Method and apparatus for analyzing composite structure
A machine learning-based method for composite structure damage tolerance analysis addresses the limitations of visual inspections by developing a depression depth meta model, enabling efficient and accurate damage prediction in composite structures.
Patent Information
- Application Number
- JP2024090572
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-07-07
- Filing Date
- 2024-06-04
- Publication Date
- 2025-01-28
AI Technical Summary
Current methods for assessing damage tolerance in complex composite structures, such as those used in aircraft, are limited by subjective visual inspections and lack accuracy in predicting damage depth and range, especially in digital models, leading to inefficiencies and high costs in physical testing.
A method utilizing a machine learning algorithm trained on past physics test datasets to develop a depression depth meta model, which estimates impact energy and simulates damage in composite structures, reducing the need for physical testing and enabling faster, more accurate damage tolerance analysis.
This approach allows for rapid and cost-effective evaluation of damage tolerance in composite structures, significantly reducing development time and costs associated with physical testing, while enhancing the precision of damage prediction.
Smart Images
Figure 2025013747000001 
Figure 2025013747000002
Abstract
Description
[Technical field]
[0001]
[0001] The present disclosure relates to the field of structural analysis, and more particularly, to structures made of composite materials for damage tolerant design. [Background technology]
[0002]
[0002] Aircraft certification requires the demonstration of durability and damage tolerance (DaDT). DaDT refers to the ability of a structure to resist damage propagation under cyclic conditions and retain sufficient strength throughout the aircraft's service life, even in the presence of damage / defects. Damage tolerance typically refers to the ability of a structure to retain sufficient strength in the presence of defects / damage, which may or may not be readily detectable.
[0003]
[0003] Current approaches to verifying damage tolerant structural designs rely on a largely subjective structural testing process that utilizes specimens exhibiting damage that must be identified by visual inspection from a predefined distance. Due to the inherent uncertainty of visual inspection, a threshold dent depth may be employed to quantify an acceptable initial damage state and verify damage tolerance. While metal-based structures can be analyzed with traditional finite element analysis, the complexity of composite structures limits the ability of digital models to predict dent depths and underlying damage extents within composite structures. Summary of the Invention
[0004]
[0004] A method of analyzing a composite structure for damage tolerance is disclosed herein. The method includes obtaining a structural configuration for a composite component and identifying at least one metric of interest for the structural configuration. A dimple depth metamodel is developed based on information obtained from the structural configuration of the composite component, the at least one metric of interest, and a machine learning algorithm. The at least one metric of interest provides a constraint for the dimple depth metamodel, and the machine learning algorithm is trained based on historical physical testing data sets of a plurality of composite components. The dimple depth metamodel for the structural configuration is utilized to estimate an impact energy to define a threshold dimple depth in the structural configuration. A computer-based simulation of the structural configuration is performed utilizing the impact energy from the dimple depth metamodel. At least one design value for the structural configuration is identified based on output data generated from the computer-based simulation.
[0005]
[0005] In one or more embodiments of the method, developing the depression depth metamodel includes providing greater weight to past physical test data sets of multiple previously analyzed physical composite components based on material or configuration similarities between the structural configuration and a corresponding one of the previously analyzed physical composite components.
[0006] In one or more embodiments of the method, the plurality of previously analyzed physical composite components includes composite components from a plurality of different vehicle platforms.
[0007] In one or more embodiments of the method, the historical physical test data set includes a plurality of test data inputs associated with corresponding test data outputs for a plurality of previously analyzed physical composite components.
[0008]
[0008] In one or more embodiments of the method, the test data input includes at least one of force-time history, displacement-time history, impact energy-time history, impactor configuration, indentation depth, material properties, associated boundary conditions, or composite structural configuration defining composite layup and shape.
[0009] In one or more embodiments of the method, the material properties include at least one of mode 1 strain energy rate, mode 2 strain energy rate, interlaminar strength, stiffness ply properties, or ply strength properties.
[0010] In one or more embodiments of the method, the test data output includes at least one of the type of damage resulting from the impact, the energy absorbed from the impact, or the measured dent depth.
[0011]
[0011] In one or more embodiments of the method, the structural configuration includes at least one of a skin joined to a reinforcing member, a skin joined to a secondary structure, or a reinforcing structure joined to a secondary structure, the secondary structure including at least one of a rib or a spar.
[0012] In one or more embodiments of the method, the at least one metric of interest includes dynamic stiffness, energy absorption rate, energy flux into the structural configuration, or initial deflection rate.
[0013]
[0013] In one or more embodiments of the method, output data generated from the computer-based simulation includes an identification of the strength of a damaged portion of a structural configuration under at least one of compressive loads, tensile loads, and shear loads.
[0014] In one or more embodiments of the method, the composite component includes a plurality of individual composite elements secured relative to one another.
[0015] In one or more embodiments of the method, the historical physical testing data set includes a calculation of a damage tolerance criterion based on the physical testing data.
[0016] In one or more embodiments of the method, the machine learning algorithm comprises a Bayesian-based algorithm or a physics-based algorithm.
[0017]
[0017] A non-transitory computer-readable medium embodying programmed instructions, which when executed by a processor, are operable to perform a method, is disclosed. The method includes obtaining a structural configuration for a composite component and identifying at least one metric of interest for the structural configuration. A dent depth metamodel is developed based on information obtained from the structural configuration of the composite component, the at least one metric of interest, and a machine learning algorithm. The at least one metric of interest provides a constraint for the dent depth metamodel, and the machine learning algorithm is trained based on historical physical testing data sets of a plurality of composite components. The dent depth metamodel for the structural configuration is utilized to estimate an impact energy to define a threshold dent depth in the structural configuration. A computer-based simulation is performed on the structural configuration utilizing the impact energy from the dent depth metamodel. At least one design value for the structural configuration is identified based on output data generated from the computer-based simulation.
[0018]
[0018] In one or more embodiments, the historical physical test dataset includes a plurality of test data inputs associated with corresponding test data outputs for a plurality of previously analyzed physical composite components, and the historical physical test dataset includes calculations for damage tolerance criteria based on the physical test data.
[0019]
[0019] In one or more embodiments, the test data input includes at least one of force-time history, displacement-time history, impact energy-time history, impactor configuration, indentation depth, material properties, associated boundary conditions, or composite structural configuration defining composite layup and shape.
[0020] In one or more embodiments, the material properties include at least one of mode 1 strain energy rate, mode 2 strain energy rate, interlaminar strength, stiffness ply properties, or ply strength properties.
[0021] In one or more embodiments, the test data output includes at least one of the type of damage resulting from the impact, the energy absorbed from the impact, or the measured dent depth.
[0022] In one or more embodiments, the at least one metric of interest includes dynamic stiffness, energy absorption rate, energy flux into the structural configuration, or initial deflection rate.
[0023]
[0023] A non-transitory computer-readable medium embodying programmed instructions, which when executed by a processor, are operable to perform a method, is disclosed. The method includes obtaining a structural configuration for a composite component and identifying at least one metric of interest for the structural configuration. A dent depth metamodel is developed based on the structural configuration of the composite component, the at least one metric of interest, and information obtained from a machine learning algorithm. The at least one metric of interest provides constraints for the dent depth metamodel, and the machine learning algorithm is trained based on historical physical test data sets of a plurality of composite components, the historical physical test data sets including calculations for damage tolerance criteria based on the physical test data. The dent depth metamodel for the structural configuration is utilized to estimate impact energy to define a threshold dent depth in the structural configuration. A computer-based simulation is performed on the structural configuration utilizing the impact energy from the dent depth metamodel. At least one design value for the structural configuration is identified based on output data generated from the computer-based simulation. [Brief description of the drawings]
[0024] [Figure 1]
[0024] An exemplary computer system is illustrated. [Diagram 2]
[0025] A method is presented to analyze composite structures for damage tolerance. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0025]
[0026] Some embodiments of the present disclosure will now be described, by way of example only, with reference to the accompanying drawings in which the same reference numbers represent the same elements or same types of elements, and in which:
[0026]
[0027] These drawings and the following description show several specific exemplary embodiments of the present disclosure. Those skilled in the art will be able to devise various arrangements (even if not explicitly described or shown herein) that embody the principles of the present disclosure and are within the scope of the present disclosure. Furthermore, any examples described herein are intended to aid in the understanding of the principles of the present disclosure and should not be construed as being limited to the specifically listed examples and conditions. As a result, it is the claims and their equivalents that limit the present disclosure, rather than the specific embodiments or examples below.
[0027]
[0028] FIG. 1 illustrates the exemplary computer system 30 as a single computer module for illustrative simplicity. The computer system 30 may be physically embodied as one or more processing nodes having a non-transitory computer-readable storage medium 34, i.e., sufficient memory for an application, and associated hardware and software, such as, but not limited to, high-speed clocks, timers, input / output circuits, buffer circuits, etc. The non-transitory computer-readable storage medium 34 may include sufficient read-only memory, e.g., magnetic or optical memory. Computer-readable code or instructions embodying the methods described below may be executed during operation of the computer system 30. To that end, the computer system 30 may include one or more processors 32, e.g., logic circuits, application specific integrated circuits (ASICs), central processing units, microprocessors, and / or other essential hardware necessary to provide the programmed functions described herein.
[0028]
[0029] The display screen 36 may be connected to or in communication with the computer readable medium 34 and the processor(s) 32 to display a structural configuration 38 of the composite component to facilitate a graphical interface for a user implementing the present disclosure as described below. In one embodiment, the structural configuration 38 for the composite component includes a three-dimensional computer-aided design (CAD) that represents the composite component in a digital domain. In one illustrated embodiment, the structural configuration 38 includes a first composite element 38A, such as a composite skin, a constructed structure (e.g., stiffener, rib, spar, frame), or a joint (e.g., bonded, co-bonded, fastened). The first composite element 38A may also be attached to a composite element, such as a composite rib or spar. The structural configuration 38 may also include acres, joints, parts, subcomponents, components, and assemblies.
[0029]
[0030] Due to the complexity of composite structures, it has been difficult to accurately predict the dimple depth in a structural configuration for a composite component in the digital domain because the dimple depth is affected by many interrelated considerations. One consideration that affects the dimple depth in a composite component includes boundary conditions.
[0030]
[0031] Another consideration that influences dent depth in composite components includes construction effects, e.g., energy absorption or load path. In construction effects, larger structures tend to have a more compliant response due to the joints, constructed attachments, etc. mentioned above, resulting in more energy being required to achieve dent depth. For this reason, the threshold impact energy observed at the flat coupon level, typically used for sizing, is much lower than the threshold impact energy observed in constructed panels at the element level. This can lead to constructed panels having greater dent resistance than the coupon level elements used for testing. Because the panels exhibit greater dent resistance than would be predicted from coupon level element testing, the panels can be stronger than required. This can result in heavier components.
[0031]
[0032] One additional consideration that affects the dimple depth in a composite component includes the layup or thickness of the material in the composite component. This adds to the complexity of analyzing composite components because the dimple depth is not consistent across the layup or uniform in thickness. Some applications of composite structures also include metal components in the stackup. Thus, parts of the structural configuration may include composite-aluminum-composite, composite-Ti-composite, and other arrangements with composite metal in the overall stack.
[0032]
[0033] Yet another consideration that influences dent depth in composite components includes the material properties of the composite. Modern toughened composites rely on toughening agents mixed into the matrix. One common toughening agent includes thermoplastics. These materials experience both elastic nonlinearity and damage when impacted, making it difficult to isolate the responses that contribute to predicting dent depth in composite components that contain them.
[0033]
[0034] Yet another consideration that affects dent depth in composite components includes the interlayer / interface between plies in a composite layup, which may be ply-to-ply, composite layup-to-composite layup, or other composite-to-rib, spar, frame, or stringer, using nano-reinforcement, adhesives, etc.
[0034]
[0035] Considerable research has been done (and may still be done) across the industry to improve damage tolerant designs by modifying ply-to-ply or composite layup-to-composite layup or other composite-to-rib / spar / frame / stringer interfaces using some form of nano-reinforcement / adhesives / etc. This can impact Inner Mold Line (IML) and Outer Mold Line (OML) simulations and even laminate acreage.
[0035]
[0036] Due to the above considerations regarding the complexity of composite structures, an exemplary testing practice seeks to develop physical screening coupons to identify a threshold impact energy that produces a threshold dent depth. This approach involves testing multiple different impact energies to converge on a threshold impact energy for each impact location. Using this approach, a threshold impact energy is identified and then applied to coupons, elements, or subcomponents for which damage tolerance has been demonstrated. While tools exist for demonstrating damage tolerance, the industry is limited by the need to perform impact investigation testing on each configuration every time a change is made. This approach results in significant costs to build and test composite components for damage tolerance.
[0036]
[0037] One feature of the present disclosure is to provide an analytical approach to identify damage tolerance of composite structures without building and testing physical hardware. The analytical approach utilizes an dent depth metamodel that is informed by historical physical test datasets through machine learning (ML) algorithms to estimate impact energy to simulate dent depths of structural configurations 38 of composite components. The approach of the present disclosure reduces development costs associated with testing physical survey panels and allows for an expansion of the design space by evaluating structural configurations of composite components in a much faster time frame in the digital domain. For example, the approach disclosed herein may reduce testing time from months to days or even hours.
[0037]
[0038] 2 illustrates an exemplary method 100 for analyzing a structural configuration 38 of a composite component for damage tolerance. The method 100 includes receiving a structural configuration 38 of the composite component (block 102). In one embodiment, the structural configuration 38 defines a composite component having a single stiffness level and is represented in a three-dimensional digital domain.
[0038]
[0039] A structural configuration 38 may identify at least one composite material forming the structural configuration 38 and at least one composite element bonded to another composite element. For example, a structural configuration 38 may include a composite element such as at least one of a skin bonded to a rigid member, a skin bonded to a secondary structure, or a stiffening structure bonded to a secondary structure including at least one of a rib or a spar. The composite elements represented within the structural configuration 38 may be bonded, bonded, fastened, or attached to one another within the structural configuration 38.
[0039]
[0040] The method 100 then identifies at least one metric of interest for the structural configuration 38 at block 104. The at least one metric of interest may include dynamic stiffness, energy absorption rate, energy flux into the composite component, or initial deflection rate. In one embodiment, the at least one metric of interest may be selected by a user of the computer system 30 for evaluation. Alternatively, the computer system 30 may pre-select the at least one metric of interest for evaluation.
[0040]
[0041] The method 100 may then develop a depression depth metamodel (block 106) based on the structural configuration (block 102), the at least one metric of interest (block 104), and information obtained from the machine learning algorithm (block 110). The at least one metric of interest (block 104) provides constraints for the development of the depression depth metamodel in block 106.
[0041]
[0042] In a machine learning algorithm at block 110, a machine learning algorithm is trained to provide information for the development of the depression depth metamodel at block 106. The machine learning algorithm may be pre-trained based on the past physical test data set from block 108 as an input training data set. In addition to the physical test data, the past physical test data set from block 108 includes a calculation for a damage tolerance criterion based on the physical test data. This calculation may be performed at block 108 by the computer system 30, or this calculation may be performed before the method 100 obtains the past physical test data set at block 108. The damage tolerance criterion may be based on at least one of calculations such as hand calculations (theoretical), design curves (theoretical, semi-empirical, empirical), or associated design criteria / guidelines that affect the structural configuration (e.g., safety margins).
[0042]
[0043] Alternatively, the machine learning algorithm may be trained on the historical physical testing data set from block 108 in conjunction with developing the depression depth metamodel for a particular structural configuration from block 102. In one embodiment, the machine learning algorithm from block 110 is a Bayesian-based algorithm, although other types of machine learning algorithms, such as physics-based machine learning algorithms, may also be used.
[0043]
[0044] The composite component historical physical testing dataset (block 108) includes many different configurations of composite components with differing material properties. In one embodiment, the historical physical testing dataset obtained in block 108 was generated from the development of multiple product assemblies, such as an aircraft platform, a land vehicle platform, or other predominantly composite platform, and was not generated specifically for evaluation of the structural configuration 38. The historical physical testing dataset may also include data collected over a period of several years using composite components of various complexities and material configurations.
[0044]
[0045] The historical physical test data set (block 108) may include a set of physical test input data associated with a corresponding set of physical test data output data for each physical test performed on different configurations of the composite component. The set of physical test input data may include at least one of force-time history, displacement-time history, impact energy-time history, impactor configuration, indentation depth, material properties, associated boundary conditions, associated environmental conditions (e.g., heat and moisture), or composite structural configuration including composite layup and geometry. The material properties from the physical test input data may include at least one of mode 1 strain energy rate, mode 2 strain energy rate, interlaminar strength, stiffness ply properties, or ply strength properties.
[0045]
[0046] The set of physical test output data may include at least one of the type of damage resulting from the impact during the physical test recorded as the input data, the energy absorbed from the impact during the physical test, or the depth of an indentation in the composite component resulting from the physical test.
[0046]
[0047] In one embodiment, the historical physical test data set from block 108 may be divided into three main categories of composite component assemblies. Each category includes the physical test input data and physical test output data identified above. The first or broadest category of composite components from the historical physical test data set (block 108) includes physical test data related to material characterization. This type of physical test data may include information related to material screening and selection, material specification development, or allowable development. The second or intermediate type of physical test data includes physical test data related to design value development. This type of physical test data may include information related to sub-component testing and structural element testing. The third or most advanced type of physical test data related to the complexity of the composite component includes physical test data with analytical evaluation. This type of physical test data may include information related to full scale testing and component testing.
[0047]
[0048] During development of the dent depth metamodel in block 106, a greater weight may be provided to past physical test data sets of the plurality of composite components having a greater similarity of at least one of material or construction between the structural configuration 38 of the composite component and a corresponding one of the plurality of composite components from the past physical testing. The machine learning algorithm may be pre-trained based on the past physical test data sets or may be trained in conjunction with developing the dent depth metamodel. Also, the machine learning algorithm may be trained at the computer system 30 or at a remote computer system 40 having greater computing power than the computer system 30. Also, the past physical test data sets may be removed from the computer system 40 or stored within the computer readable medium 34 of the computer system 30.
[0048]
[0049] The method 100 then utilizes the dimple depth metamodel developed for the structural feature 38 from block 106 to estimate the impact energy required to define a threshold dimple depth in the structural feature (block 112). In one embodiment, the threshold dimple depth corresponds to a dimple depth of 0.01 inches in the surface of the structural feature 38. In another embodiment, the threshold dimple depth corresponds to a dimple depth of 0.02 inches in the surface of the structural feature 38.
[0049]
[0050] With the impact energy required to define a threshold indentation depth in the structural configuration of the composite component, the method 100 may utilize the estimated impact energy from block 112 to perform a computer-based simulation (block 114) with the structural configuration 38. The computer-based simulation generates output data including an identification of the strength of the damaged portion of the structural configuration under at least one of a compressive load, a tensile load, or a shear load. The computer-based simulation may include a low velocity impact (LVI) simulation, a high velocity impact (HVI) simulation, an outer mold line (OML), an inner mold line (IML), or another type of impact simulation for the structural configuration.
[0050]
[0051] The computer-based simulation also includes a set of input data and a set of output data similar to the information obtained in block 108 from the historical physical test data set. The results from block 114 may be fed back recursively to block 110 to improve the machine learning model as the simulation becomes more accurate, to provide a larger training data set for the machine learning algorithm.
[0051]
[0052] Method 100 may then use information from the simulation to identify design values for structural configuration 38. Identifying design values in block 116 allows for significant savings through digital engineering as opposed to performing physical testing on composite components of varying degrees of complexity.
[0052]
[0053] The following provisions provide an example configuration of a system and method for the example method 100 of analyzing a composite structure for damage tolerance of FIG.
[0053]
[0054] Article 1. 11. A method for analyzing a composite structure for damage tolerance comprising: obtaining a structural configuration for a composite component; identifying at least one metric of interest for the structural configuration; developing a dent depth metamodel based on the structural configuration of the composite component, the at least one metric of interest, and information obtained from a machine learning algorithm, the at least one metric of interest providing a constraint for the dent depth metamodel, the machine learning algorithm being trained based on historical physical testing data sets of a plurality of composite components; utilizing the dent depth metamodel for the structural configuration to estimate an impact energy to define a threshold dent depth in the structural configuration; performing a computer-based simulation of the structural configuration using the impact energy from the dent depth metamodel; and identifying at least one design value for the structural configuration based on output data generated from the computer-based simulation.
[0054]
[0055] Article 2. 2. The method of claim 1, wherein developing the depression depth metamodel includes providing greater weighting to the past physical test data sets of a plurality of the previously analyzed physical composite components based on material or configuration similarities between the structural configuration and corresponding ones of the previously analyzed physical composite components.
[0055]
[0056] Article 3. 3. The method of claim 1 or 2, wherein the plurality of previously analyzed physical composite components includes composite components from a plurality of different vehicle platforms.
[0056]
[0057] Article 4. 4. The method of any one of clauses 1 to 3, wherein the historical physical test dataset includes a plurality of test data inputs associated with corresponding test data outputs for a plurality of previously analyzed physical composite components.
[0057]
[0058] Article 5. 5. The method of any one of clauses 1 to 4, wherein the test data input includes at least one of a force-time history, a displacement-time history, an impact energy-time history, an impactor configuration, an indentation depth, material properties, associated boundary conditions, or a composite structural configuration defining a composite layup and geometry.
[0058]
[0059] Article 6. 6. The method of claim 4 or 5, wherein the material properties include at least one of mode 1 strain energy rate, mode 2 strain energy rate, interlaminar strength, stiffness ply properties, or ply strength properties.
[0059]
[0060] Article 7. 7. The method of any one of clauses 1 to 6, wherein the test data output includes at least one of the type of damage resulting from the impact, the energy absorbed from the impact, or the measured indentation depth.
[0060]
[0061] Article 8. 8. The method of any one of clauses 1 to 7, wherein the structural configuration includes at least one of a skin joined to a stiffening member, the skin joined to a secondary structure, or a stiffening structure joined to the secondary structure, the secondary structure including at least one of a rib or a spar.
[0061]
[0062] Article 9. 9. The method of any one of clauses 1 to 8, wherein the at least one metric of interest comprises dynamic stiffness, energy absorption rate, energy flux into the structural configuration, or initial deflection rate.
[0062]
[0063] Article 10. 10. The method of any one of clauses 1 to 9, wherein output data generated from the computer-based simulation includes an identification of the strength of the damaged portion of the structural configuration under at least one of compressive, tensile, and shear loads.
[0063]
[0064] Article 11. 11. The method of any one of clauses 1 to 10, wherein the composite component comprises a plurality of individual composite elements fixed relative to one another.
[0064]
[0065] Article 12. 12. The method of any one of clauses 1 to 11, wherein the historical physical testing data set includes calculations for damage tolerance criteria based on physical testing data.
[0065]
[0066] Article 13. 13. The method of any one of clauses 1 to 12, wherein the machine learning algorithm comprises a Bayesian-based algorithm or a physics-based algorithm.
[0066]
[0067] Article 14. 11. A non-transitory computer-readable medium having programmed instructions embodied therein, the instructions, when executed by a processor, operable to perform a method comprising: obtaining a structural configuration for a composite component; identifying at least one metric of interest for the structural configuration; developing a dent depth metamodel based on the structural configuration of the composite component, the at least one metric of interest, and information obtained from a machine learning algorithm, the at least one metric of interest providing constraints for the dent depth metamodel, the machine learning algorithm being trained based on historical physical testing data sets of a plurality of composite components; utilizing the dent depth metamodel for the structural configuration to estimate an impact energy to define a threshold dent depth in the structural configuration; performing a computer-based simulation of the structural configuration using the impact energy from the dent depth metamodel; and identifying at least one design value for the structural configuration based on output data generated from the computer-based simulation.
[0067]
[0068] Article 15. The computer-readable medium of claim 14, wherein the historical physical test dataset includes a plurality of test data inputs associated with corresponding test data outputs for a plurality of previously analyzed physical composite components, and the historical physical test dataset includes calculations for damage tolerance criteria based on the physical test data.
[0068]
[0069] Article 16. 16. The computer readable medium of claim 14 or 15, wherein the test data input includes at least one of a force-time history, a displacement-time history, an impact energy-time history, an impactor configuration, an indentation depth, material properties, associated boundary conditions, or a composite structural configuration that defines a composite layup and geometry.
[0069]
[0070] Article 17. 17. The computer-readable medium of any one of clauses 14-16, wherein the material properties include at least one of mode 1 strain energy rate, mode 2 strain energy rate, interlaminar strength, stiffness ply properties, or ply strength properties.
[0070]
[0071] Clause 18. The computer-readable medium of any one of clauses 14 to 17, wherein the test data output includes at least one of the type of damage resulting from the impact, the energy absorbed from the impact, or the measured dent depth.
[0071]
[0072] Article 19. 19. The computer-readable medium of any one of clauses 14 to 18, wherein the at least one metric of interest includes dynamic stiffness, energy absorption rate, energy flux into the structural configuration, or initial deflection rate.
[0072]
[0073] Article 20. 11. A non-transitory computer readable medium having programmed instructions embodied therein, the instructions, when executed by a processor, operable to perform a method comprising: obtaining a structural configuration for a composite component; identifying at least one metric of interest for the structural configuration; developing a dent depth metamodel based on the structural configuration of the composite component, the at least one metric of interest, and information obtained from a machine learning algorithm, the at least one metric of interest providing constraints for the dent depth metamodel, the machine learning algorithm being trained based on historical physical testing data sets of a plurality of composite components, the historical physical testing data set including calculations for damage tolerance criteria based on physical testing data; utilizing the dent depth metamodel for the structural configuration to estimate an impact energy to define a threshold dent depth in the structural configuration; performing a computer-based simulation of the structural configuration using the impact energy from the dent depth metamodel; and identifying at least one design value for the structural configuration based on output data generated from the computer-based simulation.
[0073]
[0074] Although various embodiments have been described, this description is intended to be illustrative and not restrictive. It will be apparent to those skilled in the art that many more embodiments and implementations are possible within the scope of the present embodiments. Any feature of any embodiment may be used in combination with or in place of any other feature or element in any other embodiment, unless specifically limited. Thus, the present embodiments are not to be limited except in light of the appended claims and their equivalents. Also, various modifications and variations can be made within the scope of the appended claims.
Claims
1. A method (100) for analyzing a composite structure for damage tolerance, comprising: obtaining (102) a structural configuration (38) for a composite component; Identifying (104) at least one metric of interest for said structural configuration (38); developing a dimple depth metamodel (106) based on the structural configuration (38) of the composite component, the at least one metric of interest, and information obtained from a machine learning algorithm, the at least one metric of interest providing constraints for the dimple depth metamodel, the machine learning algorithm being trained (110) based on a historical physical testing dataset of a plurality of composite components (108); estimating (112) an impact energy to define a threshold dent depth in the structural configuration (38) utilizing the dent depth meta-model for the structural configuration (38); performing (114) a computer-based simulation of the structural configuration (38) utilizing the impact energy from the dent depth metamodel; and The method (100), comprising identifying (116) at least one design value for the structural configuration (38) based on output data generated from the computer-based simulation.
2. 2. The method of claim 1, wherein developing the depression depth metamodel comprises providing greater weight to the past physical test data sets of a plurality of previously analyzed physical composite components based on material or composition similarities between the structural configuration and a corresponding one of the plurality of previously analyzed physical composite components.
3. The method (100) of claim 2 , wherein the plurality of previously analyzed physical composite components comprises composite components from a plurality of different vehicle platforms.
4. The method of claim 1 , wherein the historical physical test data set comprises a plurality of test data inputs associated with corresponding test data outputs for a plurality of previously analyzed physical composite components.
5. 5. The method (100) of claim 4, wherein the test data input includes at least one of a force-time history, a displacement-time history, an impact energy-time history, an impactor configuration, an indentation depth, material properties, associated boundary conditions, or a composite structural configuration (38) that defines a composite layup and geometry.
6. The method (100) of claim 5, wherein the material properties include at least one of mode 1 strain energy rate, mode 2 strain energy rate, interlaminar strength, stiffness ply properties, or ply strength properties.
7. 6. The method (100) of claim 5, wherein the test data output includes at least one of a type of damage resulting from an impact, an energy absorbed from the impact, or a measured dent depth.
8. 2. The method (100) of claim 1, wherein the structural configuration (38) comprises at least one of a skin (38A) joined to a stiffening member (38B), the skin joined to a secondary structure, or a stiffening structure joined to the secondary structure, the secondary structure comprising at least one of a rib or a spar.
9. The method (100) of claim 1, wherein the at least one metric of interest comprises dynamic stiffness, energy absorption rate, energy flux into the structural configuration (38), or initial deflection rate.
10. 2. The method (100) of claim 1, wherein output data generated from the computer-based simulation includes an identification of a strength of the damaged portion of the structural configuration (38) under at least one of a compressive load, a tensile load, and a shear load.
11. The method (100) of claim 10, wherein the composite component comprises a plurality of individual composite elements (38A, 38B) secured relative to one another.
12. The method (100) of claim 11 , wherein the historical physical testing data set includes calculations for damage tolerance criteria based on physical testing data.
13. The method (100) of claim 1 , wherein the machine learning algorithm comprises a Bayesian-based algorithm or a physics-based algorithm.
14. A non-transitory computer readable medium having programmed instructions embodied therein, the instructions, when executed by a processor, operable to perform a method (100), the method comprising: obtaining (102) a structural configuration (38) for a composite component; Identifying (104) at least one metric of interest for said structural configuration (38); developing a dimple depth metamodel (106) based on the structural configuration (38) of the composite component, the at least one metric of interest, and information obtained from a machine learning algorithm, the at least one metric of interest providing constraints for the dimple depth metamodel, the machine learning algorithm being trained (110) based on a historical physical testing dataset of a plurality of composite components (108); estimating (112) an impact energy to define a threshold dent depth in the structural configuration (38) utilizing the dent depth meta-model for the structural configuration (38); performing (114) a computer-based simulation of the structural configuration (38) utilizing the impact energy from the dent depth metamodel; and identifying (116) at least one design value for the structural configuration (38) based on output data generated from the computer-based simulation.
15. 15. The computer-readable medium of claim 14, wherein the historical physical test data set (108) includes a plurality of test data inputs associated with corresponding test data outputs for a plurality of previously analyzed physical composite components, and the historical physical test data set includes calculations for damage tolerance criteria based on physical test data.
16. 16. The computer readable medium of claim 15, wherein the test data input includes at least one of a force-time history, a displacement-time history, an impact energy-time history, an impactor configuration, an indentation depth, material properties, associated boundary conditions, or a composite structural configuration (38) that defines a composite layup and geometry.
17. 17. The computer-readable medium of claim 16, wherein the material properties include at least one of mode 1 strain energy rate, mode 2 strain energy rate, interlaminar strength, stiffness ply properties, or ply strength properties.
18. 17. The computer readable medium of claim 16, wherein the test data output includes at least one of a type of damage resulting from an impact, energy absorbed from the impact, or a measured dent depth.
19. The computer-readable medium of claim 14 , wherein the at least one metric of interest comprises dynamic stiffness, energy absorption rate, energy flux into the structural configuration (38), or initial deflection rate.
20. A non-transitory computer readable medium having programmed instructions embodied therein, the instructions, when executed by a processor, operable to perform a method (100), the method comprising: obtaining (102) a structural configuration (38) for a composite component; Identifying (104) at least one metric of interest for said structural configuration (38); developing (106) a dent depth meta-model based on the structural configuration (38) of the composite component, the at least one metric of interest, and information obtained from a machine learning algorithm, the at least one metric of interest providing constraints on the dent depth meta-model, the machine learning algorithm being trained (110) based on historical physical testing data sets of a plurality of composite components (108), the historical physical testing data set developing (106) including calculations for damage tolerance criteria based on physical testing data; estimating (112) an impact energy to define a threshold dent depth in the structural configuration (38) utilizing the dent depth meta-model for the structural configuration (38); performing (114) a computer-based simulation of the structural configuration (38) utilizing the impact energy from the dent depth metamodel; and identifying (116) at least one design value for the structural configuration (38) based on output data generated from the computer-based simulation.