Aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion

Abstract

The application discloses an aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion, belongs to the technical field of aircraft structure health monitoring, discloses a piezoelectric wave sensor array arranged on a structure surface, first extracts damage features in units of excitation-sensing channels, then establishes multiple calibration models in units of regions, inputs damage feature values into the models to obtain multiple diagnosis results, and finally obtains quantitative diagnosis results by using a result filtering mechanism. The aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion is accurate and efficient in implementation, and can effectively improve the reliability and accuracy of the aircraft structure under the condition that data dispersity is relatively high.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aircraft structure health monitoring, and particularly relates to an aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion. BACKGROUND

[0002] Due to complex service environment and operation conditions, various forms of structural damage may occur in the service operation of an aircraft, especially for composite materials, which have complex and diverse damage forms, including delamination, fiber fracture, matrix fracture, etc. In some cases, these damages cannot be found on the surface through visual inspection, but internal damage can have a great negative impact on the load-carrying capacity of the structure. In the subsequent operation process, these damages may gradually expand and accumulate and cause sudden failure, resulting in major safety accidents of the aircraft. Therefore, carrying out structural health state monitoring and diagnosis on the aircraft structure can timely grasp whether the aircraft structure damage occurs and expands, guide and improve the operation and maintenance efficiency, and improve the operation safety of the aircraft.

[0003] Structural health monitoring technology is a technology that uses sensors, data acquisition systems and analysis algorithms to monitor and evaluate various engineering structures such as buildings, bridges, aircraft, ships and other structures in real time or periodically. The main purpose is to detect or monitor the health status of the structure, including but not limited to deformation, vibration, crack, fatigue damage, etc. of the structure, so as to timely find out the problems existing in the structure and carry out maintenance, repair or improvement, so as to ensure the safe and reliable and long-term stable operation of the structure.

[0004] Due to its high sensitivity, large monitoring range, sensitivity to small damage, and applicability to metal composites, structural health monitoring technology based on piezoelectric guided waves is widely used in aircraft structure health state monitoring. This technology mainly collects guided wave signals under different structural states, and realizes the monitoring and evaluation of structural damage by combining specific algorithms.

[0005] For aircraft structure health monitoring, it can be divided into damage alarm, damage positioning and damage quantification. Among them, the quantitative monitoring of damage is crucial for realizing efficient operation and maintenance guidance of the aircraft, improving the safety and efficiency of the structure. However, in practice, the aircraft structure is complex, the service conditions are harsh, and the damage forms are diverse, resulting in a large dispersion of the obtained data, which affects the quantitative accuracy of the damage. Therefore, how to realize the quantitative monitoring of damage under the condition of strong data dispersion is of urgent need and great significance for the safe and efficient operation of the aircraft in practice. SUMMARY

[0006] The present application aims to provide an aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion, which can realize accurate quantitative monitoring of aircraft structure damage.

[0007] The technical solution of this invention is:

[0008] A quantitative monitoring method for aircraft structural damage based on regional multi-calibration model fusion includes,

[0009] Step S1: Under the condition of structural health, acquire waveguide health signal samples;

[0010] Step S2: Obtain guided wave damage signal samples under different structural damage conditions.

[0011] Step S3: Calculate the guided wave signal state characterization factor and combine it with the actual damage size to form a state characterization factor calibration dataset.

[0012] Step S4: Determine the sub-region where the damage is located based on the acquired damage imaging location;

[0013] Step S5: Determine the location of the damage using the sub-region. K A multi-calibration model is established based on the acquired multi-calibration channels of the sensor array, the corresponding guided wave signal state characterization factor, and the damage size calibration dataset. The expression is as follows:

[0014] ,

[0015] in, β pK Indicates the first K One channel p Fitting parameters of the sub-order multiple calibration model, SI nK Indicates the first K The guided wave signal state characterization factor for the nth damage in a given channel. p express p Order fitting parameters;

[0016] Step S6: Extract the state characterization factors of the guided wave signal of the new damage to form a test sample dataset. Substitute the test sample dataset into the multi-calibration model to obtain the multi-network channel damage size quantitative diagnosis result dataset and obtain the multi-diagnosis results.

[0017] Step S7: Combining the acquired multiple diagnostic results, a final quantitative diagnostic result of the injury is obtained based on the filtering mechanism.

[0018] More specifically, adhesive arrangement on the surface of the aircraft structure x A piezoelectric guided wave sensor, using x A sensor array is formed by piezoelectric guided wave sensors. y Several excitation-sensing channels are formed to create an excitation-sensing monitoring network, wherein... yThe expression is as follows:

[0019] .

[0020] More specifically, based on the set sensor array and excitation-sensor monitoring network, the guided wave health signal of the structure in a healthy state is acquired and a guided wave health signal sample is formed; the guided wave damage signal of the structure in a damaged state is acquired and a guided wave damage signal sample is formed.

[0021] In step S3 above, based on the guided wave signals obtained under the structural health state and structural damage state, the guided wave signal state characterization factor of each excitation-sensor channel in the excitation-sensor monitoring network is extracted, and the expression is as follows:

[0022] ,

[0023] in, SI i For the first i The guided wave signal state characterization factor of an excitation-sensing channel; i To determine the order of the excitation-sensor channels in the excitation-sensor monitoring network, i =1, 2, 3… y ; h i ( t () represents the guided wave health signal acquired under structurally healthy conditions; d i ( t Guided wave damage signals acquired under structural damage conditions; f This is a method for extracting signal features.

[0024] More specifically, a total of [number] fabrications were made on the monitored structure. n For each damage, the actual size of each damage is calculated to form a damage size calibration dataset. ,in, d n Indicates the first n The actual size of each damage; based on the guided wave signal state characterization factor obtained in the excitation-sensing channel under each damage state, a state characterization factor calibration dataset is formed as follows:

[0025] ,

[0026] in, Indicates the first n The first injury y The state characterization factor of the guided wave signal of each excitation-sensing channel.

[0027] In step S5 above, a total of [number] sub-regions where the damage is located [area]. kA piezoelectric waveguide sensor, the number of multiple calibration channels obtained under a single working condition is K , and the expression is:

[0028] .

[0029] In the above step S6, after the structure is damaged, the multiple calibration channels of the sensor array in the sub-region are determined according to the damage position, and the waveguide signal state characteristic factor of each sensor array multiple calibration channel is calculated , j =1, 2, 3… K , and a test sample data set is formed, and the expression is as follows:

[0030] .

[0031] The above step S7 includes,

[0032] Step S71, a result filtering mechanism is constructed, and the data average value of the damage size is calculated according to the obtained damage size calibration data set , and the expression is as follows:

[0033] ,

[0034] According to the obtained multiple network channel damage size quantitative diagnosis result data set, the diagnosis data set filtering threshold parameter is set μ 1 and μ 2, wherein, μ 1 is a natural number less than 1, μ 2 is a natural number greater than 1. Remove the values less than and greater than in the multiple network channel damage size quantitative diagnosis result data set to obtain the updated damage size quantitative diagnosis result data set as follows:

[0035] .

[0036] The above step S7 includes,

[0037] Step S72, obtaining the quantitative diagnosis result of the damage, based on the updated damage size quantitative diagnosis result data set, calculating the final diagnosis result, obtaining the damage size D, and the expression is as follows:

[0038] ,

[0039] Wherein, The number of data in the updated damage size quantitative diagnosis result data set.

[0040] Beneficial effects, the application is a kind of aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion, combined with the piezoelectric wave sensor array arranged on the structure surface, first extract damage features with excitation-sensing channel as a unit, then establish multiple calibration model with region as a unit, and obtain multiple diagnostic results after inputting damage feature values into the model, finally obtain quantitative diagnostic results using result filtering mechanism, the implementation process is accurate and efficient, which can effectively improve the reliability and accuracy of aircraft structure under strong data dispersion.

[0041] In order to make the above features and advantages of the application more obvious and easy to understand, the following specific examples are described in detail below with the aid of the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 The flow chart of the aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion of the application.

[0043] Figure 2 The schematic diagram of the monitored structure and the position of the piezoelectric wave sensor in a specific embodiment.

[0044] Figure 3 The schematic diagram of the distribution of damage on the monitored structure in a specific embodiment.

[0045] Figure 4 The schematic diagram of the sub-regions divided by the sensor array on the surface of the monitored structure in a specific embodiment.

[0046] Figure 5 The multiple calibration model established in a specific embodiment of the application.

[0047] Figure 6 The error of the quantitative diagnostic results obtained in a specific embodiment of the application. DETAILED DESCRIPTION

[0048] In order to make the purpose and technical scheme of the embodiments of the application more clear, the technical scheme of the embodiments of the application will be described clearly and completely below with the aid of the drawings of the embodiments of the application. Obviously, the described embodiments are part of the embodiments of the application, rather than all the embodiments. Based on the described embodiments of the application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the application.

[0049] Figure 1 The flow chart of the aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion of the application. As shown in Figure 1 The aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion of the application specifically includes the following steps.

[0050] Step S1: Under the condition of structural health, obtain waveguide health signal samples.

[0051] More specifically, adhesive arrangement on the surface of the aircraft structure x One piezoelectric guided wave sensor, among which... x Greater than 4. Use x A sensor array is formed by piezoelectric guided wave sensors. y Several excitation-sensing channels are formed to create an excitation-sensing monitoring network. y The expression is as follows:

[0052] .

[0053] Based on the established sensor array and excitation-sensor monitoring network, the guided wave health signal of the structure in a healthy state is acquired and a guided wave health signal sample is formed.

[0054] For example, please refer to Figure 2 , Figure 2 This is a schematic diagram of the monitored structure and the location of the piezoelectric guided wave sensor in a specific embodiment. Figure 2 As shown, the dimensions of the monitored structure 1 are 500mm × 400mm × 2mm (length × width × thickness). Twelve piezoelectric guided wave sensors are arranged in a 3×4 pattern on the monitored structure 1, numbered PZT1-PZT12. Specifically, the piezoelectric guided wave sensors in the first row are numbered PZT1-PZT3, in the second row PZT4-PZT6, in the third row PZT7-PZT9, and in the fourth row PZT10-PZT12. The lateral spacing between adjacent piezoelectric guided wave sensors is 180mm, and the longitudinal spacing is either 140mm or 100mm. When setting up the excitation-sensing monitoring network of the sensor array, firstly, piezoelectric guided wave sensor PZT1 is used as the excitation element for the guided wave signal, and piezoelectric guided wave sensors PZT2 to PZT12 are used as the response elements for the guided wave signal; then, piezoelectric guided wave sensor PZT2 is used as the excitation element for the guided wave signal, and piezoelectric guided wave sensors PZT1, PZT3 to PZT12 are used as the response elements for the guided wave signal; and so on. Piezoelectric guided wave sensors PZT1 to PZT12 form a total of 132 excitation-sensing channels, which together form an excitation-sensing monitoring network.

[0055] Step S2: Obtain guided wave damage signal samples under different structural damage conditions.

[0056] More specifically, based on the set sensor array and excitation-sensor monitoring network, guided wave damage signals of the structure under damaged conditions are acquired and formed into guided wave damage signal samples.

[0057] Step S3: Calculate the guided wave signal state characterization factor and combine it with the actual damage size to form a state characterization factor calibration dataset.

[0058] Furthermore, based on the guided wave signals acquired under both structural health and structural damage conditions, the guided wave signal state characterization factor for each excitation-sensor channel in the excitation-sensor monitoring network is extracted, as expressed below:

[0059] ,

[0060] in, SI i For the first i The state characterization factor of the guided wave signal of each excitation-sensing channel; i To determine the order of the excitation-sensor channels in the excitation-sensor monitoring network, i =1, 2, 3… y ; h i ( t () represents the guided wave health signal acquired under structurally healthy conditions; d i ( t Guided wave damage signals acquired under structural damage conditions; f This is a method for extracting signal features.

[0061] In one specific embodiment, f The expression is:

[0062] .

[0063] Furthermore, a total of [number] fabrications were made on the monitored structure. n For each damage, the actual size of each damage is calculated to form a damage size calibration dataset. ,in, d n Indicates the first n The actual size of each damage; based on the guided wave signal state characterization factor obtained in the excitation-sensing channel under each damage state, a state characterization factor calibration dataset is formed as follows:

[0064] ,

[0065] in, Indicates the first n The first injury y The state characterization factor of the guided wave signal of each excitation-sensing channel.

[0066] For example, refer to Figure 3 , Figure 3 is a schematic diagram of the distribution of damages on the monitored structure in an embodiment. As shown in Figure 3 , nine different positions of damages are made on the surface of the monitored structure 1, and each position of damage has five different sizes of 2mm, 3mm, 5mm, 7mm and 9mm. Figure 3 The damage sizes of the damage 1, the damage 2, the damage 3, the damage 4, the damage 5, the damage 6, the damage 7 and the damage 8 form a damage size calibration data set, and the state characterization factors of the guided wave signals of the excitation-sensing channels obtained in the eight damage states form a state characterization factor calibration data set.

[0067] Step S4, according to the obtained damage imaging position, determine the sub-region where the damage is located.

[0068] Wherein, the damage imaging position is obtained by an imaging positioning algorithm, and the position of the damage occurring on the structure can be obtained according to image processing.

[0069] More specifically, according to the piezoelectric guided wave sensors arranged on the surface of the structure, the surface of the structure is divided into k sub-regions by adjacent r sensors. In an embodiment, r =4 or 6.

[0070] For example, Figure 4 is a schematic diagram of the sub-regions divided on the surface of the monitored structure according to the sensor array in an embodiment of the present application. The piezoelectric guided wave sensor PZT1, the piezoelectric guided wave sensor PZT2, the piezoelectric guided wave sensor PZT4 and the piezoelectric guided wave sensor PZT5 are between the sub-region 1, the piezoelectric guided wave sensor PZT2, the piezoelectric guided wave sensor PZT3, the piezoelectric guided wave sensor PZT5 and the piezoelectric guided wave sensor PZT6 are between the sub-region 2, the piezoelectric guided wave sensor PZT4, the piezoelectric guided wave sensor PZT5, the piezoelectric guided wave sensor PZT7 and the piezoelectric guided wave sensor PZT8 are between the sub-region 3, the piezoelectric guided wave sensor PZT5, the piezoelectric guided wave sensor PZT6, the piezoelectric guided wave sensor PZT8 and the piezoelectric guided wave sensor PZT9 are between the sub-region 4, the piezoelectric guided wave sensor PZT7, the piezoelectric guided wave sensor PZT8, the piezoelectric guided wave sensor PZT10 and the piezoelectric guided wave sensor PZT11 are between the sub-region 5, the piezoelectric guided wave sensor PZT8, the piezoelectric guided wave sensor PZT9, the piezoelectric guided wave sensor PZT11 and the piezoelectric guided wave sensor PZT12 are between the sub-region 6, and the number of excitation-sensing channels in each sub-region is 12.

[0071] Step S5, determine the sub-region where the damage is located by using the sub-region where the damage is located KThe sensor array multiple calibration channels are based on the obtained sensor array multiple calibration channels and corresponding guided wave signal state characteristic factors and damage size calibration data sets, and a multiple calibration model is established, and the expression is as follows:

[0072] ,

[0073] Among them, β pK The fitting parameters of the multiple calibration model of the nth channel of the mth order are represented by K p SI nK The guided wave signal state characteristic factor of the nth damage of the mth channel is represented by K p The fitting parameters of the mth order are represented by p

[0074] More specifically, the number of piezoelectric guided wave sensors in the sub-region where the damage is located is k The number of multiple calibration channels obtained under a single working condition is K , and the expression is as follows:

[0075] .

[0076] Step S6, extracting the guided wave signal state characteristic factor of the new damage to form a test sample data set, substituting the test sample data set into the multiple calibration model to obtain a multiple network channel damage size quantitative diagnosis result data set, and obtaining a multiple diagnosis result. Among them, the multiple diagnosis result is the diagnosis result of each channel.

[0077] More specifically, after the structure produces a new damage, the sensor array multiple calibration channels in the sub-region are determined according to the damage position, the guided wave signal state characteristic factor of each sensor array multiple calibration channel is calculated , j =1, 2, 3… K , and a test sample data set is formed, and the expression is as follows:

[0078] .

[0079] Among them, the expression of the multiple network channel damage size quantitative diagnosis result data set is

[0080] ,

[0081] The damage size diagnosis result of the mth calibration channel is represented by K

[0082] Exemplarily, in the sub-region where the damage is located Figure 3 ​​​​​In the embodiment of FIG. 9, the damage 9 is a new damage generated, and the test sample data set is constituted by the state characteristic factors of the guided wave signals of the excitation-sensing channel obtained in the state of the damage 9.

[0083] Figure 5 The multiple calibration model is established in the embodiment of the present application. According to the obtained damage size calibration data set and the sub-region where the damage 9 is located, the multiple calibration model based on the sensor array network is established by using the calibration data in the corresponding region. Figure 5 The curve in FIG. 9 is the calibration model established by the state characteristic factors of the guided wave signals of each channel of the damage 1 in the sub-region. Figure 5 The vertical line in FIG. 9 is the test sample data set constituted by the state characteristic factors of the guided wave signals obtained under the damage 9.

[0084] Step S7, in combination with the obtained multiple diagnosis results, the quantitative diagnosis result of the damage is finally obtained according to the filtering mechanism.

[0085] Further, the step S7 further includes the following steps.

[0086] Step S71, constructing the result filtering mechanism. According to the obtained damage size calibration data set, the data average value of the damage size is calculated , and the expression is as follows:

[0087] ,

[0088] According to the obtained multiple network channel damage size quantitative diagnosis result data set, the diagnosis data set filtering threshold parameter is set μ 1 and μ 2, wherein, μ 1 is a natural number less than 1, μ 2 is a natural number greater than 1. The values less than and greater than in the multiple network channel damage size quantitative diagnosis result data set are removed to obtain the updated damage size quantitative diagnosis result data set as follows:

[0089] .

[0090] Step S72, obtaining the quantitative diagnosis result of the damage. Based on the updated damage size quantitative diagnosis result data set, the final diagnosis result is calculated, and the damage size D is obtained, and the expression is as follows:

[0091] ,

[0092] wherein, K’ is the number of data in the updated damage size quantitative diagnosis result data set.

[0093] Figure 6 The error of the quantitative diagnosis result obtained in a specific embodiment of the present application is calculated according to the formula in step S72 D For a lesion with an actual diameter of 2 mm, the quantitative diagnosis error is 0.75 mm; for a lesion with an actual diameter of 3 mm, the quantitative diagnosis error is 0.06 mm; for a lesion with an actual diameter of 5 mm, the quantitative diagnosis error is 0.28 mm; for a lesion with an actual diameter of 7 mm, the quantitative diagnosis error is 0.05 mm; and for a lesion with an actual diameter of 9 mm, the quantitative diagnosis error is 0.20 mm. It can be seen that the method of the present application can accurately quantitatively calculate lesions, with a small error.

[0094] Although the present application has been disclosed with the above embodiments, it is not intended to limit the present application, and anyone with ordinary knowledge in the art can make some changes and modifications without departing from the spirit and scope of the present application, and the protection scope of the present application shall be defined by the appended patent claims.

Claims

1. An aircraft structure damage quantitative monitoring method based on regional multiple calibration model fusion, characterized in that, Comprising, Step S1, in the structure health state, obtaining a guided wave health signal sample; Step S2, in the structure damage state, obtaining a guided wave damage signal sample; Step S3, calculating a guided wave signal state characterization factor, and combining an actual damage size to form a state characterization factor calibration dataset; In the step S3, according to the guided wave signals obtained in the structure health state and the structure damage state, a guided wave signal state characterization factor of each excitation-sensing channel in the excitation-sensing monitoring network is extracted, and the expression is as follows: ， Wherein, SI i is the guided wave signal state representation factor of the ith excitation-sensing channel; i is the order of the excitation-sensing channel in the excitation-sensing monitoring network, i = 1, 2, 3…y; h i (t) is the guided wave health signal collected under the structure health state; d i (t) is the guided wave damage signal collected under the structure damage state; f is the extraction method of signal features; Step S4, determining a sub-region where the damage is located according to the obtained damage imaging position; Step S5, determining a K-sensor array multiple calibration channel using the sub-region where the damage is located, establishing a multiple calibration model based on the obtained sensor array multiple calibration channel and the corresponding guided wave signal state characterization factor and damage size calibration dataset, and the expression is as follows: ， wherein β pK represents the fitting parameters of the multi-calibration model of the Kth channel and pth order, SI nK represents the state characterization factor of the guided wave signal of the nth damage of the Kth channel, and p represents the pth order fitting parameter; d K represents the actual size of the damage of the Kth channel Step S6, extracting a guided wave signal state characterization factor of a new damage to form a test sample dataset, substituting the test sample dataset into the multiple calibration model to obtain a multiple network channel damage size quantitative diagnosis result dataset, and obtaining a multiple diagnosis result; Step S7, combining the obtained multiple diagnosis result, and finally obtaining a quantitative diagnosis result of the damage according to a filtering mechanism.

2. The method according to claim 1, wherein the method is characterized by, x piezoelectric guided wave sensors are bonded on the surface of the aircraft structure, a sensor array is formed by the x piezoelectric guided wave sensors, y excitation-sensing channels are formed and an excitation-sensing monitoring network is formed, and the expression of y is as follows: 。 3. The method according to claim 2, wherein the method further comprises: According to the set sensor array and excitation-sensing monitoring network, a guided wave health signal of the structure in the health state is obtained, and a guided wave health signal sample is formed; a guided wave damage signal of the structure in the damage state is obtained, and a guided wave damage signal sample is formed.

4. The method according to claim 3, wherein the method further comprises: n damages are co-produced on the monitored structure, the actual size of each damage is counted to form a damage size calibration dataset wherein d n represents the actual size of the nth damage; according to the excitation-sensing channel guided wave signal state characterization factors obtained under each damage state, a state characterization factor calibration dataset is formed as follows: ， wherein, represents the guided-wave signal state characterizer for the yth excite-sense channel under the nth damage.

5. The method according to claim 4, wherein the method further comprises: In the step S5, there are k piezoelectric guided wave sensors in the sub-region where the damage is located, and the number of multiple calibration channels obtained in a single working condition state is K, and the expression is as follows: 。 6. The method according to claim 5, wherein the method further comprises: In the step S6, after the structure generates a new damage, the sensor array multiple calibration channels in the sub-region are determined according to the damage position, and a guided wave signal state characteristic factor of each sensor array multiple calibration channel is calculated , j = 1, 2, 3…K, and constitute a test sample data set, and the expression is as follows: 。 7. The method according to claim 6, wherein the method further comprises: The step S7 includes, Step S71, a result filtering mechanism is constructed, and a data average value of the damage size is calculated according to the obtained damage size calibration data set The expression is as follows: ， According to the obtained multi-network channel damage size quantitative diagnosis result data set, diagnosis data set filtering threshold parameters μ1 and μ2 are set, wherein μ1 is a natural number less than 1, and μ2 is a natural number greater than 1; values less than and greater than in the multi-network channel damage size quantitative diagnosis result data set are removed, and an updated damage size quantitative diagnosis result data set is obtained as follows: 。 8. The method according to claim 7, wherein the method further comprises: The step S7 includes, Step S72, obtaining a quantitative diagnosis result of the damage, calculating a final diagnosis result based on the updated damage size quantitative diagnosis result dataset, and obtaining a damage size D, and the expression is as follows: ， wherein, The number of data in the updated lesion size quantitative diagnostic result data set.

Images