CN113702513A - Method for identifying metal material based on prediction function model - Google Patents
Method for identifying metal material based on prediction function model Download PDFInfo
- Publication number
- CN113702513A CN113702513A CN202110806763.5A CN202110806763A CN113702513A CN 113702513 A CN113702513 A CN 113702513A CN 202110806763 A CN202110806763 A CN 202110806763A CN 113702513 A CN113702513 A CN 113702513A
- Authority
- CN
- China
- Prior art keywords
- signal
- standard
- metal sample
- metal
- time domain
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 239000007769 metal material Substances 0.000 title claims abstract description 35
- 239000000523 sample Substances 0.000 claims abstract description 180
- 239000002184 metal Substances 0.000 claims abstract description 133
- 229910052751 metal Inorganic materials 0.000 claims abstract description 133
- 238000005070 sampling Methods 0.000 claims description 46
- 238000005311 autocorrelation function Methods 0.000 claims description 27
- 238000004364 calculation method Methods 0.000 claims description 18
- 239000000463 material Substances 0.000 claims description 18
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000009826 distribution Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000007796 conventional method Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 239000011541 reaction mixture Substances 0.000 claims description 3
- 238000009659 non-destructive testing Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 abstract description 2
- 230000035945 sensitivity Effects 0.000 abstract description 2
- 230000000149 penetrating effect Effects 0.000 abstract 1
- 150000002739 metals Chemical class 0.000 description 8
- 239000007822 coupling agent Substances 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 238000005520 cutting process Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 229910001220 stainless steel Inorganic materials 0.000 description 3
- 239000010935 stainless steel Substances 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000002604 ultrasonography Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011900 installation process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 238000000053 physical method Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000010970 precious metal Substances 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 125000006850 spacer group Chemical group 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
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
- G01N29/44—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
- G01N2291/00—Indexing codes associated with group G01N29/00
- G01N2291/01—Indexing codes associated with the measuring variable
- G01N2291/015—Attenuation, scattering
-
- 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
- G01N2291/0234—Metals, e.g. steel
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
Abstract
The invention belongs to the technical field of metal ultrasonic nondestructive testing, and particularly relates to a method for identifying a metal material based on a prediction function model. The invention provides a new method for metal identification, which is characterized in that an ultrasonic nondestructive testing technology based on a linear prediction coefficient algorithm is applied to the field of metal identification for the first time; the method only needs to make the ultrasonic transmitting probe contact the detected object from a certain surface, has simple and safe operation and light equipment, and has the advantages of low cost, strong penetrating power, good directionality, high sensitivity and the like.
Description
Technical Field
The invention belongs to the technical field of ultrasonic nondestructive testing, relates to a method for identifying a metal material by using a linear prediction coefficient, and particularly relates to a method for identifying a metal material based on a prediction function model.
Background
In recent years, the price of raw materials is continuously rising, precious metal materials are in short supply, and in order to gain violence, some bad merchants replace expensive metals with cheap metals, so that the production life of people is greatly influenced. Therefore, the method for realizing the authenticity identification has wide application in various fields such as industry, military, cultural relic identification and the like.
At present, there are many methods for identifying metals, such as physical methods: sensory recognition, fracture recognition, spark recognition, and the like; the chemical method comprises the following steps: titrimetric methods, gravimetric methods, volumetric methods, and the like. However, the above methods are destructive to metals themselves, and some operations are complicated, and thus are difficult to be applied in a wide range.
The ultrasonic nondestructive identification has the characteristics of non-destructiveness, simple operation, direct and quick performance, high accuracy, wide application range and the like, and becomes an important identification mode in the detection field. When ultrasonic waves are transmitted in a metal material, scattering occurs due to the influence of metal grains, and the identification can be realized by analyzing the characteristic quantity of the reflection and scattering signals carrying the metal grain information.
In the existing research technology, there is a method of identifying a metal sample by using an ultrasonic attenuation spectrum correlation coefficient method, a weighted euclidean distance method, and the like, but the existing research has certain limitations, such as: because the probe has directivity, when the probe is used for acquiring signals, the probe must be strictly controlled to be placed at the same position and in the same direction, in actual operation, the probe is difficult to achieve, if one of the signals is acquired incorrectly, the difference between the acquired signal and other signals is large, the whole experimental result is greatly influenced, and therefore, the requirements on the operation accuracy and the probe are high.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides a method for identifying a metal material based on a prediction function model, which has the advantages of easy operation, strong penetration capacity, good directionality and high sensitivity, can improve the identification accuracy of the metal material, and reduces the dependence on people and detection probes.
In order to achieve the purpose, the invention adopts the technical scheme that:
a method for identifying a metal material based on a prediction function model comprises the following steps:
(1) respectively collecting time domain signals of a standard metal sample and a metal sample to be identified according to a conventional method, and sequentially recording the time domain signals as the standard signals and the signal to be identified;
(2) taking a time domain signal of a standard metal sample, constructing a prediction function model, calculating a prediction error, and obtaining a p-order linear prediction coefficient a of the standard metal sample by using a minimum mean square error criterion and an iterative algorithmi,i=1~p;
(3) According to the type of the metal sample and the p-order linear prediction coefficient aiDetermining a threshold value delta of the standard signal by adopting a multiple standard deviation method;
(4) calculating a p-order linear prediction coefficient of each signal to be identified according to the method in the step (2);
(5) comparing the p-order linear prediction coefficients of the signals to be identified with the threshold value delta in the step (3), and if the p-order linear prediction coefficients of the signals to be identified are all within the range of the threshold value delta, obtaining the same material; if a certain order or a plurality of orders exceed the threshold range, the materials are heterogeneous materials, and identification of the metal materials is achieved.
Further, the step (2) specifically includes:
(2.1) acquiring a time domain signal of a standard metal sample by using an ultrasonic probe, wherein the amplitude is represented as X (n), and n is an ultrasonic signal sampling point;
(2.2) carrying out normalization processing on the amplitude X (n) of each sampling point of the time domain signal of the standard metal sample:
wherein: x (n)minIs the minimum amplitude in the echo signal; x (n)maxThe maximum amplitude in the echo signal, x (n), is the signal amplitude of the nth sampling point of the standard metal sample signal, namely the real signal;
(2.3) according to the standard metal sample signal amplitude of p sampling points before the nth sampling point, obtaining the signal amplitude of the nth sampling point of the standard metal sample signal through a prediction functionNamely, standard metal sample ultrasonic prediction signals;
(2.4) calculating the prediction error e (n) between the ultrasonic prediction signal and the real signal of the metal sample and the mean square error epsilon:
(2.5) calculating a using the minimum mean square error criterioniLet epsilon take the minimum value to obtain aiA system of linear equations for variables:
wherein, aiIs a linear prediction coefficient; x (n-j) represents that the amplitude of the metal sample time domain signal (x (n)) is predicted by using the amplitude of the metal sample time domain signal (j) before the amplitude of the ith metal sample time domain signal;
(2.6) constructing the autocorrelation function R of the time-domain signal of the standard metal sample by the linear equation set in the step (2.5)n(j) N is the total number of sampling data of the echo signals of the metal sample acquired by using the ultrasonic probe, so that N is more than or equal to 0 and less than or equal to N:
wherein: j is the time delay of the autocorrelation function of the ultrasonic signal of the metal specimen, Rn(i-j) means that j autocorrelation function values before the autocorrelation function value of the time-domain signal of the metal sample of the ith sampling point are used for predicting the R < th > valuen(j) A respective autocorrelation function value;
(2.7) adding Rn(j) Split into Toeplize matrices and calculate coefficients by an iterative algorithm:
iterative calculation of aiExpressed as follows:
in the formula (8), the reaction mixture is,the solution of the Toeplize matrix is formed by the autocorrelation function matrix of the metal sample time domain signal and the prediction mean square error matrix of the metal sample time domain signal; and calculating coefficients by an iterative algorithm; k is a radical ofjIs a reflection coefficient, l represents the order of each iterative calculation of the linear prediction coefficient of the metal sample time domain signal;for results obtained before the current iteration order, Rn(j-l) using the l autocorrelation function values prior to the l autocorrelation function value of the time domain signal of the jth metal sample to predict the R < th > valuen(j) A respective autocorrelation function value; epsilon(j-1)And when iterative calculation is performed, the prediction mean square error of the metal sample time-domain signal calculated by the last iteration of the current iteration is shown.
Further, the step (3) is specifically as follows:
simulating the distribution situation in the signal acquisition based on the Gaussian distribution, and calculating a threshold value delta according to the following formula (9), specifically:
wherein m is the number of standard signals; a isrIs a linear prediction coefficient of order p; note the book Representing the mean value of the m standard signals p-order linear prediction coefficients; δ is the standard deviation of the p-order linear prediction coefficients of the m standard signals.
Further, in the step 3), according to the calculated threshold value Δ, the p-order linear prediction coefficient a of the standard signaliIf the range of the threshold value delta is exceeded, the standard signal is abnormal in acquisition, the standard signal is removed by using a program algorithm, and the threshold value delta of the standard signal is recalculated after the standard signal is discarded.
Further, the conditions for acquiring the time domain signal in the step (1) are as follows: the central frequency of the ultrasonic receiving/transmitting probe is 5-6MHz, and the diameter of the wafer is 10 mm; the pulse repetition frequency of the ultrasonic pulse transmitting/receiving instrument is 100Hz, the pulse voltage is 100V, and the gain is +8 dB; the sampling frequency of the digital oscilloscope is 5G S/s, the sampling time is 20 mus, and the average sampling times is 2000 times.
Further, in the step (1), each time a signal is acquired, the probe needs to be lifted and repositioned at the same position.
Furthermore, the acquisition frequency of the standard signal is not less than 20 times, and the acquisition frequency of the signal to be identified is not less than 5 times.
Compared with the prior art, the invention has the beneficial effects that:
1. according to the method, the time domain ultrasonic reflection and scattering signals in the metal material are obtained, the p-order linear prediction coefficient of the signals is calculated to serve as the characteristic quantity of ultrasonic identification, the threshold is determined by multiple standard deviations, the method is a novel method for realizing metal identification, the error in metal detection is greatly reduced, and the accuracy of identifying the metal is improved.
2. The method overcomes the defects of the traditional metal material identification method, has no damage to metal, has the characteristics of simple operation, low cost, high efficiency, wide application range, accurate and reasonable identification result and the like, and is suitable for popularization and application.
3. According to the method, the identification program software is compiled according to the provided algorithm, so that the operation is simple, and the data is conveniently and quickly stored; the display of the recognition result is also increased, so that the method is visual and convenient to analyze; and meanwhile, the installation of the identification program software is convenient and quick.
Drawings
FIG. 1 is a diagram of an identification program operating interface;
fig. 2 is a diagram of echo signals of a metal standard sample.
Detailed Description
The present invention will now be described in detail with reference to the accompanying drawings and examples.
The invention provides a method for identifying a metal material based on a linear prediction coefficient, which comprises the following steps:
(1) acquiring time domain signals of a standard metal sample and a metal sample to be identified according to a conventional method;
the connection relation of the ultrasonic anti-counterfeiting identification device is as follows: connecting an ultrasonic pulse transmitting/receiving instrument with a computer through a digital oscilloscope, connecting a receiving/transmitting probe with the ultrasonic pulse transmitting/receiving instrument, placing the ultrasonic pulse transmitting/receiving instrument on the surface of a metal sample to be identified, and enabling the receiving/transmitting probe to be in coupling contact with the metal sample by using a coupling agent.
The central frequency of the ultrasonic receiving/transmitting probe is 5-6MHz, and the diameter of the wafer is 10 mm; the pulse repetition frequency of the ultrasonic pulse transmitting/receiving instrument is 100Hz, the pulse voltage is 100V, and the gain is +8 dB; the sampling frequency of the digital oscilloscope is 5G S/s, the sampling time is 20 mus, and the average sampling times is 2000 times.
For probe placement description: when the probe is needed to be replaced every time a signal is collected, the collection point can be marked in advance, and the probe is guaranteed to be placed at the same position on the metal surface every time. The probe can be rotated by a small angle, the specified direction of the probe is specified in advance, for example, by taking the probe line as a reference, and the acquisition of each signal needs to be within an angle interval of [ -5 °, +5 ° ] of the specified direction.
The adopted coupling agent is water, the coupling agent needs to be dripped again every time a signal is taken, the same dosage of the coupling agent is ensured as much as possible, and one drop of water can be taken in the experiment.
In order to obtain a correct identification result, the acquisition times of the signals can be determined according to actual conditions. However, the number of times of acquiring the standard signal is preferably not less than 20, the number of times of acquiring the signal to be identified is preferably not less than 3, and the acquired echo signal diagram is shown in fig. 2.
Acquiring a time domain signal of a standard metal sample by using an ultrasonic probe, wherein the amplitude is represented as X (n), and n is an ultrasonic signal sampling point; the amplitude X (n) of each sampling point of the echo signal of the standard metal sample is normalized by using the formula (1):
wherein, X (n)minIs the minimum amplitude in the echo signal; x (n)maxThe maximum amplitude in the echo signal, x (n) is the signal amplitude of the nth sampling point of the standard metal sample signal, namely a real signal;
(2) according to the standard metal sample signal amplitude of p sampling points before the nth sampling point, the signal amplitude of the nth sampling point of the standard metal sample signal is obtained through a prediction functionNamely, standard metal sample ultrasonic prediction signals;
specifically, according to the amplitude of the standard metal sample signal of p sampling points before the nth sampling point, the amplitude is as follows: x (n-1), x (n-2).. x (n-p), predicting the metal sample signal amplitude of the nth sampling point through a prediction function, and predicting the characteristic parameter a of the prediction functioniConstructing p-order prediction function as characteristic quantity of metal sample ultrasonic signal
Calculating the prediction error e (n) between the ultrasonic prediction signal and the real signal of the standard metal sample,
determining coefficient a according to the minimum mean square error criterion by calculating mean square errori:
The calculation formula (4) is applied to aiDerivative and make the derivative zero to obtain aiSystem of linear equations of
aiIs a linear prediction coefficient; x (n-j) represents that the amplitude of the metal sample time domain signal (x (n)) is predicted by using the amplitude of the metal sample time domain signal (j) before the amplitude of the ith metal sample time domain signal;
finally, the formula (5) is used for constructing an autocorrelation function R of the amplitude of the metal sample signaln(j) N is the total number of sampling data of the echo signals of the metal sample acquired by using the ultrasonic probe, so that N is more than or equal to 0 and less than or equal to N:
aithe prediction function is a linear prediction coefficient and also a parameter which enables the prediction function to meet the minimum mean square error, namely the characteristic quantity of the metal sample ultrasonic signal; wherein: j is the time delay of the autocorrelation function of the ultrasonic signal of the metal specimen, Rn(i-j) means that j autocorrelation function values before the autocorrelation function value of the time-domain signal of the metal sample of the ith sampling point are used for predicting the R < th > valuen(j) A respective autocorrelation function value;
and iterative calculation (6) is carried out to calculate the p-order linear prediction coefficient a of the time domain signal of the metal samplei(i ═ 1,2, 3.., p), specifically Rn(j) Split into Toeplize matrices and calculate coefficients by an iterative algorithm:
in the embodiment, the iterative computation is started from p ═ 1, in this case, the autocorrelation matrix of the time-domain signal of the metal sample is,
is solved outAnd ε(1)(ii) a Substituting the value into the second-order iterative calculation of p 2 to solveAnd ε(2)(ii) a The final iteration result is expressed as follows:
in the formula (8), the reaction mixture is,the solution of the Toeplize matrix is formed by the autocorrelation function matrix of the metal sample time domain signal and the prediction mean square error matrix of the metal sample time domain signal; and calculating coefficients by an iterative algorithm; k is a radical ofjIs a reflection coefficient, l represents the order of each iterative calculation of the linear prediction coefficient of the metal sample time domain signal;for results obtained before the current iteration order, Rn(j-l) using the l autocorrelation function values prior to the l autocorrelation function value of the time domain signal of the jth metal sample to predict the R < th > valuen(j) A respective autocorrelation function value; epsilon(j-1)And when iterative calculation is performed, the prediction mean square error of the metal sample time-domain signal calculated by the last iteration of the current iteration is shown.
Linear prediction coefficient of order p aiAs a characteristic quantity, k, of the ultrasonic signal of each metal specimenjIs the reflection coefficient, l represents the order of each iterative calculation of the linear prediction coefficient of the metal sample time domain signal,and calculating the total order p from the first order to complete the iteration.
E.g. first order iterationSubstituted into the second iteration and getAndand by analogy, finally obtaining a p-order linear prediction coefficient representing the time domain signal of the metal sample: a isi:
(3) According to the type of the metal sample and the p-order linear prediction coefficient aiDetermining a threshold value delta of the standard signal by adopting a multiple standard deviation method;
specifically, according to the central limit theorem, the signal acquired by each metal sample for multiple times obeys gaussian distribution, and the threshold value Δ of the standard signal is calculated by the law of raydea, specifically:
wherein m is the number of standard signals; a isrP-order linear prediction coefficients of the m groups of repeatedly acquired metal sample ultrasonic signals; note the bookIs m groupsThe mean value of the p-order linear prediction coefficients of the standard signal;and the standard deviation of the p-order linear prediction coefficients of the m groups of signals.
Meanwhile, according to the calculated threshold value delta, the p-order linear prediction coefficient a of the standard signaliIf the range of the threshold value delta is exceeded, the standard signal is abnormal in acquisition, the standard signal is removed by using a program algorithm, and the threshold value delta of the standard signal is recalculated after the standard signal is discarded.
In this embodiment, the multiple standard deviation means that the algorithm of the threshold is calculated by adding or subtracting several times of the standard deviation from the average value. The selection of the standard deviation multiple is related to metal materials, if the scattering of the grains in the standard metal sample to the ultrasound is weak, the standard deviation multiple is more accurate by 3 times, otherwise, if the scattering of the grains in the metal material to the ultrasound is strong, the fault tolerance rate can be improved by 5 times, and the error identification risk caused by the operation error of personnel is reduced. Hundreds of groups of experiments are carried out on various metals, and the fact that the universality is the best when the number of the metals is 3-5 times is found, and the identification of most metals can be met.
In this embodiment, the multiple of the standard deviation is 3, which just meets the criterion of Lauda. However, in practical application, the inevitable operation error needs a certain fault tolerance rate and the universality of different metal materials, so that improvement is performed on the basis of the Lauda criterion, namely the multiple can be manually changed according to the self characteristics of the metal materials, and the accuracy rate and the fault tolerance rate are further adjusted.
In the present embodiment, the first and second electrodes are,
(4) calculating a p-order linear prediction coefficient of each signal to be identified according to the method in the step (2);
(5) comparing the p-order linear prediction coefficient of the signal to be identified with the threshold value delta in the step (3), and if the p-order linear prediction coefficient of the signal to be identified is within the range of the threshold value delta, identifying the result as the same sample; if the coefficients of 1 order and above in the p-order linear prediction coefficients of the signal to be detected are out of the range of the threshold value delta, the identification result is 'different samples'. A method for identifying a metal material based on a linear prediction coefficient is realized.
Referring to fig. 1, the present invention is written as a recognition program according to the above mentioned algorithm, and the operation only needs 3 steps to complete the recognition: step 1, clicking a 'standard metal signal' button, and selecting a folder for storing standard metal sample signals; step 2, clicking a 'to-be-identified metal signal' button, and selecting a folder for storing a to-be-identified metal sample signal; and step 3, clicking an 'identification' button to obtain a result. If the signal to be identified comes from the standard metal sample, a green result frame shown in fig. 1 is obtained, and the same metal object is displayed; if the signal to be identified is not from a standard metal specimen, the result box will turn red and show "different metal object".
The recognition program provided by the invention also increases the specific display of the recognition result, and each time the 'recognition' button shown in figure 1 is clicked, the final recognition result, the number of the p-order linear prediction coefficients meeting the threshold value and the number of the p-order linear prediction coefficients exceeding the threshold value are given, so as to facilitate further analysis.
Because the prior art has complicated requirements on the format of a signal storage folder, 20 empty folders are required to be established for each storage, and then the acquired 20 time domain signals of the csv-format standard metal samples are sequentially placed into the 20 empty folders. The identification program software provided by the invention optimizes the signal storage mode, can automatically identify the number of data stored by putting down the probe each time and automatically perform algorithm processing, is convenient and quick to store the data, only needs to distinguish sample folders, and does not need to perform redundant folder processing.
Meanwhile, the identification program provided by the invention uses the compiler to compile the executable file, can be directly used on any computer with Windows7 and above operating systems, has the size of the executable file not exceeding 1GB, is convenient and quick in installation process, and can be installed only by selecting the installation directory.
Example 1: identification between dissimilar metallic materials of similar composition
Experimental samples: the three specifications are completely the same and have similar components, the materials are Cr17Ni2 (sample No. 1), 2Cr13 (sample No. 2) and 3Cr13 (sample No. 3) stainless steel round metal test samples, the 3 samples have the same size, the diameter is 50mm, and the thickness is 15 mm. Choose 1 sample as standard sample during the experiment, 1,2,3 samples are as waiting to discern the sample, paste the right angle spacer on the sample surface before the experiment to the position that the probe was placed when guaranteeing the signal of every turn was gathered is the same.
In order to ensure the same experimental conditions as much as possible, the couplant (water) needs to be added dropwise again every time a signal is taken, and the dosage is one drop.
The instrument device is connected: the Panametrics-NDT 5077PR ultrasonic pulse transmitting/receiving instrument is connected with a computer through a Tektronix-DPO5034B digital oscilloscope, and a receiving/transmitting probe with the center frequency of 5MHz is connected with the ultrasonic pulse transmitting/receiving instrument and then placed on the surface of a metal sample to be identified, so that the receiving/transmitting probe is in coupling contact with the metal sample by using a coupling agent (water). Referring to fig. 2, the probe transmits an ultrasonic pulse signal and receives an echo signal, and the echo signal is sampled by an oscilloscope connected to the ultrasonic pulse transmitter/receiver.
The method for identifying the metal material based on the linear prediction coefficient provided by the embodiment comprises the following steps:
the method comprises the following steps: collecting time domain signals of a standard metal sample and a metal sample to be identified according to a conventional method
At first, the couplant (two drops of water) is dripped at the right angle of the right angle positioning sheet, the probe is placed at the right angle inflection point of the right angle positioning sheet, and the probe line is aligned with the right angle inflection point, so that the position and the angle are kept consistent when the probe is placed at each time. Applying a certain pressure to the probe to ensure that the probe is tightly attached to the surface of the sample and the pressure applied to the probe every time is the same so as to obtain a stable echo signal; and the probe needs to be replaced every time a signal is taken, and the couplant is dripped again, so that the dosage of the couplant in each time is kept as consistent as possible.
And (3) acquiring a standard signal: in order to obtain more metal grain information at this point, 20 standard signals were collected at the collection point. The method selects the standard signal collected in the angle range of [ -5 °, +5 ° ] in the specified direction in consideration of the artificial operation error, can reflect the metal grain characteristics of the standard signal at the point, can not bring out the abnormal signal characteristics, and constructs a reasonable standard signal sample database.
Acquiring a signal to be identified: the method is the same as the standard signal acquisition mode and experimental conditions, and 5 times of signals to be identified are acquired within the angle range of [ -5 degrees, +5 degrees ] in the specified direction.
Step two: computing p-order linear prediction coefficient of standard signal
Firstly, cutting off initial waves of 20 standard signal data samples, wherein the amplitude of the initial wave signals is large, and the initial waves of all the signals are the same, so that in order to improve the difference of the signals, the initial waves are cut off after the data are introduced;
acquiring a time domain signal of a standard sample by using an ultrasonic probe, wherein the amplitude is represented as X (n), and n is an ultrasonic signal sampling point; the amplitude X (n) of each sampling point of the echo signal of the standard sample is normalized by using the formula (1):
wherein X (n)minIs the minimum amplitude in the echo signal; x (n)maxThe maximum amplitude value in the echo signal is obtained; x (n) is the amplitude of the sampling point subjected to normalization processing, and is also the signal amplitude of the nth sampling point of the standard metal sample signal, namely a real signal;
in this embodiment, since the sampling point is 488, the echo amplitudes x (n) of 438 metal samples are still obtained after cutting off 50 initial wave data points, and the data volume is large, the amplitudes of the first 20 sampling points are given as an example: 213.3333, respectively; 177.3333, respectively; 105.3333, respectively; 65.6667, respectively; 40.0000; 34.0000, respectively; 35.6667, respectively; 53.000073.6667, respectively; 98.0000, respectively; 100.6667, respectively; 96.0000, respectively; 66.0000, respectively; 36.3333, respectively; 9.0000, respectively; 20.000027.0000, respectively; 27.3333, respectively; 23.3333, respectively; 11.3333, obtaining the signal amplitude of the standard metal sample signal of 20 sampling points after normalization, namely the real signal is: 0.8577, respectively; 0.7128; 0.4228, respectively; 0.2631, respectively; 0.1597, respectively; 0.1356, respectively; 0.1423, respectively; 0.2121, respectively; 0.2953, respectively; 0.3933, respectively; 0.4040, respectively; 0.3852, respectively; 0.2644, respectively; 0.1450 parts by weight; 0.0349; 0.0792; 0.1074, respectively; 0.1087, respectively; 0.0926, respectively; 0.0443
Calculating p-order linear prediction coefficients and prediction mean square error of 20 groups of standard sample signals;
when selecting, the selection of p is determined by two points: firstly, the requirement that the overall mean square error of a prediction signal and a real signal is as small as possible is met; secondly, the calculation efficiency and the fault tolerance rate need to be as high as possible. In this embodiment, a large number of experiments are performed to conclude that when p is 12, the general effect on various metals is good, and the mean square error of most of the predicted signals is only 1% of the real signals. The calculation speed is high, and the calculation can be completed in about 1 second. Therefore, p is 12 th order in this embodiment;
because the sampling point is 488, the echo amplitudes x (n) of 438 metal samples are still obtained after cutting 50 initial wave data points, each amplitude corresponds to a predicted mean square error, and the predicted mean square errors of the first 20 points are given as an example because the data volume is large: 0.0090; 0.0904; 0.0106; 0.0015; 0.0012; 0.0107; 0.0033; 0.0183; 0.00003; 0.0056; 0.0089; 0.0019; 0.0004; 0.0003; 0.0022; 0.0037; 0.0007; 0.0005; 0.0005; 0.0004
Obtaining 12-order linear prediction coefficients of the 20 groups of standard sample signals according to the formulas (2) to (8);
in this embodiment, the average p-order linear prediction coefficients of the signal collected 20 times for sample No. 1 are: -0.90510.2801-0.1706-0.0612-0.00450.1366-0.1793-0.12170.08660.01100.0054-0.0300
Step three: a threshold value delta of the standard signal is calculated.
In this embodiment, the order p is 12, so 12 thresholds need to be calculated, 12 upper thresholds and 12 lower thresholds are calculated by combining the formula (9),
sample No. 1 upper threshold limit: -0.85200.3552-0.1033-0.00360.02020.1777-0.1165-0.03530.18330.08370.0326-0.0150;
sample No. 1 lower threshold: -0.95810.2050-0.2379-0.1189-0.02930.0954-0.2420-0.2082-0.0100-0.0617-0.0217-0.0450;
step four: and calculating linear prediction coefficients of the signal to be identified.
Firstly, importing 5 collected signals to be identified and preprocessing the data;
normalizing the 5 signals to be identified according to the formula (1);
respectively calculating 12-order linear prediction coefficients of the 5 signals to be identified according to the formulas (2) to (8);
fourthly, calculating the average value of the 5 groups of linear prediction coefficients; because each sampling point is sampled for multiple times, the 12-order linear prediction coefficients calculated each time are averaged;
1. the mean linear prediction coefficients of the samples to be identified in numbers 2 and 3 are shown in table 1:
linear prediction coefficient mean values of samples No. 11, 2 and 3
Step five: identification
And respectively comparing the 12 linear prediction coefficients of the 3 samples with the upper and lower limits of the threshold, and if the 12 linear prediction coefficients are in the respective threshold ranges, identifying the to-be-identified number as the same material, otherwise identifying the to-be-identified number as a different material.
It can be seen that the average values of 12 linear prediction coefficients calculated by the 5 signals to be identified of the sample No. 1 are all within the threshold range of the standard signal, so that the identification result is the same material; sample No. 2 has a linear prediction coefficient of order 7 exceeding the threshold; sample No. 3 had a linear prediction coefficient of order 9 exceeding the threshold, and sample No. 2 and sample No. 3 were both identified as dissimilar materials, with the identification results shown in tables 2-4 below.
Identification results of sample No. 21 in Table 21
Identification results of sample No. 32 in Table
Table 43 sample identification results
As can be seen from the comparison of tables 2 to 4, the method provided by the embodiment can realize accurate identification of different metal materials with similar components and extremely small metal grain difference.
Example 2: identification between metal materials of the same kind
The calculation method and the identification method of the embodiment are the same as those of the embodiment 1, and the difference is that three cylindrical metal samples, namely samples No. 4, No. 5 and No. 6, are selected from the samples, wherein the three cylindrical metal samples are made of 2Cr13, have the diameters of 50mm and the thicknesses of 15 mm.
And selecting the sample No. 4 as a standard signal, and identifying the samples No. 4, 5 and 6 to be identified. Similarly, 20 signals were collected for the standard sample and 5 signals were collected for each sample to be identified. The collection method and the identification method are the same as those of the embodiment 1.
The calculated average linear prediction coefficients of the 20 standard signals of sample No. 4 are respectively: -0.9830, 0.3205, -0.0948, -0.1125, -0.0576, 0.2070, -0.1521, -0.2180, 0.1419, -0.0367, 0.0423, -0.0055;
similarly, calculating an upper threshold and a lower threshold;
upper limit of threshold value: -0.9329, 0.4691, 0.1353, 0.1046, 0.0616, 0.2714, -0.0319, -0.0622, 0.2402, 0.0142, 0.0564, 0.0127;
lower threshold: -1.0331,0.1719, -0.3249, -0.3295, -0.1768,0.1425, -0.2723, -0.3738,0.0436, -0.0876,0.0282, -0.0237.
The average linear prediction coefficient of each signal to be recognized calculated according to the same method as the embodiment is shown in table 5.
Linear prediction coefficients of samples No. 54, 5 and 6
Comparing the average linear prediction coefficient of each signal with a threshold value to obtain an identification result: the average values of 12 linear prediction coefficients calculated by the signal to be identified of the No. 4 sample 5 times are all in the threshold range of the standard signal, so that the identification result is the same material; sample No. 5 has a linear prediction coefficient of order 8 above the threshold; sample 6 had a linear prediction coefficient of order 6 above the threshold, and sample 5 and sample 6 were identified as heterogeneous materials, with specific results in tables 6-8.
Table 64 sample identification results
Identification results of sample No. 75 in Table
Identification results of sample No. 86 in Table
As can be seen from the comparison of the data in tables 6 to 8, the method provided by the present embodiment can realize the correct identification between different objects made of the same metal material in the same batch.
Example 3: identification of metal containers of the same batch
The calculation method and the identification method of the embodiment are the same as those of the embodiment 1, and the difference is that the sample is selected from three stainless steel metal containers which are processed from the same batch and have the same material type: the size is as follows: the diameter is 190mm, the height is 255mm, the bottom thickness is 10mm, and the samples are respectively marked as No. 7, No. 8 and No. 9.
And selecting the sample No. 7 as a standard signal, and identifying the samples No. 7, 8 and 9 to be identified. Similarly, 20 signals were collected for the standard sample and 5 signals were collected for each sample to be identified. The collection method and the identification method are the same as those of the embodiment 1.
The calculated average linear prediction coefficients of the 20 standard signals of sample No. 7 are respectively: -0.9159, 0.3580, -0.1352, -0.2249, 0.0983, 0.0517, -0.0825, 0.0635, -0.2322, 0.0943, -0.0132, -0.0345;
and similarly, calculating an upper threshold and a lower threshold.
Upper limit of threshold value: -0.9060, 0.3745, -0.1190, -0.2128, 0.1175, 0.0946, -0.0303, 0.1164, -0.2111, 0.1372, 0.0139, -0.0244;
lower threshold: -0.9258,0.3414, -0.1513, -0.2370,0.0792,0.0087, -0.1347,0.0106, -0.2534,0.0515, -0.0404, -0.0447.
The calculated average linear prediction coefficients for each signal to be identified are shown in table 9.
Linear prediction coefficients of samples No. 97, 8 and 9
Comparing the amplitude root-mean-square ratio of each signal with a threshold value to obtain an identification result: the average values of 12 linear prediction coefficients calculated by 5 signals to be identified of the No. 7 sample are all within the threshold range of the standard signal, so that the identification result is the same material; sample No. 8 had a linear prediction coefficient of order 12 above the threshold; sample No. 9 had a linear prediction coefficient of order 11 above the threshold, and both sample No. 8 and sample No. 9 were identified as dissimilar materials, with the specific results shown in tables 10-12.
Identification results of sample No. 107 in Table
Sample identification results of Table 118
Sample identification results of Table 129
The comparison of the data in example 3 shows that the invention can realize the correct identification of the stainless steel metal containers processed in the same batch.
The invention relates to a method for identifying a metal material based on a linear prediction coefficient, which takes the linear prediction coefficient as a characteristic quantity to be applied to the field of anti-counterfeiting identification of the metal material for the first time, and can be applied to anti-counterfeiting identification of metal cultural relics and the like; the method optimizes the aspects of signal acquisition mode, threshold calculation and the like, can realize correct identification of different metal materials with similar components and different samples produced by the same material, calculates the linear prediction coefficient of the signal as the characteristic quantity of ultrasonic identification by acquiring the time domain ultrasonic reflection and scattering signals in the metal materials, defines the threshold by multiple standard deviations, is a new method for realizing metal identification, greatly reduces the error in metal detection, and improves the accuracy of identifying the metal.
The foregoing is a more detailed description of the invention in connection with specific preferred embodiments and it is not intended that the invention be limited to these specific details. For those skilled in the art to which the invention pertains, several simple deductions or substitutions can be made without departing from the spirit of the invention, and all shall be considered as belonging to the protection scope of the invention.
Claims (7)
1. A method for identifying a metal material based on a prediction function model is characterized by comprising the following steps:
(1) respectively collecting time domain signals of a standard metal sample and a metal sample to be identified according to a conventional method, and sequentially recording the time domain signals as the standard signals and the signal to be identified;
(2) taking a standard signal, constructing a prediction function model, calculating a prediction error, and obtaining a p-order linear prediction coefficient a of a standard metal sample by using a minimum mean square error criterion and an iterative algorithmi,i=1~p;
(3) According to the type of the metal sample and the p-order linear prediction coefficient aiDetermining a threshold value delta of the standard signal by adopting a multiple standard deviation method;
(4) calculating a p-order linear prediction coefficient of each signal to be identified according to the method in the step (2);
(5) comparing the p-order linear prediction coefficients of the signals to be identified with the threshold value delta in the step (3), and if the p-order linear prediction coefficients of the signals to be identified are all within the range of the threshold value delta, obtaining the same material; if a certain order or a plurality of orders exceed the threshold range, the materials are heterogeneous materials, and identification of the metal materials is achieved.
2. The method for identifying a metallic material based on a prediction function model according to claim 1, wherein the step (2) specifically comprises:
(2.1) acquiring a time domain signal of a standard metal sample by using an ultrasonic probe, wherein the amplitude is represented as X (n), and n is an ultrasonic signal sampling point;
(2.2) carrying out normalization processing on the amplitude X (n) of each sampling point of the time domain signal of the standard metal sample:
wherein: x (n)minIs the minimum amplitude in the echo signal; x (n)maxThe maximum amplitude in the echo signal, x (n), is the signal amplitude of the nth sampling point of the standard metal sample signal, namely the real signal;
(2.3) according to the amplitude x (n-i) of the standard metal sample signal of p sampling points before the nth sampling point, obtaining the signal amplitude of the nth sampling point of the standard metal sample signal through a prediction functionNamely, standard metal sample ultrasonic prediction signals; wherein: i is any one of the P sampling points;
(2.4) calculating the prediction error e (n) between the ultrasonic prediction signal and the real signal of the metal sample and the mean square error epsilon:
(2.5) calculating a using the minimum mean square error criterioniLet epsilon take the minimum value to obtain aiA system of linear equations for variables:
wherein, aiIs a linear prediction coefficient; x (n-j) represents the amplitude of the time domain signal of the (x) (n) th metal sample predicted by the amplitude of the time domain signal of the j metal sample before the amplitude of the time domain signal of the ith metal sample;
(2.6) the system of linear equations from step (2.5)Constructing autocorrelation function R of time domain signal of standard metal samplen(j) N is the total number of sampling data of the echo signals of the metal sample acquired by using the ultrasonic probe, so that N is more than or equal to 0 and less than or equal to N:
wherein: j is the time delay of the autocorrelation function of the ultrasonic signal of the metal specimen, Rn(i-j) represents the R < th > predicted by j autocorrelation function values before the autocorrelation function value of the metal sample time domain signal of the ith sampling pointn(j) A respective autocorrelation function value;
(2.7) adding Rn(j) Split into Toeplize matrices and calculate coefficients by an iterative algorithm:
iterative calculation of aiExpressed as follows:
in the formula (8), the reaction mixture is,the solution of the Toeplize matrix is formed by the autocorrelation function matrix of the metal sample time domain signal and the prediction mean square error matrix of the metal sample time domain signal; and calculating coefficients by an iterative algorithm; k is a radical ofjIs a reflection coefficient, l represents the order of each iterative calculation of the linear prediction coefficient of the metal sample time domain signal;for results obtained before the current iteration order, Rn(j-l) using the l autocorrelation function values prior to the l autocorrelation function value of the time domain signal of the j metal sample to predict the jRn(j) A respective autocorrelation function value; epsilon(j-1)And when iterative calculation is performed, the prediction mean square error of the metal sample time-domain signal calculated by the last iteration of the current iteration is shown.
3. The method for identifying a metal material based on a prediction function model according to claim 1, wherein the step (3) is specifically as follows:
simulating the distribution situation in the signal acquisition based on the Gaussian distribution, and calculating a threshold value delta according to the following formula (9), specifically:
wherein m is the number of standard signals; a isrIs a linear prediction coefficient of order p; note the book Representing the mean value of the m standard signals p-order linear prediction coefficients; δ is the standard deviation of the p-order linear prediction coefficients of the m standard signals.
4. The method for identifying a metal material based on a prediction function model according to claim 3, wherein in the step (3), the p-order linear prediction coefficient a of the standard signal is calculated according to the calculated threshold value ΔiIf the range of the threshold value delta is exceeded, the standard signal is abnormal in acquisition, the standard signal is removed by using a program algorithm, and the threshold value delta of the standard signal is recalculated after the standard signal is discarded.
5. The method for identifying a metallic material based on a prediction function model according to claim 1, wherein the step (1) of collecting the time domain signal is performed under the conditions of: the central frequency of the ultrasonic receiving/transmitting probe is 5-6MHz, and the diameter of the wafer is 10 mm; the pulse repetition frequency of the ultrasonic pulse transmitting/receiving instrument is 100Hz, the pulse voltage is 100V, and the gain is +8 dB; the sampling frequency of the digital oscilloscope is 5G S/s, the sampling time is 20 mus, and the average sampling times is 2000 times.
6. The method for identifying a metallic material based on a prediction function model according to claim 5, wherein in the step (1), each time the signal is acquired, the probe is lifted and replaced at the same position.
7. The method for identifying a metal material based on a prediction function model according to claim 6, wherein the number of times of acquiring the standard signal is not less than 20, and the number of times of acquiring the signal to be identified is not less than 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110806763.5A CN113702513B (en) | 2021-07-16 | 2021-07-16 | Method for identifying metal material based on predictive function model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110806763.5A CN113702513B (en) | 2021-07-16 | 2021-07-16 | Method for identifying metal material based on predictive function model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113702513A true CN113702513A (en) | 2021-11-26 |
CN113702513B CN113702513B (en) | 2024-06-25 |
Family
ID=78648764
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110806763.5A Active CN113702513B (en) | 2021-07-16 | 2021-07-16 | Method for identifying metal material based on predictive function model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113702513B (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020065664A1 (en) * | 2000-10-13 | 2002-05-30 | Witzgall Hanna Elizabeth | System and method for linear prediction |
JP2005338063A (en) * | 2004-04-28 | 2005-12-08 | Japan Science & Technology Agency | Apparatus for measuring physical characteristics of sample |
CN102590338A (en) * | 2012-01-19 | 2012-07-18 | 清华大学 | Resonance peak-based ultrasonic cavitation state identification method |
EP2927674A1 (en) * | 2014-03-31 | 2015-10-07 | Nederlandse Organisatie voor toegepast- natuurwetenschappelijk onderzoek TNO | Method and apparatus for characterizing a medium using ultrasound measurements |
CN106992011A (en) * | 2017-01-25 | 2017-07-28 | 杭州电子科技大学 | Engineering machinery sound identification method based on MF PLPCC features |
CN107703210A (en) * | 2017-08-31 | 2018-02-16 | 北京金风慧能技术有限公司 | workpiece identification method and device |
CN110797033A (en) * | 2019-09-19 | 2020-02-14 | 平安科技(深圳)有限公司 | Artificial intelligence-based voice recognition method and related equipment thereof |
CN112162041A (en) * | 2020-09-30 | 2021-01-01 | 陕西师范大学 | Method for identifying metal material based on Gaussian distribution of amplitude root mean square value |
-
2021
- 2021-07-16 CN CN202110806763.5A patent/CN113702513B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020065664A1 (en) * | 2000-10-13 | 2002-05-30 | Witzgall Hanna Elizabeth | System and method for linear prediction |
JP2005338063A (en) * | 2004-04-28 | 2005-12-08 | Japan Science & Technology Agency | Apparatus for measuring physical characteristics of sample |
CN102590338A (en) * | 2012-01-19 | 2012-07-18 | 清华大学 | Resonance peak-based ultrasonic cavitation state identification method |
EP2927674A1 (en) * | 2014-03-31 | 2015-10-07 | Nederlandse Organisatie voor toegepast- natuurwetenschappelijk onderzoek TNO | Method and apparatus for characterizing a medium using ultrasound measurements |
CN106992011A (en) * | 2017-01-25 | 2017-07-28 | 杭州电子科技大学 | Engineering machinery sound identification method based on MF PLPCC features |
CN107703210A (en) * | 2017-08-31 | 2018-02-16 | 北京金风慧能技术有限公司 | workpiece identification method and device |
CN110797033A (en) * | 2019-09-19 | 2020-02-14 | 平安科技(深圳)有限公司 | Artificial intelligence-based voice recognition method and related equipment thereof |
CN112162041A (en) * | 2020-09-30 | 2021-01-01 | 陕西师范大学 | Method for identifying metal material based on Gaussian distribution of amplitude root mean square value |
Non-Patent Citations (1)
Title |
---|
高猛: "《基于 ARM 的声纹识别模块的软件设计与实现》", 中国优秀硕士学位论文全文数据库, pages 17 - 19 * |
Also Published As
Publication number | Publication date |
---|---|
CN113702513B (en) | 2024-06-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112162041B (en) | Method for recognizing metal material based on Gaussian distribution of amplitude root mean square value | |
CN102636303B (en) | Method for measuring residual stress of thin plating layer based on surface ultrasonic waves | |
CN108490081B (en) | Alloy grain size determination method and system based on multiple parameters | |
CN104965023B (en) | Multi-modal guided wave industrial pipeline diagnostic method | |
CN107121501A (en) | A kind of turbine rotor defect classification method | |
CN104359389A (en) | Pulsed eddy current measuring method for measuring relative wall thickness variable of ferromagnetic component | |
CN109163997A (en) | A kind of rock surface strength detection method based on sonograph deep learning | |
CN110161119B (en) | Wind power blade defect identification method | |
CN1743839A (en) | Structure defect ultrasonic on-line intelligent identifying system and identifying method | |
CN110044533B (en) | Bolt pretightening force monitoring method based on combination of EMD (empirical mode decomposition) energy entropy and support vector machine | |
CN104266579A (en) | Method for extracting signal characteristic quantity for carrying out impulse eddy current detection on ferromagnetic component | |
CN110286155B (en) | Damage detection method and system for multilayer composite material | |
Sun et al. | Development of frequency-mixed point-focusing shear horizontal guided-wave EMAT for defect inspection using deep neural network | |
CN103412047A (en) | Method for identifying true and false metal by ultrasonic non-destructive detection method | |
CN109632973A (en) | A kind of ultrasound echo signal extracting method based on Based on Multiscale Matching tracking | |
CN114330435A (en) | Composite material defect detection method based on dynamic weight wavelet coefficient deep residual network | |
CN107727749A (en) | A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm | |
Lei et al. | Ultrasonic pig for submarine oil pipeline corrosion inspection | |
CN110542723A (en) | guided wave signal sparse decomposition and damage positioning-based two-stage damage position identification method | |
CN2809640Y (en) | Ultrasonic online intelligent recognition system for structural defect | |
CN113702513A (en) | Method for identifying metal material based on prediction function model | |
CN103954628B (en) | Ensemble empirical mode decomposition (EEMD) and approximate entropy combined steel tube damage monitoring method | |
CN106680375B (en) | For determining the Air Coupling supersonic detection method of the elasticity modulus of material | |
CN110308205B (en) | Metal classification method based on EMD generalized phase permutation entropy algorithm | |
CN107515251A (en) | A kind of detecting system for metal material |
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 |