CN106645391A - Multi-frequency eddy current testing system and method for evaluating carbon fiber plate defect depth - Google Patents

Abstract

The invention discloses a multi-frequency eddy current detection system and detection method for evaluating the depth of carbon fiber plate defects, wherein the multi-frequency eddy current detection system includes a composite material to be tested, a multi-frequency sinusoidal excitation signal generation module, an eddy current sensing unit, and a signal processing unit. And separation module, signal acquisition module and depth inversion module. The eddy current sensing unit is composed of transmission cable, probe control mechanism, eddy current excitation coil and eddy current detection coil. The signal processing and separation module includes a preamplifier and a multi-frequency detection signal separation module. . The multi-frequency eddy current detection system and detection method used to evaluate the depth of defects in carbon fiber boards of the present invention obtain information on the size of defects by measuring the peak induced voltage and peak frequency of defects at different depths, thereby realizing the evaluation of the depth of defects in composite boards.

Description

A multi-frequency eddy current inspection system and inspection for evaluating the depth of carbon fiber plate defects method

Technical field:

The invention relates to the technical field of non-destructive testing of CFRP composite materials, in particular to a multi-frequency eddy current testing system and testing method for evaluating the depth of defects in carbon fiber plates.

Background technique:

Carbon fiber composite material is a new type of structural material that emerged in the mid-1960s. It has superior properties unmatched by other composite materials and has been widely used in civil industry, military, aerospace and other fields. In order to ensure that composite components can meet the requirements of all directions, monitoring the internal quality of composite structures has attracted more and more attention.

At present, the methods for non-destructive testing of CFRP composite materials mainly include: ultrasonic, acoustic resistance, ray and infrared method detection, but these methods have their own shortcomings for the detection of carbon fiber composite materials, some require coupling agent, and some equipment is large in size and expensive Expensive and can only be used for off-site workshop testing.

Eddy current testing is an efficient non-contact non-destructive testing technology that requires no couplant and has a high degree of automation. Since eddy current testing is realized based on the principle of electromagnetic induction, and the probe does not need to be in contact with the tested object, it will not have any impact on the structure. Conventional eddy current testing technology uses a single frequency to work, and the amount of information obtained is limited, which cannot meet the evaluation of the depth of defects. According to the characteristics of the detected object, multi-frequency eddy current detection technology applies multiple frequencies to excite the eddy current probe at the same time, obtains the different responses of the parameters to be tested at each frequency according to the skin effect, and extracts the characteristic parameters through the quadrature lock-in amplifier and the least square method. , and then the depth information of the defect can be inverted according to the amplitude of the multi-frequency eddy current signal, which provides a basis for the quality control of composite components.

Invention content:

In view of the above problems, the purpose of the present invention is to propose a multi-frequency eddy current detection system and detection method for evaluating the depth of carbon fiber plate defects. According to the skin effect, the method achieves the recognition of CFRP composite material performance and damage by applying , easy to operate, intuitive and easy to understand test results, high sensitivity, suitable for online and in-service testing.

The present invention adopts the following technical scheme: a multi-frequency eddy current detection system for evaluating the depth of carbon fiber plate defects, which is characterized in that it includes a composite material to be tested, a multi-frequency sinusoidal excitation signal generation module, an eddy current sensing unit, a transmission cable, Signal processing and separation module, signal acquisition module and depth inversion module. The eddy current sensing unit is composed of probe controller, eddy current excitation coil and eddy current detection coil. The signal processing and separation module includes preamplifier and multi-frequency detection signal separation module, the output of the multi-frequency sinusoidal excitation signal generation module is sent to the eddy current excitation coil and the multi-frequency detection signal separation module respectively by the transmission cable, the probe control mechanism is connected with the eddy current excitation coil and the eddy current detection coil, and the output signal of the eddy current detection coil is passed through After the amplified output of the preamplifier is connected to one input of the multi-frequency detection signal separation module, the signal acquisition module collects the signal output by the multi-frequency detection signal separation module and sends it to the depth inversion module.

Further, the pre-amplifier amplifies and filters the electrical signal output by the detection coil and outputs it. The multi-frequency detection signal separation module receives the electrical signal output by the pre-amplifier on the one hand, and receives the multi-frequency sinusoidal The signal source reference signal output by the excitation signal generation module, if the detection signal and the reference signal have the same frequency, the multi-frequency detection signal separation module outputs a low-frequency signal after low-pass filtering, otherwise, there is no output signal.

Further, the multi-frequency detection signal separation module is implemented by a quadrature lock-in amplifier module, which includes a phase-sensitive detector and a low-pass filter, and the signal output by the eddy current detection coil is amplified together with the reference signal Send it to a phase-sensitive detector, and use a low-pass filter to filter out high-frequency items to obtain the amplitude and phase information of the detection signal.

Further, the signal acquisition module sends the separated output analog voltage signal to the depth inversion module for processing and analysis through A/D conversion, and extracts the maximum value ΔV of the induced voltage variation and the frequency f _g corresponding to the peak point .

Further, the depth inversion module establishes the maximum value ΔV of the voltage change and the frequency corresponding to the peak point through the backpropagation neural network according to the extracted maximum value ΔV of the induced voltage change and the frequency f _g corresponding to the peak point The relationship between f _g and the corresponding defect depth D: D=F(ΔV, f _g ).

Further, the transmission cable adopts a coaxial cable, which is composed of an inner conductor, an insulating layer, an aluminum foil shield, and a braided shielding box sheath, the length of which is less than 1m, and the parasitic capacitance is less than 100pF/m.

Further, both the eddy current excitation coil and the eddy current detection coil are hollow circular coils arranged left and right to form a transmitting-receiving eddy current probe, and the eddy current excitation coil and the eddy current detection coil are separated from each other.

Further, the probe control mechanism uses a flexible spring mechanism to compress the probe so that the eddy current probe is always attached to the surface of the composite material to be tested, and the eddy current excitation coil converts the electrical signal into a magnetic signal and couples it to the composite material to be tested In the eddy current detection coil, the changing magnetic signal is received and converted into an electrical signal.

The present invention also adopts the following technical solution: a multi-frequency eddy current detection method for evaluating the depth of carbon fiber plate defects, including the following steps:

Step 1: Extract the characteristic parameters of the eddy current signal in the composite material to be tested, obtain the first internal defect by means of a straight line scan, induce the voltage change curve at different frequencies, and extract the peak voltage ΔV and peak value in the amplitude-distance graph The frequency f _g corresponding to the point;

Step 2: Using the same method, extract the peak voltage ΔV ^* caused by the second internal defect and the frequency f _g ^* corresponding to the peak point;

Step 3: Draw a curve for the two sets of extracted signals, the peak voltage ΔV increases with the increase of the defect depth, and the frequency corresponding to the peak point also increases;

Step 4: The depth inversion module is carried out by constructing a neural network according to the extracted maximum value ΔV of the induced voltage variation and the frequency f _g corresponding to the peak point, including:

Construct the input layer, the variables of the input layer include the maximum value ΔV of the output induced voltage change, and the frequency f _g corresponding to the peak point;

Build a hidden layer, the hidden layer includes 3 or more neurons;

Construct the output layer, the output layer is the defect depth D; construct the reverse neural network;

Substituting the maximum value ΔV of the induced voltage change caused by the defect of the composite material to be tested and the frequency f _g corresponding to the peak point into the input layer; substituting the defect depth D into the output layer, performing neural network training, and obtaining training results;

Step 5: According to the training results, establish the relationship between the maximum value of the voltage change ΔV, the frequency f _g corresponding to the peak point, and the corresponding defect depth D: D=F(ΔV, f _g );

Step 6: Use the multi-frequency eddy current testing system to obtain the signal change caused by the defect in the composite material to be tested, extract the characteristic parameters: the maximum value of the induced voltage change ΔV and the frequency f _g corresponding to the peak point, and substitute it into the trained neural network In the input layer, the depth information of the defects in the composite material to be tested is obtained.

Further, the multi-frequency sinusoidal excitation signal generation module is used to generate a multi-frequency signal to excite the eddy current probe, wherein a dual-frequency excitation signal is generated, and the specific steps include:

Two-way DDS chip AD9958 is used to generate four-way output signals;

One of the outputs of each AD9958 is sent to the input of the subsequent multi-frequency detection signal separation module, and the other two outputs are mixed to generate dual-frequency signals through addition;

The dual-frequency signal after frequency mixing is output by the drive circuit to excite the eddy current excitation coil to generate a magnetic field.

The present invention has the following beneficial effects: the multi-frequency eddy current detection method used in the present invention to evaluate the depth of carbon fiber plate defects includes: generation of multi-frequency eddy current signals, acquisition of eddy current signals in composite material structures, separation of multi-frequency detection signals, noise removal and improvement Offset, feature signal extraction, and corresponding peak frequencies at different depths are obtained to evaluate the defect depth in the composite material structure to be tested. Among them, the multi-frequency eddy current signal is generated synchronously, the single-frequency sinusoidal signal is generated by DDS, and the adder is used to mix the signal; the signal transmission adopts coaxial cable to reduce noise interference and transmission loss; the probe control mechanism adopts a flexible spring mechanism Press the probe to make it stick to the surface of the composite material to be tested to reduce the interference of lift-off; the signal processing and separation module outputs the output signal of the detection probe after amplification, filtering, and frequency separation; the signal acquisition and depth inversion module is used for Obtain the output signal of lock-in amplification, and extract the characteristic parameters for defect inversion. The multi-frequency eddy current method obtains the depth information of the defect in the composite material to be tested by constructing the relationship between the maximum value Δv of the induced voltage change caused by the defect, the frequency f _g corresponding to the peak point, and the depth D of the defect. Provide basis for quality control of material components.

Description of drawings:

Fig. 1 is a structural block diagram of a multi-frequency eddy current detection system for evaluating the depth of a carbon fiber plate defect according to the present invention.

Fig. 2 is a schematic diagram of the positions of internal defects of the composite material structure in the present invention.

Fig. 3 is a schematic diagram of multi-frequency eddy current generation and signal processing in the present invention.

Fig. 4 is a line scan result diagram of the induced voltage amplitude variation of the first internal crack defect at different frequencies in the present invention.

Fig. 5 is a graph showing the relationship between characteristic parameters of induced voltage and scanning frequency of internal defects of different depths in the present invention.

Fig. 6 is a schematic diagram of an inverse neural network in the present invention.

detailed description:

The solution of the present invention will be described in detail below in conjunction with the accompanying drawings. The following examples are carried out on the premise of the technical solution of the present invention, and specific implementation schemes and operation processes are provided, but the protection scope of the present invention is not limited to the following examples.

As shown in FIG. 1 , it is a structural block diagram of a multi-frequency eddy current testing system for assessing the depth of a carbon fiber plate defect provided by the present invention. In order to make the above purpose and features of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

The multi-frequency eddy current detection system used to evaluate the depth of carbon fiber plate defects in the present invention includes a composite material to be tested 101, a multi-frequency sinusoidal excitation signal generation module 102, an eddy current sensing unit 103, a transmission cable 104, a signal processing and separation module 108, a signal An acquisition module 111 and a depth inversion module 112 . The eddy current sensing unit 103 is composed of a transmission cable 104 , a probe control mechanism 105 , an eddy current excitation coil 106 and an eddy current detection coil 107 , and the signal processing and separation module 108 includes a preamplifier 109 and a multi-frequency detection signal separation module 110 .

Wherein, the output of the multi-frequency sinusoidal excitation signal generating module 102 is sent to the eddy current excitation coil 106 and the multi-frequency detection signal separation module 110 respectively by the transmission cable 104, and the probe control mechanism 105 is connected with the eddy current excitation coil 106 and the eddy current detection coil 107 for Adjust the position and lift-off of the eddy current probe; the output signal of the eddy current detection coil 107 is amplified and output by the preamplifier 109 and connected to one input of the multi-frequency detection signal separation module 110; the signal acquisition module 111 is collected by the multi-frequency detection signal separation module 110 output The signal is sent to the depth inversion module 112.

The preamplifier 109 amplifies and filters the electrical signal output by the detection coil 107 and outputs it. The multi-frequency detection signal separation module 110 receives the electrical signal output by the preamplifier 109 on the one hand, and receives the electrical signal output by the multi-frequency signal on the other hand. The signal source reference signal output by the sinusoidal excitation signal generation module 102, if the detection signal has the same frequency as the reference signal, then the multi-frequency detection signal separation module 110 outputs a low-frequency signal (approximately a DC signal) after low-pass filtering, otherwise, there is no output Signal.

The multi-frequency detection signal separation module 110 is realized by a quadrature lock-in amplifier module, which mainly includes a phase-sensitive detector and a low-pass filter, and the signal output by the eddy current detection coil 107 is amplified and sent together with the reference signal The phase-sensitive detector uses a low-pass filter to filter out high-frequency items to obtain the amplitude and phase information of the detection signal.

The signal acquisition module 111 sends the separated and output analog voltage signal to the depth inversion module 112 for processing and analysis through A/D conversion, and extracts the maximum value ΔV of the induced voltage variation and the frequency f _g corresponding to the peak point.

The depth inversion module 112 establishes the maximum value ΔV of the voltage change, the frequency _{f g corresponding} to the peak point and the corresponding The relationship between the defect depth D: D=F(ΔV, f _g );

According to the relationship between the voltage variation and the defect depth at different peak frequencies, the defect depth in the composite material structure to be tested is evaluated.

The composite material to be tested 101 takes the form of crack damage as the research object, and the carbon fiber plate is made of crack defects before the test. This embodiment takes two internal crack defects of different depths (respectively a first internal crack defect and a second internal crack defect) as an example, as shown in FIG. 2 .

The multi-frequency sinusoidal excitation signal generation module 102 is used to generate a multi-frequency signal to excite the eddy current probe. As an optional implementation, this embodiment takes the generation of a dual-frequency excitation signal as an example, and the specific steps include:

Two-way DDS chip AD9958 is used to generate four-way output signals;

One of the outputs of each AD9958 is respectively sent to the input of the subsequent multi-frequency detection signal separation module 110, and the other two outputs are mixed to generate a dual-frequency signal through addition;

The mixed frequency signal is output by the driving circuit and then excites the eddy current excitation coil 106 to generate a magnetic field, as shown in FIG. 3 .

The transmission cable 104 adopts a coaxial cable, which is composed of an inner conductor, an insulating layer, an aluminum foil shield, and a braided shielding box sheath. Its length is less than 1m, and its parasitic capacitance is less than 100pF/m, which can effectively reduce noise interference and transmission loss. The signal is transmitted through the transmission cable 104. The probe control mechanism 105 is responsible for adjusting the detection height of the two probes. The flexible spring mechanism is used to compress the probes. Through three-dimensional adjustment, the eddy current probes are always attached to the surface of the composite material to be tested 101, reducing lift-off effect on the test results. The eddy current excitation coil 106 converts the electrical signal into a magnetic signal and couples it into the composite material DUT 101 , and the eddy current detection coil 107 receives the changing magnetic signal and converts it into an electrical signal.

Both the eddy current excitation coil 106 and the eddy current detection coil 107 are hollow circular coils arranged left and right to form a transmitting-receiving eddy current probe; there is a certain distance between the eddy current excitation coil 106 and the eddy current detection coil 107 to prevent the magnetic field of the eddy current detection coil 107 from being affected The interference of the magnetic field of the eddy current excitation coil 106 affects the detection accuracy.

The multi-frequency eddy current detection method for evaluating the depth of carbon fiber plate defects in the present invention comprises the following steps:

Step 1: extract the characteristic parameters of the eddy current signal in the composite material to be tested 101, and obtain the first internal defect 1 by means of linear scanning, as shown in Figure 2, the induced voltage variation curves at different frequencies, the corresponding scanning results are as follows As shown in Figure 4, the peak voltage ΔV in the amplitude-distance graph and the frequency f _g corresponding to the peak point are extracted;

Step 2: Using the same method, extract the peak voltage ΔV ^* caused by the second internal defect 2 and the frequency f _g ^* corresponding to the peak point;

Step 3: Draw a curve for the two sets of extracted signals, as shown in Figure 5, the peak voltage ΔV increases with the increase of the defect depth, and the frequency corresponding to the peak point also increases.

Step 4: The depth inversion module 112 proceeds by constructing a neural network according to the extracted maximum value ΔV of the induced voltage variation and the frequency f _g corresponding to the peak point, specifically including:

Constructing an input layer, the variables of the input layer include the maximum value ΔV of the output induced voltage variation, and the frequency f _g corresponding to the peak point;

Build a hidden layer, the hidden layer includes 3 or more neurons;

Construct the output layer, the output layer is the defect depth D; the constructed reverse neural network is shown in Figure 6;

Substituting the maximum value ΔV of the induced voltage variation caused by the defect of the composite board to be tested and the frequency f _g corresponding to the peak point into the input layer; substituting the defect depth D into the output layer, and performing neural network training to obtain training result;

Step 5: According to the training results, establish the relationship between the maximum value of the voltage change ΔV, the frequency f _g corresponding to the peak point, and the corresponding defect depth D: D=(ΔV, f _g );

Step 6: Use the multi-frequency eddy current testing system to obtain the signal change caused by the defect in the composite material to be tested 101, extract the characteristic parameters: the maximum value of the induced voltage change ΔV and the frequency f _g corresponding to the peak point, and substitute it into the trained neural network In the input layer of the network, the depth information of defects in the composite material to be tested 101 is obtained.

The above is only a preferred embodiment of the present invention, it should be pointed out that for those of ordinary skill in the art, some improvements can also be made without departing from the principle of the present invention, and these improvements should also be regarded as the invention. protected range.

Claims (10)

1. a kind of multi frequency detection system for assessing carbon fiber board depth of defect, it is characterised in that：Including composite Test specimen (101) to be measured, multifrequency sine pumping signal occur module (102), currents sensing unit (103), transmission cable (104), Signal transacting and separation module (108), signal acquisition module (111) and Depth Inverse module (112), the currents sensing unit (103) by probe controlling organization (105), vortex excitation coil (106) and test coil of eddy current (107) constitute, signal transacting and Separation module (108) includes preamplifier (109) and multi-frequency testing signal separation module (110), the multifrequency sine excitation The output of signal generating module (102) delivers to respectively vortex excitation coil (106) and multi-frequency testing signal by transmission cable (104) Separation module (110), probe controlling organization (105) is connected with vortex excitation coil (106) and test coil of eddy current (107), whirlpool Stream detection coil (107) output signal it is premenstrual put amplifier (109) amplify output after with multi-frequency testing signal separation module (110) input all the way is connected, and signal acquisition module (111) gathers the letter exported by multi-frequency testing signal separation module (110) Number and send into Depth Inverse module (112).

2. it is used to as claimed in claim 1 assess the multi frequency detection system of carbon fiber board depth of defect, it is characterised in that： Electric signal that the preamplifier (109) exports detection coil (107) is amplified, export after filtering, the multi-frequency testing Signal separation module (110) on the one hand receives the electric signal exported by preamplifier (109), on the other hand receives by multifrequency just There is the signal source reference signal of module (102) output in string pumping signal, if detection signal and reference signal same frequency, multifrequency Detection signal separation module (110) output obtains low frequency signal after LPF, conversely, non-output signal.

3. it is used to as claimed in claim 2 assess the multi frequency detection system of carbon fiber board depth of defect, it is characterised in that： The multi-frequency testing signal separation module (110) is realized that orthogonal lock-in-amplifier module includes phase sensitivity by orthogonal lock-in-amplifier module Wave detector and low pass filter, the signal of test coil of eddy current (107) output is amplified to send into together phase sensitivity with reference signal Wave detector, is filtered after high frequency item using low pass filter, obtains the amplitude and phase information of detection signal.

4. it is used to as claimed in claim 3 assess the multi frequency detection system of carbon fiber board depth of defect, it is characterised in that： The analog voltage signal Jing A/D exported after separation conversions are sent into Depth Inverse module (112) by the signal acquisition module (111) In processed, analyzed, extract the maximum Δ V and corresponding frequency f of peak point of induced voltage variable quantity_g。

5. it is used to as claimed in claim 4 assess the multi frequency detection system of carbon fiber board depth of defect, it is characterised in that： Maximum Δ V and peak point corresponding frequency f of the Depth Inverse module (112) according to the induced voltage variable quantity for extracting_g, By reverse transmittance nerve network, maximum Δ V, corresponding frequency f of peak point of voltage variety are set up_gWith correspondence defect depth Relation between degree D：D=F (Δ V, f_g)。

6. it is used to as claimed in claim 5 assess the multi frequency detection system of carbon fiber board depth of defect, it is characterised in that： The transmission cable (104) adopts coaxial cable, is made up of inner wire, insulating barrier, aluminum foil shielded, braid shielded box sheath, its Length is less than 1m, and parasitic capacitance is less than 100pF/m.

7. it is used to as claimed in claim 6 assess the multi frequency detection system of carbon fiber board depth of defect, it is characterised in that： Vortex excitation coil (106) and test coil of eddy current (107) adopt open circles annulus, and left-right situs are constituted to be sent out - receiving type eddy current probe is penetrated, the vortex excitation coil (106) and test coil of eddy current (107) are separated at intervals.

8. it is used to as claimed in claim 7 assess the multi frequency detection system of carbon fiber board depth of defect, it is characterised in that： Probe controlling organization (105) is compressed using flexure spring mechanism to be popped one's head in, and makes eddy current probe to be tested with composite all the time Part (101) surface is fitted, and vortex excitation coil (106) converts electrical signals to magnetic signal and to be coupled to composite to be measured In test specimen (101), test coil of eddy current (107) receives the magnetic signal of change and is converted to electric signal.

9. a kind of multi frequency detection detection method for assessing carbon fiber board depth of defect, it is characterised in that：Including as follows Step：

Step one：The eddy current signal characteristic parameter in composite test specimen (101) to be measured is extracted, is obtained by straight line scanning mode First internal flaw (1), at different frequencies induced voltage change curve, extracts the crest voltage in amplitude-distance Curve figure Δ V and corresponding frequency f of peak point_g；

Step 2：Using same procedure, the crest voltage Δ V caused by the second internal flaw (2) is extracted^*It is corresponding with peak point Frequency f_g ^*；

Step 3：The two groups of signals for extracting are drawn into curve, crest voltage Δ V increases with the increase of depth of defect, correspondence The frequency of peak point also increases therewith；

Step 4：Depth Inverse module (112) is corresponding according to the maximum Δ V and peak point of the induced voltage variable quantity for extracting Frequency f_g, carry out by the way of neutral net using being built, specifically include：

Input layer is built, the variable of input layer includes the maximum Δ v of output induced voltage variable quantity, the corresponding frequency of peak point f_g；

Hidden layer is built, hidden layer includes the neuron of 3 or more than 3；

Output layer is built, output layer is depth of defect D；Build reverse neural network；

The maximum Δ V of the induced voltage variable quantity that will be caused by composite test specimen to be measured (101) defect is corresponding with peak point Frequency f_gSubstitute into input layer；Depth of defect D is substituted into into output layer, neural metwork training is carried out, training result is obtained；

Step 5：According to training result, maximum Δ V, corresponding frequency f of peak point of voltage variety are set up_gWith correspondence defect Relation between depth D：D=F (Δ V, f_g)；

Step 6：Obtain the signal caused by defect in composite test specimen (101) to be measured using multi frequency detection system to become Change, extract characteristic parameter：The maximum Δ V of induced voltage variable quantity and corresponding frequency f of peak point_g, substitute into trained In neural network input layer, the depth information of defect in composite test specimen to be measured (101) is obtained.

10. it is used to as claimed in claim 9 assess the multi frequency detection detection method of carbon fiber board depth of defect, its feature It is：There is module (102) as multiple-frequency signal excitation eddy current probe is produced in the multifrequency sine pumping signal, wherein producing double Frequency pumping signal, concrete steps include：

Using two-way DDS chip AD9958, four tunnel output signals are produced；

Each AD9958's wherein exports all the way the input for being respectively fed to follow-up multi-frequency testing signal separation module (110), separately Outer two-way output then produces two-frequency signal by add operation Jing mixing；

Excitation vortex excitation coil (106) Jing after drive circuit output of two-frequency signal after mixing produces magnetic field.