analog-to-digital conversion, frequency spectrum conversion, band-pass filtering and acceleration envelope processing on the vibration signal to obtain envelope information, wherein the envelope information comprises vibration acceleration, vibration speed and acceleration envelope value.
3. The method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the obtaining of the fluctuation frequency factor according to the difference between adjacent elements in the envelope information sequence and the element mean value comprises:
obtaining the fluctuation frequency factor according to a fluctuation frequency factor formula, wherein the fluctuation frequency factor formula comprises:
wherein the content of the first and second substances,
for the purpose of said wave frequency factor,
is the mean value of the elements in the envelope information sequence,
for the first in the envelope information sequence
The value of each of the elements is,
for the number of elements in the envelope information sequence,
as a function of absolute values.
4. The method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the obtaining of the fluctuation amplitude factor according to the element range and the element mean value in the envelope information sequence comprises:
and obtaining the element value difference between the maximum element value and the element average value in the envelope information sequence, and taking the ratio of the element range of the envelope information sequence to the element value difference as the fluctuation amplitude factor.
5. The method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the obtaining of the first volatility factor according to the fluctuation amplitude and the fluctuation frequency of the audio information sequence comprises:
obtaining the first volatility factor according to a first volatility factor formula, the first volatility factor formula comprising:
wherein, the first and the second end of the pipe are connected with each other,
for the purpose of the first volatility factor,
is the average of the elements of the sequence of audio information,
is the variance of an element of the sequence of audio information,
for the number of elements in the sequence of audio information,
for the first in the audio information sequence
And (4) each element.
6. The method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the step of grouping the plurality of gearboxes according to the feature matrix to obtain a plurality of gearbox groups comprises the following steps:
and clustering the research samples by using the characteristic matrix as a research sample by using a quantum clustering algorithm to obtain a plurality of clustering clusters, wherein each clustering cluster corresponds to one gear box group.
7. The method for obtaining and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the extracting the vibration characteristics of the characteristic matrix to obtain the vibration characteristic variation sequence of each gearbox in the gearbox group in time sequence comprises:
dividing the characteristic matrix into a plurality of determinants with the same size according to time sequence; and obtaining the row and column values of the determinant, taking the absolute values of the row and column values as the vibration characteristics, and arranging the vibration characteristics according to a time sequence to obtain the vibration characteristic change sequence.
8. The method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the predicting the future vibration characteristic change sequence of the gearbox to be evaluated according to the corresponding CEEMD-GRU information by the prediction network comprises:
obtaining a real-time characteristic matrix and a real-time vibration characteristic change sequence of the gearbox to be evaluated, matching the real-time characteristic matrix with the characteristic matrix of the central sample of each gearbox group, and obtaining a best-matched gearbox group corresponding to the gearbox to be evaluated; and inputting the real-time vibration characteristic change sequence of the gearbox to be evaluated into the CEEMD-GRU information prediction network corresponding to the mostly matched gearbox group to obtain the future vibration characteristic change sequence.
9. The method for acquiring and evaluating the local damage vibration envelope signal of the metallurgical transmission mechanism according to claim 1, wherein the step of judging whether the gearbox to be evaluated is abnormal or not according to the sizes and the element positions of the elements in the future vibration characteristic change sequence comprises the following steps of:
if a certain element exists in the future vibration characteristic change sequence and is larger than a preset characteristic threshold value, obtaining a time sequence position of the element, obtaining an abnormal time point according to the time sequence position, and if a time period from the abnormal time point to a real-time point is smaller than a preset time length threshold value, determining that the gear box to be evaluated is abnormal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211169217.6A CN115235762A (en) | 2022-09-26 | 2022-09-26 | Method for acquiring and evaluating local damage vibration envelope signal of metallurgical transmission mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211169217.6A CN115235762A (en) | 2022-09-26 | 2022-09-26 | Method for acquiring and evaluating local damage vibration envelope signal of metallurgical transmission mechanism |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115235762A true CN115235762A (en) | 2022-10-25 |
Family
ID=83667124
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211169217.6A Pending CN115235762A (en) | 2022-09-26 | 2022-09-26 | Method for acquiring and evaluating local damage vibration envelope signal of metallurgical transmission mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115235762A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102768115A (en) * | 2012-06-27 | 2012-11-07 | 华北电力大学 | Method for dynamically monitoring health status of wind turbine gearbox in real time |
| CN109708907A (en) * | 2018-11-30 | 2019-05-03 | 浙江胄天科技股份有限公司 | A kind of equipment fault feature extracting method based on envelope information |
| CN110610035A (en) * | 2019-08-28 | 2019-12-24 | 武汉科技大学 | Rolling bearing residual life prediction method based on GRU neural network |
| CN111709490A (en) * | 2020-06-24 | 2020-09-25 | 河北工业大学 | Fan health state assessment method based on GRU neural network |
| CN112651290A (en) * | 2020-10-22 | 2021-04-13 | 国家电网有限公司 | Hydroelectric generating set vibration trend prediction method and system |
| CN113255848A (en) * | 2021-07-08 | 2021-08-13 | 浙江大学 | Water turbine cavitation sound signal identification method based on big data learning |
| CN113339204A (en) * | 2021-05-01 | 2021-09-03 | 北京工业大学 | Wind driven generator fault identification method based on hybrid neural network |
-
2022
- 2022-09-26 CN CN202211169217.6A patent/CN115235762A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102768115A (en) * | 2012-06-27 | 2012-11-07 | 华北电力大学 | Method for dynamically monitoring health status of wind turbine gearbox in real time |
| CN109708907A (en) * | 2018-11-30 | 2019-05-03 | 浙江胄天科技股份有限公司 | A kind of equipment fault feature extracting method based on envelope information |
| CN110610035A (en) * | 2019-08-28 | 2019-12-24 | 武汉科技大学 | Rolling bearing residual life prediction method based on GRU neural network |
| CN111709490A (en) * | 2020-06-24 | 2020-09-25 | 河北工业大学 | Fan health state assessment method based on GRU neural network |
| CN112651290A (en) * | 2020-10-22 | 2021-04-13 | 国家电网有限公司 | Hydroelectric generating set vibration trend prediction method and system |
| CN113339204A (en) * | 2021-05-01 | 2021-09-03 | 北京工业大学 | Wind driven generator fault identification method based on hybrid neural network |
| CN113255848A (en) * | 2021-07-08 | 2021-08-13 | 浙江大学 | Water turbine cavitation sound signal identification method based on big data learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lu et al. | Dominant feature selection for the fault diagnosis of rotary machines using modified genetic algorithm and empirical mode decomposition | |
| CN105971901B (en) | A kind of centrifugal pump fault diagnostic method based on complete overall experience mode decomposition and random forest | |
| Dou et al. | A rule-based intelligent method for fault diagnosis of rotating machinery | |
| US20200364386A1 (en) | Soft sensing method and system for difficult-to-measure parameters in complex industrial processes | |
| CN106153179A (en) | Medium-speed pulverizer vibrating failure diagnosis method | |
| Hameed et al. | Comparative analysis of fuzzy classifier and ANN with histogram features for defect detection and classification in planetary gearbox | |
| Qu et al. | Feature selection for damage degree classification of planetary gearboxes using support vector machine | |
| Zhang et al. | Multi-model ensemble deep learning method for intelligent fault diagnosis with high-dimensional samples | |
| Zhao et al. | Rolling bearing fault diagnosis based on CEEMD and time series modeling | |
| CN116304751B (en) | Operation data processing method for overhauling motor train unit components | |
| CN104863840A (en) | Reciprocating compressor intelligent diagnosis method based on EMD-PCA | |
| Zheng et al. | An improved empirical wavelet transform and refined composite multiscale dispersion entropy-based fault diagnosis method for rolling bearing | |
| CN101532902A (en) | Quality-monitoring and fault-diagnosing method for cold-punching mold | |
| CN115034263A (en) | Variable-scale evolution self-adaptive noise elimination method, noise elimination system and fault diagnosis system | |
| CN115235762A (en) | Method for acquiring and evaluating local damage vibration envelope signal of metallurgical transmission mechanism | |
| Isham et al. | Iterative variational mode decomposition and extreme learning machine for gearbox diagnosis based on vibration signals | |
| Ding et al. | A fault feature extraction method of motor bearing using improved LCD | |
| Zurita et al. | A review of vibration machine diagnostics by using artificial intelligence methods | |
| CN111693279A (en) | Mechanical fault diagnosis method based on MPGA parametric resonance sparse decomposition | |
| CN116610990A (en) | Method and device for identifying hidden danger of breaker based on characteristic space differentiation | |
| CN115753067A (en) | Digital twin-based method for diagnosing rotary mechanical fault under strong noise | |
| Wang et al. | Fault Diagnosis Method of Wind Turbine Gearbox Based on Fusion Multispectrogram and Improved CNN Neural Network | |
| CN113639985A (en) | Mechanical fault diagnosis and state monitoring method based on optimized fault characteristic frequency spectrum | |
| Tastimur et al. | A New Bearing Fault Diagnosis Method using Envelope based Feature Extraction | |
| Liu et al. | Bearing fault detection in varying operational conditions based on empirical mode decomposition and random forest |
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 |