Dynamic evaluation method for ultimate bearing capacity of layered damage composite material cylindrical pressure-resistant shell

Abstract

The invention discloses a method for dynamically evaluating the ultimate bearing capacity of a layered damage composite material cylindrical pressure-resistant shell in the field of deep sea equipment structure health monitoring, which comprises the following steps of: establishing a nonlinear relationship between a critical buckling load of an unlayered composite material cylindrical shell and a critical buckling load of a layered composite material cylindrical shell, and calculating an attenuation coefficient of each group of samples; measuring inner and outer layer strain values of the to-be-evaluated composite cylindrical shell, detecting extreme value statistics and cumulative sum based on change points of the inner and outer strain values, judging that the position of the point is layered, constructing a layering point set, calculating five layering characteristic parameters of the to-be-evaluated composite cylindrical shell based on the layering point set, and substituting the five layering characteristic parameters into an attenuation coefficient formula, obtaining a critical load of the to-be-evaluated composite cylindrical shell; according to the method, through combination of dual-surface strain cooperative monitoring and a regression model, prediction of critical buckling loads in a service state is realized, environments such as shell thickness fluctuation, salt spray corrosion, interface fracture parameter measurement errors and impact loads are compatible, strain gauge monitoring and limit load online evaluation are fused, and a dual evaluation system is formed.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of deep-sea equipment structure health monitoring, in particular to the key component pressure hull of a submersible, and specifically to a method for dynamically evaluating the ultimate bearing capacity of a layered damage composite pressure hull. BACKGROUND

[0002] Submersibles play an important role in deep-sea resource exploration, marine ecological monitoring, military security, etc. The pressure hull is a key component of the submersible, including manned cabins, adjustable ballast water tanks, electronic equipment cabins, etc. As a pressure vessel that withstands hydrostatic external pressure, it plays an important role in protecting non-pressure equipment, ensuring the health and safety of crew members, and providing positive buoyancy. Advanced materials play a crucial role in improving the ultimate bearing capacity of the hull. Composite materials have high specific strength, high specific stiffness, corrosion resistance, and fatigue resistance, and have been widely used in the field of submersibles. However, during the manufacturing and actual use of such hulls, the adhesive interface of the composite material is prone to debonding, resulting in complex nonlinear instability and degradation of the ultimate bearing capacity of the hull.

[0003] With the widespread use of composite materials in deep-sea equipment, the problem of bearing capacity degradation caused by delamination damage has become increasingly prominent. Existing technologies for delamination damage mainly have the following shortcomings: 1) Test detection methods rely on laboratory environments and are difficult to implement real-time monitoring in service conditions, such as the method disclosed in the document with publication number CN116577410A and title "A method for identifying delamination damage in composite cylindrical shell structure"; 2) Finite element analysis methods have limitations in handling manufacturing defects and complex load coupling effects, such as the method disclosed in the document with publication number CN105740566B and title "A finite element method for predicting interlaminar damage and delamination in layered composite materials". In addition, existing test methods ignore initial geometric eccentricity or thickness fluctuations; it is difficult to quantitatively analyze the effects of seawater infiltration, long-term moisture absorption, and salt mist corrosion on interlaminar adhesion; it is difficult to determine the deviation between the parameters obtained from the test, such as interlaminar stiffness, interlaminar strength, and interface fracture toughness, and the actual values; it is also difficult to realistically simulate the random water pressure fluctuations and impact loads encountered by composite cylindrical pressure hulls in marine environments. In particular, it is worth noting that the asymmetric strain field evolution caused by delamination damage, such as outer layer tension / inner layer compression, lacks an effective dynamic identification mechanism. The present application patent analyzes the delamination by monitoring the strain of the service hull, avoiding the calculation method that does not consider hull thickness fluctuations, salt mist corrosion, interface fracture parameter measurement errors, impact loads, etc., resulting in a large error in the delamination prediction results. SUMMARY

[0004] The purpose of the present invention is to address the problems and defects of the above-mentioned known technologies and provide a dynamic assessment method for the ultimate bearing capacity of a composite pressure hull with delamination damage, so as to realize real-time and accurate monitoring of the delamination damage of the ultimate bearing capacity of the composite pressure hull under service conditions.

[0005] The technical solution adopted by the method for dynamically evaluating the ultimate bearing capacity of a layered damaged composite cylindrical pressure hull of the present invention comprises the following steps:

[0006] Step 1): Calculate the critical buckling load P of the non-delaminated composite cylindrical shell NASA ;

[0007] Step 2): Select the five layer characteristic parameters that have a greater impact on the critical buckling load: layer area, layer depth, minimum distance from the layer boundary to the end, shape factor, and number of layers. Design and generate a test plan to cover all test conditions and obtain Z groups of samples. Use the finite element method to perform buckling analysis and obtain the critical buckling load of the layered composite cylindrical shell.

[0008] Step 3): Establish the critical buckling load P NASA and P cr Nonlinear relationship: P cr =K·P NASA , calculate the attenuation coefficient of each group of samples X a and X b are five different hierarchical characteristic parameters, B0 is the constant regression coefficient, B aa is the quadratic regression coefficient, B ab is the cross-term regression coefficient;

[0009] Step 4): For the composite cylindrical shell to be evaluated, patch-type fiber Bragg grating strain sensors are arranged at intervals along the annular direction on its outer surface, and strain gauges are arranged on its inner surface at positions corresponding to the outer surface to measure the inner and outer layer strain values ​​ε in , ε out ;

[0010] Step 5): Use the ideal strain values ​​of the inner and outer layers to correct the strain values ​​ε of the inner and outer layers in , ε out , to eliminate the seawater pressure error; the strain values ​​of the inner and outer layers ε in , ε out Perform denoising and smoothing to obtain smoothed internal and external strain values

[0011] Step 6): Calculate the internal and external strain values The extreme value statistics T(x) and the cumulative sum S of the change point detection k , when the extreme value statistic of change point detection T(x)>1.5ε y And the cumulative sum Sk >1.8ε y , determine that the point position delaminates, construct a delamination point set according to the delamination point, ε y is the ideal strain value of the inner and outer layers;

[0012] Step 7): based on the five delamination characteristic parameters of the delamination point set of the composite cylinder shell to be evaluated, the five delamination characteristic parameters are substituted into the attenuation coefficient formula in step 3), and the critical load of the composite cylinder shell to be evaluated is obtained.

[0013] The beneficial effects of the application after adopting the above technical scheme are:

[0014] (1) Combined with double-surface strain cooperative monitoring and regression model, the critical buckling load in service state is predicted, and the risk of delamination failure of the composite pressure shell under complex working conditions is reduced.

[0015] (2) The strain gauge monitoring and the limit load online evaluation are combined to form a double evaluation system, and the false alarm risk is reduced.

[0016] (3) The ultrasonic flaw detection and the strain calculation model are used at the same time, the results of material performance degradation can be used for the improvement of the calculation model, and the feasibility of the scheme is improved.

[0017] (4) The application establishes the relationship between strain and delamination parameters, and brings the real-time parameters into the database to calculate the limit load that the shell can withstand. The real-time acquisition of strain and delamination parameters in the service shell, the method itself can be compatible with shell thickness fluctuation, salt spray corrosion, interface fracture parameter measurement error, impact load and other environments. There is no need to consider special working conditions in the calculation process, which greatly reduces the error of the method. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 is an axial side view of a delamination parameter of a composite cylindrical pressure shell;

[0019] Figure 2 is a cross-sectional view of a delamination parameter of a composite cylindrical pressure shell;

[0020] Figure 3 is an axial side view of a strain gauge distribution of a composite cylindrical pressure shell;

[0021] Figure 4 is an end view of a strain gauge distribution of a composite cylindrical pressure shell;

[0022] Figure 5 is a flow chart of the dynamic evaluation method for the limit bearing capacity of the delamination damage composite pressure shell;

[0023] In the figure: 1-delamination damage (indicated by a dashed line), 2-composite cylindrical pressure shell, 3-strain sensor. DETAILED DESCRIPTION

[0024] See also Figures 1-5 , the present invention specifically adopts the following steps:

[0025] Step 1: Establish a prediction formula for non-delaminated composite cylindrical pressure shells and obtain the critical buckling load of non-delaminated composite cylindrical shells.

[0026] The critical buckling load of a composite cylindrical pressure shell (hereinafter referred to as "composite cylindrical shell") is affected by material parameters and geometric parameters, such as elastic modulus, Poisson's ratio, thickness, length, radius, etc. Therefore, the critical buckling load P of a non-delaminated composite cylindrical shell is first calculated based on the NASA mathematical formula. NASA , the high accuracy of this formula can reduce the error rate of the regression model in step 4. NASA's mathematical formula is as follows:

[0027]

[0028] in,

[0029]

[0030] Where F is the design safety factor, ranging from 1 to 5. m is the axial wave number of the cylindrical shell, n is the circumferential wave number of the cylindrical shell. L is the length of the cylindrical shell, and R is the outer radius of the composite cylindrical shell. ij ,B ij ,D ij is the element in the i-th row and j-th column of the stiffness matrices A, B, and D. A, B, and D are stiffness matrices solved based on classical laminate theory. Classical laminate theory divides the composite material into several layers (the i-th layer is also the i-th row in the stiffness matrix), then calculates the stiffness matrix layer by layer, and integrates the layers to obtain the overall stiffness matrix. The formulas for calculating the stiffness matrices A, B, and D are as follows:

[0031]

[0032] in is the transformed stiffness matrix of the i-th layer obtained by the following transformation formula (3), Z i , Z i-1 It represents the distance between the mid-plane of the i-th layer and the i-1-th layer and the neutral plane of the laminate (i.e. the center plane of the total thickness).

[0033]

[0034] In formula (3), T is the matrix transpose, Q is the constitutive matrix, and the expression of the constitutive matrix Q is as follows:

[0035]

[0036] where E1 is the Young's modulus in the fiber direction, E2 is the Young's modulus perpendicular to the fiber direction, G 12 is the shear modulus in the horizontal plane, v 12 is the major Poisson's ratio, which is the ratio of the lateral strain perpendicular to the fiber direction to the axial strain in the fiber direction when a tensile stress is applied in the fiber direction.v 21 is the minor Poisson's ratio, which is the ratio of the lateral strain in the fiber direction to the axial strain perpendicular to the fiber direction when a tensile stress is applied perpendicular to the fiber direction. This part of the parameters needs to be obtained by tensile test according to ASTM or GB / T standards.

[0037] In equation (3), T (i) is the transformation matrix T (i) of the i-th layer, and T (i) is expressed as follows:

[0038]

[0039] In equation (5), θ is the fiber angle of the i-th layer.

[0040] Second step: design of test scheme for layered composite cylindrical shell

[0041] According to existing research, five layered characteristic parameters that have a greater impact on the critical buckling load of the cylindrical shell are selected, including the layered area, the layered depth, the minimum distance between the layered boundary and the end, the shape factor, and the number of layers, as shown in Figure 1 , Figure 2 The Box-behnken design is used to generate the test scheme, and three level values (i.e., the maximum value, the intermediate value, and the minimum value) of each layered characteristic parameter are set to cover all possible test cases, obtaining Z sets of parameter combinations, which are also Z sets of samples.

[0042] Third step: numerical simulation analysis of layered composite cylindrical shell

[0043] For the test scheme designed in the second step, buckling analysis is performed based on the finite element method (FEM), and the virtual crack closure technique (VCCT) in the ABAQUS software is used to simulate the initial layering of the composite cylindrical shell. According to the Chinese Register of Shipping Submersible Classification Specification, the first-order modal defect nonlinear Riks method is introduced for numerical analysis, and the critical buckling load P cr of the layered composite cylindrical shell is obtained. This critical buckling load P cr is taken as the numerical result, and in the fourth step, the response surface method is used to establish the nonlinear relationship between the critical buckling load P cr and the critical buckling load P NASA of the composite cylindrical shell without initial layering obtained in the first step.

[0044] Fourth step: establishment of regression model (database) by response surface method.

[0045] The nonlinear relationship between the critical buckling load P of the layered composite cylindrical shell and the initial layered parameters is established according to the following formula (6) cr P NASA The regression calculation is performed using the Design-Expert software. The attenuation coefficient K of each group of samples, that is, the regression coefficient of the regression model, is calculated by formula (6), and is input into the Design-Expert software for regression analysis. Finally, a quadratic polynomial response surface model about K is obtained through regression analysis, as shown in the following formula (7). After obtaining the quadratic polynomial response surface model, formula (6) and formula (7) are taken as a database, and the real-time layered parameters are brought into the database to calculate the real-time bearing capacity of the shell after the eighth step.

[0046] P cr = K·P NASA (6)

[0047]

[0048] where X a and X b are five different layered parameters, which can be taken as: layered area S, layered depth h, minimum distance l from the layered boundary to the end, shape factor f, and number of layers N d . B0 is the constant term regression coefficient, B aa is the quadratic term regression coefficient, and B ab is the cross term regression coefficient, which are obtained by regression calculation. The detailed expansion of formula (7) is shown in the tenth step.

[0049] Step 5: Regression analysis evaluation.

[0050] In order to verify the accuracy and reliability of the regression model, the following statistical and error analysis is performed in the Design-Expert software:

[0051] (1) Significance test

[0052] The F test and p value analysis are performed on each regression coefficient, that is, the attenuation coefficient K, to determine whether the variable and the interaction term are significant; if the p value is less than 0.05, it is considered that the variable has a significant effect on the response value. Otherwise, it can be considered that the variable has little effect on the result, and the influence law analysis of the variable can be omitted.

[0053] (2) Fitting degree evaluation

[0054] The determination coefficient R 2 is calculated, and generally R 2 > 0.95 indicates good fitting

[0055]

[0056] in is the critical buckling load P of the i-th data point cr ; is the i-th response value predicted by the response surface model after removing the i-th data point (the result of formula (7) in the fourth step); Z is the number of samples (the Z group parameter combination in the second step); P cr average value.

[0057] (3) Verification of predictive ability

[0058] An additional 10% of the experimental schemes were selected for random parameters to simulate and compare the model predictions with the simulation results. The error between the regression model results and the numerical simulation results was calculated. An error less than 10% indicates that the regression model is accurate.

[0059] Step 6: While executing step 2, collaboratively monitor the strains of both surfaces.

[0060] like Figure 2 、 Figure 3 As shown, for the composite cylindrical shell to be evaluated, patch-type fiber Bragg grating strain sensors 3 are arranged on its outer surface at intervals of 10° to 30° along the circumferential direction. At the same time, metal foil strain gauges are arranged on the inner surface of the pressure shell at positions corresponding to the outer surface, forming an internal and external symmetrical strain size monitoring network to achieve accurate detection of the strain on the inner and outer surfaces of the pressure shell. Before the strain sensor is arranged, the bonding area is first pre-treated, cleaned with alcohol and polished with sandpaper, and then high-performance epoxy structural adhesive is applied to the preset position and the strain sensor is bonded. A laser positioning system is used to ensure that the positions of the internal and external strain sensors correspond accurately, with a spatial registration error of ≤50μm, a sampling frequency of 10-30Hz, and a synchronization accuracy of the internal and external strain sensor data of <1ms. The adjacent measurement points are equidistant and set to a. The measured strains of the inner and outer layers of the composite pressure shell are the inner layer strain value ε and the inner layer strain value ε, respectively. in and the outer layer strain value ε out .

[0061] Step 7: Eliminate seawater pressure error

[0062] In order to reduce the internal and external strain measurement errors caused by seawater pressure, the strain values ​​of the non-layered composite cylindrical pressure hull were collected through a hydrostatic test. The noise-free strain values ​​measured inside and outside the shell during the hydrostatic test were taken as the ideal strain values, i.e., the ideal strain value of the inner layer ε y，in and the ideal strain value of the outer layer ε y，out , the inner layer strain value ε used for actual detection in and the outer layer strain value ε out Strain correction is used to eliminate systematic errors caused by sea waves, currents, and impacts, thus avoiding incorrect determination of delamination damage in step 9.

[0063] The seawater pressure test procedure is as follows: Using a closed static pressure chamber or hydraulic loading device, a multi-level constant pressure in the range of 0–10 MPa is applied, maintaining each pressure level for 5–10 minutes. Once the strain data stabilizes, the simultaneous strain responses of the patch-type fiber Bragg grating sensor and the metal foil strain gauge are recorded. Based on the obtained test data, strain-pressure response curves are established for each, and a pressure-induced intrinsic response model is constructed.

[0064] Step 8: Strain signal denoising and smoothing

[0065] The strain data is processed based on the Kalman filter algorithm. First, the inner layer strain value ε measured on the inner and outer surfaces of the shell is processed using the Kalman filter. in and the outer layer strain value ε out The signal is smoothed to reduce the interference of random noise. Its state space model includes the state equation and the observation equation, as shown in formulas (9) and (10).

[0066]

[0067] Formula (9) is the state equation of the state space model of Kalman filter, ε k is the noise-free internal and external strain state vector at the current moment k, which can be expressed as ε in the seventh step. y，in and ε y，out As a reference, the strain value ε applied to the inner layer in and the outer layer strain value ε out ; ε k-1 is the noise-free internal and external strain state vector at the previous moment; w k is the process noise, represents w k Obeys normal distribution, the noise mean is 0, Q is the process noise covariance matrix; v k Observation noise, Indicates v k It obeys the normal distribution, the noise mean is 0, R is the observation noise covariance matrix, in and is the output variance of the sensor measurement process, which can be calibrated by referring to the sensor manual. Q is the process noise covariance matrix. Through the strain signal collection in step 7, the strain change rate variance is calculated, such as: Var represents variance, and Cov represents covariance. Experience has shown that sensors can be noisy, so the value can be between 2 and 60.

[0068] Formula (11) is the observation equation of the state space model of Kalman filter. kis the real-time strain measurement result at the current moment k;

[0069]

[0070] z k yes:

[0071]

[0072] After the eighth step of iterative update, the smoothed internal strain value is obtained and external strain value,

[0073] Step 9: Determination of layered damage.

[0074] Next, the internal strain value is identified by the change point detection algorithm and external strain value, Signal trend change. When the tensile strain on the outer surface of the pressure hull increases (the upper layer of the stratification deforms outward) and the compressive strain on the inner surface increases (the lower layer of the stratification deforms inward), and the strain shows an unstable growth, it indicates that the pressure hull has been delaminated. The strain change trend is detected by the mean change point detection extreme value statistic T(x), see formula (13). By accumulating and S k The (CUSUM) algorithm, see formula (14), detects unstable strain growth.

[0075]

[0076] Formula (13) is the change point detection equation, is the value of the strain signal at time k, x is the assumed change point position, y is the number of data points of the strain signal, and c is the data length. Contains the output of step 8 Therefore, formula (13) needs to be brought into the smoothed internal and external strain signals for calculation. y When , it indicates that the strain at the measuring point on the shell has a sudden change, and delamination may occur. y is the strain value without delamination under the ideal uniform external pressure condition obtained in the seventh step, that is, the ideal strain value of the inner layer ε y，in and the ideal strain value of the outer layer ε y，out .

[0077]

[0078] Formula (14) is the CUSUM formula, S k is the cumulative sum at time k; μ0 is the expected mean strain when there is no stratification; when S k When the set threshold is exceeded, a change point is determined. k>1.8ε y Significant changes can be detected.

[0079] When the strain measurement of a certain point meets the above two determination results, it is determined that delamination occurs at the point position. According to the sensor nodes that meet the conditions, a point set P = {P1, P2, P3,..., Pg} is constructed, where g is the number of points that delaminate. g The inner and outer layer strain values corresponding to the delamination point set P are recorded and and the coordinate positions P of all point sets m，L are used to calculate the delamination feature parameters.

[0080] Step 10: Calculate the delamination area S, delamination depth h, minimum distance l from the end of the delamination boundary, shape factor f, and number of delaminations N of the composite cylinder shell to be evaluated d These five delamination feature parameters.

[0081] 1. Delamination depth h calculation.

[0082] Due to the impact of external force or initial manufacturing defects, the pressure-resistant shell delaminates, causing the strain generated by the upper and lower layers of the shell to differ from the strain when it is not delaminated. By comparing this strain ratio relationship, the delamination depth can be calculated. The inner layer strain value and the outer layer strain value corresponding to the delamination point set determined in the ninth step are calculated using formula (15) to obtain the delamination depth h parameter.

[0083]

[0084] Formula (15) is the delamination thickness estimation formula, where t is the total thickness of the shell, is the inner layer strain value of the shell corresponding to the delamination point set is the outer layer strain value of the shell corresponding to the delamination point set. After calculation, the delamination depth h is output.

[0085] 2. Delamination area S calculation

[0086] The envelope area surrounded by the coordinate positions P m，L that meet the delamination conditions in the ninth step is calculated using the convex hull algorithm.

[0087]

[0088] where S is the delamination area, A s represents the area of a small triangle or grid element that constitutes the delamination area, and g is the number of points that delaminate. After calculation, the delamination area S is output.

[0089] 3. Minimum distance l from the end of the delamination boundary calculation

[0090] l is calculated by the following equation:

[0091] l = min(|y pm -y bottom |, |y pm -y top |) (16)

[0092] where y pm represents the y coordinate of the delamination boundary point, which is obtained by the convex hull algorithm on the delamination point coordinates P m，L filtered in the ninth step. y bottom represents the y axis coordinate of the bottom end of the pressure-resistant shell, and y top represents the y axis coordinate of the top end of the pressure-resistant shell. The pressure-resistant shell is arranged axially along the y axis. min represents the minimum value. After the calculation is completed, the distance l of the delamination from the end surface is output.

[0093] 4. Delamination shape factor f calculation

[0094] In order to accurately describe the shape of the delamination, equation (16) is used for calculation

[0095]

[0096] where P f is the perimeter of the delamination region, which is calculated by the boundary point set coordinates obtained by the convex hull algorithm in step 3. π is the ratio of the circumference to the diameter. f is the shape factor, which is used to quantify the compactness of the delamination region, f = 1 represents a circular delamination (most compact), f < 1 represents a more elongated or irregular delamination region. When f > 0.7, it is circular, and when f < 0.3, it is linear. After the calculation is completed, the delamination shape factor is output.

[0097] 5. Delamination number N d calculation

[0098] The number of delaminations W is determined by judging whether the points meeting the delamination condition in the ninth step form multiple independent regions in space. The point set P is constructed for the sensor nodes meeting the delamination condition, and the spatial clustering or adjacency judgment method is used to divide the continuously distributed points into multiple independent delamination regions. Each non-intersecting connected region is determined as an independent delamination. The number of delaminations is calculated by the following equation:

[0099]

[0100] where N d is the total number of delaminations, and δ v is the judgment function of whether the vth region constitutes an independent delamination:

[0101]

[0102] By calculating the spatial distance between any two points, and setting the minimum discriminant distance threshold (such as 1.5a), it is judged whether the points belong to the same piece of layered region, and finally the number of all connected regions is counted as the number of layers. The final output is the number of layers N d .

[0103] Eleventh step: limit load dynamic prediction.

[0104] Substitute the five layering characteristic parameters calculated in the tenth step into the database obtained in the fourth step, that is, substitute into the calculation formula of the attenuation coefficient K. Specifically:

[0105]

[0106] The critical load P of the evaluated composite cylinder shell is calculated again cr = K·P NASA .

[0107] It is worth noting that the accuracy of real-time delamination parameters measured by strain gauges is affected by the density of strain gauge measurement points. There is a certain error in real-time measurement of delamination parameters, which reduces the critical load of the delamination pressure shell by a small amount and is close to the working load of the pressure shell. Further accurate detection is needed in the detection center.

[0108] Twelfth step: high-precision detection of delamination parameters.

[0109] The method of ultrasonic detection is used to detect the depth of the delamination area to be evaluated. Through the propagation and reflection of high-frequency sound waves in the material, cracks, pores, delamination and other defects are identified. Provide accurate delamination defect parameters. Use a thickness measuring instrument to measure the thickness of the delamination pressure shell and obtain the real thickness distribution of the shell. Perform three-dimensional laser scanning on the delamination pressure shell to obtain the real three-dimensional profile of the pressure shell.

[0110] Thirteenth step: accurate calculation of critical load.

[0111] Based on the delamination parameters measured in the above twelve steps, a reverse modeling is performed to obtain an accurate numerical model of the delamination pressure shell. The RIKS analysis method recommended by China Classification Society is used to accurately calculate the critical load of the pressure shell. The critical load P of the evaluated composite cylinder shell is calculated cr Compare the numerical value with the actual measurement result to realize online calibration of material performance degradation and optimize the delamination parameter measurement theory.

Claims

1. A dynamic evaluation method for the ultimate bearing capacity of a composite cylindrical pressure hull with layered damage, characterized by: The following steps are involved: Step 1): Calculate the critical buckling load P of the non-delaminated composite cylindrical shell NASA ; Step 2): Select the five layer characteristic parameters that have a greater impact on the critical buckling load, namely, layer area, layer depth, minimum distance from the layer boundary to the end, shape factor, and number of layers, and design a test plan to cover all test conditions to obtain Z groups of samples; use the finite element method to perform buckling analysis and obtain the critical buckling load P of the layered composite cylindrical shell. cr ; Step 3): Establish the critical buckling load P NASA and P cr Nonlinear relationship: P cr =K·P NASA , calculate the attenuation coefficient of each group of samples X a and X b are five different hierarchical characteristic parameters, B0 is the constant regression coefficient, B aa is the quadratic regression coefficient, B ab is the cross-term regression coefficient; Step 4): For the composite cylindrical shell to be evaluated, patch-type fiber Bragg grating strain sensors are arranged at intervals along the annular direction on its outer surface, and strain gauges are arranged on its inner surface at positions corresponding to the outer surface to measure the inner and outer layer strain values ​​ε in , ε out ; Step 5): Use the ideal strain values ​​of the inner and outer layers to correct the strain values ​​ε of the inner and outer layers in , ε out , to eliminate the seawater pressure error; the strain values ​​of the inner and outer layers ε in , ε out Perform denoising and smoothing to obtain smoothed internal and external strain values Step 6): Calculate the internal and external strain values The extreme value statistics T(x) and the cumulative sum S of the change point detection k , when the extreme value statistic of change point detection T(x)>1.5ε y And the cumulative sum S k >1.8ε y , determine the point location where stratification occurs, and construct a stratified point set based on the point where stratification occurs, ε y is the ideal strain value of the inner and outer layers; Step 7): Calculate the five layer characteristic parameters of the composite cylindrical shell to be evaluated based on the layer point set, and substitute these five layer characteristic parameters into the attenuation coefficient formula in step 3) to obtain the critical load of the composite cylindrical shell to be evaluated.

2. The method for dynamic assessment of ultimate bearing capacity according to claim 1, characterized in that: In step 7), the five hierarchical feature parameters are: Layer depth t is the total thickness of the shell, are the inner and outer layer strain values ​​corresponding to the layered point set; Layered area A s is the area of ​​the small triangles or grid cells that constitute the delamination region, g is the number of points where delamination occurs; The minimum distance between the layer boundary and the end is l = min(|y pm -y bottom |,|y pm -y top |), min is the minimum value, y pm is the y coordinate of the layer boundary point, y bottom is the y-axis coordinate of the bottom end of the pressure shell, top is the y-axis coordinate of the top of the pressure hull, and the pressure hull is arranged axially along the y-axis direction; Layered form factor P f is the perimeter of the delamination area, and π is the ratio of pi; Number of layers δ v Is the decision function for whether the vth region constitutes an independent layer:

3. The method for dynamic assessment of ultimate bearing capacity according to claim 1, characterized in that: In step 6), the extreme value statistics of change point detection Cumulative Sum is the value of the strain signal at time k, Include x is the position of the change point, y is the number of data points of the strain signal, c is the data length, and μ0 is the expected mean strain value when there is no stratification.

4. The method for dynamic assessment of ultimate bearing capacity according to claim 1, characterized in that: In step 1), the critical buckling load of the non-layered composite column shell is F is the design safety factor, ranging from 1 to 5, m is the axial wave number of the cylindrical shell, n is the circumferential wave number of the cylindrical shell, L is the length of the cylindrical shell, R is the outer radius of the composite cylindrical shell, A ij ,B ij ,D ij is the element in the i-th row and j-th column of the stiffness matrices of A, B, and D.

5. The method for dynamic assessment of ultimate bearing capacity according to claim 4, characterized in that: Stiffness matrix Stiffness matrix Stiffness matrix Z i , Z i-1 Represents the distance between the mid-surface of the i-th layer and the i-1-th layer and the center plane of the total thickness, the constitutive matrix T is the matrix transpose, E1 is the Young's modulus in the fiber direction, E2 is the Young's modulus perpendicular to the fiber direction, G 12 is the shear modulus on the horizontal plane, v 12 Poisson's ratio in the principal direction, v 21 is the Poisson’s ratio in the secondary direction, θ is the fiber angle of the i-th layer.

6. The method for dynamic assessment of ultimate bearing capacity according to claim 1, characterized in that: In step 5), the Kalman filter is used to calculate the inner and outer layer strain values ​​ε in , ε out Perform denoising and smoothing.

7. The method for dynamic assessment of ultimate bearing capacity according to claim 1, characterized in that: In step 5), the strain value of the non-delaminated composite cylindrical shell is collected by the hydraulic pressure test, and the noise-free strain value inside and outside the shell measured by the hydraulic pressure test is used as the ideal strain value ε of the inner and outer layers y,in , ε y,out The process of the seawater pressure test is as follows: a multi-level constant pressure in the range of 0 to 10 MPa is applied through a closed static pressure chamber or a hydraulic loading device, and each level of pressure is maintained for 5 to 10 minutes. After the strain data stabilizes, the synchronous strain responses of the patch fiber Bragg grating sensor and the metal foil strain gauge are recorded respectively. Based on the obtained test data, their respective strain-pressure response curves are established.

8. The method for dynamic assessment of ultimate bearing capacity according to claim 1, characterized in that: In step 3), the regression model of the attenuation coefficient K is subjected to F test and p-value analysis. If the p-value is less than 0.05, the variable has a significant effect on the response value, otherwise the variable has no significant effect; Then evaluate the goodness of fit and calculate the coefficient of determination When R 2 >0.95 indicates good fit, P cr(i) is the critical buckling load P of the i-th data point cr ; is the predicted i-th response value after removing the i-th data point, Z is the number of samples, P cr Finally, an additional 10% of the experimental schemes are selected for random parameters to simulate, and the model prediction values ​​are compared with the simulation results. The error between the results of the regression model and the results of the numerical simulation is calculated. An error less than 10% indicates the accuracy of the regression model.

9. The method for dynamic assessment of ultimate bearing capacity according to any one of claims 1 to 8, characterized in that: In step 7), after obtaining the critical load of the composite column shell to be evaluated, an ultrasonic detection method is used to actually detect the delamination area to be evaluated, identify defects, and provide delamination defect parameters; a thickness gauge is used to actually measure the shell thickness to obtain the true thickness distribution, and a three-dimensional laser scan is performed on the layered pressure shell to obtain the true three-dimensional true contour.

10. The method for dynamic assessment of ultimate bearing capacity according to claim 9, characterized in that: Based on the actual measured layer parameters, reverse modeling is performed to obtain the numerical model of the layered pressure shell and calculate the critical load of the pressure shell; the critical load P of the composite column shell to be evaluated is calculated. cr By comparing with the actual measurement results, online calibration of material performance degradation can be achieved.

Images