CN113657221A - Power plant equipment state monitoring method based on intelligent sensing technology - Google Patents
Power plant equipment state monitoring method based on intelligent sensing technology Download PDFInfo
- Publication number
- CN113657221A CN113657221A CN202110889067.5A CN202110889067A CN113657221A CN 113657221 A CN113657221 A CN 113657221A CN 202110889067 A CN202110889067 A CN 202110889067A CN 113657221 A CN113657221 A CN 113657221A
- Authority
- CN
- China
- Prior art keywords
- power plant
- sound
- fault
- method based
- monitoring method
- 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
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000005516 engineering process Methods 0.000 title claims abstract description 28
- 238000012544 monitoring process Methods 0.000 title claims abstract description 27
- 230000005236 sound signal Effects 0.000 claims abstract description 70
- 230000002159 abnormal effect Effects 0.000 claims abstract description 51
- 238000003745 diagnosis Methods 0.000 claims abstract description 45
- 238000001514 detection method Methods 0.000 claims abstract description 44
- 230000008569 process Effects 0.000 claims abstract description 12
- 230000005856 abnormality Effects 0.000 claims description 18
- 230000006870 function Effects 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 18
- 238000012706 support-vector machine Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 14
- 238000000605 extraction Methods 0.000 claims description 10
- 230000008447 perception Effects 0.000 claims description 10
- 230000009467 reduction Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000001228 spectrum Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 4
- 230000009977 dual effect Effects 0.000 claims description 3
- 238000009432 framing Methods 0.000 claims description 3
- 230000002045 lasting effect Effects 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 5
- 230000008859 change Effects 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000005389 magnetism Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000037433 frameshift Effects 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
Images
Classifications
-
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Abstract
The invention provides a power plant equipment state monitoring method based on an intelligent sensing technology, which comprises the following steps: s1, collecting a sound signal in the operation process of equipment; s2, carrying out abnormity detection on the sound signal by an abnormity detection model, and inputting an abnormal signal to a fault diagnosis model when the abnormal signal is detected; and S3, carrying out fault diagnosis on the abnormal signal by the fault diagnosis model. According to the scheme, reliability judgment is carried out according to the probability value of the fault type of the fault diagnosis model, the judgment result with low reliability is output to wait for manual calibration, the fault diagnosis model is continuously optimized in the using process, the fault type judgment accuracy can be continuously improved along with continuous use of the model, and the problem that fault diagnosis signals are missed in practical use due to insufficient fault type data can be solved.
Description
Technical Field
The invention belongs to the technical field of power plant equipment state monitoring, and particularly relates to a power plant equipment state monitoring method based on an intelligent sensing technology.
Background
Currently, the equipment of the power plant is often detected, and the equipment is usually periodically checked and maintained, mainly depending on manual experience and based on historical data. The method has certain disadvantages of limited maintenance effect and high cost because the fault interval discreteness of most of equipment and spare parts is large. Fixed cycle polling, if the cycle is too frequent, not only can generate significant labor costs, but also may generate unnecessary and even damaging maintenance activities. Secondly, if the period is not frequent enough, a larger risk of equipment failure is brought, and a larger loss is brought. Therefore, the real-time parameters of the equipment operation are monitored and analyzed through advanced technical means to judge whether the equipment has abnormity or faults, the fault positions and reasons and the degradation trend of the faults, so that a reasonable overhaul opportunity is determined, the accidents are eliminated in a bud state, the maintenance cost is effectively reduced, and the accident outage rate is reduced, which is very necessary.
The power plant production environment has very large noise, the sound of different equipment is mixed in a relatively concentrated space, and as the sound signals are not processed by a good technical means in the past, the abnormal sound of the equipment can be heard only by inspection personnel with abundant professional experience, and along with the development of the power plant equipment towards high, precise and sharp directions, the running state of the equipment is difficult to be effectively and accurately judged only by the inspection personnel. In a practical system, the running state of the equipment under different working conditions is different, and when the working conditions are changed, if a fault occurs, the running state of the equipment is changed. Although the mechanism of sound signal generation of the equipment in the state change process is fuzzy, the sound signal often has non-stationary characteristics, so that the sound signal can be analyzed and processed by adopting a statistical model theory. The device state often causes the change of the sound signal structure and the appearance of different sound wave patterns in the changing process, and as shown in fig. 1, the running state of the device can be judged through the change of the device state sound signal characteristics, and even the device state sound signal characteristics are used for judging the type and the occurrence position of the device fault.
In recent years, with the development and successful industrial application of technologies such as predictive control, kernel partial least squares, neural networks, support vector machines and the like and machine learning algorithms, a model of equipment fault diagnosis can be established by fully utilizing sufficient historical operating data of a power plant, and online analysis and intelligent early warning of equipment states are realized. In the actual production, the failure signal is judged to be a normal signal by mistake or by mistake, which may bring a large loss. Fault diagnosis models trained on historical data must discriminate these new types of faults into a certain class of existing classes. And normal operation data in the operation of the equipment is far higher than fault type data, and the model easily causes misjudgment or missed judgment on the very small fault data.
In order to accurately judge and early warn the fault type of the power plant equipment, a new fault diagnosis method is required to be found to solve the problem that fault signals are missed and misjudged by a fault diagnosis model in actual use due to insufficient fault type data.
Disclosure of Invention
The invention aims to solve the problems and provides a power plant equipment state monitoring method based on an intelligent sensing technology.
In order to achieve the purpose, the invention adopts the following technical scheme:
a power plant equipment state monitoring method based on an intelligent perception technology comprises the following steps:
s1, collecting a sound signal in the operation process of equipment;
s2, carrying out abnormity detection on the sound signal by an abnormity detection model, and inputting an abnormal signal to a fault diagnosis model when the abnormal signal is detected;
and S3, carrying out fault diagnosis on the abnormal signal by the fault diagnosis model.
In the above power plant equipment state monitoring method based on the intelligent perception technology, the anomaly detection model is a single-class support vector machine, and the anomaly detection model is obtained by training in advance in the following way:
A1. acquiring first sample data;
A2. preprocessing the sound signal in the first sample data;
A3. extracting the characteristics of the preprocessed sound signals to form characteristic vectors;
A4. carrying out dimension reduction processing on the feature vector;
A5. establishing an anomaly detection model by utilizing a feature training single-class support vector machine subjected to dimension reduction;
in step S2, the acquired sound signal is subjected to data processing including preprocessing, feature extraction, and dimension reduction, and then input to the abnormality detection model to perform abnormality detection on the sound signal by the abnormality detection model.
In the method for monitoring the state of the power plant equipment based on the intelligent sensing technology, the sound signals in the first sample data are normal signals;
or, the sound signals in the first sample data include normal signals marked with normal labels and abnormal signals marked with abnormal labels.
In the above power plant equipment state monitoring method based on the intelligent sensing technology, step a2 specifically includes:
A21. filtering the sound signal, and filtering low-frequency interference signals below 50Hz by using a high-pass filter;
A22. the sound is segmented and a sliding window framing process is used for each segment of sound.
In the above power plant equipment condition monitoring method based on the intelligent sensing technology, in step a3, the extracted features include a margin factor, a pulse factor, a skew factor in the time domain, and a barycentric frequency, a mean square frequency, and a frequency feature in the frequency spectrum.
In the above power plant equipment state monitoring method based on the intelligent sensing technology, in step a5, an ant colony algorithm is used to perform parameter optimization to obtain an optimal kernel function parameter and a balance parameter of a single-class support vector machine, the optimal kernel function parameter and the balance parameter are substituted into the support vector machine to obtain a trained anomaly detection model, and the initial problem of an objective function is as follows:
using a gaussian kernel function:
the conditions are satisfied: h (x)i,xj)＝y(xi)Ty(xj) (3)
(1) The lagrangian dual problem of formula (la) is:
solve out each lambdaiThen, the discriminant function is obtained as:
(5) in the formula
In the above formulas, l is the number of training samples;
sigma is the required optimal kernel function parameter;
ν epsilon (0, 1) is a balance parameter, is a percentage parameter estimation which is predefined through the ant colony algorithm and represents a compromise between a support vector and a wrong component;
rho is compensation of a required hyperplane in the feature space;
omega is a normal vector of a required hyperplane in a feature space;
ξiis a relaxation variable;
λi、λjis Lagrange multiplier, xi、xjIs the sample in the original space.
In the above power plant equipment state monitoring method based on the intelligent sensing technology, in step S3, the fault diagnosis model is used to output a fault type and a fault probability value determination of the corresponding fault type;
and step S3 is followed by:
s4, judging whether the fault probability value is higher than a preset probability value, if so, directly outputting a fault type, otherwise, executing a step S5;
and S5, outputting the judged fault type and the fault probability value, and waiting for a worker to give a manual calibration result.
In the above power plant equipment state monitoring method based on the intelligent sensing technology, in step S5, after the staff gives the artificial calibration result, the calibration result is used as the given label of the corresponding sound signal, and the sound signal with the given label is input to the abnormality detection model and/or the fault diagnosis model for training and updating.
In the above power plant equipment state monitoring method based on the intelligent sensing technology, in step S1, sound signals in the operation process of the equipment are collected at a first collection frequency;
in step S2, when an abnormal signal is detected, acquiring a sound signal during the operation of the device at a second acquisition frequency, and simultaneously inputting the abnormal signal and the sound signal acquired at the second acquisition frequency to the fault diagnosis model;
the second acquisition frequency is greater than the first acquisition frequency.
In the above power plant equipment state monitoring method based on the intelligent sensing technology, in step S2, the abnormality detection model simultaneously detects the sound signals during the operation of the second collection frequency collection equipment, and when the sound signals lasting for the preset times/time are all detected as normal sound signals, the first collection frequency is recovered.
The invention has the advantages that:
1. the problem that the fault diagnosis signal is missed in actual use due to insufficient fault type data can be solved;
2. and according to the probability value of the fault type of the fault diagnosis model, reliability judgment is carried out, a judgment result with low reliability is output to wait for manual calibration, the fault diagnosis model is continuously optimized in the using process, and the fault type judgment accuracy can be continuously improved along with the continuous use of the model.
Drawings
FIG. 1 is a diagram of sound waveforms for a device in various operating states;
FIG. 2 is a flow chart of the training of the anomaly detection model in the present invention;
FIG. 3 is a flow chart of the training of the fault diagnosis model in the present invention;
FIG. 4 is a flow chart of a method for determining the state of a power plant based on the intelligent sensing technology;
FIG. 5 is a graph of AUC (area Under cut) as a model for evaluating the abnormality detection by the ROC curve.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
As shown in fig. 2, the embodiment discloses a power plant equipment state monitoring method based on an intelligent sensing technology, and firstly, a trained anomaly detection model and a fault diagnosis model are prepared.
The anomaly detection model adopts a single-class support vector machine, and is mainly obtained by the following training:
A1. acquiring first sample data; the sound signals in the first sample data may all be normal signals; a normal signal with a normal tag and an abnormal signal with an abnormal tag may also be included.
A2. Preprocessing a sound signal in the first sample data, firstly, filtering the sound signal, and filtering a low-frequency interference signal below 50Hz by using a high-pass filter; and dividing the sound into 6s sections, and performing frame division processing on each section of sound by using a sliding window, wherein the window length is 0.25s, and the window is shifted by 50 percent, namely the window length is shifted by half each time.
A3. Extracting the characteristics of the preprocessed sound signals to form characteristic vectors: extracting the statistical characteristics of all frame signals in the time domain and the characteristics of all frame signals in the frequency spectrum distribution, such as margin factors, pulse factors, skew factors in the time domain, center-of-gravity frequency, mean square frequency and frequency characteristics in the frequency spectrum, setting fs as a sampling frequency, sampling and dispersing a continuous sound signal s (t) per second into s (N), setting the frame length for framing every 6 seconds of signals as N, frame shift as N/2 and total frame number as NF. The ith frame signal si(n) statistical characteristics in the time domain, margin factors, and an extraction formula:
the pulse factor, the extraction formula is:
the skewness factor, the extraction formula is:
s (N) (1, 2.., N) is a time-domain signal, and N is a signal sample length.
The signal power spectrum reflects the random distribution of signal energy, i.e. the characteristics of the signal are analyzed from the frequency components in the signal and the energy size of the frequency components. Performing Fourier transform on the signal s (n) to obtain distribution information of the signal s (n) on a frequency spectrum, wherein the formula of the fast Fourier transform is as follows:
wherein s isi(k) Is the ith frame signal siSTFT (short time fourier transform) of (n).
The ith frame signal si(k) The characteristics in the frequency domain are such that,
the gravity center frequency, the extraction formula is:
mean square frequency, the extraction formula is:
frequency characteristics, the extraction formula is:
y (K) 1,2, K) is the spectral value of the signal s (n), K is the number of spectral lines, f is the number of spectral lineskIs the frequency value. And forming a 6-dimensional vector by the characteristic parameters of the signals in the time domain and the frequency domain to serve as the characteristic vector of a section of sound signal.
A4. And (3) performing dimensionality reduction on the feature vector: finding out main features and replacing original data with the main features;
A5. establishing an anomaly detection model by utilizing a feature training single-class support vector machine subjected to dimensionality reduction, specifically comprising the following steps of:
normalizing the characteristic parameters subjected to dimension reduction; performing parameter optimization by using an ant colony algorithm to obtain an optimal kernel function parameter and a balance parameter of a single-class support vector machine, and constructing and solving an optimal problem; the initial problem of the objective function is:
using a gaussian kernel function:
the conditions are satisfied: h (x)i,xj)＝y(xi)Ty(xj) (10)
(8) The lagrangian dual problem of formula (la) is:
solve out each lambdaiThen, the discriminant function is obtained as:
(12) in the formula
In the above formulas, l is the number of training samples;
sigma is the required optimal kernel function parameter;
ν epsilon (0, 1) is a balance parameter, is a predefined percentage parameter estimation and represents a compromise between a support vector and a wrong component;
rho is compensation of a required hyperplane in the feature space;
omega is a normal vector of a required hyperplane in a feature space;
ξiis a relaxation variable;
λi、λjis Lagrange multiplier, xi、xjIs the sample in the original space.
And substituting the optimal value of the model parameters (v, sigma) of the single-class support vector machine obtained by the ant colony algorithm optimization into the support vector machine to obtain the trained anomaly detection model.
The fault diagnosis model adopts a one-dimensional convolutional neural network formed by an input layer, a feature extraction layer and a classification layer. As shown in fig. 3, the trained fault diagnosis model can be obtained by inputting the abnormal signal for marking the fault type into the fault diagnosis model and training. The abnormal signal can be prepared in advance by a worker, or the second sample data can be detected by an abnormal detection model, namely the second sample data is processed and then input into the judgment function (12) to obtain a training sample set D of the abnormal signal, then the worker classifies the data in the training sample set according to different fault types and respectively marks the data as D1,D2,...,DkAnd making classification marks.
The first sample data and the second sample data may use different sample data, or may use the same sample data.
When the device is put into use, as shown in fig. 4, the acquired original signals of the device are input to the abnormality detection model for abnormality judgment, and then the abnormal signals are input to the fault diagnosis model for fault classification. The specific method comprises the following steps:
s1, collecting sound signals in the operation process of equipment at a first collection frequency, wherein the first collection frequency can be once in 10 minutes and is collected for 6s each time; can use magnetism to inhale the sound signal of sound sensor collection equipment operation in-process on the equipment that awaits measuring, magnetism is inhaled the mounting means and can directly is acquireed the inside sound of equipment, has reduced the influence of environmental noise to the signal, preferably utilizes the intelligent sound sensor of imitative ear hearing to through using thing networking 5G technique, solve the problem that sound signal collection transmission is delayed.
S2, after the collected sound signals are processed by preprocessing, feature extraction and dimension reduction processing, inputting the processed sound signals into an abnormality detection model so as to perform abnormality detection on the sound signals by the abnormality detection model, and when the abnormal signals are detected, inputting the abnormal signals into a fault diagnosis model;
and S3, carrying out fault diagnosis on the abnormal signal by the fault diagnosis model.
Since the abnormal detection model cannot distinguish the normal signal from the abnormal signal by one hundred percent, some normal signals of the equipment running in the extreme working condition may be misjudged as abnormal signals (the fault diagnosis model cannot judge the abnormal signals as the normal signals, but can judge the normal signals as the abnormal signals), so the detected data may include known fault types and abnormal types (unknown fault types or normal), and the judgment condition is added after the fault identification model in the embodiment to distinguish the abnormal types.
Specifically, in step S3, the fault diagnosis model not only outputs the fault type, but also outputs a fault probability value of the corresponding fault type;
and step S3 is followed by:
s4, judging whether the fault probability value is higher than a preset probability value, if so, taking the preset probability value as a critical point which can be accurately judged, if not, judging that the fault type is accurately judged, directly outputting the fault type, and otherwise, executing the step S5;
and S5, considering that the fault type cannot be accurately judged, outputting the judged fault type and the fault probability value, and waiting for a worker to give a manual calibration result.
And when the staff gives a man-made calibration result, taking the calibration result as a given label of the corresponding sound signal, and inputting the sound signal with the given label into the abnormality detection model and/or the fault diagnosis model for training and updating. If the abnormal sound signals are calibrated to be normal, the related sound signals are placed into a normal sample database, the abnormal detection model is trained and updated, and if the abnormal sound signals are calibrated to be abnormal, the sound signals are placed into a fault sample database, and the fault diagnosis model is trained and updated.
Preferably, when the abnormal signal is detected, collecting the sound signal in the running process of the equipment at a second collection frequency, and simultaneously inputting the abnormal signal and the sound signal collected at the second collection frequency to the fault diagnosis model; the second acquisition frequency may be once in 1 minute for 6 seconds. The data acquisition method has the advantages that the data are acquired once within 10-30 minutes under normal conditions, the acquisition frequency is increased when the data are judged to be abnormal preliminarily, the requirements of saving energy consumption and avoiding frequent acquisition can be met, the continuous acquisition of the data in the abnormal condition can be met, and the data acquisition method has similar inspection effect.
Further, the abnormality detection model simultaneously detects the sound signals during the operation of the second collection frequency collection device, and when the sound signals lasting for a preset number of times/time are all detected as normal sound signals, the first collection frequency is recovered. When the sound signal is abnormal, the sound signal is continuously collected, so that the fault state can be diagnosed by the model, and when the sound signal is normal, the sound signal is not required to be collected frequently and is recovered to be collected at a long interval.
And evaluating the OCSVM abnormity detection model by using an ROC-AUC model evaluation index. Firstly, manually labeling test data, and manually identifying normal sound and abnormal sound data. The test data of the tag is then input into the ROC-AUC evaluation program, and the results of ROC acquisition are shown in fig. 5. In the ROC result graph, the closer the curve is to the upper right corner of the coordinate system, the higher the accuracy of the model is proved, and the better the effect is. Considering that the ROC curve cannot intuitively explain the performance of a classifier, and the AUC represents the area under the ROC curve, the ROC curve is mainly used for measuring the generalization performance of the model, namely the classification effect, the value is taken as a numerical value of [0,1], the closer the value is to 1, the evaluated model has comparability, and quantitative comparison can be carried out. The AUC value of the model is 0.891, which proves that the model can well complete the task in the sound abnormal detection of specific situations.
According to the scheme, the fault diagnosis is carried out after the abnormal detection, and the fault diagnosis model is trained by using only abnormal data, so that the influence of an unbalanced sample on the classification model identification accuracy is avoided; the abnormal detection model can detect all abnormal signals, so that the problem that the fault diagnosis model fails to judge the fault in actual use due to insufficient fault type data can be well solved; and whether the model is reliable or not is judged according to the probability value output by diagnosis, the unreliable result is waited for manual calibration and the training model is further updated, and the model diagnosis accuracy can be continuously improved along with the continuous use of the model.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.
Although the terms acoustic signature, anomaly detection model, fault diagnosis model, one-class support vector machine, normal signature, abnormal signature, etc. are used more often herein, the possibility of using other terms is not excluded. These terms are used merely to more conveniently describe and explain the nature of the present invention; they are to be construed as being without limitation to any additional limitations that may be imposed by the spirit of the present invention.
Claims (10)
1. A power plant equipment state monitoring method based on an intelligent perception technology is characterized by comprising the following steps:
s1, collecting a sound signal in the operation process of equipment;
s2, carrying out abnormity detection on the sound signal by an abnormity detection model, and inputting an abnormal signal to a fault diagnosis model when the abnormal signal is detected;
and S3, carrying out fault diagnosis on the abnormal signal by the fault diagnosis model.
2. A power plant equipment state monitoring method based on intelligent perception technology according to claim 1, characterized in that the anomaly detection model is a single-class support vector machine, and the anomaly detection model is obtained by training in advance in the following way:
A1. acquiring first sample data;
A2. preprocessing the sound signal in the first sample data;
A3. extracting the characteristics of the preprocessed sound signals to form characteristic vectors;
A4. carrying out dimension reduction processing on the feature vector;
A5. establishing an anomaly detection model by utilizing a feature training single-class support vector machine subjected to dimension reduction;
in step S2, the acquired sound signal is subjected to data processing including preprocessing, feature extraction, and dimension reduction, and then input to the abnormality detection model to perform abnormality detection on the sound signal by the abnormality detection model.
3. A power plant equipment state monitoring method based on intelligent perception technology according to claim 2, characterized in that the sound signals in the first sample data are all normal signals;
or, the sound signals in the first sample data include normal signals marked with normal labels and abnormal signals marked with abnormal labels.
4. The power plant equipment state monitoring method based on the intelligent perception technology as claimed in claim 2, wherein the step A2 is specifically as follows:
A21. filtering the sound signal, and filtering low-frequency interference signals below 50Hz by using a high-pass filter;
A22. the sound is segmented and a sliding window framing process is used for each segment of sound.
5. The power plant equipment condition monitoring method based on the intelligent perception technology as claimed in claim 4, wherein in the step A3, the extracted features include margin factor, impulse factor, skew factor in time domain, and barycentric frequency, mean square frequency, and frequency feature in frequency spectrum.
6. The power plant equipment state monitoring method based on the intelligent perception technology according to claim 2, wherein in the step A5, an ant colony algorithm is used for parameter optimization to obtain an optimal kernel function parameter and a balance parameter of a single-class support vector machine, the optimal kernel function parameter and the balance parameter are substituted into the support vector machine to obtain a trained anomaly detection model, and an initial problem of an objective function is as follows:
using a gaussian kernel function:
the conditions are satisfied: h (x)i,xj)＝y(xi)Ty(xj) (3)
(1) The lagrangian dual problem of formula (la) is:
solve out each lambdaiThen, the discriminant function is obtained as:
(5) in the formula
In the above formulas, l is the number of training samples;
sigma is the required optimal kernel function parameter;
ν epsilon (0, 1) is a balance parameter, is a percentage parameter estimation which is predefined through the ant colony algorithm and represents a compromise between a support vector and a wrong component;
rho is compensation of a required hyperplane in the feature space;
omega is a normal vector of a required hyperplane in a feature space;
ξiis a relaxation variable;
λi、λjis pulling aGlanz multiplier, xi、xjIs the sample in the original space.
7. The power plant equipment state monitoring method based on the intelligent perception technology as claimed in any one of claims 1-6, wherein in step S3, the fault diagnosis model is used for outputting fault type and fault probability value judgment of the corresponding fault type;
and step S3 is followed by:
s4, judging whether the fault probability value is higher than a preset probability value, if so, directly outputting a fault type, otherwise, executing a step S5;
and S5, outputting the judged fault type and the fault probability value, and waiting for a worker to give a manual calibration result.
8. The power plant equipment state monitoring method based on the intelligent perception technology as claimed in claim 7, wherein in step S5, after the staff member gives the artificial calibration result, the calibration result is used as the given label of the corresponding sound signal, and the sound signal with the given label is input to the abnormality detection model and/or the fault diagnosis model for training and updating.
9. A power plant equipment state monitoring method based on intelligent sensing technology according to any one of claims 1-6, characterized in that in step S1, sound signals during the operation of the equipment are collected at a first collection frequency;
in step S2, when an abnormal signal is detected, acquiring a sound signal during the operation of the device at a second acquisition frequency, and simultaneously inputting the abnormal signal and the sound signal acquired at the second acquisition frequency to the fault diagnosis model;
the second acquisition frequency is greater than the first acquisition frequency.
10. A power plant equipment state monitoring method based on smart sensing technology according to claim 9, characterized in that in step S2, the abnormality detection model simultaneously detects the sound signals during the operation of the second collection frequency collection equipment, and when the sound signals lasting for the preset times/time are all detected as normal sound signals, the first collection frequency is recovered.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110889067.5A CN113657221A (en) | 2021-08-04 | 2021-08-04 | Power plant equipment state monitoring method based on intelligent sensing technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110889067.5A CN113657221A (en) | 2021-08-04 | 2021-08-04 | Power plant equipment state monitoring method based on intelligent sensing technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113657221A true CN113657221A (en) | 2021-11-16 |
Family
ID=78490281
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110889067.5A Pending CN113657221A (en) | 2021-08-04 | 2021-08-04 | Power plant equipment state monitoring method based on intelligent sensing technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113657221A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114692674A (en) * | 2022-02-21 | 2022-07-01 | 武汉船用电力推进装置研究所（中国船舶重工集团公司第七一二研究所） | Ship propulsion system fault judgment method and system |
| CN115588265A (en) * | 2022-12-12 | 2023-01-10 | 华能酒泉风电有限责任公司 | Intelligent monitoring system of wind power plant |
| CN116384980A (en) * | 2023-05-25 | 2023-07-04 | 杭州青橄榄网络技术有限公司 | Repair reporting method and system |
| CN116720073A (en) * | 2023-08-10 | 2023-09-08 | 江苏金恒信息科技股份有限公司 | Abnormality detection extraction method and system based on classifier |
| CN117409816A (en) * | 2023-12-14 | 2024-01-16 | 湖南华夏特变股份有限公司 | Equipment fault detection method and system based on sound signals |
| WO2024045387A1 (en) * | 2022-08-31 | 2024-03-07 | 南方电网调峰调频发电有限公司储能科研院 | "listening-vibration" combined gas leakage monitoring method for sealed device in power plant |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102122133A (en) * | 2011-01-21 | 2011-07-13 | 北京工业大学 | Self-adaption wavelet neural network abnormity detection and fault diagnosis classification system and method |
| CN107505133A (en) * | 2017-08-10 | 2017-12-22 | 滁州学院 | The probability intelligent diagnosing method of rolling bearing fault based on adaptive M RVM |
| US20200251127A1 (en) * | 2019-02-06 | 2020-08-06 | Hitachi, Ltd. | Abnormal sound detection device and abnormal sound detection method |
| CN112087703A (en) * | 2020-08-14 | 2020-12-15 | 陕西千山航空电子有限责任公司 | Audio acquisition fault diagnosis method for audio monitor |
| CN112233695A (en) * | 2020-10-13 | 2021-01-15 | 张鹏 | Oiling machine abnormal sound analysis and fault early warning system based on artificial intelligence and big data |
-
2021
- 2021-08-04 CN CN202110889067.5A patent/CN113657221A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102122133A (en) * | 2011-01-21 | 2011-07-13 | 北京工业大学 | Self-adaption wavelet neural network abnormity detection and fault diagnosis classification system and method |
| CN107505133A (en) * | 2017-08-10 | 2017-12-22 | 滁州学院 | The probability intelligent diagnosing method of rolling bearing fault based on adaptive M RVM |
| US20200251127A1 (en) * | 2019-02-06 | 2020-08-06 | Hitachi, Ltd. | Abnormal sound detection device and abnormal sound detection method |
| CN112087703A (en) * | 2020-08-14 | 2020-12-15 | 陕西千山航空电子有限责任公司 | Audio acquisition fault diagnosis method for audio monitor |
| CN112233695A (en) * | 2020-10-13 | 2021-01-15 | 张鹏 | Oiling machine abnormal sound analysis and fault early warning system based on artificial intelligence and big data |
Non-Patent Citations (2)
| Title |
|---|
| 向昌盛;周子英;: "支持向量分类机的参数选择方法研究", 计算机技术与发展, no. 09 * |
| 庄严;白振林;许云峰;: "基于蚁群算法的支持向量机参数选择方法研究", 计算机仿真, no. 05 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114692674A (en) * | 2022-02-21 | 2022-07-01 | 武汉船用电力推进装置研究所（中国船舶重工集团公司第七一二研究所） | Ship propulsion system fault judgment method and system |
| WO2024045387A1 (en) * | 2022-08-31 | 2024-03-07 | 南方电网调峰调频发电有限公司储能科研院 | "listening-vibration" combined gas leakage monitoring method for sealed device in power plant |
| CN115588265A (en) * | 2022-12-12 | 2023-01-10 | 华能酒泉风电有限责任公司 | Intelligent monitoring system of wind power plant |
| CN115588265B (en) * | 2022-12-12 | 2023-04-07 | 华能酒泉风电有限责任公司 | Intelligent monitoring system of wind power plant |
| CN116384980A (en) * | 2023-05-25 | 2023-07-04 | 杭州青橄榄网络技术有限公司 | Repair reporting method and system |
| CN116384980B (en) * | 2023-05-25 | 2023-08-25 | 杭州青橄榄网络技术有限公司 | Repair reporting method and system |
| CN116720073A (en) * | 2023-08-10 | 2023-09-08 | 江苏金恒信息科技股份有限公司 | Abnormality detection extraction method and system based on classifier |
| CN117409816A (en) * | 2023-12-14 | 2024-01-16 | 湖南华夏特变股份有限公司 | Equipment fault detection method and system based on sound signals |
| CN117409816B (en) * | 2023-12-14 | 2024-03-26 | 湖南华夏特变股份有限公司 | Equipment fault detection method and system based on sound signals |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113657221A (en) | Power plant equipment state monitoring method based on intelligent sensing technology | |
| CN113400652B (en) | 3D printer monitoring and diagnosis knowledge base device and system based on vibration signals | |
| CN107436597B (en) | A kind of chemical process fault detection method based on sparse filtering and logistic regression | |
| CN108444708A (en) | The method for building up of rolling bearing intelligent diagnostics model based on convolutional neural networks | |
| Yan et al. | Fisher’s discriminant ratio based health indicator for locating informative frequency bands for machine performance degradation assessment | |
| CN113305645B (en) | Numerical control machine tool cutter residual life prediction method based on hybrid neural model | |
| CN101718634A (en) | Equipment state comprehensive dynamic alarming method based on multivariate probability model | |
| KR102427205B1 (en) | Apparatus and method for generating training data of artificial intelligence model | |
| CN111665066B (en) | Equipment fault self-adaptive upper and lower early warning boundary generation method based on convolutional neural network | |
| CN111964909A (en) | Rolling bearing operation state detection method, fault diagnosis method and system | |
| CN112660745A (en) | Intelligent diagnosis method and system for carrier roller fault and readable storage medium | |
| CN107844067A (en) | A kind of gate of hydropower station on-line condition monitoring control method and monitoring system | |
| CN112173636B (en) | Method for detecting faults of belt conveyor carrier roller by inspection robot | |
| CN105626502A (en) | Plunger pump health assessment method based on wavelet packet and Laplacian Eigenmap | |
| CN111678699B (en) | Early fault monitoring and diagnosing method and system for rolling bearing | |
| CN113177646A (en) | Power distribution equipment online monitoring method and system based on self-adaptive edge proxy | |
| CN114118219A (en) | Data-driven real-time abnormal detection method for health state of long-term power-on equipment | |
| CN109598309B (en) | Detection system and monitoring method of metal packaging punching machine | |
| CN116415127A (en) | Method, system and medium for paper quality assessment | |
| CN115496108A (en) | Fault monitoring method and system based on manifold learning and big data analysis | |
| KR20210050274A (en) | Equipment failure prediction system and method | |
| KR20210006832A (en) | Method and apparatus for machine fault diagnosis | |
| CN116720073A (en) | Abnormality detection extraction method and system based on classifier | |
| CN109115271B (en) | Digit control machine tool remote monitoring system | |
| CN115310497A (en) | Abnormity identification method for bearing of numerical control machine tool |
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 |