CN106680371A - Electric porcelain body online flaw detection device and electric porcelain body online flaw detection method - Google Patents

Electric porcelain body online flaw detection device and electric porcelain body online flaw detection method Download PDF

Info

Publication number
CN106680371A
CN106680371A CN201611264803.3A CN201611264803A CN106680371A CN 106680371 A CN106680371 A CN 106680371A CN 201611264803 A CN201611264803 A CN 201611264803A CN 106680371 A CN106680371 A CN 106680371A
Authority
CN
China
Prior art keywords
porcelain body
electroceramics
electroceramics porcelain
flaw detection
knock
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201611264803.3A
Other languages
Chinese (zh)
Other versions
CN106680371B (en
Inventor
王莹
周强
陈颖
脱羚
杜晞盟
胡江涛
李烨
李敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shaanxi University of Science and Technology
Original Assignee
Shaanxi University of Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shaanxi University of Science and Technology filed Critical Shaanxi University of Science and Technology
Priority to CN201611264803.3A priority Critical patent/CN106680371B/en
Publication of CN106680371A publication Critical patent/CN106680371A/en
Application granted granted Critical
Publication of CN106680371B publication Critical patent/CN106680371B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/04—Analysing solids
    • G01N29/045—Analysing solids by imparting shocks to the workpiece and detecting the vibrations or the acoustic waves caused by the shocks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
    • G01N29/4454—Signal recognition, e.g. specific values or portions, signal events, signatures
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
    • G01N29/4481—Neural networks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2291/00—Indexing codes associated with group G01N29/00
    • G01N2291/02—Indexing codes associated with the analysed material
    • G01N2291/023—Solids
    • G01N2291/0232—Glass, ceramics, concrete or stone
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2291/00—Indexing codes associated with group G01N29/00
    • G01N2291/02—Indexing codes associated with the analysed material
    • G01N2291/028—Material parameters
    • G01N2291/0289—Internal structure, e.g. defects, grain size, texture
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2291/00—Indexing codes associated with group G01N29/00
    • G01N2291/26—Scanned objects
    • G01N2291/269—Various geometry objects

Abstract

The invention discloses an electric porcelain body online flaw detection device and an electric porcelain body online flaw detection method. The electric porcelain body online flaw detection device comprises a detection unit, a drive unit, a collection unit and a treatment unit, wherein a signal is transmitted to the drive unit and the collection unit when the detection unit detects an electric porcelain body; the electric porcelain body is knocked through the drive unit, a knocking sound signal generated by the electric porcelain body is collected by combining the collection unit; the collection unit converts the collected knocking sound of the electric porcelain body into a digital signal to transmit the digital signal to the treatment unit; and the treatment unit identifies the flaw type of the electric porcelain body by adopting a default algorithm according to the collected knocking sound of the electric porcelain body. The flaw detection in the electric porcelain body can be automatically completed only by arranging the electric porcelain body online flaw detection device on a production line of the electric porcelain body.

Description

A kind of electroceramics porcelain body line flaw detection device and method
Technical field
Electrotechnical ceramics production field of the present invention, more particularly to plant electroceramics porcelain body line flaw detection device and method.
Background technology
At present, during electroceramics porcelain body is widely used in being directly connected to the power system of our daily electricity consumptions and railway transportation, Such as, the transmission line of electricity of different voltage class, transformer station, electric equipment.Electroceramics porcelain body is used as very important in power system One composition part, once there is such as stomata, crackle, long-term use can severely impact the performance of electroceramics and cause power train The paralysis of system, therefore, carrying out flaw detection is carried out to electroceramics porcelain body and is just particularly important in electroceramics porcelain body production process.
Method for detection fault detection main at present is mainly carried out by artificial mode, i.e. the surface by manually visually carrying out Carrying out flaw detection.Although current method can also realize the surface inspection detection of ceramic porcelain body, there is efficiency in this method Lowly, automaticity is low and cannot realize the detection of the latent defect of ceramagnet.
The content of the invention
It is an object of the invention to provide a kind of electroceramics porcelain body line flaw detection measurement apparatus and method, solve existing Electroceramics porcelain body flaw detection inefficiency, automaticity are low and cannot realize the detection of the latent defect of ceramagnet.
In order to achieve the above object, the technical solution adopted by the present invention is as follows：
A kind of electroceramics porcelain body line flaw detection measurement apparatus that the present invention is provided, including detection unit, driver element, collection list Unit and processing unit, wherein, the detection unit is used to detect electroceramics porcelain body and controls driver element to tap detected electricity Porcelain porcelain body；The collector unit is used to collect the knock that driver element taps the generation of electroceramics porcelain body, and the electricity that will be collected into Porcelain porcelain body knock is converted to digital data transmission to the processing unit；The processing unit is according to the electroceramics porcelain for being collected Body knock, the type of impairment of electroceramics porcelain body is recognized using preset algorithm.
Preferably, the detection unit includes optoelectronic switch, and electroceramics porcelain body is detected by optoelectronic switch.
Preferably, the driver element includes tapping subelement and amplification subelement, wherein, the percussion subelement includes Hammer is tapped, it is described to tap the surface that hammer is arranged on electroceramics porcelain body；The amplification subelement includes power amplifier, the power The knock that amplifier will hit against the generation of electroceramics porcelain body is amplified.
Preferably, the collector unit includes sound collecting subelement and sound conversion subunit, wherein, the sound is received Collection subelement includes microphone, and the sound conversion subunit includes sound collection card, and the sound collection card receives microphone The knock for collecting is converted to digital data transmission to processing unit.
Preferably, the processing unit includes five rank butterworth filters and processor, wherein, it is fertile by five rank Barts Hereby wave filter is filtered treatment to electroceramics porcelain body knock, obtains filtering voice signal；The processor is by using default Algorithm carries out processing the type of impairment for obtaining electroceramics porcelain body to the filtering voice signal for obtaining.
A kind of measuring method of electroceramics porcelain body line flaw detection, comprises the following steps：
S1：Electroceramics porcelain body is detected by detection unit, and a knocking is sent to driver element, while single to collecting Unit sends one and collects enable signal；
S2：After the driver element receives knocking, by its internal percussion subelement in default dynamics scope Interior percussion electroceramics porcelain body；
S3：The knock that will hit against the generation of electroceramics porcelain body by the amplification subelement in the driver element is amplified, Electroceramics porcelain body knock is produced to amplify signal；
S4：Electroceramics porcelain body knock is collected by collector unit and amplifies signal, and the electroceramics porcelain body sound is amplified into letter Number digital data transmission is converted to processing unit；
S5：The processing unit is filtered treatment to electroceramics porcelain body knock data signal, obtains filtering voice signal, The filtering acoustic information for obtaining is extracted using the first preset algorithm according to the filtering voice signal, filtering sound letter is obtained Number corresponding flaw detection characteristic information, the damage of electroceramics porcelain body is obtained further according to the flaw detection characteristic information for obtaining using the second preset algorithm Hinder type.
Preferably, first preset algorithm uses fuzzy wavelet algorithm, and it is concretely comprised the following steps：
S101：The morther wavelet in fuzzy wavelet conversion is determined first, and the conversion of yardstick a is carried out to morther wavelet；
S102：Then obscured to the filtering voice signal that obtains after filtering and by the morther wavelet that yardstick a is converted Change is processed, and obtains the first language variable S of the filtering voice signal after Fuzzy processingjIt is corresponding with the first language variable First is subordinate to angle valueAnd the second language variable M of the morther wavelet converted by yardstick aiWith the second language variable pair Second for answering is subordinate to angle value
S103：Then according to predetermined fuzzy matching rule to first language variable SjWith second language variable MiCarry out mould Paste matching, obtain between the filtering voice signal after the Fuzzy processing and the morther wavelet converted by yardstick a With linguistic variable Rij；
S104：Finally according to the first language variable SjCorresponding first is subordinate to angle valueWith the second language variable MiCorresponding second is subordinate to angle valueTo the matching language variable RijSharpening treatment is carried out, filtering voice signal pair is obtained The flaw detection characteristic information T for answeringa。
Preferably, second preset algorithm uses radial base neural net algorithm, the radial base neural net Algorithm is specifically that the corresponding flaw detection characteristic information T of voice signal will be filtered under yardstick aaAs the input quantity of input layer, by Φi(|| Ta-ci| |) as hidden layer activation primitive, using the type of impairment y of electroceramics porcelain body as output layer output quantity, wherein, it is described The non-linear relation of the type of impairment y of electroceramics porcelain body is：
Wherein, | | Ta-ci| | represent TaWith ciBetween Euclidean distance, ciIn representing i-th data of hidden layer neuron Center value, wiIt is output weights.
Preferably, the activation primitive Φ of the hidden layeri(||Ta-ci| |) mathematical definition be：
Wherein, δiIt is i-th extension constant of hidden layer neuron.
Preferably, before step S101, it is necessary first to determine the data center value c of i-th hidden layer neuroni, expand Exhibition constant δiWith output weight wi, realization is fitted to radial base neural net, wherein it is determined that i-th hidden layer neuron Data center value ci, extension constant δiWith output weight wiProcess it is specific as follows：
First, choose equivalent amount without the electroceramics porcelain body for damaging, there is the electroceramics porcelain body of stomata and split in the presence of inside The electroceramics porcelain body of line several, respectively using without the electroceramics porcelain body for damaging as first kind sample；There is the electroceramics porcelain body of stomata As Equations of The Second Kind sample；There is the electroceramics porcelain body of underbead crack as the 3rd class sample；
Second, first kind sample, Equations of The Second Kind sample and the 3rd class sample are obtained respectively by fuzzy wavelet algorithm respectively Corresponding first kind characteristic quantity, Equations of The Second Kind characteristic quantity and the 3rd category feature amount；
3rd, the first kind characteristic quantity of acquisition, Equations of The Second Kind characteristic quantity, the 3rd category feature amount are input into RBFNN respectively, pass through Network training obtains c1,c2,c3...ch、δ1,δ2,δ3...δhAnd w1,w2,w3...wh, complete the training of radial base neural net Process.
Compared with prior art, the beneficial effects of the invention are as follows：
The present invention provide a kind of electroceramics porcelain body line flaw detection measurement apparatus and method, including detection unit, driver element, Collector unit and processing unit, wherein, when detection unit detects electroceramics porcelain body, letter is sent to driver element and collector unit Number；Electroceramics porcelain body is tapped by driver element, the percussion acoustical signal produced by electroceramics porcelain body is collected in conjunction with collector unit, it is described The electroceramics porcelain body knock that collector unit will be collected into is converted to digital data transmission to the processing unit；The processing unit According to the electroceramics porcelain body knock for being collected, the type of impairment of electroceramics porcelain body is recognized using preset algorithm.The structure only needs peace On the production line of electroceramics porcelain body, you can be automatically performed the detection of electroceramics porcelain body latent defect.
Brief description of the drawings
Fig. 1 is measuring system schematic diagram；
Fig. 2 is measuring method flow chart；
Fig. 3 is radial base neural net structure chart；
Wherein, 101, detection unit 102, driver element 103, collector unit 104, processing unit.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, invention is described further, those skilled in the art can have Content disclosed by this specification understands other advantages of the invention and effect easily.The present invention can also be by different in addition Specific embodiment be embodied or practiced, the various details in this specification can also based on different viewpoints with application, Without departing from it is of the invention spirit under carry out various modifications or change.It should be noted that in the case where not conflicting, it is real below Applying the feature in example and embodiment can be combined with each other.
It should be noted that the diagram provided in following examples only schematically illustrates basic structure of the invention Think, so being only shown in associated component in the present invention in diagram rather than component count during according to actual implementation, shape and size Draw, the kenel of each component, quantity and ratio are a kind of random change during its actual implementation, and its assembly layout kenel Can be with increasingly complex.
As shown in figure 1, a kind of electroceramics porcelain body line flaw detection measurement apparatus that the present invention is provided, including detection unit 101, drive Moving cell 102, collector unit 103 and processing unit 104.Wherein, the detection unit 101 is used to detect electroceramics porcelain body and drive Unit 102 taps detected electroceramics porcelain body；The collector unit 103 taps electroceramics porcelain for collecting driver element 102 The knock that body is produced, and the electroceramics porcelain body knock that will be collected into is converted to digital data transmission to the processing unit 104； The processing unit 104 recognizes the type of impairment of electroceramics porcelain body using preset algorithm according to the electroceramics porcelain body knock.
The detection unit 101 includes optoelectronic switch, and electroceramics porcelain body is detected by optoelectronic switch；
The driver element 102 includes tapping subelement and amplification subelement, wherein, the percussion subelement includes tapping Hammer, it is described to tap the surface that hammer is arranged on electroceramics porcelain body；The amplification subelement includes power amplifier, the power amplification The knock that device will hit against the generation of electroceramics porcelain body is amplified；
The collector unit 103 includes sound collecting subelement and sound conversion subunit, wherein, sound collecting Unit includes microphone, and the sound conversion subunit includes sound collection card, and be collected into for microphone by the sound collection card Knock be converted to digital data transmission to processing unit 104；
The processing unit 104 includes five rank butterworth filters and processor, wherein, the processing unit 104 leads to Cross five rank butterworth filters and treatment is filtered to electroceramics porcelain body knock, filtering voice signal is obtained, afterwards by place Reason device carries out processing the type of impairment for obtaining electroceramics porcelain body to the filtering voice signal for obtaining.
A kind of measuring method of electroceramics porcelain body line flaw detection, comprises the following steps：
S1：When the detection unit 101 detects electroceramics porcelain body, the detection unit 101 sends one to driver element 102 Individual knocking, signal is enabled while sending one to collector unit 103 and collecting；
S2：After the driver element 102 receives knocking, by its internal percussion subelement in default dynamics model Enclose interior percussion electroceramics porcelain body；
S3：The knock that will hit against the generation of electroceramics porcelain body by the amplification subelement in the driver element 102 is put Greatly, electroceramics porcelain body knock is produced to amplify signal；
S4：Electroceramics porcelain body knock is collected by the sound collecting subelement in collector unit 103 and amplifies signal；
S5：The electroceramics porcelain body sound is amplified into signal by the sound conversion subunit in collector unit 103 to be converted to Digital data transmission is to processing unit 103；
S6：The processing unit 104 is filtered treatment to electroceramics porcelain body knock data signal, obtains filtering sound letter Number；
S7：The processing unit 104 recognizes the type of impairment of electroceramics porcelain body by preset algorithm.
Specifically, in step S1, the detection unit 101 includes optoelectronic switch, and electroceramics porcelain body is realized by optoelectronic switch Detection.
The knocking and collection enable signal and are high level signal.
Specifically, in step S2, the default dynamics scope is configured according to different electroceramics porcelain bodies.
Specifically, in step S6, the wave filter for using is five rank butterworth filters, and the filtering process refers to filter Except the electroceramics porcelain body sound amplifies the ambient noise signal in signal, to obtain filtering voice signal.
Specifically, in step S7, the preset algorithm includes the first preset algorithm and the second preset algorithm, wherein, it is described First preset algorithm uses fuzzy wavelet algorithm, and second preset algorithm uses radial base neural net algorithm.
After step S6, the filtering acoustic information for obtaining is extracted by the first preset algorithm first, filtered The corresponding flaw detection characteristic information of voice signal, then the type of impairment that electroceramics porcelain body is obtained by the second preset algorithm.
First preset algorithm is concretely comprised the following steps：
S101：The morther wavelet in fuzzy wavelet conversion is determined first, and the conversion of yardstick a is carried out to morther wavelet；
S102：Then obscured to the filtering voice signal that obtains after filtering and by the morther wavelet that yardstick a is converted Change is processed, and obtains the first language variable S of the filtering voice signal after Fuzzy processingjIt is corresponding with the first language variable First is subordinate to angle valueAnd the second language variable M of the morther wavelet converted by yardstick aiWith the second language variable pair Second for answering is subordinate to angle value
S103：Then according to predetermined fuzzy matching rule to first language variable SjWith second language variable MiCarry out mould Paste matching, obtain between the filtering voice signal after the Fuzzy processing and the morther wavelet converted by yardstick a With linguistic variable Rij；
S104：Finally according to the first language variable SjCorresponding first is subordinate to angle valueWith the second language variable MiCorresponding second is subordinate to angle valueTo the matching language variable RijSharpening treatment is carried out, filtering voice signal pair is obtained The flaw detection characteristic information T for answeringa。
For step S101, the morther wavelet specifically refers to Maxican-hat morther wavelets；
Further, the yardstick a conversion, refers specifically to carry out yardstick a conversion and translation transformation to morther wavelet.
For step S103, the predetermined fuzzy matching rule such as table 1 obtains the filtering sound after the Fuzzy processing Matching language variable between message number and the morther wavelet by yardstick a conversion；
Table 1
From table 1 it follows that the predetermined fuzzy matching rule is the matrix of I × J, wherein, I is represented under yardstick a The second language variable number of morther wavelet, J represents the first language variable number of filtering voice signal under yardstick a, SjRepresent chi J-th first language variable of filtering voice signal, M under degree aiI-th second language variable of morther wavelet under yardstick a is represented, RijRepresent MiWith SjBetween matching language variable.
For step S104, the formula of the sharpening treatment is：
In formula (1), TaTo filter the corresponding flaw detection characteristic information of voice signal under yardstick a, a represents the chi of small echo Degree,It is the first language variable S that voice signal is filtered under yardstick ajFirst be subordinate to angle value,It is female small under yardstick a The second language variable M of rippleiSecond be subordinate to angle value,Represent the M under yardstick aiWith SjBetween matching language variable.
As shown in figure 3, second preset algorithm uses radial base neural net (RBFNN, Radical Basis Function Neural network) algorithm, the radial base neural net algorithm is specifically sound letter will to be filtered under yardstick a Number corresponding flaw detection characteristic information TaAs the input quantity of input layer, by Φi(||Ta-ci| |) as hidden layer activation primitive, Using the type of impairment y of electroceramics porcelain body as output layer output quantity, wherein, the type of impairment y's of the electroceramics porcelain body is non-linear Relation is：
Wherein, | | Ta-ci| | represent TaWith ciBetween Euclidean distance, ciIn representing i-th data of hidden layer neuron Center value, wiIt is output weights；The Φi() chooses Gaussian functions, then：
Wherein, δiIt is i-th extension constant of hidden layer neuron, δiIt is smaller, Φi(||Ta-ci| |) width it is smaller.
Radial base neural net puts into the course of work：That is, the corresponding flaw detection characteristic information T of voice signal will be filteredaIt is logical Radial base neural net is crossed, the type of impairment y of corresponding electroceramics porcelain body can be obtained.
The type of impairment y of electroceramics porcelain body is obtained to measure, it is necessary first to before step S101, determine to imply for i-th The data center value c of layer neuroni, extension constant δiWith output weight wi, realize being fitted radial base neural net.
Wherein it is determined that the data center value c of i-th hidden layer neuroni, extension constant δiWith output weight wiProcess It is specific as follows：
Choose first without the electroceramics porcelain body 20 for damaging, as first kind sample；There is the electroceramics porcelain body of stomata in selection 20, as Equations of The Second Kind sample；There is the electroceramics porcelain body 20 of underbead crack in selection, used as the 3rd class sample；
Secondly, respectively by first kind sample, Equations of The Second Kind sample and the 3rd class sample respectively by fuzzy wavelet algorithm, obtain Corresponding first kind characteristic quantity, Equations of The Second Kind characteristic quantity and the 3rd category feature amount, now first kind characteristic quantity correspondence do not damage There are 20 electroceramics porcelain bodies of stomata, the 3rd category feature amount correspondence underbead crack in 20 electroceramics porcelain bodies, Equations of The Second Kind characteristic quantity correspondence 20 electroceramics porcelain bodies；
First kind characteristic quantity, Equations of The Second Kind characteristic quantity, the 3rd category feature amount are input into RBFNN respectively, are obtained by network training Take c1,c2,c3...ch, δ1,δ2,δ3...δhAnd w1,w2,w3...wh, complete the training process of radial base neural net.Its meter Calculating formula is：
The type of impairment of the electroceramics porcelain body includes：Wound or internal injury；Wherein wound includes：Alligatoring, surface are drawn Wound；Internal injury includes：Internal porosity and underbead crack.
More than, only presently preferred embodiments of the present invention is not intended to limit the scope of the present invention.

Claims (10)

1. a kind of electroceramics porcelain body line flaw detection measurement apparatus, it is characterised in that：Including detection unit (101), driver element (102), collector unit (103) and processing unit (104), wherein, the detection unit (101) is for detecting electroceramics porcelain body and controlling Driver element (102) processed taps detected electroceramics porcelain body；The collector unit (103) is for collecting driver element (102) The knock that tapped electroceramics porcelain body is produced, and the electroceramics porcelain body knock that will be collected into is converted to digital data transmission to described Processing unit (104)；The processing unit (104) is recognized according to the electroceramics porcelain body knock for being collected using preset algorithm The type of impairment of electroceramics porcelain body.
2. a kind of electroceramics porcelain body line flaw detection measurement apparatus according to claim 1, it is characterised in that：The detection unit (101) including optoelectronic switch, electroceramics porcelain body is detected by optoelectronic switch.
3. a kind of electroceramics porcelain body line flaw detection measurement apparatus according to claim 1, it is characterised in that：The driver element (102) including tapping subelement and amplifying subelement, wherein, the percussion subelement includes tapping to be hammered into shape, and the hammer that taps is set In the surface of electroceramics porcelain body；The amplification subelement includes power amplifier, and the power amplifier will hit against electroceramics porcelain body The knock of generation is amplified.
4. a kind of electroceramics porcelain body line flaw detection measurement apparatus according to claim 1, it is characterised in that：The collector unit (103) including sound collecting subelement and sound conversion subunit, wherein, the sound collecting subelement includes microphone, institute Stating sound conversion subunit includes sound collection card, and the knock that microphone is collected into is converted to numeral by the sound collection card Signal transmission gives processing unit (104).
5. a kind of electroceramics porcelain body line flaw detection measurement apparatus according to claim 1, it is characterised in that：The processing unit (104) including five rank butterworth filters and processor, wherein, electroceramics porcelain body is tapped by five rank butterworth filters Sound is filtered treatment, obtains filtering voice signal；The processor is believed the filtering sound for obtaining by using preset algorithm Number carry out processing the type of impairment for obtaining electroceramics porcelain body.
6. a kind of measuring method of electroceramics porcelain body line flaw detection, it is characterised in that comprise the following steps：
S1：Electroceramics porcelain body is detected by detection unit, and a knocking is sent to driver element, while being sent out to collector unit Send one to collect and enable signal；
S2：After the driver element receives knocking, struck in the range of default dynamics by its internal percussion subelement Hit electroceramics porcelain body；
S3：The knock that will hit against the generation of electroceramics porcelain body by the amplification subelement in the driver element is amplified, and produces Electroceramics porcelain body knock amplifies signal；
S4：Electroceramics porcelain body knock is collected by collector unit and amplifies signal, and the electroceramics porcelain body sound is amplified into signal and turned Digital data transmission is changed to processing unit；
S5：The processing unit is filtered treatment to electroceramics porcelain body knock data signal, obtains filtering voice signal, according to The filtering voice signal is extracted using the first preset algorithm to the filtering acoustic information for obtaining, and obtains filtering voice signal pair The flaw detection characteristic information answered, the damage class of electroceramics porcelain body is obtained further according to the flaw detection characteristic information for obtaining using the second preset algorithm Type.
7. a kind of electroceramics porcelain body line flaw detection measuring method according to claim 6, it is characterised in that described first presets Algorithm uses fuzzy wavelet algorithm, and it is concretely comprised the following steps：
S101：The morther wavelet in fuzzy wavelet conversion is determined first, and the conversion of yardstick a is carried out to morther wavelet；
S102：Then carried out at obfuscation to the filtering voice signal that obtains after filtering and by the morther wavelet that yardstick a is converted Reason, obtains the first language variable S of the filtering voice signal after Fuzzy processingjWith the first language variable corresponding first It is subordinate to angle valueAnd the second language variable M of the morther wavelet converted by yardstick aiIt is corresponding with the second language variable Second is subordinate to angle value
S103：Then according to predetermined fuzzy matching rule to first language variable SjWith second language variable MiCarry out fuzzy Match somebody with somebody, to obtain and match language between the filtering voice signal after the Fuzzy processing and the morther wavelet by yardstick a conversion Speech variable Rij；
S104：Finally according to the first language variable SjCorresponding first is subordinate to angle valueWith the second language variable MiIt is right Second for answering is subordinate to angle valueTo the matching language variable RijSharpening treatment is carried out, filtering voice signal is obtained corresponding Flaw detection characteristic information Ta。
8. a kind of electroceramics porcelain body line flaw detection measuring method according to claim 6, it is characterised in that：Described second presets Algorithm uses radial base neural net algorithm, and the radial base neural net algorithm is specifically that will filter sound under yardstick a The corresponding flaw detection characteristic information T of signalaAs the input quantity of input layer, by Φi(||Ta-ci| |) as the activation letter of hidden layer Number, using the type of impairment y of electroceramics porcelain body as output layer output quantity, wherein, the non-thread of the type of impairment y of the electroceramics porcelain body Sexual intercourse is：
y = Σ i = 1 h w i Φ i ( | | T a - c i | | )
Wherein, | | Ta-ci| | represent TaWith ciBetween Euclidean distance, ciI-th data center's value of hidden layer neuron is represented, wiIt is output weights.
9. a kind of electroceramics porcelain body line flaw detection measuring method according to claim 8, it is characterised in that：The hidden layer Activation primitive Φi(||Ta-ci| |) mathematical definition be：
Φ i ( | | T a - c i | | ) = e | | T a - c i | | 2 δ i 2
Wherein, δiIt is i-th extension constant of hidden layer neuron.
10. a kind of electroceramics porcelain body line flaw detection measuring method according to claim 8, it is characterised in that：In step S101 Before, it is necessary first to determine the data center value c of i-th hidden layer neuroni, extension constant δiWith output weight wi, it is right to realize Radial base neural net is fitted, wherein it is determined that the data center value c of i-th hidden layer neuroni, extension constant δiWith it is defeated Go out weight wiProcess it is specific as follows：
First, the electroceramics porcelain body without damage of equivalent amount is chosen, be there is the electroceramics porcelain body of stomata and be there is underbead crack Electroceramics porcelain body several, respectively using without the electroceramics porcelain body for damaging as first kind sample；There is the electroceramics porcelain body conduct of stomata Equations of The Second Kind sample；There is the electroceramics porcelain body of underbead crack as the 3rd class sample；
Second, first kind sample, Equations of The Second Kind sample and the 3rd class sample are obtained into correspondence respectively by fuzzy wavelet algorithm respectively First kind characteristic quantity, Equations of The Second Kind characteristic quantity and the 3rd category feature amount；
3rd, the first kind characteristic quantity of acquisition, Equations of The Second Kind characteristic quantity, the 3rd category feature amount are input into RBFNN respectively, by network Training obtains c1,c2,c3...ch、δ1,δ2,δ3...δhAnd w1,w2,w3...wh, complete the training of radial base neural net Journey.
CN201611264803.3A 2016-12-30 2016-12-30 Online flaw detection device and method for electroceramic body Active CN106680371B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201611264803.3A CN106680371B (en) 2016-12-30 2016-12-30 Online flaw detection device and method for electroceramic body

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201611264803.3A CN106680371B (en) 2016-12-30 2016-12-30 Online flaw detection device and method for electroceramic body

Publications (2)

Publication Number Publication Date
CN106680371A true CN106680371A (en) 2017-05-17
CN106680371B CN106680371B (en) 2019-12-20

Family

ID=58849628

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201611264803.3A Active CN106680371B (en) 2016-12-30 2016-12-30 Online flaw detection device and method for electroceramic body

Country Status (1)

Country Link
CN (1) CN106680371B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109387565A (en) * 2018-10-12 2019-02-26 山东理工大学 A method of brake block internal flaw is detected by analysis voice signal
CN114324580A (en) * 2021-12-03 2022-04-12 西安交通大学 Intelligent knocking detection method and system for structural defects

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1402015A (en) * 2002-09-13 2003-03-12 西安交通大学 Method and device for diagnosis of insulation ageing of electric machine based on small wave conversion
CN202614725U (en) * 2012-04-16 2012-12-19 中南林业科技大学 Device for losslessly and intelligently detecting flaws in wood material
CN203385697U (en) * 2013-08-29 2014-01-08 廖文祥 Ceramic detection equipment
CN104198578A (en) * 2014-07-30 2014-12-10 国家电网公司 Automatic flaw-detecting and testing system of porcelain post insulator and testing method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1402015A (en) * 2002-09-13 2003-03-12 西安交通大学 Method and device for diagnosis of insulation ageing of electric machine based on small wave conversion
CN202614725U (en) * 2012-04-16 2012-12-19 中南林业科技大学 Device for losslessly and intelligently detecting flaws in wood material
CN203385697U (en) * 2013-08-29 2014-01-08 廖文祥 Ceramic detection equipment
CN104198578A (en) * 2014-07-30 2014-12-10 国家电网公司 Automatic flaw-detecting and testing system of porcelain post insulator and testing method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
冷劲松 等: "复合材料结构敲击法无损检测灵敏度研究", 《复合材料学报》 *
臧春艳 等: "振动法检测支柱绝缘子缺陷的研究", 《高压电器》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109387565A (en) * 2018-10-12 2019-02-26 山东理工大学 A method of brake block internal flaw is detected by analysis voice signal
CN114324580A (en) * 2021-12-03 2022-04-12 西安交通大学 Intelligent knocking detection method and system for structural defects

Also Published As

Publication number Publication date
CN106680371B (en) 2019-12-20

Similar Documents

Publication Publication Date Title
CN106841403A (en) A kind of acoustics glass defect detection method based on neutral net
CN109142547B (en) Acoustic online nondestructive testing method based on convolutional neural network
CN104965023B (en) Multi-modal guided wave industrial pipeline diagnostic method
CN106405339A (en) Power transmission line fault reason identification method based on high and low frequency wavelet feature association
CN101231847B (en) Structural vibration noise initiative control method and apparatus
WO2021068983A1 (en) Smart acoustic information recognition-based welded weld impact quality determination method and system
CN106680371A (en) Electric porcelain body online flaw detection device and electric porcelain body online flaw detection method
CN107329932A (en) The time-frequency domain Modal Parameters Identification decomposed based on nonlinear frequency modulation component
CN100571452C (en) Loudspeaker simple tone detecting method
CN104535646A (en) Method for detecting imperfection of food grains
CN105043767A (en) Rolling ball bearing outer ring spalling fault double-impact feature extraction method and rolling ball bearing outer ring spalling fault double-impact feature extraction system
CN107886078A (en) A kind of Threshold Denoising method based on layered self-adapting threshold function table
CN107462797A (en) A kind of automobile starting voltage analog system and method
TWI804093B (en) Method and system for underground pipe leak detection
CN108278438A (en) A kind of duct noise active control device and method
CN101738436B (en) System and method for enhancing signal-to-noise ratio of medium plate automatic flaw detection signal
CN114970610A (en) Power transformer state identification method and device based on gram angular field enhancement
CN205209628U (en) Outdoor wireless ambient noise monitoring system
CN106595835A (en) Sound beacon signal extraction method and system based on dislocation superposition algorithm
CN202433457U (en) Control device for electromagnetic valve online detection platform
CN110196069B (en) Sensor compensation system and compensation method thereof
Li et al. A Robust Classification Method for Power Transformer Status Recognition Based on Sound Signals
CN206919927U (en) A kind of area surroundings detecting and controlling system based on wireless network sensor
CN106353406A (en) Wind turbine generation set bolt breakage monitoring device
CN111582132A (en) Improved EEMD and PCNN-based gas leakage signal noise reduction method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant