Failure suppression method around connection holes of aircraft CFRP panels based on assembly stress equilibrium control

Abstract

The present invention belongs to the field of aviation manufacturing engineering / aircraft assembly, and proposes a method for suppressing failure around the connection holes of aircraft CFRP wall panels based on assembly stress balance control. The method obtains the stress cloud map and maximum stress around the hole under different node deformations, constructs a deformation-stress characteristic mapping model to predict the stress around the hole, and obtains the damage extension mechanism of the area around the hole under non-equilibrium stress based on the dynamic cross-scale finite element method, and determines the stress concentration tolerance; the current process parameters and geometric parameters are judged and optimized to form a control method for failure. The present invention solves the problem that the fitting gap in the assembly of CFRP wall panels with holes causes forced assembly and affects the fatigue life of the structure. A deformation-stress prediction model around the hole of CFRP wall panels with holes that is independent of the loading mode is constructed using a deep learning algorithm. The stress in the area around the hole can be predicted based on the actual deformation at the production site, and real-time perception of the stress state of the wall panel during assembly is realized.

Description

Technical Field

[0001] The invention relates to the field of aviation manufacturing engineering / aircraft assembly, and in particular to a method for suppressing failure around aircraft CFRP panel connection holes based on assembly stress balance control. Background Art

[0002] The panel structure with CFRP laminate as the main material is a typical load-bearing structure on the aircraft. It plays a very important role in maintaining the overall shape of the aircraft and bearing aerodynamic loads. The CFRP panel is large in size and thin in thickness. The curvature of different areas varies greatly, so the overall rigidity is weak. The work processes such as curing molding, fixture clamping, and hole making processing will cause the size of the CFRP panel to deviate from the design value, and it is easy to form a gap at the fitting surface during assembly. Although tightening the bolts connecting the panel and other load-bearing parts of the aircraft can eliminate the gap, it causes a forced connection, which causes non-balanced stress in the area around the hole on the panel. When the CFRP panel is forced to connect, there is a fitting gap between the panel and the wing beam. After tightening the high-locking bolts, the fitting gap disappears, but lateral overload and thickness misalignment will be formed around the nail hole, causing non-balanced stress in the area around the hole.

[0003] When the main engine factories are currently assembling wall panels, the workers mainly adopt the methods of clamping force, padding compensation, grinding, etc. to solve the problem of fitting gaps in CFRP wall panels.

[0004] The clamping force method is to measure the clamping force required to clamp the wall panel and compare it with a fixed clamping force value. If it is less than the fixed clamping force value, no measures will be taken and direct assembly will be carried out; if it exceeds the clamping force value, the wall panel holes will be ground or padded according to the gap size to reduce the stress concentration level around the holes.

[0005] Padding compensation is to fill the gap between CFRP parts with gaskets of specific thickness and shape to compensate the gap between the connected parts, reduce the deformation of the wall during connection, and achieve the effect of reducing the stress concentration in the area around the hole. Usually, when the gap is 0.12-1.54mm, appropriate gaskets can be selected for gap compensation according to the specific gap thickness.

[0006] Grinding is another measure to reduce the gap between the wall panels. The raised parts of the wall panels when they are fitted with another connected part are ground flat so that the area around the raised parts can naturally fit with the other connected part, avoiding a forced connection.

[0007] The current methods used by OEMs to solve the problem of gap in CFRP panel assembly, such as fixed clamping force, gap filling compensation, and grinding, have the following two defects:

[0008] First, the values ​​of the clamping force standard, pad thickness, and grinding size in the existing technical solutions are often based on experience, lack of standardization, and contain subjective assumptions;

[0009] Second, the existing measures lack feedback on their effects. The unbalanced stress caused by forced connection is still difficult to eliminate, and the compensation effect and accuracy are difficult to guarantee, and a feedback loop has not been formed. Summary of the invention

[0010] The present invention is used for failure suppression of connecting holes in aircraft composite material wall panels assembly, and mainly solves the problem that when CFRP wall panels have a fitting gap during aircraft assembly, forced connection generates non-equilibrium stress around the hole, which easily causes failure of the area around the hole after being loaded, thereby affecting the bearing capacity and service life of the aircraft structure.

[0011] The technical solution of the present invention is a method for suppressing failure around the connection holes of aircraft CFRP panel based on assembly stress balance control. The method is aimed at the area around the hole where there is a fitting gap during the assembly process of the CFRP panel. Usually, a rectangular area with deformation around the hole is selected. This area will produce non-balanced stress during forced assembly, affecting the service status of the structure.

[0012] The specific steps are as follows:

[0013] Step 1: Obtain a model training data set; establish a finite element model of the target object, convert process parameters and geometric parameters into the form of boundary conditions and loads, and input them into the finite element model; automatically change the boundary conditions and loads for the model multiple times through a pre-written program and input them into the finite element model to change the process parameters and geometric parameters; export the hole circumference stress cloud map and maximum stress under different node deformations after the finite element model simulation is completed to obtain a model training data set;

[0014] Step 2, construct a deformation-stress characteristic mapping model that is independent of the loading mode; select a data set with a proportion r as the training set for deep learning training, and the remaining proportion (1-r) of the data is used as the verification set; use the deep learning network for data training to establish the mapping relationship between the deformation of each node and the stress cloud map around the hole and the maximum stress; use the deformation of each node as input data to obtain the deformation-stress characteristic mapping model F, which is only related to the actual deformation of each node; the deformation-stress characteristic mapping model obtained by training can realize the rapid generation of the stress distribution map of the hole area under the parameter after reading the input deformation.

[0015] Step 3: Model verification: Use the data in the verification set to verify the validity of the deformation-stress characteristic mapping model, input the node deformation in the verification set into the deformation-stress characteristic mapping model, and perform error analysis on the output predicted maximum hole circumference stress information and the maximum hole circumference stress obtained by finite element analysis; adjust and iterate the deformation-stress characteristic mapping model parameters according to the error results to obtain the final deformation-stress characteristic mapping model to predict the hole circumference stress level;

[0016] Step 4: Based on the dynamic cross-scale finite element method, the damage expansion mechanism of the hole area under non-equilibrium stress is obtained, and the stress concentration tolerance is determined accordingly; forced assembly will cause stress concentration in the hole area. If the aircraft assembly enters the next link without treatment, the hole area will carry non-equilibrium stress into the service state and bear the service load. As a result, the load of the hole stress concentration area will be greater than the expected design value, which is prone to failure within the original service life or maintenance cycle, affecting the overall stability of the aircraft structure.

[0017] Step 5: Determine and optimize the current process parameters and geometric parameters to form a control method for failure; compare the stress value σ predicted by the deformation-stress characteristic mapping model i The upper limit of stress concentration σ d , when σ i ≤σ d When σ , the structural strength under this process parameter and geometric parameter meets the requirements, and it is judged that the process parameter and geometric parameter are feasible; when σ i >σ d When the process parameters or geometric parameters of the target area are changed, the adjusted process parameters and geometric parameters are used to predict the stress again and compared with the upper limit of stress concentration, until the stress predicted by the adjusted deformation-stress characteristic mapping model is less than the upper limit of stress concentration, that is, the expected life of the current structure meets the design expectations, and the optimization of the entire assembly process parameters is completed.

[0018] The process parameters include tightening sequence P1, gap filling compensation P2 and tightening torque P3; the geometric parameters include fitting gap G1, gasket thickness G2 and bolt spacing G3;

[0019] P1 = {sq1, sq2, sq3 ...}, sq1, sq2, sq3 ... is a specific connection order;

[0020] P2 = {0, 1}, 0 and 1 represent no gap compensation and gap compensation respectively;

[0021] P3={tq1, tq2, tq3…}, tq1, tq2, tq3… are different torque values;

[0022] G1={gp1,gp2,gp3…}, gp1,gp2,gp3…are different gap amounts;

[0023] G2={gd1,gd2,gd3…}, gd1,gd2,gd3…are different gasket thicknesses;

[0024] G3={gw1,gw2,gw3…}, gw1,gw2,gw3…are different bolt spacings.

[0025] The deformation-stress characteristic mapping model realizes reading the input node deformation amount and generating a node stress distribution diagram of the hole periphery area under the parameter;

[0026]

[0027] σ i is the node stress; δ i (x, y, z) is the node deformation; i is the node number.

[0028] The step 4 is specifically as follows:

[0029] S4.1: Propose a dynamic cross-scale finite element method;

[0030] Based on the secondary development function development program of finite element software, according to the stress threshold σ th Identify stress concentration units, and use a microscopic fiber matrix model for the stress concentration units. Use a laminate model for other units. The unit set of the two is expressed as:

[0031] Stress concentration unit node set: SetMi = {σ i |σ i ≥σ th};

[0032] Other unit node sets: SetMa = {σ i |σ i <σ th}, where i is the node number;

[0033] When the stress conditions in the target area change, the secondary development function development program updates the area range using the stress concentration unit to achieve dynamic cross-scale finite element simulation;

[0034] S4.2: Quantitative characterization of hole wall damage;

[0035] A refined gradient model of the hole wall interface based on statistical principles is established to reduce the stiffness of the damaged hole wall interface and simulate the influence of the hole wall damage generated during the hole making process on the fatigue expansion mode of the area around the hole. The specific process is to traverse the elements in the stress concentration unit set SetMi to determine whether its stiffness reaches the damage limit. When the damage limit is reached, the stiffness is reduced according to the following formula:

[0036] E=(1-D m )E0

[0037] Among them, E is the stiffness after reduction; E0 is the stiffness before reduction; D m is the damage coefficient;

[0038] S4.3: Determination of stress concentration tolerance;

[0039] Take the structural design load as the input condition, set the assembly gap to 0, repeat the above steps S4.1-S4.2, and obtain the maximum failure stress σ of the node around the hole d As the upper limit of stress concentration, the stress concentration tolerance of each node in the current stress concentration unit node set is [0,σ d ).

[0040] The target area is the area where there is deformation around the hole, and the area is set as a rectangular area, and the area of ​​the rectangular area is A. def ,

[0041] A def =(L+m)(W+n)

[0042] Among them, L is the maximum length of the area around the hole where deformation occurs, W is the maximum width of the area around the hole where deformation occurs, m and n are the boundary margins of the long side and the wide side, respectively, and the values ​​are 10%-30% of L and W.

[0043] In the step 5, the process parameters or geometric parameters of the target area are changed by adding pads for compensation, adjusting the bolt connection sequence or adjusting the bolt tightening torque.

[0044] This patent application proposes a finite element simulation method that takes into account both accuracy and efficiency for the failure mechanism of the hole area under the combined action of assembly non-equilibrium stress and expected design load, and studies the relationship between the non-equilibrium stress level and the expected fatigue life of the structure.

[0045] The beneficial effects of the present invention are as follows: the problem that the impact of the process parameters used in the assembly site of aircraft CFRP wall panels on the assembly state cannot be perceived in real time at the assembly site is solved, and the real-time prediction of the stress around the hole of the CFRP laminate with holes is realized; the dynamic cross-scale calculation method is used in the study of the damage mechanism and fatigue life of the hole area under various load stress concentrations, taking into account both the calculation accuracy and efficiency. At the same time, the initial damage around the hole is taken into account, which better simulates the extension of the damage around the hole in real situations; a process-geometry optimization method within the stress concentration tolerance that meets the fatigue life requirements is proposed, which solves the problem that the existing technology remains at the theoretical level and is difficult to implement at the assembly site. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flow chart for constructing the deformation-stress feature mapping model;

[0047] Figure 2 The figure is a flow chart of the failure suppression method around the connection holes of aircraft CFRP panel based on assembly stress balance control. DETAILED DESCRIPTION

[0048] First, the present invention targets the area around the hole where there is a gap in the assembly process of the CFRP wall panel. Usually, the area around the hole where there is deformation is selected. This area will produce non-equilibrium stress during forced assembly, affecting the service state of the structure. The deformation area is described by a certain rectangular local area, and the area of ​​the rectangle is recorded as A def , which is specifically defined as:

[0049] A def =(L+m)(W+n)

[0050] Among them, L is the maximum length of the area around the hole where deformation exists, W is the maximum width of the area around the hole where deformation exists, m and n are the boundary margins of the long side and the wide side, respectively, and the values ​​are 10%-30% of L and W.

[0051] Secondly, a stress prediction model based on deep learning is established for the above rectangular area, that is, a deformation-stress characteristic mapping model that is independent of the loading mode. The construction process of the deformation-stress characteristic mapping model is as follows: Figure 2 shown.

[0052] Step 1: Obtaining the model training data set. Obtain the geometric features of the selected rectangular area, establish a three-dimensional model, convert the assembly process parameters and geometric parameters into boundary conditions and loading methods, and perform finite element simulation calculations. The process parameters include tightening sequence (P1), whether to compensate for gap filling (P2) and tightening torque (P3); the geometric parameters include fitting gap (G1), gasket thickness (G2) and bolt spacing (G3). The above parameter values ​​can be expressed as:

[0053] P1 = {sq1, sq2, sq3 ...}, sq1, sq2, sq3 ... is a specific connection order;

[0054] P2 = {0, 1}, 0 and 1 represent no gap compensation and gap compensation respectively;

[0055] P3={tq1, tq2, tq3…}, tq1, tq2, tq3… are different torque values;

[0056] G1={gp1,gp2,gp3…}, gp1,gp2,gp3…are different gap amounts;

[0057] G2={gd1,gd2,gd3…}, gd1,gd2,gd3…are different gasket thicknesses;

[0058] G3={gw1,gw2,gw3…}, gw1,gw2,gw3…are different bolt spacings.

[0059] The above process parameters and geometric parameters are combined according to the actual production and then simulated and calculated. The hole stress cloud map, maximum stress and other information after the simulation are exported to obtain the model training data set, named SetD, as the input for the next step.

[0060] Step 2: Construction of a deformation-stress characteristic mapping mechanism that is independent of the loading mode. SetD includes the deformation of each node in the target area in three-dimensional space δ i (x, y, z), the angle of each layer, the stress of each node and other information. SetD is processed, and the data with a proportion of r is selected as the training set input model, and the remaining data with a proportion of (1-r) is used as the validation set. A deep neural network is used for data training to establish a mapping relationship model (F) between the deformation and stress of the target area. The node stress is σ i , then the relationship can be expressed as:

[0061]

[0062] Among them, i is the node number, which is automatically traversed by the program. Since the model uses the deformation of each node in the target area as input data, and does not use information such as the loading method to eliminate the gap as input, the obtained model is independent of the loading method and is only related to the actual deformation. The deformation-stress correlation model obtained by training can quickly generate the stress distribution diagram of the hole area under this parameter after reading the input deformation.

[0063] Step 3: Model verification. Use the data in the verification set to verify the validity of the deformation-stress characteristic mapping model. Input the node deformation in the verification set into the deformation-stress characteristic mapping model, and perform error analysis on the maximum stress value predicted by the deformation-stress characteristic mapping model and the maximum stress obtained by finite element analysis. Adjust the parameters of the deformation-stress characteristic mapping model iteratively according to the error results to achieve better results.

[0064] Step 4: Based on the dynamic cross-scale finite element calculation method, the damage expansion mechanism of the area around the hole under non-equilibrium stress is revealed, and the stress concentration tolerance is determined accordingly. Forced assembly will cause stress concentration in the area around the hole. If the aircraft assembly enters the next link without treatment, the area around the hole will carry non-equilibrium stress into the service state and bear the service load. As a result, the load-bearing capacity of the stress concentration area around the hole will be greater than the expected design value, which is prone to failure within the original service life or maintenance cycle, affecting the overall stability of the aircraft structure. A finite element simulation method that takes into account both accuracy and efficiency is proposed for the failure mechanism of the area around the hole under the combined action of assembly non-equilibrium stress and expected design load, and the relationship between the non-equilibrium stress level and the expected fatigue life of the structure is studied.

[0065] S4.1 proposed a dynamic cross-scale finite element method. In order to determine the area that needs to be simulated using a microscopic model, a related program was developed based on the secondary development function of the finite element software. th Identify stress concentration units, and use a microscopic fiber matrix model for stress concentration areas, and a laminate model for other areas. The unit set of the two is expressed as:

[0066] Mesoscopic unit model node set: SetMi = {σ i |σ i ≥σ th};

[0067] Macro unit model node set: SetMa = {σ i |σ i <σ th}, where i is the node number.

[0068] When the stress conditions in the target area change, the program will automatically update the area using the mesoscopic model to achieve dynamic cross-scale finite element simulation, taking into account both accuracy and computational efficiency;

[0069] S4.2 Quantitative characterization of hole wall damage. A refined gradient model of the hole wall interface based on statistical principles is established to reduce the stiffness of the damaged hole wall interface and simulate the effect of the hole wall damage generated during the hole making process on the fatigue expansion mode of the hole area. The specific process is to traverse the elements in the stress concentration unit set SetMi to determine whether its stiffness reaches the damage limit. If it reaches the limit, the stiffness is reduced according to the following formula:

[0070] E=(1-D m )E0

[0071] Among them, E is the stiffness after reduction; E0 is the stiffness before reduction; D m is the damage coefficient, and its value is determined based on calculation experience or experimental calibration.

[0072] S4.3 Determination of stress concentration tolerance. Take the structural design load as the input condition, set the assembly gap to 0, repeat the above simulation calculation process, and obtain the stress concentration upper limit σ of the current structure specific unit. d .

[0073] Finally, the current process parameters and geometric parameters are determined and optimized to form a control method for failure. Figure 2 As shown, the stress value σ predicted by the deformation-stress characteristic mapping model is compared i The upper limit of stress concentration tolerance σ d , if σ i ≤σ d , indicating that the strength of the structure under this process geometric parameter meets the requirements, and the process geometric parameter is feasible; if σ i >σ d , it is necessary to change the process-geometry parameters of this area by adding pads for compensation, adjusting the bolt connection sequence, adjusting the bolt tightening torque, etc. Use the adjusted process-geometry parameters to predict stress again and compare it with the stress concentration tolerance, repeat the above process until the adjusted predicted stress is less than the stress concentration tolerance, that is, the expected life of the current structure meets the design expectations, and the optimization of the entire assembly process parameters is completed.

Claims

1. A method for suppressing failure around aircraft CFRP panel connection holes based on assembly stress balance control, characterized in that: The specific steps are as follows: Step 1: Obtain a model training data set; establish a finite element model of the target area, convert process parameters and geometric parameters into the form of boundary conditions and loads, and input them into the finite element model; change the process parameters and geometric parameters by changing the boundary conditions and loads multiple times and inputting them into the finite element model; After the finite element model simulation is completed, the hole circumference stress cloud diagram and the maximum stress at different node deformations are exported to obtain the model training data set; Step 2: construct a deformation-stress characteristic mapping model that is independent of the loading mode; For the model training data set, a data with a proportion of r is selected as the training set for deep learning training, and the remaining proportion of data (1-r) is used as the validation set; Use deep learning network to train data and establish the mapping relationship between the deformation of each node and the stress cloud map around the hole and the maximum stress; Using the deformation of each node as input data, the deformation-stress characteristic mapping model F is obtained, which is only related to the actual deformation of each node; Step 3: Model verification: Use the data in the verification set to verify the validity of the deformation-stress characteristic mapping model, input the node deformation in the verification set into the deformation-stress characteristic mapping model, and perform error analysis on the output predicted maximum hole circumference stress information and the maximum hole circumference stress obtained by finite element analysis; adjust and iterate the deformation-stress characteristic mapping model parameters according to the error results to obtain the final deformation-stress characteristic mapping model to predict the hole circumference stress level; Step 4: Based on the dynamic cross-scale finite element method, the damage propagation mechanism of the hole area under non-equilibrium stress is obtained, and the stress concentration tolerance is determined accordingly; Step 5: Determine and optimize the current process parameters and geometric parameters to form a control method for failure; Comparison of stress values ​​σ predicted by deformation-stress characteristic mapping model i The upper limit of stress concentration σ d , when σ i ≤σ d When σ , the structural strength under this process parameter and geometric parameter meets the requirements, and it is judged that the process parameter and geometric parameter are feasible; when σ i >σ d When the stress is predicted, the process parameters or geometric parameters of the target area are changed, and the adjusted process parameters and geometric parameters are used to predict the stress again and compared with the upper limit of stress concentration, until the stress predicted by the adjusted deformation-stress characteristic mapping model is less than the upper limit of stress concentration, that is, the expected life of the current structure meets the design expectations, and the optimization of the entire assembly process parameters is completed.

2. The method for suppressing failure around the connection hole of aircraft CFRP panel based on assembly stress balance control according to claim 1, characterized in that: The process parameters include tightening sequence P1, gap filling compensation P2 and tightening torque P3; the geometric parameters include fitting gap G1, gasket thickness G2 and bolt spacing G3; P1 = {sq1, sq2, sq3 ...}, sq1, sq2, sq3 ... is a specific connection order; P2 = {0, 1}, 0 and 1 represent no gap compensation and gap compensation respectively; P3={tq1, tq2, tq3…}, tq1, tq2, tq3… are different torque values; G1={gp1,gp2,gp3…}, gp1,gp2,gp3…are different gap amounts; G2={gd1,gd2,gd3…}, gd1,gd2,gd3…are different gasket thicknesses; G3={gw1,gw2,gw3…}, gw1,gw2,gw3…are different bolt spacings.

3. The method for suppressing failure around the connection hole of aircraft CFRP panel based on assembly stress balance control according to claim 2, characterized in that: The deformation-stress characteristic mapping model realizes reading the input node deformation amount and generating a node stress distribution diagram of the hole periphery area under the parameter; σ i is the node stress; δ i (x, y, z) is the node deformation; i is the node number.

4. The method for suppressing failure around the connection hole of aircraft CFRP panel based on assembly stress balance control according to claim 1, characterized in that: The step 4 is specifically as follows: S4.1: Propose a dynamic cross-scale finite element method; Based on the secondary development function development program of finite element software, according to the stress threshold σ th Identify stress concentration units, and use a microscopic fiber matrix model for the stress concentration units. Use a laminate model for other units. The unit set of the two is expressed as: Stress concentration unit node set: SetMi = {σ i |σ i ≥σ th }; Other unit node sets: SetMa = {σ i |σ i <σ th }, where i is the node number; When the stress conditions in the target area change, the secondary development function development program updates the area range using the stress concentration unit to achieve dynamic cross-scale finite element simulation; S4.2: Quantitative characterization of hole wall damage; A refined gradient model of the hole wall interface based on statistical principles is established to reduce the stiffness of the damaged hole wall interface and simulate the influence of the hole wall damage generated during the hole making process on the fatigue expansion mode of the area around the hole. The specific process is to traverse the elements in the stress concentration unit set SetMi to determine whether its stiffness reaches the damage limit. When the damage limit is reached, the stiffness is reduced according to the following formula: E(1-D m )E0 Among them, E is the stiffness after reduction; E0 is the stiffness before reduction; D m is the damage coefficient; S4.3: Determination of stress concentration tolerance; Take the structural design load as the input condition, set the assembly gap to 0, repeat the above steps S4.1-S4.2, and obtain the maximum failure stress σ of the node around the hole d As the upper limit of stress concentration, the stress concentration tolerance of each node in the current stress concentration unit node set is [0,σ d ).

5. The method for suppressing failure around the connection hole of aircraft CFRP panel based on assembly stress balance control according to claim 4, characterized in that: The target area is the area where there is deformation around the hole, and the area is set as a rectangular area, and the area of ​​the rectangular area is A. def , A def =(L+m)(W+n) Among them, L is the maximum length of the area around the hole where deformation occurs, W is the maximum width of the area around the hole where deformation occurs, m and n are the boundary margins of the long side and the wide side, respectively, and the values ​​are 10%-30% of L and W.

6. The method for suppressing failure around the connection holes of aircraft CFRP panel based on assembly stress balance control according to claim 1, characterized in that: In the step 5, the process parameters or geometric parameters of the target area are changed by adding pads for compensation, adjusting the bolt connection sequence or adjusting the bolt tightening torque.

Images