CN117129565A - Concrete filled steel tube void knocking force detection method based on energy ratio and GWO-SVM - Google Patents
Concrete filled steel tube void knocking force detection method based on energy ratio and GWO-SVM Download PDFInfo
- Publication number
- CN117129565A CN117129565A CN202311069317.6A CN202311069317A CN117129565A CN 117129565 A CN117129565 A CN 117129565A CN 202311069317 A CN202311069317 A CN 202311069317A CN 117129565 A CN117129565 A CN 117129565A
- Authority
- CN
- China
- Prior art keywords
- void
- knocking
- steel pipe
- energy
- energy ratio
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000011800 void material Substances 0.000 title claims abstract description 85
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 82
- 239000010959 steel Substances 0.000 title claims abstract description 82
- 238000001514 detection method Methods 0.000 title abstract description 31
- 230000007547 defect Effects 0.000 claims abstract description 43
- 238000000034 method Methods 0.000 claims abstract description 42
- 241000282461 Canis lupus Species 0.000 claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000012706 support-vector machine Methods 0.000 claims abstract description 22
- 238000012360 testing method Methods 0.000 claims abstract description 20
- 238000013145 classification model Methods 0.000 claims abstract description 14
- 238000005457 optimization Methods 0.000 claims abstract description 5
- 238000001228 spectrum Methods 0.000 claims description 20
- 238000010079 rubber tapping Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000005070 sampling Methods 0.000 claims description 8
- 239000000463 material Substances 0.000 claims description 4
- 230000010354 integration Effects 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 claims description 3
- 229920005989 resin Polymers 0.000 claims description 3
- 239000011347 resin Substances 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 description 6
- 238000010801 machine learning Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 238000001931 thermography Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000009659 non-destructive testing Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- XAGFODPZIPBFFR-UHFFFAOYSA-N aluminium Chemical compound [Al] XAGFODPZIPBFFR-UHFFFAOYSA-N 0.000 description 1
- 229910052782 aluminium Inorganic materials 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000005056 compaction Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 239000007822 coupling agent Substances 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 239000003822 epoxy resin Substances 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000011229 interlayer Substances 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 229920000647 polyepoxide Polymers 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 239000002023 wood Substances 0.000 description 1
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
- G01N29/4409—Processing the detected response signal, e.g. electronic circuits specially adapted therefor by comparison
- G01N29/4418—Processing the detected response signal, e.g. electronic circuits specially adapted therefor by comparison with a model, e.g. best-fit, regression analysis
-
- 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
- G01N29/4445—Classification of defects
Landscapes
- Physics & Mathematics (AREA)
- Biochemistry (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Acoustics & Sound (AREA)
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
Abstract
The method for detecting the concrete filled steel tube void knocking force based on the energy ratio and GWO-SVM comprises the following steps: 1. preparing a non-hollow steel pipe concrete member and a plurality of hollow steel pipe concrete members with different height hollow defects, and arranging knocking measuring points outside each hollow steel pipe concrete member; 2. respectively knocking the non-hollow steel pipe concrete member and knocking measuring points of each hollow steel pipe concrete member for multiple times, and respectively obtaining original knocking force pulse signals of the non-hollow steel pipe concrete member and the knocking measuring points for multiple times; 3. extracting the energy ratio value characteristic of each original knocking force pulse signal to obtain a characteristic data set; 4. dividing the obtained characteristic data set into a training set and a testing set, wherein the training set is used for training the void classification model through a gray wolf algorithm optimization support vector machine, and the testing set is used for testing the accuracy of the trained model on the void classification. The method can effectively improve the robustness of the classification model, avoid the lack of the robustness of the sound vibration signal to the noise and simplify the detection equipment.
Description
Technical Field
The invention belongs to the field of nondestructive testing of a concrete filled steel tube structure, and particularly relates to a concrete filled steel tube void knocking force detection method based on an energy ratio and a GWO-SVM (gray wolf algorithm optimization support vector machine).
Background
The steel pipe Concrete (English name is Concrete-filled steel tube, CFST for short) structure has the characteristics of high strength, high ductility, high energy absorption capacity and the like, and is widely used in high-rise buildings and large-span structures. The special structural characteristics of the CFST structure improve the compressive strength of core concrete, and effectively delay the local buckling of the external steel pipe under compression, so that better compressive resistance and earthquake resistance are obtained. However, in practical engineering, environmental temperature differences and construction processes often cause separation of the inner wall of the steel pipe from the core concrete, thereby forming void defects. Studies have shown that when the void fraction exceeds 0.05%, the strength, stiffness and ductility of the CFST structure will be significantly reduced. Therefore, the void detection is an essential link for the health monitoring and maintenance of the CFST structure.
The CFST structure is preferably detected by a nondestructive detection method (Non destructive testing, NDT) including an infrared thermal imaging method, an ultrasonic method and an impact echo method. The methods achieve certain effects in aerospace, bridge and other applications. However, they also suffer from certain drawbacks. For example, patent number CN201911296095.5 discloses a method for identifying the void defect of a void concrete filled steel tube member based on infrared thermal imaging, which is limited by the structural surface defect, auxiliary thermal instruments are required to be erected for interlayer and void defect detection, the detection process is tedious and time-consuming, the influence of environmental temperature is large, and the detection equipment is expensive. Patent number CN202211020985.5 discloses a method and equipment for detecting and evaluating the compaction state in a steel tube concrete arch bridge tube based on ultrasonic waves, wherein transducers are required to be symmetrically arranged on the surface of the structure, a coupling agent is required to be used, and the noise resistance is poor. Patent number CN201610340911.8 discloses a method for detecting concrete filled steel tube void defect based on HHT feature extraction, which uses an impact echo method, but the impact echo method has the disadvantage that complex stress wave of non-uniform medium is difficult to process, and the detection accuracy depends on the coupling condition between the transducer and the structure. Therefore, the simple, efficient and accurate CFST structure void defect NDT technology is developed, and the method has important engineering significance for improving the structural health monitoring level and guaranteeing the structural safety.
Tapping is commonly used for the diagnosis of internal breaks and cavities in structures. The traditional knocking method uses coins or light hammers to knock the surface of the structure, and judges the damage inside the structure according to the perception of the detector to the echo. The method has high detection speed, can be used in any environment, but the evaluation index is subjective. Along with the development of intelligent materials and sensing technology, the precise instrument replaces subjective receiving of echo by people, the knocking method is greatly improved, and then the detection mode of collecting acoustic wave by a sound pressure sensor and collecting force echo by a piezoelectric technology is developed. However, by means of simple analysis of signals by instruments and meters, an accurate damage identification model cannot be provided, and detection accuracy and efficiency are to be improved. In recent years, machine learning algorithms have been continuously developed and improved, and excellent application effects are obtained in various fields. The intelligent detection method based on machine learning provides a new thought and power for the development of the NDT technology. Many documents show that the detection of bolt loosening, internal voids of structures, wood voids, concrete moisture content, and the like by combining a machine learning algorithm with a knock sound wave signal is efficient and accurate. However, the knocking sound wave signal lacks the capability of resisting noise interference, and various noises are unavoidable in practical engineering. The knocking force echo signal depends on the impedance of the tested structure and the used hammer, and has excellent noise resistance compared with the sound wave signal due to the characteristic of transient pulse. The existing research shows that the knocking force echo method can effectively check the internal damage of the thin wall or the layered structure, but the accurate damage evaluation model research is lacking at present, and the knocking force echo method is particularly suitable for the void detection of the CFST structure.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a steel tube concrete void tapping force detection method based on an energy ratio and GWO-SVM, and aims to identify whether a CFST structure has void and a void category by combining a tapping force signal and a machine learning algorithm.
In order to achieve the above object, the present invention is specifically as follows:
the method for detecting the concrete filled steel tube void knocking force based on the energy ratio and GWO-SVM comprises the following steps:
step 1, preparing a non-hollow steel pipe concrete member and a plurality of hollow steel pipe concrete members with different height hollow defects, and respectively arranging knocking measuring points outside each hollow steel pipe concrete member;
step 2, knocking the non-hollow steel pipe concrete member and knocking measuring points of each hollow steel pipe concrete member in the step 1 for multiple times respectively, and obtaining original knocking force pulse signals of the non-hollow steel pipe concrete member and the knocking measuring points for multiple times respectively;
step 3, extracting the energy ratio value characteristics of each original knocking force pulse signal in the step 2 to obtain a characteristic data set;
and 4, dividing the characteristic data set obtained in the step 3 into a training set and a testing set, wherein the training set is used for training the void classification model through a gray wolf algorithm optimization support vector machine, and the testing set is used for testing the accuracy of the trained void classification model on the void classification.
Further, the preparation method of the artificial steel tube void component with the void defects at different heights comprises the following steps:
step 1-1, preparing a plurality of void defect molds with different heights, respectively marking contour lines of the void defect molds with one height on the inner and outer surfaces at the longitudinal middle points of each steel pipe correspondingly, and marking the center point of the contour lines on the outer surface of the steel pipe as a knocking measuring point;
and 1-2, correspondingly fixing the void defect dies with corresponding heights in the outline of the void defect die marked with the inner surface of the steel pipe with corresponding heights, pouring concrete from top to bottom after the bottom end of each steel pipe is provided with a steel plate, and carrying out outdoor regular watering maintenance for 28 days after pouring is finished.
Further, the void defect mold is arc-shaped and is made of resin materials.
Further, the method for extracting the energy ratio value feature in the step 3 includes the following steps:
step 3-1: performing discrete Fourier transform on the original knocking force pulse signals acquired in the step 2;
step 3-2: solving an energy spectrum density value ESD of an original knocking force pulse signal after discrete Fourier transform;
step 3-3: taking the logarithm of the energy spectrum density value to calculate as the amplitude of the energy spectrum curve, and giving an energy spectrum curve;
step 3-4: dividing an energy spectrum curve into N subintervals on a selected frequency interval, and calculating the energy integral of each subinterval;
step 3-5: the energy integral of the first subinterval is recorded as A, and the energy integral of the rest subintervals is recorded as B i I=1, 2,..n-1, the energy ratio value ER is calculated.
Further, the discrete Fourier transform formula described in step 3-1 is:
in the formula (1), X s (ω) represents the discrete fourier transformed tapping force pulse signal; x (n) represents an original tapping force pulse signal; ω represents frequency; n represents the nth sampling point; t (T) s Representing a sampling point time interval;
the calculation formula of the energy spectrum density value ESD in the step 3-2 is as follows:
ESD＝|X s (ω)| 2 (2)
the calculation formula of the energy integral of each subinterval in the step 3-4 is as follows:
in the formula (3), a and b respectively represent upper and lower limit frequencies of the energy integration interval;
the calculation formula of the energy ratio value in the step 3-5 is as follows:
in the formula (4), ER represents an energy ratio value; b (B) i Representing the energy integral of the remainder interval.
Further, the method for training the void classification model by optimizing the support vector machine by the training set in the step 4 through the wolf algorithm comprises the following steps:
step 4-1, initializing the positions of the gray wolf population after inputting a training set and a testing set;
step 4-2, defining a gray wolf algorithm to optimize a support vector machine fitness function, and initializing support vector machine parameters C and gamma;
step 4-3, calculating individual fitness values of the wolf clusters and marking the first three bits with the largest fitness values as alpha, beta and delta from large to small respectively;
step 4-4, updating the positions of the rest wolf individuals through the positions of alpha, beta and delta;
step 4-5, judging whether the maximum iteration times are reached, if the maximum iteration times are reached, stopping searching, returning to the position of the alpha wolf, respectively assigning the position value of the alpha wolf to the parameters C and gamma of the support vector machine to obtain the optimized parameters C and gamma of the support vector machine, and outputting a trained void classification model; otherwise, returning to the step 4-3 to continue execution.
Further, the formula for updating the positions of the rest wolf group individuals from the positions of alpha, beta and delta in the step 4-4 is as follows:
in the formula (6), X 1 、X 2 、X 3 Positions of alpha, beta and delta wolf respectively, D α 、D β 、D δ The distances between the prey and alpha, beta, delta wolf, respectively.
THE ADVANTAGES OF THE PRESENT INVENTION
1. The steel tube concrete void tapping force detection method based on the energy ratio and the GWO-SVM combines a tapping force signal and a machine learning algorithm, verifies the effectiveness and the accuracy through experiments, and can efficiently detect the void inside the CFST structure in actual engineering.
2. The detection method can construct a feature set with high latitude, ensure the training depth of the gray wolf algorithm optimization support vector machine model, simultaneously compress the feature calculation value to be between 0 and 1, and improve the generalization capability of the model.
3. Compared with a thermal imaging detection method, the method does not need to erect auxiliary thermal instruments, reduces labor cost and simplifies detection procedures; compared with an ultrasonic detection method, the detection method of the knocking force signal has excellent anti-noise capability, and can perform detection work in a noise environment.
Drawings
FIG. 1 is a flow chart of a method for detecting the concrete filled steel tube void striking force based on the energy ratio and GWO-SVM.
Fig. 2 is a schematic diagram of a contour line and a knocking point of a void defect mold on the outer surface of a steel pipe of the concrete filled steel tube member in the detection method of fig. 1.
FIG. 3 is a schematic diagram of a void defect mold of different heights in the inspection method of FIG. 1.
FIG. 4 is a schematic view of the void defect mold of FIG. 3 at different heights secured to a void concrete filled steel tube member.
Fig. 5 is a flow chart of a method of extracting energy ratio features in the detection method of fig. 1.
FIG. 6 is a schematic diagram of the energy spectrum interval division of the energy ratio feature extraction method of FIG. 5.
Fig. 7 is a schematic diagram of the operation of the gray wolf algorithm optimized support vector machine in the detection method of fig. 1.
FIG. 8 is a confusion matrix for the trained deairing classification model of FIG. 7 performing a test of two deairing classification randomly. Wherein fig. 8 (a) is a first void classification test; fig. 8 (b) is a second classification test.
In the figure: 1: a steel pipe; 2: a contour line of the defect mold is emptied; 3: knocking a measuring point; 4: concrete; 5: a defect mold is emptied; 6: removing the hollow steel pipe concrete member; 7. sample zero; 8. sample one; 9. sample two; 10. sample III; 11. sample IV; 12. sample five.
Detailed Description
The invention is further illustrated in the following drawings and detailed description, which are not intended to limit the scope of the invention.
As shown in fig. 1, the concrete filled steel tube void and knocking force detection method based on the energy ratio and GWO-SVM provided by the embodiment comprises the following steps:
and step 1, preparing a non-hollow steel pipe concrete member and five hollow steel pipe concrete members 6 with different height hollow defects, and respectively arranging knocking measuring points 3 outside each hollow steel pipe concrete member 6.
Specifically, according to the research content of the invention, a steel pipe 1, a steel plate arranged at the bottom end of the steel pipe 1 and concrete filled in the steel pipe 1 are adopted to prepare a non-void steel pipe concrete member; and preparing five void defect molds 5 with different heights by adopting a resin material, wherein the width W of each void defect mold 5 is 50mm, the thickness B is 3mm, the heights H are 10mm, 20mm, 30mm, 40mm and 50mm respectively, and the five void defect molds 5 are shown in figure 3.
As shown in fig. 2, the inner and outer surfaces at the longitudinal midpoints of the five steel pipes 1 respectively correspond to the contour lines 2 of the void defect molds with the marking heights H of 10mm, 20mm, 30mm, 40mm and 50mm, and the center points of the contour lines 2 of the void defect molds are respectively marked on the outer surfaces of the five steel pipes 1 as knocking measuring points 3;
and 1-2, correspondingly fixing the void defect dies 5 with the heights H of 10mm, 20mm, 30mm, 40mm and 50mm in the outline 2 of the void defect die marked with the inner surface of the steel pipe 1 with the corresponding height by using epoxy resin, respectively installing steel plates at the bottom ends of the steel pipes 1, pouring concrete 4 from top to bottom, properly vibrating during pouring to ensure the compactness of a non-void area, and periodically watering and curing for 28 days outdoors after pouring to obtain five void steel pipe concrete members 6 with void defects with different heights.
The non-void concrete filled steel tube member and the void concrete filled steel tube member 6 with five void defects at different heights are sequentially marked as a sample zero 7, a sample one 8, a sample two 9, a sample three 10, a sample four 11 and a sample five 12, wherein the sample zero 7 is the non-void concrete filled steel tube member, and the samples one 8 to five 12 are the void concrete filled steel tube members 6 with the void defects at different heights. Sample zero 7 through sample five 12 as shown in fig. 4.
And 2, respectively carrying out 100 times of knocking on the non-hollow steel pipe concrete member and the knocking measuring points of each hollow steel pipe concrete member 6 in the step 1 through the knocking force hammer, respectively collecting original knocking force pulse signals of the non-hollow steel pipe concrete member and the knocking measuring points 3 times of knocking 100 times through a data collector, collecting and storing the original knocking force pulse signals in a notebook computer, and collecting 600 samples in total.
The knocking force hammer is derived from Beijing eastern vibration and noise technology institute, and is of the model IEPE, and mainly comprises an aluminum hammer head, a force sensor and a connecting handle, wherein the measuring range is 25kN, and the sensitivity is 0.2mV/N.
The data acquisition instrument is derived from Beijing eastern vibration and noise technology institute, the model is INV3062V, is connected with a notebook computer and is used for collecting electric signals of the force sensor, the electric signals are converted into mechanical signals by a computer end, and the sampling frequency is set to be 51.2kHz.
Step 3, extracting the energy ratio value characteristics of each original knocking force pulse signal in the step 2 to obtain a characteristic data set;
as shown in fig. 5, the method for extracting the energy ratio value feature includes the following steps:
step 3-1: performing discrete Fourier transform on the original knocking force pulse signals acquired in the step (2) through the step (1);
in the formula (1), X s (ω) represents the discrete fourier transformed tapping force pulse signal; x (n) represents the original striking forceA pulse signal; ω represents frequency; n represents the nth sampling point; t (T) s Representing the sampling point time interval.
Step 3-2: solving an energy spectrum density value ESD of an original knocking force pulse signal after discrete Fourier transform, wherein a calculation formula of the energy spectrum density value ESD is as follows:
ESD＝|X s (ω)| 2 (2)
step 3-3: taking the logarithm of the energy spectrum density value to calculate as the amplitude of the energy spectrum curve, and giving an energy spectrum curve; the calculation formula of the energy integral of each subinterval is as follows:
in the formula (3), a and b represent upper and lower limit frequencies of the energy integration interval to be obtained, respectively.
Step 3-4: dividing an energy spectrum curve into N subintervals on a selected frequency interval, and calculating the energy integral of each subinterval;
in this example, in order to ensure that the frequency range of feature extraction contains enough sensitive information, the frequency range of 0-8kHz is taken as the frequency range of feature calculation, and the energy spectrum curve is divided at intervals of 0.5kHz, as shown in fig. 6, for a total of 16 subintervals;
step 3-5: the energy integral of the first subinterval is recorded as A, and the energy integral of the rest subintervals is recorded as B i I=1, 2,..n-1, calculating an energy ratio value ER, the energy ratio value having a calculation formula:
15×600 feature data sets were calculated by the above procedure, and the same type of sample was assigned the same label (label=0, 1,2,3,4, 5). Non-void concrete filled steel tube member of sample zero 7 is corresponded to label 0, label 1 is corresponded to sample one 8 of height 10mm, label 2 is corresponded to sample two of height 20mm 9, label 3 is corresponded to sample three of height 30mm 10, label 4 is corresponded to sample four of height 40mm 11, label 5 is corresponded to void concrete filled steel tube member 6 of sample five of height 50mm 12, and sample zero 7 to sample five 12 are respectively expressed as by sample matrix:
S＝[X 1 X 2 … X 15 Y] (5)
in the formula (5), X i Is a feature vector, and Y is a label vector.
Sample data for samples zero 7 through five 12 are shown in table 1:
table 1 partial sample data of this example
And 4, dividing the 15 multiplied by 600 feature data sets obtained in the step 3 into a training set and a test set by using a hierarchical sampling method, wherein 70% of samples are used as the training set (total 420 sample data), 30% of samples are used as the test set (total 180 sample data), the training set is used for training a void classification model by optimizing a support vector machine through a gray wolf algorithm, and the test set is used for testing the accuracy of the trained model on the void classification.
As shown in fig. 7, the method for training the void classification model by optimizing the support vector machine through the wolf algorithm by the training set comprises the following steps:
step 4-1, initializing the positions of the gray wolf population after inputting a training set and a testing set;
step 4-2, defining a gray wolf algorithm to optimize a support vector machine fitness function, and initializing support vector machine parameters C and gamma;
step 4-3, calculating individual fitness values of the wolf clusters and marking the first three bits with the largest fitness values as alpha, beta and delta from large to small respectively;
and 4-4, updating the positions of the rest wolf individuals through the positions of alpha, beta and delta, wherein the specific formula is as follows:
in the formula (6), X 1 、X 2 、X 3 Positions of alpha, beta and delta wolf respectively, D α 、D β 、D δ The distances between the prey and alpha, beta, delta wolf, respectively.
Step 4-5, judging whether the maximum iteration times are reached, if the maximum iteration times are reached, stopping searching, returning to the position of the alpha wolf, respectively assigning the position value of the alpha wolf to the parameters C and gamma of the support vector machine to obtain the optimized parameters C and gamma of the support vector machine, and outputting a trained void classification model; otherwise, returning to the step 4-3 to continue execution.
In this particular embodiment, the test set is randomly subjected to two-time deairing class classification, see fig. 8. In fig. 8 (a), only 1 defect of 10mm depth is erroneously identified by the model as a defect of 20mm depth. In fig. 8 (b), 1 defect 20mm deep is erroneously identified by the model as a defect 10mm deep, and 1 defect 40mm deep is erroneously identified by the model as a defect 50mm deep.
Experiments prove that the average recognition precision of the detection method provided by the invention to the CFST structure void type reaches 98.51 percent, the detection method for the CFST structure void has higher innovation in the process of feature extraction, can efficiently detect the internal void of the CFST structure, simplifies detection equipment, and has application value in practical engineering.
Claims (7)
1. The method for detecting the concrete filled steel tube void knocking force based on the energy ratio and GWO-SVM is characterized by comprising the following steps of:
step 1, preparing a non-hollow steel pipe concrete member and a plurality of hollow steel pipe concrete members with different height hollow defects, and respectively arranging knocking measuring points outside each hollow steel pipe concrete member;
step 2, knocking the non-hollow steel pipe concrete member and knocking measuring points of each hollow steel pipe concrete member in the step 1 for multiple times respectively, and obtaining original knocking force pulse signals of the non-hollow steel pipe concrete member and the knocking measuring points for multiple times respectively;
step 3, extracting the energy ratio value characteristics of each original knocking force pulse signal in the step 2 to obtain a characteristic data set;
and 4, dividing the characteristic data set obtained in the step 3 into a training set and a testing set, wherein the training set is used for training the void classification model through a gray wolf algorithm optimization support vector machine, and the testing set is used for testing the accuracy of the trained void classification model on the void classification.
2. The method for detecting the void tapping force of the concrete-filled steel tube based on the energy ratio and GWO-SVM according to claim 1, wherein the method for preparing the void concrete-filled steel tube member with the void defects at different heights comprises the following steps:
step 1-1, preparing a plurality of void defect molds with different heights, respectively marking contour lines of the void defect molds with one height on the inner and outer surfaces at the longitudinal middle points of each steel pipe correspondingly, and marking the center point of the contour lines on the outer surface of the steel pipe as a knocking measuring point;
and 1-2, correspondingly fixing the void defect dies with corresponding heights in the outline of the void defect die marked with the inner surface of the steel pipe with corresponding heights, pouring concrete from top to bottom after the bottom end of each steel pipe is provided with a steel plate, and carrying out outdoor regular watering maintenance for 28 days after pouring is finished.
3. The method for detecting the void knocking force of the concrete filled steel tube based on the energy ratio and the GWO-SVM according to claim 2, wherein the void defect die is arc-shaped and is made of a resin material.
4. The method for detecting the concrete filled steel tube void tapping force based on the energy ratio and GWO-SVM according to claim 1, wherein the method for extracting the energy ratio value characteristics in the step 3 comprises the following steps:
step 3-1: performing discrete Fourier transform on the original knocking force pulse signals acquired in the step 2;
step 3-2: solving an energy spectrum density value ESD of an original knocking force pulse signal after discrete Fourier transform;
step 3-3: taking the logarithm of the energy spectrum density value to calculate as the amplitude of the energy spectrum curve, and giving an energy spectrum curve;
step 3-4: dividing an energy spectrum curve into N subintervals on a selected frequency interval, and calculating the energy integral of each subinterval;
step 3-5: the energy integral of the first subinterval is recorded as A, and the energy integral of the rest subintervals is recorded as B i I=1, 2,..n-1, the energy ratio value ER is calculated.
5. The method for detecting the concrete filled steel tube void and knock force based on the energy ratio and GWO-SVM according to claim 4, wherein the discrete Fourier transform formula in the step 3-1 is as follows:
in the formula (1), X s (ω) represents the discrete fourier transformed tapping force pulse signal; x (n) represents an original tapping force pulse signal; ω represents frequency; n represents the nth sampling point; t (T) s Representing a sampling point time interval;
the calculation formula of the energy spectrum density value ESD in the step 3-2 is as follows:
ESD＝|X s (ω)| 2 (2)
the calculation formula of the energy integral of each subinterval in the step 3-4 is as follows:
in the formula (3), a and b respectively represent upper and lower limit frequencies of the energy integration interval;
the calculation formula of the energy ratio value in the step 3-5 is as follows:
in formula (4), ER representsAn energy ratio value; b (B) i Representing the energy integral of the remainder interval.
6. The method for detecting the concrete filled steel tube void and knocking force based on the energy ratio and the GWO-SVM according to claim 4, wherein the training set in the step 4 is used for training a void classification model by optimizing a support vector machine through a gray wolf algorithm, and comprises the following steps:
step 4-1, initializing the positions of the gray wolf population after inputting a training set and a testing set;
step 4-2, defining a gray wolf algorithm to optimize a support vector machine fitness function, and initializing support vector machine parameters C and gamma;
step 4-3, calculating the fitness value of the wolf population individuals and respectively marking the first three bits with the maximum fitness value as alpha, beta and delta from large to small;
step 4-4, updating the positions of the rest wolf individuals through the positions of alpha, beta and delta;
step 4-5, judging whether the maximum iteration times are reached, if the maximum iteration times are reached, stopping searching, returning to the position of the alpha wolf, respectively assigning the position value of the alpha wolf to the parameters C and gamma of the support vector machine to obtain the optimized parameters C and gamma of the support vector machine, and outputting a trained void classification model; otherwise, returning to the step 4-3 to continue execution.
7. The method for detecting the concrete filled steel tube void tapping force based on the energy ratio and GWO-SVM according to claim 6, wherein the formula for updating the positions of the rest wolf group individuals in the positions of alpha, beta and delta in the step 4-4 is as follows:
in the formula (6), X 1 、X 2 、X 3 Positions of alpha, beta and delta wolf respectively, D α 、D β 、D δ The distances between the prey and alpha, beta, delta wolf, respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311069317.6A CN117129565A (en) | 2023-08-23 | 2023-08-23 | Concrete filled steel tube void knocking force detection method based on energy ratio and GWO-SVM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311069317.6A CN117129565A (en) | 2023-08-23 | 2023-08-23 | Concrete filled steel tube void knocking force detection method based on energy ratio and GWO-SVM |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117129565A true CN117129565A (en) | 2023-11-28 |
Family
ID=88860890
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311069317.6A Pending CN117129565A (en) | 2023-08-23 | 2023-08-23 | Concrete filled steel tube void knocking force detection method based on energy ratio and GWO-SVM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117129565A (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108648763A (en) * | 2018-04-04 | 2018-10-12 | 深圳大学 | Personal computer usage behavior monitoring method based on acoustic channels and system |
| CN110390957A (en) * | 2018-04-19 | 2019-10-29 | 半导体组件工业公司 | Method and apparatus for speech detection |
| CN110469832A (en) * | 2013-01-07 | 2019-11-19 | 玻点太阳能有限公司 | For selectively being used to include the system and method for improving oil recovery process from solar thermal collector and heater production steam |
| CN111476263A (en) * | 2019-12-27 | 2020-07-31 | 江苏科技大学 | Bearing defect identification method based on SDAE and improved GWO-SVM |
| CN114462627A (en) * | 2022-03-16 | 2022-05-10 | 兰州理工大学 | Method for diagnosing abnormity of top-blown smelting system based on Hui wolf algorithm and support vector machine |
| CN115356397A (en) * | 2022-08-23 | 2022-11-18 | 福建农林大学 | Steel pipe concrete structure void defect diagnosis method and device based on sound signals |
| CN115616078A (en) * | 2022-10-19 | 2023-01-17 | 广西大学 | Concrete filled steel tube void detection method based on sound vibration characteristics |
| CN116401571A (en) * | 2022-12-05 | 2023-07-07 | 广西大学 | Concrete filled steel tube abnormal shape void identification method based on knocking sound wave and MiniRoccket |
-
2023
- 2023-08-23 CN CN202311069317.6A patent/CN117129565A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110469832A (en) * | 2013-01-07 | 2019-11-19 | 玻点太阳能有限公司 | For selectively being used to include the system and method for improving oil recovery process from solar thermal collector and heater production steam |
| CN108648763A (en) * | 2018-04-04 | 2018-10-12 | 深圳大学 | Personal computer usage behavior monitoring method based on acoustic channels and system |
| CN110390957A (en) * | 2018-04-19 | 2019-10-29 | 半导体组件工业公司 | Method and apparatus for speech detection |
| CN111476263A (en) * | 2019-12-27 | 2020-07-31 | 江苏科技大学 | Bearing defect identification method based on SDAE and improved GWO-SVM |
| CN114462627A (en) * | 2022-03-16 | 2022-05-10 | 兰州理工大学 | Method for diagnosing abnormity of top-blown smelting system based on Hui wolf algorithm and support vector machine |
| CN115356397A (en) * | 2022-08-23 | 2022-11-18 | 福建农林大学 | Steel pipe concrete structure void defect diagnosis method and device based on sound signals |
| CN115616078A (en) * | 2022-10-19 | 2023-01-17 | 广西大学 | Concrete filled steel tube void detection method based on sound vibration characteristics |
| CN116401571A (en) * | 2022-12-05 | 2023-07-07 | 广西大学 | Concrete filled steel tube abnormal shape void identification method based on knocking sound wave and MiniRoccket |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101561379B (en) | Tap-scanning method for detecting structural damages | |
| CN104634879B (en) | A kind of Metal Material Fatigue load test and fatigue damage Nondestructive Detection method | |
| CN102928514B (en) | Frequency characteristic-based nondestructive detection method of stress waves of wood | |
| WO2006011484A1 (en) | Buried pipe examining method | |
| CN107505399A (en) | A kind of intelligent reinforced bar sleeve grouting defect inspection method based on vocal print feature | |
| CN116401571A (en) | Concrete filled steel tube abnormal shape void identification method based on knocking sound wave and MiniRoccket | |
| CN103837603B (en) | Method for using residual stress gradient calibrating block | |
| CN115616078B (en) | Steel pipe concrete void detection method based on sound vibration characteristics | |
| CN114692302B (en) | Fatigue crack detection method and system based on Gaussian mixture model | |
| CN115659737A (en) | Beam member damage identification method based on curvature mode and discrete wavelet transform | |
| Liu et al. | Depth detection of void defect in sandwich-structured immersed tunnel using elastic wave and decision tree | |
| CN101701882B (en) | Fast identification method of tower structure stiffness | |
| CN1258078C (en) | Nondestructive pressure testing method and its device based on Rayleigh surface wave | |
| CN112014471B (en) | Plate structure multi-mode lamb wave topological gradient imaging method based on virtual sensor | |
| CN112285651A (en) | Method and system for identifying position of composite impact source based on feature extraction matrix similarity | |
| CN111443131B (en) | Method for detecting grouting compactness of steel bar sleeve | |
| CN117129565A (en) | Concrete filled steel tube void knocking force detection method based on energy ratio and GWO-SVM | |
| Zhong et al. | A composite beam integrating an in-situ FPCB sensor membrane with PVDF arrays for modal curvature measurement | |
| KR102611457B1 (en) | AI-based defect detection system inside concrete members | |
| CN115183965B (en) | Tunnel lining earthquake accumulated damage evaluation method suitable for vibrating table test | |
| Gao et al. | Percussion-based concrete fiber content recognition using homologous heterogeneous data fusion and denoising deep learning network | |
| Carreira et al. | Determination of the e/g ratio of wood logs using transverse vibration | |
| CN113237582A (en) | Wall internal stress detection method and detection system for engineering acceptance | |
| CN111597736A (en) | Construction quality detection method for assembled concrete beam column joint | |
| CN117470752B (en) | Method for detecting prestress grouting content in steel pipe truss body |
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 |