A method for handling waveform aliasing in ultrasonic testing of thermal barrier coatings

Abstract

This invention discloses a method for processing aliased waveforms in ultrasonic testing of thermal barrier coatings, relating to the field of nondestructive testing. First, the ultrasonic echo signal of the thermal barrier coating is acquired. Second, an overcomplete atom dictionary conforming to the characteristics of the ultrasonic echo signal is established based on sparse representation. An adaptive global update artificial bee colony algorithm is used to improve the orthogonal matching algorithm. The ultrasonic echo waveform signal is projected onto the overcomplete atom dictionary using the improved orthogonal matching algorithm. The optimal atom with the highest correlation coefficient with the echo waveform signal is selected in each iteration to continuously approximate the original signal, achieving signal reconstruction and decomposition, and separating the echo signals from different interfaces. The waveform aliasing processing method for ultrasonic testing of thermal barrier coatings provided by this invention can effectively separate the echo waveforms aliased between multiple interfaces of the thermal barrier coating, thereby realizing the detection and evaluation of defects in the thermal barrier coating.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing of materials, and specifically to a method for processing waveform aliasing in ultrasonic testing of thermal barrier coatings. Background Technology

[0002] Thermal barrier coatings (TBCs) are widely used in gas turbine engines, significantly improving turbine inlet temperature and service life of hot-end components. However, due to long-term operation in high-temperature and high-pressure environments, TBCs are prone to defects such as poor adhesion or detachment, severely affecting engine stability and safety. Therefore, establishing a method for detecting TBC detachment defects has become a critical issue that urgently needs to be addressed.

[0003] Ultrasonic testing technology is highly sensitive to defects, fast, and convenient for defect localization. It is also simple to operate and widely used in defect detection for various materials. However, when applied to the detection of defects in thermal barrier coatings, the thin physical structure of the coating inevitably leads to aliasing of ultrasonic echoes from the interfaces between the ceramic layer and the adhesive layer, and between the adhesive layer and the substrate. This makes it difficult to extract echo information from defects, resulting in significant challenges in detecting defects in thermal barrier coatings. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method for processing waveform aliasing in ultrasonic testing of thermal barrier coatings, separating echoes from different interfaces of the thermal barrier coating, and achieving high-precision detection and evaluation of defects in the thermal barrier coating. Specifically:

[0005] Step 1: Obtain the ultrasonic echo signal of the thermal barrier coating;

[0006] Step 2: Establish an overcomplete atom dictionary that conforms to the characteristics of ultrasonic echo signals based on sparse representation;

[0007] Step 3: Improve the orthogonal matching algorithm by adopting an adaptive global update artificial bee colony algorithm;

[0008] Step 4: Project the echo waveform signal onto the overcomplete atom dictionary using the improved orthogonal matching algorithm. Select the optimal atom with the largest correlation coefficient with the echo waveform signal in each iteration to continuously approximate the original signal, thereby realizing signal reconstruction and decomposition and separating the echo signals from different interfaces.

[0009] Step 5: Defect assessment of thermal barrier coating based on the separated interface echo signal.

[0010] Compared with existing technologies, this invention addresses the problem that defect echoes are submerged in aliased waveforms, making defects inconspicuous and difficult to detect. It proposes a method for processing aliased waveforms in ultrasonic testing of thermal barrier coatings, which separates the defect waveform from the aliased waveform, thereby achieving high-precision detection and evaluation of defects in thermal barrier coatings. Attached Figure Description

[0011] The embodiments illustrated in conjunction with the accompanying drawings will be described in detail below. The accompanying drawings are merely some embodiments of the present invention, in which:

[0012] Figure 1 The overall flowchart shows the method for handling waveform aliasing during ultrasonic testing of thermal barrier coatings.

[0013] Figure 2 A model diagram of a thermal barrier coating with built-in defects;

[0014] Figure 3 An exploded view of the ultrasonic echo of a thermal barrier coating with a defect length of 3 mm.

[0015] Figure 4 The following waveforms of different interfaces of thermal barrier coating were obtained by decomposition for defect lengths of 0-5 mm: (a) ceramic layer / adhesive layer interface echo, (b) adhesive layer / substrate interface echo;

[0016] Figure 5 Decomposition of the thermal barrier coating at different interfaces with defect lengths of 0-5 mm yielded defect length-amplitude diagrams: (a) ceramic layer / adhesive layer interface, (b) adhesive layer / substrate interface. Detailed Implementation

[0017] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments. The present invention proposes a method for processing waveform aliasing in ultrasonic testing of thermal barrier coatings. Figure 1 The flowchart illustrates the method. First, ultrasonic echo signals from the thermal barrier coating are acquired. Second, an overcomplete atom dictionary conforming to the characteristics of ultrasonic echo signals is established based on sparse representation. An adaptive global update artificial bee colony algorithm is used to improve the orthogonal matching algorithm. Finally, the echo signals are projected onto the overcomplete atom dictionary using the improved orthogonal matching algorithm. The optimal atom with the highest correlation coefficient to the echo signal is selected in each iteration to continuously approximate the original signal, thus reconstructing and decomposing the echo signals at each interface. High-precision detection of defects in the thermal barrier coating is achieved by analyzing the decomposed echo information from each interface.

[0018] Step 1: Obtain the ultrasonic echo signal of the thermal barrier coating;

[0019] A two-dimensional model of a thermal barrier coating with built-in defects was created using COMSOL, such as... Figure 2As shown, the model is 5mm long and 2mm thick, consisting of three layers: a ceramic layer (0.26mm), an adhesive layer (0.2mm), and a substrate (1.54mm). The defect depth is set to 0.41mm (from top to bottom), and the lengths are set to 0mm, 1mm, 2mm, 3mm, 4mm, and 5mm.

[0020] A sine function modulated by a Gaussian function is used as the ultrasonic excitation signal source, and its formula is:

[0021]

[0022] In the formula, the value of A is 1×10 -6 Its size limits the range of amplitude variation, t is the excitation duration, and m is the pulse time T0, i.e., 1×10 -7 s and n are adjustment parameters that can adjust the shape of the Gaussian pulse signal waveform. In this paper, n is set to 0.5, and f0 refers to the center frequency of the ultrasonic pulse signal, which is set to 10MHz.

[0023] A mapping-based mesh generation method is adopted. Furthermore, to ensure the accuracy of ultrasonic wave propagation simulation and effectively capture the propagation path of elastic waves, the mesh size generally needs to meet certain requirements.

[0024] ΔX≤λ / 5 (1-2)

[0025] In the formula, λ is the wavelength of the ultrasonic wave, and ΔX is the grid size, meaning the largest cell size should not exceed one-fifth of the ultrasonic wavelength. The grid size for this model is set to λ / 6, where λ is 3 × 10⁻⁶. -4 ;

[0026] Line excitation was used as the excitation loading method, and the "Specify Displacement" option in the solid mechanics module was used for setting. The excitation line source was located at the exact center of the top of the ceramic layer, with a length of 0.4 mm. The transient solver was selected to study the propagation process of ultrasonic waves in the model, and the time step and solution time were set. The time step needed to satisfy the following formula:

[0027]

[0028] In the formula, ΔX is the grid size set in this paper, and c Lmax This refers to the speed of sound, which travels the fastest in the model. The formula for calculating the speed of sound is:

[0029]

[0030] In the formula, E is Young's modulus, ρ is material density, and ν is Poisson's ratio. Substituting these parameters, the longitudinal wave velocity of each layer can be calculated, and the sound velocity of the ceramic layer c... L1 = 3429.97 m / s, sound velocity of the adhesive layer c L2= 5287.05 m / s, ground sound velocity c L3 =4183.72m / s, then c Lmax =c L1 The calculated time step is T0 / 20, where T0 is 1×10. -7 The solution time needs to be greater than twice the time it takes for the probe to receive the echo from the ground. In this paper, the solution time is set to 0-30T0.

[0031] For the constructed two-dimensional thermal barrier coating model, built-in defect control groups with different lengths were set up, namely 0mm, 1mm, 2mm, 3mm, 4mm, and 5mm. Calculations were performed according to the set solution time to obtain the corresponding ultrasonic echo waveform signals of the defects.

[0032] Step 2: Establish an overcomplete atom dictionary that conforms to the characteristics of ultrasonic echo signals based on sparse representation;

[0033] A Nakagami atomic model is used to construct an overcomplete atomic dictionary D that conforms to the characteristics of ultrasonic echo signals. Unlike general overcomplete dictionary D which treats the ultrasonic scattering coefficient m as a constant without considering material scattering, this method uses the variation of the ultrasonic scattering coefficient m to characterize the different scattering situations of ultrasonic waves propagating through the thermal barrier coating, considering the coarse grains and severe scattering in the thermal barrier coating. This serves as a supplement to the overcomplete dictionary D. Atoms with parameters m within a preset range are selected, and the atom with the highest correlation coefficient to the original signal is chosen. The selected m value reflects the actual scattering situation of ultrasonic waves in the current thermal barrier coating, thereby improving the accuracy of subsequent signal decomposition and waveform reconstruction. The Nakagami atomic model formula is as follows:

[0034]

[0035] In the formula: Γ(m) is the parameter vector; K is the energy normalization factor; m determines the shape of the density function, which can be used to describe various scattering situations of ultrasound. When m = 0.5, the density function corresponds to a half-Gaussian distribution; when m = 1, the density function corresponds to a Rayleigh distribution; when m > 1, the density function corresponds to a Rice distribution; Ω is the time-width factor; fc is the signal center frequency; τ is the time-shift factor, which determines the specific arrival time of the time-domain signal. τ is the signal phase; u(t-τ) is the step function.

[0036] Step 3: Improve the orthogonal matching algorithm by adopting an adaptive global update artificial bee colony algorithm;

[0037] The Nakagami function is used instead of the original OMP algorithm to find the specific atom d used in the optimal atom formula. This is changed to a parameter optimization function, which transforms the process of OMP searching for the optimal atom in the discrete dictionary into an optimization problem of finding the extremum in the continuous parameter space.

[0038] An adaptive artificial bee colony algorithm is adopted to solve the optimization problem with the Nakagami function, and a two-dimensional global update method is added during the update process of candidate solutions.

[0039] Step 4: Project the echo waveform signal onto the overcomplete atom dictionary using the improved orthogonal matching algorithm. Select the optimal atom with the largest correlation coefficient with the echo waveform signal in each iteration to continuously approximate the original signal, thereby realizing signal reconstruction and decomposition and separating the echo signals from different interfaces.

[0040] Initialization: Signal estimation y (0) =0, residual r (0) =y, iteration number k=0;

[0041] When k = k + 1, execute the entire loop;

[0042] Initialization of solution distribution: via formula Initialize the distribution of solutions within the feasible region of g; and preset the maximum number of iterations max for the improved artificial bee colony algorithm and the maximum number of search iterations for candidate solutions;

[0043] Through formula Search for candidate solutions; and use the formula fit(g) = |<r k-1 ,g r >|, g∈D to obtain the function value of each solution (D is the overcomplete atom dictionary that conforms to the characteristics of ultrasonic echo signals constructed in step 2, g r The Nakagami function is used to replace the specific atom g used in the original OMP algorithm;

[0044] The number of times a solution is searched is calculated using a formula. If the number of times a solution is searched exceeds the maximum number of times a candidate solution can be searched, that solution is discarded, and the formula is used instead. (In this study, m=2) Two-dimensional global update to supplement new solutions;

[0045] Through formula and Select a region of possible candidate solutions and search it; calculate the number of times a solution has been searched. If the number of times a solution has been searched exceeds the maximum number of times a candidate solution can be searched, discard that solution and use the formula... (In this study, m=2) Two-dimensional global update to supplement new solutions;

[0046] Update the optimal solution g iIncrement the iteration count by 1. If the iteration count of the improved artificial bee colony algorithm is greater than the maximum, then stop the iteration.

[0047] Use the formula y (k) =argmin||yD (k) C (k) ||2 Update signal estimation y (k) (y is the ultrasonic echo waveform signal obtained in step 1), and the updated residual signal r (k) =yy (k) ;

[0048] If the number of iterations reaches a preset value or the residual decreases to a set threshold, the iteration stops, and the decomposition result of the output signal y is given. Where k is the total number of iterations; otherwise, return to step (4.2). Figure 3 This is the ultrasonic echo decomposition diagram of a thermal barrier coating with a defect length of 3mm.

[0049] Step 5: Defect assessment of thermal barrier coating based on the separated interface echo signal;

[0050] Based on the waveform aliasing processing algorithm in step 4, the echo waveforms of the ceramic layer / binder layer separated under a defect length of 0-5mm are summarized. Figure 4 (a) and the echo waveform at the adhesive layer / substrate interface ( Figure 4 (b)). Echo amplitude values ​​for different defect lengths were extracted to obtain the ceramic layer / adhesive layer echo defect length-amplitude diagram. Figure 5 (a) and the echo defect length-amplitude diagram of the adhesive layer / substrate interface (a) Figure 5 (b)). From Figure 5 (a) shows that since there are no defects in the ceramic layer and the adhesive layer, and the defects are located between the adhesive layer and the substrate, the echo amplitude of the ceramic layer / adhesive layer is not related to the defect length and remains basically stable. Figure 5 (b) It can be seen that the interface echo amplitude is the lowest when there are no defects, and the defect echo amplitude increases significantly with the increase of defect length. Therefore, as long as the aliased echo of the sample under test is separated, and the amplitude of the separated adhesive layer / substrate interface echo is compared with that of a standard defect-free sample, the defects of the thermal barrier coating can be quickly and effectively evaluated.

[0051] Finally, it is necessary to state that the specific embodiments described above are merely illustrative of or explanations of the principles of the present invention, and do not constitute a limitation thereof. Any non-essential improvements and adjustments made by those skilled in the art based on the above description of the present invention are within the scope of protection of the present invention.

Claims

1. A method for processing waveform aliasing in ultrasonic testing of thermal barrier coatings, characterized in that, The method includes the following steps: Step 1: Obtain the ultrasonic echo signal of the thermal barrier coating; Step 2: Establish an overcomplete atom dictionary that conforms to the characteristics of ultrasonic echo signals based on sparse representation; Step 3: Improve the orthogonal matching algorithm by adopting an adaptive global update artificial bee colony algorithm. The specific steps are as follows: Step 3.1: Replace the specific atom d used in the original OMP algorithm to find the optimal atom formula with the Nakagami function, and change it to a parameter optimization function. This transforms the process of OMP searching for the optimal atom in the discrete dictionary into an optimization problem of finding the extremum in the continuous parameter space. Step 3.2: Adaptive artificial bee colony algorithm is used to solve the optimization problem with Nakagami function. A two-dimensional global update method is added in the process of updating candidate solutions to improve efficiency and accuracy. Step 4: Project the echo waveform signal onto an overcomplete atom dictionary using an improved orthogonal matching algorithm. Select the optimal atom with the highest correlation coefficient with the echo waveform signal in each iteration to continuously approximate the original signal, thereby reconstructing and decomposing the signal and separating the echo signals from different interfaces. The specific steps are as follows: Step 4.1: Initialization: Signal estimation y (0) =0, residual r (0) =y, iteration number k=0; Step 4.2: k = k + 1, execute the entire loop; Step 4.3: Initialize the distribution of solutions: using the formula Initialize the distribution of solutions within the feasible region of g, and preset the maximum number of iterations max for the improved artificial bee colony algorithm and the maximum number of search times for candidate solutions; Step 4.4: Using the formula Search for candidate solutions; and use the formula fit(g) = |<r k-1 ,g r >|, g∈D to obtain the function value of each solution, where D is the overcomplete atom dictionary that conforms to the characteristics of ultrasonic echo signals constructed in step 2, and g r The Nakagami function is used to replace the specific atom g used in the original OMP algorithm; Step 4.5: Calculate the number of times a solution has been searched. If the number of times a solution has been searched is greater than the maximum number of times a candidate solution has been searched, then discard that solution and use the [presumably a different solution]. Select a region of possible candidate solutions and search it; calculate the number of times a solution has been searched. If the number of times a solution has been searched exceeds the maximum number of times a candidate solution can be searched, discard that solution and use the formula... A new solution is provided for the two-dimensional global update. Step 4.7: Update the optimal solution g i Increment the iteration count by 1. If the iteration count of the improved artificial bee colony algorithm is greater than the maximum, then stop the iteration. Step 4.8: Using formula y (k) =argmin||yD (k) C (k) ||2 Update signal estimation y (k) and residual signal r (k) =yy (k) , where y is the ultrasonic echo waveform signal obtained in step 1; Step 4.9: If the number of iterations reaches a preset value or the residual decreases to a set threshold, stop the iteration and output the decomposition result of the signal y. Where k is the total number of iterations; otherwise, return to step 4.

2. Step 5: Defect assessment of thermal barrier coating based on the separated interface echo signal.

2. The method for processing waveform aliasing in ultrasonic testing of thermal barrier coatings according to claim 1, characterized in that, In step 1, the ultrasonic detection echo signal of the thermal barrier coating is obtained, and the specific steps are as follows: A two-dimensional model of the thermal barrier coating with built-in defects was built using COMSOL, with dimensions and material properties consistent with the measured sample. A sine function modulated by a Gaussian function was used as the ultrasonic excitation signal source, with a center frequency of 10MHz for the ultrasonic pulse signal. A mapping mesh generation method was used, with the maximum cell size not exceeding one-fifth of the ultrasonic wavelength. The mesh size for the model was set to λ / 6, where λ is 3×10⁻⁶. -4 Line excitation was used as the excitation loading method, with the excitation line source located at the center of the top of the ceramic layer and a length of 0.4 mm. Built-in defect control groups of different lengths were set, with defect lengths of 0 mm, 1 mm, 2 mm, 3 mm, 4 mm, and 5 mm, respectively. The corresponding ultrasonic echo waveform signal of the defect was obtained by calculation according to the set solution time.

3. The method for processing waveform aliasing in ultrasonic testing of thermal barrier coatings according to claim 1, characterized in that, In step 2, an overcomplete atom dictionary conforming to the characteristics of ultrasonic echo signals is established based on sparse representation. The specific steps are as follows: Considering the large grain size and severe scattering of the thermal barrier coating, the change in the ultrasonic scattering coefficient m is used to characterize the different scattering situations of ultrasonic waves propagating in the thermal barrier coating when constructing an overcomplete atom dictionary using the Nakagami atomic model, as a supplement to the overcomplete dictionary. Atoms with parameters m within a preset range are screened, and the atom with the largest correlation coefficient with the original signal is selected. The corresponding m value is then used to reflect the actual scattering situation of ultrasonic waves in the current thermal barrier coating, thereby improving the accuracy of subsequent signal decomposition and waveform reconstruction.

4. The method for processing waveform aliasing in ultrasonic testing of thermal barrier coatings according to claim 1, characterized in that, In step 5, the thermal barrier coating defect assessment is performed based on the separated interface echo signal. The specific steps are as follows: By separating the ceramic layer / binder layer echo waveforms and the binder layer / substrate interface echo waveforms under different defect lengths, the echo amplitudes under different defect lengths are extracted to obtain the defect length-amplitude relationship of different interfaces. Based on the defect length-amplitude relationship, the tested sample is compared with the standard sample to evaluate the defects of the thermal barrier coating.

Images