CN114217147A - Acoustic fingerprint early warning device for large power transformer - Google Patents
Acoustic fingerprint early warning device for large power transformer Download PDFInfo
- Publication number
- CN114217147A CN114217147A CN202111343854.6A CN202111343854A CN114217147A CN 114217147 A CN114217147 A CN 114217147A CN 202111343854 A CN202111343854 A CN 202111343854A CN 114217147 A CN114217147 A CN 114217147A
- Authority
- CN
- China
- Prior art keywords
- noise
- transformer
- fingerprint
- state
- measuring point
- 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
- 230000005236 sound signal Effects 0.000 claims abstract description 52
- 238000012544 monitoring process Methods 0.000 claims abstract description 33
- 238000012545 processing Methods 0.000 claims abstract description 19
- 238000000926 separation method Methods 0.000 claims abstract description 13
- 239000011159 matrix material Substances 0.000 claims abstract description 11
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 9
- 238000004458 analytical method Methods 0.000 claims abstract description 6
- 238000001514 detection method Methods 0.000 claims abstract description 5
- 230000002401 inhibitory effect Effects 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 22
- 238000009826 distribution Methods 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 14
- 238000010276 construction Methods 0.000 claims description 10
- 238000001228 spectrum Methods 0.000 claims description 8
- 239000013598 vector Substances 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000003745 diagnosis Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 238000011161 development Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000000007 visual effect Effects 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000035772 mutation Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 abstract description 4
- 238000004804 winding Methods 0.000 description 7
- 230000006872 improvement Effects 0.000 description 6
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 239000003921 oil Substances 0.000 description 3
- 238000005259 measurement Methods 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 229910000976 Electrical steel Inorganic materials 0.000 description 1
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000005389 magnetism Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 239000000725 suspension Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R29/00—Arrangements for measuring or indicating electric quantities not covered by groups G01R19/00 - G01R27/00
- G01R29/26—Measuring noise figure; Measuring signal-to-noise ratio
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses an acoustic fingerprint early warning device for a large power transformer, wherein an acoustic signal detection module is used for detecting the running state of the transformer, monitoring variables of a fault physics model and comparing the variables with calculated values of the model to obtain sound data of the power transformer, an acoustic signal analysis module is used for analyzing the statistical regularity characteristics of target source signals, separating the collected sound data into source signals, summarizing and gradually finding an optimal separation matrix in the separation process, enabling the separated signals to be independent mutually so as to extract the target sound signals in a noise environment, a data processing module is used for adding a white noise sequence for a limited time into the signals, inhibiting modal aliasing of EMD (empirical mode decomposition) and eliminating white noise by solving the mean value of the decomposition results for the limited time, a fingerprint model dynamically extracts monitoring data in the monitoring process to form a transformer noise fingerprint of a certain period, the noise fingerprint is visually displayed through the three-dimensional model and the pseudo color chart, so that the equipment state can be conveniently and visually analyzed.
Description
Technical Field
The invention relates to the field of transformer monitoring, in particular to an acoustic fingerprint early warning device for a large power transformer.
Background
At present, in a plurality of transformer monitoring means, a non-electric quantity monitoring method represented by noise, temperature and vibration is widely applied because the non-electric quantity monitoring method does not generate direct electric or magnetic connection with tested equipment and has little influence on the safe operation of a power system. The generation of sound is caused by vibration of particles, and the main source of noise of the transformer is vibration of an oil tank, a winding, an iron core, a cooling device fan and the like, so that the insulation condition in the transformer can be judged by detecting the vibration of a transformer box body.
Generally, a transformer operates in an alternating electromagnetic environment, periodic changes of an electric field and a magnetic field can cause periodic vibration of an iron core and a winding of the transformer, the internal vibration is transmitted to a transformer oil tank through multiple paths, and finally the vibration of the oil tank is represented as noise.
Disclosure of Invention
In order to solve the technical problems, the invention provides an acoustic fingerprint early warning device for a large power transformer, which can reasonably utilize noise signals generated during the operation of the power transformer and analyze the noise signals to achieve the purposes of real-time online monitoring, fault diagnosis and early warning of the transformer. The monitoring result is displayed by applying an acoustic principle, friendly storage and interaction strategies, the running state of the transformer can be comprehensively mastered, the equipment management efficiency is improved, and the method has strong guidance for timely finding the hidden trouble of equipment and carrying out targeted maintenance.
The invention provides an acoustic fingerprint early warning device for a large power transformer, which comprises:
the sound signal detection module is used for detecting the running state of the transformer, monitoring the variable of the fault physics model and comparing the variable with the calculated value of the model to obtain the sound data of the power transformer;
the acoustic signal analysis module is used for analyzing the statistical rule characteristics of the target source signal, separating the source signal from the acquired acoustic data, summarizing and gradually searching for an optimal separation matrix in the separation process, and enabling the separated signals to be mutually independent so as to extract the target acoustic signal in a noise environment;
the data processing module is used for adding a white noise sequence of limited times into the signal, inhibiting modal aliasing of EMD and eliminating the white noise by averaging the decomposition results of the limited times;
the characteristic extraction module is used for carrying out FastICA separation on a target sound signal, separating a plurality of sound signals to be processed, respectively carrying out CEEMDAN decomposition on the separated sound signals to be processed, decomposing the sound signals to be processed into a plurality of IMFs by the CEEMDAN, combining the IMFs of all orders to form a total energy-benefiting matrix H, then solving singular values of the total energy-benefiting matrix H, solving the singular values to form a fault sound signal singular spectrum of the power transformer, decomposing the signals into frequency bands with the frequencies from high to low by the CEEMDAN, solving the energy of the IMFs of all orders and solving the energy entropy to obtain the CEEMDAN energy entropy of the fault sound signal of the power transformer, carrying out Hilbert conversion on the IMFs of all orders, solving the marginal spectrum, carrying out normalization processing on the solved entropy values, and finally forming characteristic vectors by the characteristic vectors for classification and identification;
the fingerprint model construction module is used for preprocessing the acquired noise signals to form description parameters and formulating a model construction process, processing the obtained state monitoring data, performing normalization processing and interpolation calculation on the three-dimensional heterogeneous data, and constructing a three-dimensional model and a pseudo color map for describing the state characteristics of the equipment so as to realize visual data display of the multi-dimensional characteristics and the development trend of the monitoring data.
As a further improvement of the technical scheme, the fingerprint model building module comprises an electrical equipment layer, a monitoring object layer and an application layer, a model building rule is formulated according to the combination of expert knowledge, corresponding state monitoring data are retrieved according to the rule to build a noise fingerprint of the transformer, and the application layer is divided into global noise and measuring point noise and is respectively used for integrating noise state description of the transformer and local noise state description of a noise measuring point of the transformer;
the noise fingerprints give independent diagnosis results of the noise state of the transformer so as to realize expression description or comparison explanation of noise monitoring information.
As a further improvement of the technical scheme, the measuring point noise fingerprint is used for describing the local noise state of each measuring point of the transformer, the noise state is described by selecting a measuring point noise main frequency, a measuring point noise main frequency amplitude, a measuring point noise secondary frequency amplitude, a measuring point noise main frequency multiplication amplitude, a measuring point noise secondary frequency multiplication and a measuring point noise secondary frequency multiplication amplitude, and selected characteristic indexes are used for finely dividing the components of the noise signal when the noise signal intensity of each measuring point is reflected, wherein a pseudo color map is presented in a two-dimensional form, and image recognition is carried out according to the local shape characteristic, the color characteristic and the texture characteristic of the pseudo color map so as to evaluate the noise state of the transformer.
As a further improvement of the technical scheme, the global noise fingerprint is used for describing the macroscopic noise state of the whole transformer and is used for describing the noise state of the whole transformer by the global noise fingerprint, and the whole noise sound pressure mean value, the whole noise sound pressure level maximum value and the whole noise sound power level are selected to describe the noise state of the whole transformer, wherein the pseudo-color chart is presented in a two-dimensional form, and image recognition is carried out according to the local shape characteristic, the color characteristic and the texture characteristic of the pseudo-color chart to evaluate the noise state of the transformer.
As a further improvement of the technical scheme, the fingerprint model building module further comprises a noise fingerprint comparison unit, and the noise fingerprint comparison unit is used for comparing noise fingerprint descriptions of different working conditions and different measuring points by storing the noise fingerprint and connecting the noise fingerprint with the actual working condition of the transformer, and counting and inducing the change rule of the noise fingerprint under the normal operation and fault state of the transformer so as to obtain a preliminary state judgment basis.
As a further improvement of the technical scheme, the noise fingerprint comparison unit comprises a normal-abnormal noise fingerprint comparison process and a measuring point-measuring point noise fingerprint comparison process, and the noise fingerprint in the normal-abnormal noise fingerprint comparison process has the following characteristics:
the main frequency of the measured point noise and the main frequency multiplication distribution have consistency, and the current transformer noise source is mainly alternating current, magnetic field change and vibration or rotation of practical commercial power equipment;
the noise signal content with the secondary frequency and the secondary frequency multiplication as the central frequency has obvious difference compared with the main frequency and the main frequency multiplication;
the noise loudness level of the transformer is stable, and no obvious mutation point appears in the noise fingerprint of a preset scale.
As a further improvement of the above technical solution, the early warning device further includes:
and the sound signal classification module is used for classifying and identifying fault sound signals of the transformer in a fault sound signal model training stage and a to-be-detected sound signal identification stage, acquiring various fault sound signals by simulating various faults of the distribution transformer and actually measuring various fault sound signals on site, extracting characteristic values of the fault sound signals in the characteristic value extraction process for SVDD model training, and identifying the to-be-detected sound signals by adopting a trained SVDD distribution transformer fault noise signal model.
The invention provides an acoustic fingerprint early warning device for a large power transformer, which has the following beneficial effects compared with the prior art:
the device carries out fault diagnosis on equipment by arranging the acoustic signal detection module, the acoustic signal analysis module, the data processing module and the fingerprint model construction module under the condition of environmental interference, and for the non-stationarity problem that an interference sound source and a fault acoustic signal exist around a distribution transformer, the body area and the processing of the acoustic signal are analyzed according to a quick independent component, and the device has the characteristic of quickly separating the independent sound source and adopts a CEEMDAN algorithm to process the non-stationary acoustic signal with a body area target sound source. The problem that modal aliasing and Gaussian white noise introduced by adopting an EEMD algorithm are difficult to completely eliminate is solved by a self-adaptive white noise adding method, signal reconstruction is good, and technical support is provided for extracting and processing fault sound signals of the distribution transformer by extracting and processing the sound signals. The transformer state online monitoring can effectively avoid the reduction of the equipment operation reliability and the economic loss caused by excessive maintenance or loss of equipment due to regular maintenance, a plurality of valuable description parameters can be obtained by carrying out acoustic calculation on the collected noise original data, and the description parameters of the lock are summarized according to rules, so that the noise state description of a certain transformer at a certain moment can be obtained.
According to the basic principle of acoustic calculation and the general characteristics of transformer noise, a plurality of acoustic parameters for describing the transformer noise are determined, the acoustic parameters include sound pressure level and sound power level parameters which are common in the existing noise monitoring, and main frequency, secondary frequency, main frequency multiplication, secondary frequency multiplication and corresponding amplitude parameters of the noise in frequency domain analysis are specifically refined, the unexpected overtone of fundamental tones in the composite noise is enhanced, and the method is suitable for describing the noise caused by alternating current and magnetic fields. The fingerprint model dynamically extracts monitoring data in the monitoring process to form a transformer noise fingerprint of a certain time period, the noise fingerprint is visually displayed through the three-dimensional model and the pseudo color chart, the equipment state can be conveniently and visually analyzed, and the means for distinguishing the transformer states is enriched by a noise comparison sample formed by special fingerprints.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
Fig. 1 is a block diagram of an early warning apparatus according to an embodiment of the present invention;
FIG. 2 is a flow chart of feature extraction provided by an embodiment of the present invention;
fig. 3 is a flowchart of fingerprint model construction according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention.
It will be understood that when an element is referred to as being "secured to" another element, it can be directly on the other element or intervening elements may also be present. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present. In contrast, when an element is referred to as being "directly on" another element, there are no intervening elements present. The terms "vertical," "horizontal," "left," "right," and the like as used herein are for illustrative purposes only.
In the present invention, unless otherwise expressly stated or limited, the terms "mounted," "connected," "secured," and the like are to be construed broadly and can, for example, be fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; either directly or indirectly through intervening media, either internally or in any other relationship. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
Referring to fig. 1, 2 and 3, the invention provides an acoustic fingerprint early warning device for a large power transformer, which includes:
the sound signal detection module is used for detecting the running state of the transformer, monitoring the variable of the fault physics model and comparing the variable with the calculated value of the model to obtain the sound data of the power transformer;
the acoustic signal analysis module is used for analyzing the statistical rule characteristics of the target source signal, separating the source signal from the acquired acoustic data, summarizing and gradually searching for an optimal separation matrix in the separation process, and enabling the separated signals to be mutually independent so as to extract the target acoustic signal in a noise environment;
the data processing module is used for adding a white noise sequence of limited times into the signal, inhibiting modal aliasing of EMD and eliminating the white noise by averaging the decomposition results of the limited times;
the characteristic extraction module is used for carrying out FastICA separation on a target sound signal, separating a plurality of sound signals to be processed, respectively carrying out CEEMDAN decomposition on the separated sound signals to be processed, decomposing the sound signals to be processed into a plurality of IMFs by the CEEMDAN, combining the IMFs of all orders to form a total energy-benefiting matrix H, then solving singular values of the total energy-benefiting matrix H, solving the singular values to form a fault sound signal singular spectrum of the power transformer, decomposing the signals into frequency bands with the frequencies from high to low by the CEEMDAN, solving the energy of the IMFs of all orders and solving the energy entropy to obtain the CEEMDAN energy entropy of the fault sound signal of the power transformer, carrying out Hilbert conversion on the IMFs of all orders, solving the marginal spectrum, carrying out normalization processing on the solved entropy values, and finally forming characteristic vectors by the characteristic vectors for classification and identification;
the fingerprint model construction module is used for preprocessing the acquired noise signals to form description parameters and formulating a model construction process, processing the obtained state monitoring data, performing normalization processing and interpolation calculation on the three-dimensional heterogeneous data, and constructing a three-dimensional model and a pseudo color map for describing the state characteristics of the equipment so as to realize visual data display of the multi-dimensional characteristics and the development trend of the monitoring data.
In this embodiment, distribution transformer's body noise signal mainly because the vibration of iron core and winding produces, distribution transformer iron core receives the effect of magnetostriction and electromagnetic force and produces the vibration to produce the noise signal, the magnetostriction power takes the lead, the factor that influences iron core noise signal includes temperature, magnetism steel sheet coating, silicon steel sheet material etc. distribution transformer winding receives the effect of electromagnetic force vibrations and produces the acoustic signal, when the winding becomes flexible the back, the pretightning force reduces, winding natural frequency is close the electric wire netting frequency doubling and can make the winding vibration obviously increase.
It should be noted that, in order to effectively extract the distribution transformer fault acoustic signal, a relatively high-efficiency blind source separation algorithm FastICA is selected to extract the distribution transformer fault acoustic signal, and the fault acoustic signal is processed by adopting a CEEMDAN method which can adapt to non-stationary signal characteristics and has a small reconstruction error. The method for extracting the fault acoustic signal characteristics of the distribution transformer comprises the following specific steps:
s10: performing FastICA separation on the acquired distribution transformer acoustic signals to separate a plurality of acoustic signals to be processed;
s11: respectively carrying out CEEMDAN decomposition on the separated acoustic signals to be processed;
s12: CEEMDAN decomposes the acoustic signal to be processed into a plurality of IMFs, combines the IMFs of each order to form a very-high-energy matrix H,
s13: then solving singular values, wherein the solved singular values form a singular spectrum of a fault sound signal of the power transformer;
s14: the CEEMDAN decomposes the signal into frequency bands with frequencies from high to low, and energy of each order of IMF is obtained and energy entropy is obtained to obtain the CEEMDAN energy entropy of the fault sound signal of the power transformer;
s15: carrying out Hilbert transformation on each order of IMF to obtain a marginal spectrum;
s16: normalizing the obtained entropy values;
s17: and finally, forming a characteristic vector by using the characteristic vector for classification and identification.
The process of constructing the transformer noise fingerprint model comprises the following steps:
s20: preprocessing the collected noise signals to form description parameters, making a model structure, and processing the obtained state monitoring data;
s21: carrying out normalization processing and interpolation calculation on the three-dimensional heterogeneous data;
s22: and constructing a three-dimensional model and a pseudo color chart describing the state characteristics of the equipment so as to realize visual data display of the multi-dimensional characteristics and the development trend of the monitoring data.
Optionally, the fingerprint model building module comprises an electrical equipment layer, a monitoring object layer and an application layer, a model building rule is formulated according to the combination of expert knowledge, corresponding state monitoring data are retrieved according to the rule to build a transformer noise fingerprint, and the application layer is divided into global noise and measuring point noise and is respectively used for integrating the noise state description of the transformer and the local noise state description of the transformer noise measuring point;
the noise fingerprints give independent diagnosis results of the noise state of the transformer so as to realize expression description or comparison explanation of noise monitoring information.
In the embodiment, the measuring point noise fingerprint is used for describing the local noise state of each measuring point of the transformer, the measuring point noise dominant frequency amplitude, the measuring point noise secondary frequency amplitude, the measuring point noise dominant frequency multiplication amplitude, the measuring point noise secondary frequency multiplication and the measuring point noise secondary frequency multiplication amplitude are selected to describe the noise state, selected characteristic indexes are used for finely dividing the components of the noise signal when the noise signal intensity of each measuring point is reflected, wherein a pseudo color map is presented in a two-dimensional form, and image recognition is carried out according to the local shape characteristic, the color characteristic and the texture characteristic of the pseudo color map so as to evaluate the noise state of the transformer.
It should be noted that the global noise fingerprint is used for describing a macroscopic noise state of the whole transformer and describing the noise state of the whole transformer by the global noise fingerprint, and the whole noise sound pressure mean value, the whole noise sound pressure level maximum value and the whole noise sound power level are selected to describe the noise state of the whole transformer, wherein the pseudo color chart is presented in a two-dimensional form, and image recognition is performed according to the local shape feature, the color feature and the texture feature of the pseudo color chart to evaluate the noise state of the transformer. The individual devices are distinguished from each other due to differences in manufacturing processes, operating conditions and levels of life cycles, but at certain points in time, the description is unique and can be regarded as a fingerprint of transformer noise information, which is defined as a noise fingerprint, like a human fingerprint. When the selected characteristic indexes reflect the noise signal intensity of each measuring point, the components of the noise signals are divided more finely, and the identification capability of different noise states is improved.
Optionally, the fingerprint model building module further includes a noise fingerprint comparison unit, and compares noise fingerprints at different working conditions and different measuring point noise fingerprint descriptions by storing the noise fingerprints and associating the noise fingerprints with the actual working conditions of the transformer, and counts and summarizes a change rule of the noise fingerprints in normal operation and fault states of the transformer to obtain a preliminary state judgment basis.
In this embodiment, the noise fingerprint comparison unit includes a normal-abnormal noise fingerprint comparison process and a measurement point-measurement point noise fingerprint comparison process, and the noise fingerprint in the normal-abnormal noise fingerprint comparison process has the following characteristics: the main frequency of the measured point noise and the main frequency multiplication distribution have consistency, and the current transformer noise source is mainly alternating current, magnetic field change and vibration or rotation of practical commercial power equipment; the noise signal content with the secondary frequency and the secondary frequency multiplication as the central frequency has obvious difference compared with the main frequency and the main frequency multiplication; the noise loudness level of the transformer is stable, and no obvious mutation point appears in the noise fingerprint of a preset scale. The primary frequency, the secondary frequency, the primary frequency multiplication and the secondary frequency multiplication of a typical noise signal of the transformer are all below 1000Hz, and the noise above 1000Hz has the characteristics of small occurrence probability and low signal amplitude. The meaning of the secondary index value of the measuring point noise fingerprint is the amplitude of the noise description parameter, namely the primary frequency amplitude, the secondary frequency amplitude, the primary frequency multiplication amplitude and the secondary frequency multiplication amplitude, the measuring point noise description parameter value for 15 minutes is extracted to carry out four sample application strip function interpolation, and the obtained measuring point noise fingerprint three-dimensional model can visually present the noise characteristics of the measuring point of the transformer and the change trend of the measuring point noise fingerprint within 15 minutes.
Optionally, the early warning device further includes:
and the sound signal classification module is used for classifying and identifying fault sound signals of the transformer in a fault sound signal model training stage and a to-be-detected sound signal identification stage, acquiring various fault sound signals by simulating various faults of the distribution transformer and actually measuring various fault sound signals on site, extracting characteristic values of the fault sound signals in the characteristic value extraction process for SVDD model training, and identifying the to-be-detected sound signals by adopting a trained SVDD distribution transformer fault noise signal model.
In the embodiment, three common discharge faults of the transformer, namely needle plate discharge, suspension discharge and creeping discharge, are simulated, the discharge sound signal characteristics of the transformer are extracted, different types of discharge faults are identified through SVDD, and the result shows that the FastICA algorithm can well extract the discharge fault sound signal, different discharge faults are identified through the SVDD classification method, the identification effect is good, and the accuracy and the efficiency of transformer fault diagnosis can be improved through the effectiveness of identification of the discharge faults of the transformer.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.
The above examples are merely illustrative of several embodiments of the present invention, and the description thereof is more specific and detailed, but not to be construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention.
Claims (7)
1. The utility model provides a large-scale power transformer acoustics fingerprint early warning device which characterized in that, early warning device includes:
the sound signal detection module is used for detecting the running state of the transformer, monitoring the variable of the fault physics model and comparing the variable with the calculated value of the model to obtain the sound data of the power transformer;
the acoustic signal analysis module is used for analyzing the statistical rule characteristics of the target source signal, separating the source signal from the acquired acoustic data, summarizing and gradually searching for an optimal separation matrix in the separation process, and enabling the separated signals to be mutually independent so as to extract the target acoustic signal in a noise environment;
the data processing module is used for adding a white noise sequence of limited times into the signal, inhibiting modal aliasing of EMD and eliminating the white noise by averaging the decomposition results of the limited times;
the characteristic extraction module is used for carrying out FastICA separation on a target sound signal, separating a plurality of sound signals to be processed, respectively carrying out CEEMDAN decomposition on the separated sound signals to be processed, decomposing the sound signals to be processed into a plurality of IMFs by the CEEMDAN, combining the IMFs of all orders to form a total energy-benefiting matrix H, then solving singular values of the total energy-benefiting matrix H, solving the singular values to form a fault sound signal singular spectrum of the power transformer, decomposing the signals into frequency bands with the frequencies from high to low by the CEEMDAN, solving the energy of the IMFs of all orders and solving the energy entropy to obtain the CEEMDAN energy entropy of the fault sound signal of the power transformer, carrying out Hilbert conversion on the IMFs of all orders, solving the marginal spectrum, carrying out normalization processing on the solved entropy values, and finally forming characteristic vectors by the characteristic vectors for classification and identification;
the fingerprint model construction module is used for preprocessing the acquired noise signals to form description parameters and formulating a model construction process, processing the obtained state monitoring data, performing normalization processing and interpolation calculation on the three-dimensional heterogeneous data, and constructing a three-dimensional model and a pseudo color map for describing the state characteristics of the equipment so as to realize visual data display of the multi-dimensional characteristics and the development trend of the monitoring data.
2. The acoustic fingerprint early warning device for the large-scale power transformer as claimed in claim 1, wherein the fingerprint model construction module comprises an electrical equipment layer, a monitoring object layer and an application layer, a model construction rule is formulated according to the combination of expert knowledge, corresponding state monitoring data are retrieved through the rule to construct a transformer noise fingerprint, and the application layer is divided into global noise and measuring point noise and is respectively used for integrating noise state description of the transformer and local noise state description of a transformer noise measuring point;
the noise fingerprints give independent diagnosis results of the noise state of the transformer so as to realize expression description or comparison explanation of noise monitoring information.
3. The acoustic fingerprint early warning device for the large-scale power transformer as claimed in claim 2, wherein the measuring point noise fingerprint is used for describing the local noise state of each measuring point of the transformer, the noise state is described by selecting the measuring point noise dominant frequency, the measuring point noise dominant frequency amplitude, the measuring point noise secondary frequency amplitude, the measuring point noise dominant frequency multiplication amplitude, the measuring point noise secondary frequency multiplication and the measuring point noise secondary frequency multiplication amplitude, and the selected characteristic indexes are used for finely dividing the components of the noise signal when the noise signal intensity of each measuring point is reflected, wherein the pseudo color chart is presented in a two-dimensional form, and image recognition is carried out according to the local shape characteristic, the color characteristic and the texture characteristic of the pseudo color chart so as to evaluate the noise state of the transformer.
4. The acoustic fingerprint early warning device for the large power transformer as recited in claim 2, wherein the global noise fingerprint is used for describing a macroscopic noise state of the whole transformer and describing a noise state of the whole transformer by the global noise fingerprint, and a whole machine noise sound pressure mean value, a whole machine noise sound pressure level maximum value and a whole machine noise sound power level are selected to describe the noise state of the whole transformer, wherein the pseudo-color chart is presented in a two-dimensional form, and image recognition is performed according to local shape features, color features and texture features of the pseudo-color chart to evaluate the noise state of the transformer.
5. The acoustic fingerprint early warning device for the large-scale power transformer as claimed in claim 2, wherein the fingerprint model building module further comprises a noise fingerprint comparison unit, which compares the noise fingerprint descriptions of different working conditions and different measuring points by storing the noise fingerprint and connecting the noise fingerprint with the actual working condition of the transformer, and counts and summarizes the change rules of the noise fingerprint under the normal operation and fault state of the transformer to obtain a preliminary state judgment basis.
6. The acoustic fingerprint early warning device for the large-scale power transformer as claimed in claim 5, wherein the noise fingerprint comparison unit comprises a normal-abnormal noise fingerprint comparison process and a measuring point-measuring point noise fingerprint comparison process, and the noise fingerprint in the normal-abnormal noise fingerprint comparison process has the following characteristics:
the main frequency of the measured point noise and the main frequency multiplication distribution have consistency, and the current transformer noise source is mainly alternating current, magnetic field change and vibration or rotation of practical commercial power equipment;
the noise signal content with the secondary frequency and the secondary frequency multiplication as the central frequency has obvious difference compared with the main frequency and the main frequency multiplication;
the noise loudness level of the transformer is stable, and no obvious mutation point appears in the noise fingerprint of a preset scale.
7. The acoustic fingerprint early warning device for the large-scale power transformer according to claim 1, wherein the early warning device further comprises:
and the sound signal classification module is used for classifying and identifying fault sound signals of the transformer in a fault sound signal model training stage and a to-be-detected sound signal identification stage, acquiring various fault sound signals by simulating various faults of the distribution transformer and actually measuring various fault sound signals on site, extracting characteristic values of the fault sound signals in the characteristic value extraction process for SVDD model training, and identifying the to-be-detected sound signals by adopting a trained SVDD distribution transformer fault noise signal model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111343854.6A CN114217147A (en) | 2021-11-14 | 2021-11-14 | Acoustic fingerprint early warning device for large power transformer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111343854.6A CN114217147A (en) | 2021-11-14 | 2021-11-14 | Acoustic fingerprint early warning device for large power transformer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114217147A true CN114217147A (en) | 2022-03-22 |
Family
ID=80697088
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111343854.6A Pending CN114217147A (en) | 2021-11-14 | 2021-11-14 | Acoustic fingerprint early warning device for large power transformer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114217147A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118471265A (en) * | 2024-07-09 | 2024-08-09 | 宁波惠康工业科技股份有限公司 | Refrigerator fault acoustic signal recognition system and method based on feature mining |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108362966A (en) * | 2018-02-12 | 2018-08-03 | 广东电网有限责任公司电力科学研究院 | A kind of oil-immersed type transformer high-precision noise on-line monitoring method and system |
| CN113158769A (en) * | 2021-03-03 | 2021-07-23 | 安徽大学 | CEEMDAN and FastICA-based electromechanical device bearing vibration signal denoising method |
-
2021
- 2021-11-14 CN CN202111343854.6A patent/CN114217147A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108362966A (en) * | 2018-02-12 | 2018-08-03 | 广东电网有限责任公司电力科学研究院 | A kind of oil-immersed type transformer high-precision noise on-line monitoring method and system |
| CN113158769A (en) * | 2021-03-03 | 2021-07-23 | 安徽大学 | CEEMDAN and FastICA-based electromechanical device bearing vibration signal denoising method |
Non-Patent Citations (3)
| Title |
|---|
| 周颖涛等人。: "减少模态混叠的改进EEMD算法", 《重庆理工大学学报》, pages 112 - 113 * |
| 梁端丹等人: "独立分量分析及其应用研究", 《现代电子技术》, pages 17 - 19 * |
| 舒畅等。: "基于CEEMDAN的配电变压器放电故障噪声诊断方法", 《高电压技术》, pages 1 - 9 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118471265A (en) * | 2024-07-09 | 2024-08-09 | 宁波惠康工业科技股份有限公司 | Refrigerator fault acoustic signal recognition system and method based on feature mining |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107122790B (en) | Non-invasive load identification algorithm based on hybrid neural network and ensemble learning | |
| CN109974782A (en) | Equipment fault early-warning method and system based on big data sensitive features optimum option | |
| Übeyli | ECG beats classification using multiclass support vector machines with error correcting output codes | |
| CN110490071A (en) | A kind of substation's Abstraction of Sound Signal Characteristics based on MFCC | |
| CN110534118A (en) | Transformer/reactor method for diagnosing faults based on Application on Voiceprint Recognition and neural network | |
| CN105275833A (en) | CEEMD (Complementary Empirical Mode Decomposition)-STFT (Short-Time Fourier Transform) time-frequency information entropy and multi-SVM (Support Vector Machine) based fault diagnosis method for centrifugal pump | |
| CN106934421A (en) | Converting station electric power transformer fault detecting system and detection method based on 2DPCA and SVM | |
| CN113255777B (en) | Equipment fault early warning method and system based on multi-mode sensitive feature selection fusion | |
| Hu et al. | Copula-based statistical health grade system against mechanical faults of power transformers | |
| CN105411565A (en) | Heart rate variability feature classification method based on generalized scale wavelet entropy | |
| CN105320969A (en) | A heart rate variability feature classification method based on multi-scale Renyi entropy | |
| CN115809407A (en) | Breaker fault diagnosis method and system based on vibration characteristic analysis | |
| Esfahanian et al. | On contour-based classification of dolphin whistles by type | |
| CN113283310A (en) | System and method for detecting health state of power equipment based on voiceprint features | |
| Zhao et al. | A novel deep fuzzy clustering neural network model and its application in rolling bearing fault recognition | |
| CN118244110B (en) | Industrial robot motor fault early warning method and system | |
| CN114217147A (en) | Acoustic fingerprint early warning device for large power transformer | |
| Gong et al. | An integrated health condition detection method for rolling bearings using time-shift multi-scale amplitude-aware permutation entropy and uniform phase empirical mode decomposition | |
| Li et al. | The Application of AE Signal in Early Cracked Rotor Fault Diagnosis with PWVD and SVM. | |
| CN113221937A (en) | Emergency processing system and method based on artificial intelligence judgment | |
| Chen et al. | Gear fault diagnosis model combined with MED-LMD-Hypersphere multiclass SVM | |
| CN115455684A (en) | Oil-immersed transformer non-contact state evaluation method and system based on noise analysis | |
| CN115223576A (en) | Method and system for extracting and identifying controllable precision of transformer voiceprint characteristics based on MFCC | |
| CN112731208B (en) | Low-voltage line fault and abnormity on-line monitoring method, equipment and medium | |
| CN112417966B (en) | Multi-dimensional machine fault detection method based on audio signal characteristics and AI algorithm |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220322 |