Cloud model-based buffer structure and design method of shipboard voyage data recorder protection memory

Abstract

The present application relates to the technical field of ship equipment protection structure, especially to a shipborne voyage data recorder protection storage buffer structure based on cloud model and a design method thereof, the method comprising: collecting shipborne impact load and extracting features, using a three-parameter forward cloud model to model uncertainty and grade the load; designing a nested three-level multi-layer buffer structure matching the impact grade, including an outer yield energy absorption layer, a middle platform energy absorption layer and an inner flexible isolation layer; establishing a mapping relationship model of load cloud model features to buffer structure mechanical response to predict protection performance, and optimizing structure parameters based on the model, and finally completing the design through test verification and iterative correction. The present application solves the problems that traditional methods cannot handle load randomness and buffer structure response is single, realizes the hierarchical and controllable absorption of complex impact energy, and significantly improves the reliability and adaptability of the shipborne voyage data recorder protection storage.

Description

Technical Field

[0001] This invention relates to the technical field of protective structures for ship equipment, and in particular to a cloud-based protective storage buffer structure and design method for a shipborne navigation data recorder. Background Technology

[0002] As a core device for ship navigation safety and accident tracing, the shipborne navigation data recorder needs to operate for a long time in a complex shipborne environment with high impact, high vibration, multi-field coupling and transient compression. Its protective storage body needs to withstand multimodal loads such as drop, side collision and ship hull knock. These loads are characterized by abrupt changes in peak acceleration, uneven impact pulse width and drift of frequency band energy concentration area, which puts forward strict requirements for energy absorption, response stability and load graded absorption of the buffer structure of the protective storage body.

[0003] In existing technologies, the buffer structure of the protective storage body of shipborne navigation data recorders mostly adopts traditional single-layer or simple multi-layer energy-absorbing structures. The design process does not consider the randomness and distribution characteristics of impact loads, and the structural parameters are designed based solely on experience or static mechanical analysis, which has many shortcomings:

[0004] (1) The uncertain distribution characteristics of impact loads cannot be accurately described, and the optimization of structural parameters lacks reliable quantitative basis, resulting in poor adaptability of buffer structures to different levels and types of impact loads;

[0005] (2) The load transfer chain of the energy-absorbing structure has no controllable deformation zone, which is prone to short-term high stress penetration leading to local yielding of the storage core, and cannot achieve graded absorption and path redistribution of high energy input;

[0006] (3) The arrangement of buffer units lacks matching with the distribution of impact loads, and different buffer layers have no distinguishable response sensitivity. Under complex impact conditions, energy distribution is disordered, and the stability and protection reliability of the buffer structure are low.

[0007] (4) The mapping relationship between impact load and the mechanical response of the buffer structure has not been established, making it impossible to accurately predict and optimize the protective performance of the buffer structure. The protective structure has poor robustness under extreme impact scenarios.

[0008] (5) The material and configuration design of the buffer structure did not take into account the influence of environmental factors such as temperature and humidity on mechanical performance, resulting in large fluctuations in protective performance under different shipboard environments.

[0009] Some existing impact protection structures are also used in industrial equipment, but they are all based on deterministic load design and do not introduce uncertainty modeling methods. They cannot adapt to the random characteristics of impact loads in the shipboard environment, and they do not design graded energy absorption structures for the limited installation space of shipboard navigation data recorders. Therefore, they are difficult to meet the protection requirements of shipboard equipment and there is room for improvement. Summary of the Invention

[0010] To address the aforementioned technical issues, this application provides a cloud-based protective storage buffer structure and design method for shipborne navigation data recorders.

[0011] The above-mentioned objective of this application is achieved through the following technical solution:

[0012] A design method for a protective storage buffer structure of a shipborne navigation data recorder based on a cloud model, comprising the following steps:

[0013] The original time-domain signal of the multimodal impact load in the environment where the shipborne navigation data recorder is located is collected, and the peak acceleration, pulse width and energy density are extracted from the original time-domain signal as the characteristic parameter sequence of the impact load.

[0014] Based on the aforementioned feature parameter sequence, a three-parameter positive cloud model is used to model the uncertainty of shipborne impact loads, resulting in a cloud model that characterizes the random distribution of impact loads. The cloud model is expressed by three feature quantities: expected value, entropy value, and hyperentropy value. Based on the cloud droplet clustering results, the impact loads are divided into multiple impact levels.

[0015] Based on multiple impact levels, a nested three-level multi-layer buffer structure is designed. The buffer structure consists of an outer yield energy-absorbing layer, a middle platform energy-absorbing layer, and an inner flexible isolation layer from the outside to the inside. Different impact levels are matched with different levels of the buffer structure to achieve response sensitivity matching, so that each level has a differentiated load triggering threshold and energy absorption characteristics.

[0016] Based on the buffer structure, a mechanical model of each layer of material is established, and the cloud model characteristics and impact level are used as inputs to construct a mapping relationship model from the cloud distribution of impact load to the mechanical response of the buffer structure, so as to predict the stress, strain and energy absorption of the buffer structure under different impact loads.

[0017] By adopting the above technical solution, the characteristic parameters of multimodal impact loads are collected and extracted, providing a data foundation for subsequent uncertainty modeling. Furthermore, a three-parameter positive cloud model is innovatively introduced to model the uncertainty of impact loads. Expectation, entropy, and hyperentropy are used to accurately characterize the random distribution of shipborne impact loads, and impact levels are objectively classified based on cloud droplet clustering. This fundamentally solves the problem that traditional deterministic design methods cannot handle load randomness, providing quantitative input conditions that conform to the real load environment for buffer structure design. Based on this, a nested three-level multi-layer buffer structure matching different impact levels is designed. Through differentiated design and response matching of the outer yield energy-absorbing layer, the middle platform energy-absorbing layer, and the inner flexible isolation layer, the hierarchical, orderly absorption, and optimized control of the impact energy transmission path are achieved. Finally, by establishing a mapping relationship model from load cloud distribution to structural mechanical response, accurate and rapid prediction of the buffer structure's protective performance under different impact conditions is achieved. This method forms a complete design chain of "load uncertainty characterization - structural hierarchical matching - performance mapping prediction", which significantly improves the adaptability of buffer structures to complex random impact environments, the reliability of protection, and the scientific nature of design.

[0018] In a preferred embodiment, this application can be further configured as follows: based on the feature parameter sequence, an uncertainty model is performed on the shipborne impact load using a three-parameter positive cloud model to obtain a cloud model characterizing the random distribution characteristics of the impact load. The cloud model is expressed using three feature quantities: expected value, entropy value, and hyperentropy value. Furthermore, based on the cloud droplet clustering results, the impact load is divided into multiple impact levels, specifically including:

[0019] The original sequence of impact loads is slid through a fixed sampling window to extract local extreme points to form a feature sequence. The mean of the feature sequence is calculated as the expected value, and the entropy and hyperentropy values ​​are generated based on the three-parameter positive cloud model algorithm.

[0020] Based on the expected value, entropy value, and hyperentropy value, a cloud droplet sequence that follows a cloud distribution is generated using a two-level random number engine, and the membership degree of each feature parameter is calculated using an exponential function. The cloud droplets are then clustered based on the membership degree.

[0021] The stability of the cloud model is verified by analyzing the parameter drift trajectory and discrete consistency of the cloud droplet sequence. Based on the cloud droplet clustering results, the impact load is divided into multiple impact levels corresponding to different cloud droplet aggregation areas, and a unique number is assigned to each impact level.

[0022] By employing the aforementioned technical solutions, through fixed sampling window sliding and local extremum extraction, the characteristics of non-stationary impact signals can be effectively captured, providing a stable feature sequence for the cloud model. By calculating the mean as the expectation and using algorithms to generate entropy and hyperentropy, the digital characteristics describing the randomness and volatility of the load can be accurately obtained. By generating cloud droplet sequences through a two-level random number engine and calculating membership degrees for clustering, a fuzzy yet precise partitioning of the load feature space based on the data itself is achieved. Finally, by analyzing the stability of the cloud model and numbering and classifying the clustering results, the reliability and usability of the load model are ensured. This series of steps transforms the impact load from a continuous, random time-domain signal into discrete, classifiable design conditions in a scientific and objective manner, laying a solid foundation for subsequent precise matching between the structure and the load.

[0023] In a preferred embodiment, this application can be further configured as follows: a nested three-level multi-layer buffer structure is designed based on multiple impact levels. The buffer structure, from the outside in, consists of an outer yield energy-absorbing layer, a middle plateau energy-absorbing layer, and an inner flexible isolation layer. Different impact levels are matched with different levels of the buffer structure to achieve response sensitivity matching, so that each level has differentiated load triggering thresholds and energy absorption characteristics. Specifically, this includes:

[0024] The primary energy-absorbing zone is constructed using easily yieldable energy-absorbing material, and its design parameters are matched with the low-level impact loads to absorb the initial peak energy of the impact load.

[0025] A honeycomb or corrugated partitioned structure with varying thickness is adopted to form multiple secondary energy absorption zones with different yield plateaus. Each partition is matched with a medium-level impact load for graded load absorption and path redistribution.

[0026] A flexible isolation zone is constructed using low-modulus buffer material and is positioned close to the core storage body being protected. It is matched with the high-level or residual impact loads to isolate residual loads and control contact stress.

[0027] The outer yield energy-absorbing layer, the middle platform energy-absorbing layer, and the inner flexible isolation layer are nested together, and geometric transitions are made between each layer to limit stress concentration.

[0028] By employing the aforementioned technical solutions, the outer layer, constructed with easily yieldable energy-absorbing materials and matched to low-level impacts, ensures rapid activation and efficient absorption of peak energy in the initial stages of impact, acting as the "first line of defense" to mitigate peak energy. The middle layer, designed with a honeycomb or corrugated structure of varying thickness, forms multiple secondary energy-absorbing zones with different yield plates, matched to medium-level impacts. This provides stable reaction forces within a large deformation range and performs graded energy absorption and stress path redistribution for medium-intensity impacts, preventing overall structural failure. The inner layer, constructed with low-modulus flexible materials and positioned close to the core storage unit, is specifically designed to match and isolate high-level or residual impacts, flexibly controlling the contact stress and acceleration ultimately transmitted to the storage unit, providing a fine "last barrier." Finally, nested combination and geometric transition design optimize interlayer load transfer and limit stress concentration. This design enables the buffer structure to have clear functional zoning and sequential triggering mechanisms, achieving controllable, efficient dissipation and attenuation of impact energy across different layers.

[0029] In a preferred embodiment, this application can be further configured as follows: A mechanical model of each layer of material is established based on the buffer structure, and cloud model characteristics and impact level are used as inputs to construct a mapping model from the cloud distribution of impact load to the mechanical response of the buffer structure, in order to predict the stress, strain, and energy absorption of the buffer structure under different impact loads, specifically including:

[0030] The mechanical parameters of each layer of material were obtained through quasi-static compression tests. Segmented stress-strain constitutive models were established for the different mechanical behaviors of the outer yield energy-absorbing layer, the middle plateau energy-absorbing layer and the inner flexible isolation layer.

[0031] The segmented mechanical model is assigned to the corresponding buffer layer, and a global finite element model of the three-level multi-layer buffer structure is established, and the contact relationship between the layers is defined.

[0032] The obtained impact levels and their corresponding cloud model characteristics are bound together to form a load input matrix, and then mapped to the corresponding element set of the finite element model according to the time series.

[0033] Based on the linear weight superposition method, a response mapping function is constructed with cloud model feature quantities and impact level as inputs and peak stress of each layer of the buffer structure as output. The weight coefficients of the response mapping function are iteratively corrected according to the error between the output of the finite element simulation model and the mapping prediction result until the mapping accuracy meets the requirements.

[0034] By adopting the above technical solutions, material parameters are obtained through quasi-static tests and a piecewise constitutive model is established, providing high-fidelity material behavior input for finite element simulation. By establishing a global finite element model and correctly defining interlayer contact, a virtual prototype capable of simulating complex mechanical behaviors is constructed. By binding impact levels with cloud model characteristic quantities to form a load matrix and mapping it to the model, the uncertain load spectrum is transformed into a series of specific simulation conditions. Finally, by constructing and iteratively correcting the response mapping function based on the linear weight superposition method, a surrogate model capable of quickly and accurately predicting key structural responses (such as peak stress) is established. This mapping model replaces the time-consuming and laborious traditional trial-and-error method or extensive full-condition simulations, enabling designers to efficiently and cost-effectively evaluate and compare the protective performance of different buffer schemes during the design phase, providing a powerful quantitative analysis tool for subsequent optimization design.

[0035] In a preferred embodiment, this application can be further configured as follows: after establishing a mechanical model of each layer of material based on the buffer structure, and using cloud model characteristics and impact level as inputs to construct a mapping relationship model from the impact load cloud distribution to the mechanical response of the buffer structure, so as to predict the stress, strain and energy absorption of the buffer structure under different impact loads, the cloud model-based design method for the shipborne navigation data recorder protective storage buffer structure further includes:

[0036] Based on the mapping relationship model, the thickness and geometric configuration parameters of the outer yield energy-absorbing layer, the middle platform energy-absorbing layer and the inner flexible isolation layer are optimized, so that the optimized buffer structure can achieve a controllable triggering sequence and stable energy distribution under different impact levels.

[0037] The optimized buffer structure is subjected to multi-level impact tests, and test response data is collected. A protection performance rating is generated based on the test response data. The cloud model parameters and / or buffer structure parameters are iteratively corrected according to the protection performance rating until the protection requirements are met.

[0038] By adopting the above technical solution, based on the established response mapping model, and aiming to maximize total energy absorption and minimize transmission acceleration, the thickness, configuration, and other parameters of each layer of the buffer structure are automated. This allows the optimized structure to achieve the expected and controllable triggering sequence and stable energy distribution ratio under different impact levels, thus theoretically achieving optimal or near-optimal protective performance. Furthermore, multi-level impact tests are conducted on the optimized physical prototype to verify the protective performance with real test data and generate a quantitative rating. If the performance fails to meet the standards, an iterative correction process is initiated to adjust the model parameters or structural parameters. This complete closed loop of "design-prediction-optimization-verification-iteration" ensures that the final design result is not only theoretically optimal but also physically verified, meeting stringent actual protection requirements and greatly improving the reliability and success rate of the design.

[0039] In a preferred embodiment, this application can be further configured such that: based on the mapping relationship model, the thickness and geometric configuration parameters of the outer yield energy-absorbing layer, the middle plateau energy-absorbing layer, and the inner flexible isolation layer are optimized, specifically including:

[0040] Using the maximization of the total energy absorption of the buffer structure and the minimization of the peak acceleration transferred to the core memory as the multi-objective optimization function, the thickness of the outer yield energy-absorbing layer, the pore size and wall thickness of the cell cells of the middle platform energy-absorbing layer, the thickness of the inner flexible isolation layer, and the density and modulus parameters of each layer material are set as the design variables to be optimized.

[0041] Based on the installation space limitations of the shipborne navigation data recorder, geometric constraints are set for the overall dimensions of the buffer structure; based on the mechanical properties of each layer of material, performance constraints are set for the yield stress and compaction strain in the segmented stress-strain model.

[0042] By using the constructed response mapping function as a surrogate model, and combining it with the defined objective optimization function, design variables, and applied constraints, the optimal combination of buffer structure parameters that satisfies the objective optimization function is obtained.

[0043] By adopting the above technical solution and defining a pair of mutually constraining objective functions—"maximizing total energy absorption" and "minimizing peak acceleration transferred to the core memory"—the core contradiction in buffer structure design is grasped, ensuring that the optimization direction simultaneously considers its own energy absorption efficiency and the protection effect on the core equipment. By setting the key dimensions and material properties of each layer as design variables, sufficient search space is provided for the optimization algorithm to find the optimal combination. By imposing installation space constraints and material performance constraints, the feasibility and rationality of the optimization results in engineering are guaranteed. Finally, by using an efficient response mapping function as a surrogate model for optimization, the computational cost is significantly reduced while ensuring accuracy, enabling efficient implementation of simulation-based parameter optimization. This method can systematically search for the optimal solution that balances multiple performance objectives under given constraints, avoiding the one-sidedness of relying on experience, and realizing the refinement and optimization of buffer structure design.

[0044] In a preferred embodiment, this application can be further configured such that: the iterative correction of cloud model parameters and / or buffer structure parameters based on the protection performance rating until the protection requirements are met specifically includes:

[0045] If the protection performance rating is lower than the preset protection performance index, the test response data is first compared with the predicted data of the response mapping function, and the weighting coefficient of the response mapping function is corrected.

[0046] Based on the revised response mapping function, the parameters of the buffer structure are re-optimized to generate new buffer structure parameters.

[0047] The second protection performance rating is obtained based on the new buffer structure parameters. If the second protection performance rating is lower than the preset protection performance index, the parameters of the cloud model are adjusted and the impact level is reclassified until the protection performance rating reaches the preset protection performance index.

[0048] By adopting the above technical solution and comparing experimental and predicted data, the weight coefficients of the response mapping function were corrected, improving the prediction accuracy of the surrogate model and solving the problem of "inaccurate model". Subsequently, the structural parameters were re-optimized based on the corrected model to generate an improved design scheme. If the performance still does not meet the requirements, it indicates that there may be a deviation in the initial load perception, and the cloud model parameters are adjusted and the impact level is reclassified to correct the "inaccurate input" problem from the source. This progressive iterative correction strategy, from "correcting the response model" to "correcting the load model", constructs a robust self-improvement mechanism. It ensures that the design process can continuously converge based on experimental feedback, ultimately achieving a high degree of matching between the protective performance of the buffer structure and the real complex shipboard impact environment.

[0049] Secondly, the above-mentioned inventive objective of this application is achieved through the following technical solutions:

[0050] A cloud-based protective storage buffer structure for a shipborne navigation data recorder is disclosed. The buffer structure is positioned between the recorder's outer shell and the core storage unit. The buffer structure is a nested, three-tiered, multi-layered energy-absorbing structure, comprising, along the load transfer direction: an outer yielding energy-absorbing layer composed of easily yielding energy-absorbing material, used to absorb the initial peak energy of the impact load; a middle platform energy-absorbing layer adjacent to the inner side of the outer yielding energy-absorbing layer, employing an energy-absorbing configuration with gradually varying thickness and multiple yielding platforms, used for graded absorption of the impact load and energy path redistribution; and an inner flexible isolation layer adjacent to the inner side of the middle platform energy-absorbing layer and close to the core storage unit, composed of a low-modulus buffer material, used for flexibly isolating residual loads and controlling the contact stress acting on the core storage unit.

[0051] By adopting the above technical solution, the structure employs a nested three-level multi-layer configuration, consisting of a yielding energy-absorbing layer, a plateau energy-absorbing layer, and a flexible isolation layer from the outside in. The materials and configurations of each layer are specifically designed for impacts of different energy levels. The outer layer yields first to absorb the initial high-energy impact, the middle layer stably absorbs medium-energy impacts and redistributes the load path through its plateau characteristics, and the inner layer flexibly isolates residual impacts and vibrations. This "graded response, collaborative work" design concept enables the structure to dissipate and attenuate impacts of varying intensities from low to high in an orderly and efficient manner, avoiding premature failure, insufficient energy absorption, or stress penetration problems that may occur in single-layer or homogeneous structures under complex impacts. This buffer structure possesses high energy absorption efficiency, stable mechanical response, and excellent resistance to multiple impacts, providing reliable protection for the core storage of the shipborne navigation data recorder.

[0052] Preferably, the middle platform energy-absorbing layer is a honeycomb partitioned structure. The cell pore size and cell wall thickness of the honeycomb structure vary in a gradient on a cross section parallel to the direction of impact load transmission, thereby forming multiple energy-absorbing partitions with different compressive yield strengths. Furthermore, the outer yield energy-absorbing layer and the middle platform energy-absorbing layer, as well as the middle platform energy-absorbing layer and the inner flexible isolation layer, are geometrically transitioned by rounded corners.

[0053] By adopting the above technical solution, the energy-absorbing layer of the middle platform is specifically designed as a honeycomb partitioned structure with gradient changes in cell pore size and wall thickness. This gradient design allows different areas of the honeycomb structure to collapse sequentially under pressure due to differences in local stiffness / strength, thus naturally forming multiple energy-absorbing partitions with different compressive yield strengths, achieving more refined and graded absorption of impact energy. Simultaneously, by using rounded corners for geometric transitions between layers, the stress flow lines at the interlayer interfaces are effectively smoothed, significantly reducing the risk of stress concentration caused by geometric abrupt changes. This prevents the structure from tearing or prematurely failing at the joints under impact, ensuring the integrity and stability of the overall buffer structure during deformation, and further improving the product's reliability and durability.

[0054] In summary, this application includes at least one of the following beneficial technical effects:

[0055] 1. By collecting and extracting characteristic parameters of multimodal impact loads, a data foundation was provided for subsequent uncertainty modeling. Furthermore, a three-parameter positive cloud model was innovatively introduced to model the uncertainty of impact loads. Expectation, entropy, and hyperentropy were used to accurately characterize the random distribution of shipborne impact loads. Impact levels were objectively classified based on cloud droplet clustering, fundamentally solving the problem that traditional deterministic design methods cannot handle load randomness. This provides quantitative input conditions that conform to the real load environment for buffer structure design. Based on this, a nested three-level multi-layer buffer structure matching different impact levels was designed. Through differentiated design and response matching of the outer yield energy-absorbing layer, the middle platform energy-absorbing layer, and the inner flexible isolation layer, the hierarchical, orderly absorption, and optimized control of the impact energy transmission path were achieved. Finally, by establishing a mapping relationship model from load cloud distribution to structural mechanical response, accurate and rapid prediction of the buffer structure's protective performance under different impact conditions was achieved. This method forms a complete design chain of "load uncertainty characterization - structural hierarchical matching - performance mapping prediction", which significantly improves the adaptability of buffer structures to complex random impact environments, the reliability of protection, and the scientific nature of design.

[0056] 2. By using a fixed sampling window sliding mechanism and local extremum extraction, the characteristics of non-stationary impact signals can be effectively captured, providing a stable feature sequence for the cloud model. By calculating the mean as the expectation and using algorithms to generate entropy and hyperentropy, the numerical characteristics describing the randomness and volatility of the load can be accurately obtained. By generating cloud droplet sequences through a two-level random number engine and calculating membership degrees for clustering, a fuzzy yet precise partitioning of the load feature space based on the data itself is achieved. Finally, by analyzing the stability of the cloud model and assigning numbers and levels to the clustering results, the reliability and usability of the load model are ensured. This series of steps transforms the impact load from a continuous, random time-domain signal into discrete, classifiable design conditions in a scientific and objective manner, laying a solid foundation for the subsequent accurate matching of structure and load.

[0057] 3. By employing an easily yieldable energy-absorbing material for the outer layer and matching it with low-level impacts, rapid activation and efficient absorption of peak energy are ensured in the initial stages of impact, acting as the "first line of defense" to reduce peak energy. A honeycomb or corrugated middle layer structure with varying thickness is designed to form multiple secondary energy-absorbing zones with different yield plates, matching them with medium-level impacts. This provides stable reaction force within a large deformation range and performs graded energy absorption and stress path redistribution for medium-intensity impacts, preventing overall structural failure. The inner layer, constructed with a low-modulus flexible material close to the core storage unit, is specifically designed to match and isolate high-level or residual impacts, flexibly controlling the contact stress and acceleration ultimately transmitted to the storage unit, providing a fine "last barrier." Finally, nested combination and geometric transition design optimize interlayer load transfer and limit stress concentration. This design enables the buffer structure to have clear functional zoning and sequential triggering mechanisms, achieving controllable, efficient dissipation and attenuation of impact energy across different layers.

[0058] 4. The middle platform energy-absorbing layer is specifically designed as a honeycomb-like partitioned structure with gradient changes in cell pore size and wall thickness. This gradient design allows different areas of the honeycomb structure to collapse sequentially under pressure due to differences in local stiffness / strength, naturally forming multiple energy-absorbing zones with different compressive yield strengths. This achieves more refined and graded absorption of impact energy. Simultaneously, by using rounded corners for geometric transitions between layers, the stress flow lines at the interlayer interfaces are effectively smoothed, significantly reducing the risk of stress concentration caused by geometric abrupt changes. This prevents the structure from tearing or prematurely failing at the joints under impact, ensuring the integrity and stability of the overall buffer structure during deformation, further improving the product's reliability and durability. Attached Figure Description

[0059] Figure 1 This is a flowchart of a cloud-based protective storage buffer structure design method for a shipborne navigation data recorder according to one embodiment of this application;

[0060] Figure 2 This is a flowchart illustrating the implementation of step S20 in the cloud-based protective storage buffer structure design method for a shipborne navigation data recorder in one embodiment of this application.

[0061] Figure 3 This is a flowchart illustrating the implementation of step S30 in the cloud-based protective storage buffer structure design method for a shipborne navigation data recorder in one embodiment of this application.

[0062] Figure 4 This is a flowchart illustrating the implementation of step S40 in the cloud-based protective storage buffer structure design method for a shipborne navigation data recorder in one embodiment of this application.

[0063] Figure 5 This is another implementation flowchart of the design method for the protective storage buffer structure of a shipborne navigation data recorder based on a cloud model in one embodiment of this application;

[0064] Figure 6 This is a flowchart illustrating the implementation of step S50 in the cloud-based protective storage buffer structure design method for a shipborne navigation data recorder in one embodiment of this application.

[0065] Figure 7 This is a flowchart illustrating the implementation of step S60 in the cloud-based protective storage buffer structure design method for a shipborne navigation data recorder in one embodiment of this application.

[0066] Figure 8 This is a cross-sectional view of a cloud-based protective storage buffer structure for a shipborne navigation data recorder according to an embodiment of this application. Detailed Implementation

[0067] The present application will be further described in detail below with reference to the accompanying drawings.

[0068] In one embodiment, such as Figure 1 As shown, this application discloses a design method for a protective storage buffer structure of a shipborne navigation data recorder based on a cloud model, which specifically includes the following steps:

[0069] S10: Collect the original time-domain signal of the multimodal impact load in the environment where the shipborne navigation data recorder is located, and extract the peak acceleration, pulse width and energy density from the original time-domain signal as the characteristic parameter sequence of the impact load.

[0070] Specifically, this step is the data foundation preparation stage for uncertainty modeling and buffer design. The shipboard environment is complex; navigation data recorders may be subjected to impacts from various sources, such as equipment drops, ship collisions, and explosive shocks. These impact loads manifest as transient signals with different intensities, durations, and frequency components in the time domain—i.e., "multimodal impact loads." Acquiring the raw time-domain signals of these loads is to obtain the most realistic environmental excitation data. Raw signals alone are insufficient for quantitative analysis and modeling; therefore, it is necessary to extract key characteristic parameters that can effectively characterize the impact's destructive force. Peak acceleration reflects the instantaneous maximum intensity of the impact and is directly related to the inertial force on the structure; pulse width (i.e., the duration of the impact pulse) relates to the time history of energy input and affects the dynamic response characteristics of the structure; energy density (which can be estimated by integrating the square of the acceleration) comprehensively reflects the energy magnitude of the impact load. Constructing these three parameters into a characteristic parameter sequence provides multi-dimensional input samples for subsequent characterization of its random distribution using a cloud model.

[0071] S20: Based on the aforementioned feature parameter sequence, the uncertainty of shipborne impact load is modeled using a three-parameter positive cloud model to obtain a cloud model characterizing the random distribution characteristics of the impact load. The cloud model is expressed by three feature quantities: expected value, entropy value, and hyperentropy value. The impact load is divided into multiple impact levels based on the cloud droplet clustering results.

[0072] Specifically, this step is one of the core innovations of this invention, aiming to solve the problem that traditional deterministic design methods cannot handle the randomness of loads. Because shipborne impact loads are affected by various uncertainties, their characteristics (peak acceleration, pulse width, etc.) are not fixed values, but rather random variables that fluctuate within a certain range and conform to a certain probability distribution. The three-parameter positive cloud model is a mathematical model for handling the uncertainty transformation between qualitative concepts and quantitative data, and is particularly suitable for describing data with fuzziness and randomness. The expected value reflects the average level or central value of the characteristic parameters; the entropy value reflects the degree of dispersion of the characteristics, i.e., the magnitude of uncertainty; and the hyperentropy value reflects the uncertainty of the entropy value itself, used to characterize the "thick tail" or fluctuation characteristics of the random distribution. Through these three numerical features, the random distribution characteristics of impact loads can be comprehensively and concisely characterized. The "cloud droplets" generated after modeling are random data points that conform to the distribution of this cloud model. By performing cluster analysis on the cloud droplets in the feature space, impact loads with similar characteristics can be grouped into the same category, thereby objectively and data-drivenly dividing the continuous load spectrum into multiple discrete impact levels, providing clear input conditions for the subsequent design of differentiated buffer structure levels.

[0073] S30: Based on the multiple impact levels, a nested three-level multi-layer buffer structure is designed. The buffer structure consists of an outer yield energy-absorbing layer, a middle platform energy-absorbing layer, and an inner flexible isolation layer from the outside to the inside. Different impact levels are matched with different levels of the buffer structure to ensure that each level has a differentiated load triggering threshold and energy absorption characteristics.

[0074] Specifically, this step is the stage of structural innovation based on the load analysis results. For the different impact levels identified in step S20, this invention abandons the traditional uniform buffer design and proposes a graded response, collaborative "nested three-level multi-layer buffer structure." Its design concept is to allow different structural levels to specifically cope with different levels of impact. The outer yield energy-absorbing layer, as the first line of defense, uses materials with low yield strength and high plastic deformation capacity (such as porous metals and polymer foams). Its design goal is to match low-level impacts, absorbing a large amount of energy through plastic yielding in the initial stage of the impact, rapidly attenuating the impact peak. The middle plateau energy-absorbing layer is the core energy-absorbing unit, employing a structure with long yield plateau stress (such as gradually thickened honeycomb or corrugated plates). Its "plateau" characteristic means that it can provide a nearly constant reaction force within a large compression stroke, thereby achieving stable and controllable energy absorption. By designing it as a partitioned structure, it can be matched with multiple medium-level impacts, achieving graded and sequential energy absorption. The innermost flexible isolation layer is the innermost protective barrier, made of a flexible material with low elastic modulus and high damping (such as silicone or rubber). Its function is not to absorb large amounts of energy, but to further filter and isolate the residual impact (the residual part of a high-level impact or high-frequency vibration) after being attenuated by the first two layers. It also flexibly controls the contact stress that is finally transmitted to the core storage body, avoiding damage caused by local stress concentration. "Response sensitivity matching" between impact level and structural layer refers to adjusting the material parameters and geometry of each layer so that its yield or activation threshold corresponds to the load range of the corresponding impact level, thereby achieving orderly and efficient dissipation of impact energy.

[0075] S40: Based on the buffer structure, establish a mechanical model of each layer of material, and use the cloud model characteristic quantity and impact level as input to construct a mapping relationship model from the cloud distribution of impact load to the mechanical response of the buffer structure, so as to predict the stress, strain and energy absorption of the buffer structure under different impact loads.

[0076] Specifically, this step is a crucial one in establishing a digital design analysis model. To achieve accurate design, it's necessary to move from qualitative matching to quantitative prediction. First, establishing the mechanical model of each material layer involves obtaining its stress-strain curves through material experiments (such as quasi-static / dynamic compression tests) and mathematically describing them using appropriate constitutive models (such as elastoplastic models or collapsible foam models). This forms the microscopic basis for simulating structural mechanical behavior. Second, constructing a mapping relationship model is the core objective of this step. Its inputs are the load uncertainty information (cloud model characteristic quantities) and impact level classification obtained in step S20, and its output is the dynamic response of the buffer structure (stress, strain, energy absorption). This essentially establishes a proxy model or rapid prediction tool for the "load-structure response." In practical implementation, the finite element method can be used: different impact levels and their corresponding cloud model characteristics (such as generating random load samples by combining entropy and hyperentropy with the expected value as a benchmark) are applied as load conditions to the finite element model of the buffer structure for extensive simulation calculations. Then, machine learning (such as response surface methodology and neural networks) or mathematical statistics methods are used to fit a mapping function from the input (load characteristics, level) to the output (structural response) from the massive simulation data. This model enables designers to quickly evaluate the protective performance of the designed buffer structure under random impact load spectrum without conducting physical tests, providing an efficient analytical tool for optimization design.

[0077] In this embodiment, by collecting and extracting characteristic parameters of multimodal impact loads, a data foundation is provided for subsequent uncertainty modeling. Furthermore, a three-parameter positive cloud model is innovatively introduced to model the uncertainty of impact loads. Expectation, entropy, and hyperentropy are used to accurately characterize the random distribution of shipborne impact loads. Impact levels are objectively classified based on cloud droplet clustering, fundamentally solving the problem that traditional deterministic design methods cannot handle load randomness. This provides quantitative input conditions that conform to the real load environment for buffer structure design. Based on this, a nested three-level multi-layer buffer structure matching different impact levels is designed. Through differentiated design and response matching of the outer yield energy-absorbing layer, the middle platform energy-absorbing layer, and the inner flexible isolation layer, the hierarchical, orderly absorption, and optimized control of the impact energy transmission path are achieved. Finally, by establishing a mapping model from load cloud distribution to structural mechanical response, accurate and rapid prediction of the buffer structure's protective performance under different impact conditions is achieved. This method forms a complete design chain of "load uncertainty characterization - structural hierarchical matching - performance mapping prediction", which significantly improves the adaptability of buffer structures to complex random impact environments, the reliability of protection, and the scientific nature of design.

[0078] In one embodiment, such as Figure 2As shown, in step S20, based on the feature parameter sequence, the uncertainty of the shipborne impact load is modeled using a three-parameter positive cloud model to obtain a cloud model characterizing the random distribution characteristics of the impact load. The cloud model is expressed by three feature quantities: expected value, entropy value, and hyperentropy value. Based on the cloud droplet clustering results, the impact load is divided into multiple impact levels, specifically including:

[0079] S21: Slide the original sequence of impact loads with a fixed sampling window, extract local extreme points to form a feature sequence, calculate the mean of the feature sequence as the expected value, and generate entropy and hyperentropy values ​​based on the three-parameter positive cloud model algorithm.

[0080] Specifically, this sub-step involves calculating the digital features of the cloud model. Directly calculating global statistics from the original time-domain signal may fail due to signal non-stationarity. Therefore, a fixed sampling window sliding method is used, where the signal is gradually slid across a window of fixed time length, and analysis is performed within each window. This better captures the local time-varying characteristics of the load. Within each window, local extreme points (peaks or troughs) are extracted, and the amplitudes of these extreme points, their corresponding time differences (used to calculate pulse width), etc., constitute a feature sequence. This sequence reflects the intensity and temporal distribution of the impact event within that window. The mean of this feature sequence is calculated, which is the expected value of the cloud model, representing the average level of the feature within the window. Entropy and hyperentropy values ​​need to be estimated using a reverse cloud generation algorithm for the cloud model. This algorithm typically calculates entropy (En) and hyperentropy (He) through mathematical transformations based on the sample variance and higher-order moments of the feature sequence. Entropy characterizes the discrete range of the feature around the expected value, while hyperentropy characterizes the volatility of this discreteness itself, together defining the random distribution "cloud" of the load features.

[0081] S22: Based on the expected value, entropy value, and hyperentropy value, a cloud droplet sequence that follows the cloud distribution is generated through a two-level random number engine, and the membership degree of each feature parameter is calculated using an exponential function. The cloud droplets are then clustered based on the membership degree.

[0082] Specifically, this sub-step involves forward cloud generation and cloud droplet analysis. After obtaining the three numerical features—expectation (Ex), entropy (En), and hyperentropy (He)—a forward cloud generator can be run to simulate the random distribution of loads. A two-stage random number engine is the core of the forward cloud generator: the first stage generates a normally distributed random number with En as the expectation and He as the standard deviation, serving as the "random realization of entropy," En'; the second stage generates a normally distributed random number with Ex as the expectation and |En'| as the standard deviation, the result of which is a cloud droplet (a specific feature value sample). This process is repeated to generate a large number of cloud droplets, forming a cloud droplet sequence. An exponential function (usually the exponential form of a normal distribution) is used to calculate the membership degree of each cloud droplet to a qualitative concept (such as "moderate impact") defined by (Ex, En, He), with membership degrees between 0 and 1, representing the degree of conformity. Then, based on the cloud droplet's position in the feature space and its membership degree, a clustering algorithm (such as fuzzy C-means clustering) is used to group the cloud droplets. Cloud droplets with high membership are more likely to be classified as core classes, which enables soft partitioning of random payload data.

[0083] S23: Analyze the parameter drift trajectory and discrete consistency of the cloud droplet sequence to verify the stability of the cloud model; based on the cloud droplet clustering results, divide the impact load into multiple impact levels corresponding to different cloud droplet aggregation areas, and assign a unique number to each impact level.

[0084] Specifically, this sub-step involves model validation and load rating. In engineering applications, it is necessary to ensure that the established cloud model is stable and representative. Analyzing parameter drift trajectories involves observing whether the calculated values ​​(Ex, En, He) change significantly across different data subsets or time series; the discrete consistency test assesses the uniformity and rationality of cloud droplet distribution. These two tests verify the stability of the cloud model, ensuring it can reliably characterize the overall statistical properties of the load. After passing the tests, the clustering results obtained in step S22 are adopted. Each cluster corresponds to a cloud droplet aggregation region in the feature space, where cloud droplets have similar load characteristics. Therefore, each cluster can naturally be defined as an impact level; for example, clusters with high centroid values ​​are "high-level impacts," and those with low centroid values ​​are "low-level impacts." A unique number (e.g., L1, L2, L3) is assigned to each level to facilitate identification, matching, and retrieval by the computer system in subsequent steps, thereby transforming the continuous, random impact load environment into discrete, classifiable design conditions.

[0085] In one embodiment, such as Figure 3As shown, in step S20, a nested three-level multi-layer buffer structure is designed based on multiple impact levels. The buffer structure consists of an outer yield energy-absorbing layer, a middle plateau energy-absorbing layer, and an inner flexible isolation layer, from the outside in. Different impact levels are matched with different levels of the buffer structure to ensure that each level has differentiated load triggering thresholds and energy absorption characteristics. Specifically, this includes:

[0086] S31: The primary energy absorption zone is constructed using easily yieldable energy-absorbing material, and its design parameters are matched with the low-level impact loads to absorb the initial peak energy of the impact load.

[0087] Specifically, this sub-step defines the design of the outer yield-absorbing layer. Yield-absorbing materials typically refer to materials with low yield strength but high plastic deformation capacity, such as open-cell or closed-cell aluminum foam, polyurethane foam, and extruded aluminum tubes. Their "yield-easy" characteristic means they can enter the plastic deformation stage at relatively low stress levels, thus quickly initiating energy absorption. The "primary energy-absorbing zone" is positioned as the outermost layer of the buffer structure. Design parameters mainly include the material type, density, thickness, and relative density (for porous materials). Matching to low-level impact loads means adjusting these design parameters so that the compaction initiation stress (or plateau stress) of this layer is lower than or equal to the typical stress level of low-level impact loads. Thus, when a low-level impact occurs, the energy is mainly dissipated by this layer through plastic crushing, thereby protecting the internal structure; for higher-level impacts, after the layer completes its own crushing and energy absorption, the impact load is transferred to the next layer. Its core function is to absorb the initial peak energy and prevent high strain rate loads from directly impacting the interior.

[0088] S32: It adopts a honeycomb or corrugated partitioned structure with varying thickness to form multiple secondary energy absorption zones with different yield plateaus. Each partition is matched with a medium-level impact load to perform graded load absorption and path redistribution.

[0089] Specifically, this sub-step defines an innovative configuration for the intermediate plateau energy-absorbing layer. Honeycomb or corrugated partitioned structures are two typical high-efficiency energy-absorbing configurations. They provide a long and stable compaction stroke and near-constant reaction force (i.e., yield plateau) through structural buckling and folding. Gradual thickness variation is a key design feature; for example, the cell wall thickness of a honeycomb structure gradually increases along a certain direction. This gradation leads to spatial variations in structural stiffness / strength, so that under compression, different regions sequentially enter the yielding and compaction stages, forming multiple secondary energy-absorbing zones with different yield plateaus. The plateau stress level of each partition can be precisely controlled through its local thickness. Matching to medium-level impact loads means setting the plateau stress thresholds of these partitions to correspond to the load levels of several defined medium-level impacts. When the impact load increases, the buffer structure does not reach its limit simultaneously as a whole, but rather starts from the partition with the lower plateau stress and sequentially triggers the partitions with the higher plateau stress, achieving graded energy absorption. Simultaneously, this sequential deformation also changes the internal stress propagation path, achieving energy path redistribution, avoiding stress concentration, and improving energy absorption efficiency.

[0090] S33: A flexible isolation zone is constructed using low-modulus buffer material and is positioned close to the protected core storage body. It is matched with the high-level or residual impact loads to isolate residual loads and control contact stress.

[0091] Specifically, this sub-step defines the design of the inner flexible isolation layer. Low-modulus buffer materials refer to materials with low elastic modulus, such as rubber, silicone, gel, or low-density foam. They typically have high damping characteristics and large recoverable deformation capacity. The flexible isolation zone is positioned close to the core storage component (such as a solid-state drive module) that needs protection. Its mechanism of action differs from the two layers mentioned above, which primarily dissipate energy through plasticity; it mainly absorbs and dissipates energy through elastic deformation and viscoelastic damping. Matching high-level or residual impact loads means that its design objective is not to withstand the main energy of the impact (this has already been borne by the first two layers), but rather to handle residual impacts (which may be high-frequency vibrations, overload peaks, or quasi-static pressure) that may still cause damage to the core storage component after being attenuated by the first two layers. Through its flexibility, this layer can further extend the load application time, reduce the peak acceleration transmitted to the storage component, and control contact stress through surface contact, preventing damage to the storage component package or internal chips due to excessive local pressure.

[0092] S34: The outer yield energy-absorbing layer, the middle platform energy-absorbing layer, and the inner flexible isolation layer are nested together, and geometric transitions are made between each layer to limit stress concentration.

[0093] Specifically, this step involves completing the overall structural assembly and detail optimization. Nested assembly refers to the three-layer structure being tightly stacked together in space, from the outside in, filling the annular or box-shaped space between the recorder's outer casing and the core storage unit, forming a unified structure. Geometric transitions between layers are a crucial design detail. This means avoiding sharp right angles or abrupt cross-sectional changes at the interface between two materials or at structural junctions, instead using rounded corners, chamfers, or ramps for smooth transitions. The direct purpose of this is to limit stress concentration. Under impact loads, geometric discontinuities easily lead to stress concentration, becoming the origin of structural failure (such as cracking). By optimizing geometric transitions, stress flow lines can be smoothly transmitted, ensuring that each layer deforms collaboratively according to the design sequence, fully utilizing the energy absorption and protective functions of each layer, and improving the reliability and durability of the overall buffer structure.

[0094] In one embodiment, such as Figure 4 As shown, in step S40, a mechanical model of each layer of material is established based on the buffer structure. The cloud model characteristics and impact level are used as inputs to construct a mapping model from the cloud distribution of the impact load to the mechanical response of the buffer structure, in order to predict the stress, strain, and energy absorption of the buffer structure under different impact loads. Specifically, this includes:

[0095] S41: Obtain the mechanical parameters of each layer of material through quasi-static compression tests, and establish segmented stress-strain constitutive models for the different mechanical behaviors of the outer yield energy-absorbing layer, the middle plateau energy-absorbing layer and the inner flexible isolation layer.

[0096] Specifically, this step is fundamental to obtaining a high-fidelity material model. Accurate simulation predictions depend on accurate input parameters. The quasi-static compression test compresses a buffer material sample at a low strain rate to obtain its complete stress-strain curve. Key mechanical parameters, such as elastic modulus, yield stress, plateau stress, compaction strain, and compaction strength, can be extracted from this curve. Due to the significant differences in the mechanical responses of the three layers, separate piecewise stress-strain constitutive models need to be established. For example, for the outer yield energy-absorbing layer (such as foam metal), its constitutive model typically includes an elastic segment, a yield plateau segment, and a compaction segment; for the middle honeycomb structure, an anisotropic elastoplastic model may be considered; and for the inner flexible material, a hyperelastic model (such as the Mooney-Rivlin model) or a viscoelastic model may be used. Piecewise modeling means using different mathematical formulas to describe the material's behavior at different strain stages, thereby more realistically reproducing its complex mechanical response in finite element simulations.

[0097] S42: Assign the established segmented mechanical model to the corresponding buffer layer, establish the overall finite element model of the three-level multi-layer buffer structure, and define the contact relationship between the layers.

[0098] Specifically, this sub-step involves constructing a virtual simulation prototype. In the finite element analysis software, firstly, based on the design scheme in step S30, a three-dimensional geometric model of the three-layer buffer structure, outer shell, and core storage is established and meshed. Then, the segmented mechanical model established in step S41 is assigned to the corresponding buffer layer elements in the form of material property cards. Next, defining the contact relationships between layers is crucial, as it determines how loads and deformations are transferred between layers. Typically, the layers are defined as having "face-to-face contact" or "shared nodes," and appropriate friction coefficients are set. Correctly setting the contact relationships ensures that in impact simulation, the deformation of the outer layer will compress the middle layer, then the inner layer, and finally act on the core storage, realistically simulating the load transfer chain in the physical world.

[0099] S43: Bind the obtained impact levels and their corresponding cloud model characteristics to form a load input matrix, and map it to the corresponding element set of the finite element model according to the time series.

[0100] Specifically, this sub-step transforms the uncertain load into simulated load conditions. The result obtained from step S20 is several impact levels and their corresponding cloud model digital characteristics (Ex, En, He). For simulation, specific load time histories need to be generated by sampling from these cloud models. Binding refers to associating each impact level with a set (Ex, En, He). The load input matrix is ​​a data structure where each row may represent an impact condition, containing: the impact level number, and one or more sets of random characteristic parameter values ​​(such as peak acceleration A, pulse width T) generated from the cloud model of that level through a forward cloud generator. In finite element simulation, the abstract load characteristics need to be transformed into specific excitations acting on the model. Time-series mapping refers to constructing a classic impact pulse waveform (such as a half-sine wave or a rectangular wave) as an acceleration load based on the load characteristics (such as A and T). The corresponding element set refers to applying this acceleration load as an inertial force to all elements belonging to the buffer structure and core storage, or to the constraint surface connected to the mounting base, to simulate the basic excitation.

[0101] S44: Based on the linear weight superposition method, a response mapping function is constructed with cloud model feature quantities and impact level as inputs and peak stress of each layer of the buffer structure as output. The weight coefficients of the response mapping function are iteratively corrected according to the error between the output of the finite element simulation model and the mapping prediction result until the mapping accuracy meets the requirements.

[0102] Specifically, this sub-step involves constructing and training an efficient surrogate model (response mapping function). Performing a complete finite element transient dynamic simulation for each load condition defined from the load input matrix is ​​computationally expensive. For rapid prediction, a simplified yet sufficiently accurate mathematical model is needed. The linear weighted superposition method is a feasible approach. Its basic idea is that the structural response (e.g., peak stress in a certain layer) can be approximated as a linear weighted sum of various input features (e.g., peak acceleration, pulse width, energy density, impact level coding, etc.). First, based on a batch of finite element simulation sample data (input-output pairs), a set of weighting coefficients is initially determined using methods such as multiple linear regression, thus obtaining a preliminary response mapping function. Then, the prediction accuracy of this mapping function is tested using new simulation data, calculating the error between the predicted peak stress and the simulated peak stress. If the error does not meet the requirements, the weighting coefficients are iteratively corrected using more simulation data or more advanced fitting algorithms (e.g., ridge regression, neural networks) until the prediction error (e.g., root mean square error) of the mapping function on the validation set meets the preset accuracy requirements. The resulting mapping function only requires input of the cloud characteristics and level of the load to instantly output a reliable prediction of the structural response, greatly improving the efficiency of optimization design.

[0103] In one embodiment, such as Figure 5 As shown, after step S40, the design method for the protective storage buffer structure of the shipborne navigation data recorder based on the cloud model further includes:

[0104] S50: Based on the mapping relationship model, the thickness and geometric configuration parameters of the outer yield energy-absorbing layer, the middle platform energy-absorbing layer and the inner flexible isolation layer are optimized so that the optimized buffer structure can achieve a controllable triggering sequence and stable energy distribution under different impact levels.

[0105] Specifically, this step involves automatically optimizing the initial design scheme under the guidance of a digital model. The initial design (step S30) is based on experience or a rough matching. The mapping relationship model established in step S40 can quickly evaluate the protective performance under any set of design parameters. The design variables are the parameters to be optimized, including the thickness of each layer of material and the cell geometry parameters of the middle layer honeycomb / corrugated structure (such as wall thickness, pore size, corrugation angle, etc.). The optimization goal is to ensure that the buffer structure, under different impact levels from low to high, can be triggered in the intended order of the outer layer, each zone of the middle layer, and the inner layer (i.e., "controllable triggering order"), and that the proportion of energy absorbed by each layer to the total impact energy is relatively stable (i.e., "stable energy distribution"), avoiding premature failure or insufficient energy absorption of any layer. By calling optimization algorithms (such as genetic algorithms, sequential quadratic programming), multiple virtual "experiments" are conducted on the mapping relationship model to search for the optimal combination of design parameters that can achieve the above goals, thereby obtaining a theoretically better buffer structure scheme.

[0106] S60: Conduct multi-level impact tests on the optimized buffer structure, collect test response data, generate a protection performance rating based on the test response data, and iteratively correct the cloud model parameters and / or buffer structure parameters according to the protection performance rating until the protection requirements are met.

[0107] Specifically, this step is a crucial link in completing the "design-simulation-optimization-verification" closed loop, ensuring the effectiveness of the theoretical design in practice. The design scheme obtained after optimization in step S50 is fabricated into a physical prototype. On an impact test bench, a series of impact tests, from low to high, are applied to the prototype according to the multi-level impacts defined in step S20. Test response data is collected through sensors, such as the acceleration and strain at different locations of the buffer structure, and the acceleration ultimately transmitted to the core memory simulation block. A protection performance rating is generated based on the test response data, which refers to quantitatively scoring the test results according to predetermined evaluation criteria (such as whether the peak acceleration of the memory block is below the allowable value, whether the deformation of the buffer structure is stable and controllable, and whether each level is triggered sequentially). If the protection performance rating meets the standard, the design is complete. If it does not meet the standard, an iterative correction process is initiated: first, it is analyzed whether the load model is inaccurate or the structural design is poor. If the deviation between the simulation prediction and the test results is large, it may be necessary to return to step S20 and use new test data to correct the cloud model parameters (Ex, En, He) to more accurately describe the real load environment. If the problem stems from a structural response mechanism, return to step S50 and re-optimize the buffer structure parameters based on the updated load model or the corrected mapping model. This cycle continues until the experimental performance of the physical prototype meets all protection requirements, achieving final design approval based on physical verification.

[0108] In one embodiment, such as Figure 6 As shown, in step S50, based on the mapping relationship model, the thickness and geometric configuration parameters of the outer yield energy-absorbing layer, the middle platform energy-absorbing layer, and the inner flexible isolation layer are optimized, specifically including:

[0109] S51: Using the maximization of the total energy absorption of the buffer structure and the minimization of the peak acceleration transferred to the core memory as the multi-objective optimization function, the thickness of the outer yield energy-absorbing layer, the pore size and wall thickness of the middle platform energy-absorbing layer, the thickness of the inner flexible isolation layer, and the density and modulus parameters of each layer material are set as the design variables to be optimized.

[0110] Specifically, this sub-step clarifies the mathematical definition of the optimization problem. Optimization requires clear objectives and adjustable variables. The multi-objective optimization function includes two aspects: first, maximizing total energy absorption, aiming to improve the ability of the buffer structure to dissipate impact energy; second, minimizing the peak acceleration transmitted to the core storage, which is the ultimate goal of protection and directly related to the survivability of the storage. These two objectives sometimes conflict and need to be balanced. Design variables are parameters that can be adjusted during the optimization process, and they directly determine the performance of the structure. These include: 1) Geometric dimensions: the thickness of the outer yield energy-absorbing layer, the pore size and wall thickness of the cellular units in the middle plateau energy-absorbing layer (these two parameters together determine the plateau stress), and the thickness of the inner flexible isolation layer; 2) Material properties: the density and modulus parameters of each layer of material (such as elastic modulus and yield modulus). In optimization, these parameters vary within a given upper and lower limit range.

[0111] S52: Based on the installation space limitations of the shipborne navigation data recorder, set geometric constraints for the overall dimensions of the buffer structure; based on the mechanical properties of each layer of material, set performance constraints for the yield stress and compaction strain in the segmented stress-strain model.

[0112] Specifically, this sub-step involves applying practical engineering constraints. Optimization cannot be arbitrary; it must conform to realistic limitations. Geometric constraints are the most fundamental. The buffer structure must fit within the limited installation space between the shipborne navigation data recorder's casing and the core storage unit; therefore, its overall dimensions, such as outer diameter and total thickness, have clear upper limits. Performance constraints ensure that the structure retains reasonable mechanical behavior after optimization. For example, setting a lower limit for the yield stress in the piecewise stress-strain model ensures the structure has sufficient supporting stiffness to prevent failure under normal vibration; setting an upper limit for its compaction strain ensures the material provides an effective energy-absorbing plateau before reaching that strain, preventing premature "compaction" and loss of buffering effect. These constraints limit the optimization search space to a feasible and reasonable region.

[0113] S53: Using the constructed response mapping function as a surrogate model, combined with the defined objective optimization function, design variables, and applied constraints, the optimal combination of buffer structure parameters that satisfies the objective optimization function is obtained.

[0114] Specifically, this sub-step involves performing optimization. Since directly calling the finite element model for optimization involves enormous computational costs, the response mapping function constructed in step S44 is used as a surrogate model (or meta-model, response surface model). This surrogate model can approximately predict the objective function values ​​(total energy absorption and transfer acceleration) corresponding to any set of design variables with extremely low computational cost. An optimization algorithm (such as the multi-objective genetic algorithm NSGA-II) searches on the surrogate model. In each generation, the algorithm generates a series of design variable combinations, quickly evaluates their objective function values ​​using the surrogate model, and checks whether all constraints are satisfied. Through selection, crossover, mutation, and other operations, iteratively generates better-performing design schemes. Finally, the algorithm outputs a Pareto optimal solution set, where each solution represents an optimal trade-off among multiple objectives. Designers can then select an optimal buffer structure parameter combination from this set based on their actual emphasis (whether energy absorption or shock absorption is more important) as the optimization output of step S50.

[0115] In one embodiment, such as Figure 7 As shown, in step S60, the cloud model parameters and / or buffer structure parameters are iteratively corrected according to the protection performance rating until the protection requirements are met. This specifically includes:

[0116] S61: If the protection performance rating is lower than the preset protection performance index, the test response data and the predicted data of the response mapping function are compared first, and the weighting coefficient of the response mapping function is corrected.

[0117] Specifically, this sub-step is the first step in the iterative correction process, focusing on correcting inaccuracies in the prediction model. When experimental results fail to meet standards, it indicates a deviation in the "design-simulation" stage. The first step is to diagnose the source of this deviation. This involves comparing the experimental response data with the predicted data from the response mapping function. For example, comparing the peak acceleration of the storage chamber measured in the experiment with the peak acceleration predicted by the mapping function under the same impact level. If the deviation is systematic and significant, it indicates that the surrogate model (response mapping function) constructed in step S44 is insufficiently accurate and fails to accurately reflect the true "load-structure response" relationship. In this case, correcting the weight coefficients of the response mapping function is the most direct correction method. Specifically, the new input-output data pairs obtained from this experiment are added to the existing training sample library, regression fitting is performed again, and the weight coefficients in the mapping function are updated. This allows the updated mapping function to simultaneously fit historical simulation data and new experimental data, thereby improving its prediction fidelity.

[0118] S62: Based on the modified response mapping function, the parameters of the buffer structure are re-optimized to generate new buffer structure parameters.

[0119] Specifically, this sub-step involves redesigning using the updated model. After completing step S61 and obtaining the revised response mapping function (i.e., a more accurate surrogate model), the design needs to be re-optimized based on this more reliable model. That is, the optimization process of step S50 is repeated, but this time using the updated mapping function. The optimization algorithm will re-search for buffer structure parameters that can satisfy the multi-objective optimization function and all constraints on the new, more physically realistic model. This generates a new combination of buffer structure parameters. This new scheme is the theoretically optimal solution after correcting for model biases, and its performance prediction is expected to be closer to the subsequent actual experimental results.

[0120] S63: Obtain a second protection performance rating based on the new buffer structure parameters. If the second protection performance rating is lower than the preset protection performance index, adjust the parameters of the cloud model and reclassify the impact level until the protection performance rating reaches the preset protection performance index.

[0121] Specifically, this sub-step involves a deeper iteration, revising the load perception model. The new parameters obtained in step S62 are used to create a second-generation physical prototype, which undergoes another impact test to obtain a second protection performance rating. If the rating still fails to meet the requirements, the problem may not be limited to the structural response model but may stem from insufficient understanding of the load environment itself (i.e., an inaccurate cloud model). For example, the randomness of the load in the actual environment (entropy, hyperentropy) may be greater than initially estimated, or the impact level classification may be unreasonable. In this case, the parameters of the cloud model need to be adjusted, i.e., the expected value, entropy, and hyperentropy are recalculated using more comprehensive experimental data. Then, the impact level is reclassified based on the new cloud model. This is equivalent to returning to the starting point of step S20, but based on more realistic load information. Subsequently, based on the new load level, structural matching (S30), constructing a new mapping relationship (S40), optimization on the new mapping relationship (S50), and retesting (S60) are required. This large cycle of "load model revision - structural design update" will continue until the final prototype's test performance reaches the preset protection performance indicators.

[0122] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0123] In one embodiment, such as Figure 8As shown, a cloud-based protective storage buffer structure for a shipborne navigation data recorder is provided. The buffer structure is positioned between the recorder's outer shell and the core storage unit. The buffer structure is a nested, three-tiered, multi-layered energy-absorbing structure, comprising, along the load transfer direction: an outer yielding energy-absorbing layer made of easily yielding energy-absorbing material, used to absorb the initial peak energy of the impact load; a middle platform energy-absorbing layer adjacent to the inner side of the outer yielding energy-absorbing layer, employing a gradient thickness and multiple yielding platforms for graded absorption of the impact load and energy path redistribution; and an inner flexible partition... The delamination layer, adjacent to the inner side of the middle platform energy-absorbing layer and close to the core memory, is made of a low-modulus buffer material. It is used to flexibly isolate residual loads and control the contact stress acting on the core memory. The middle platform energy-absorbing layer is specifically a honeycomb partitioned structure. The cell pore size and cell wall thickness of this honeycomb structure vary in gradient on a cross section parallel to the impact load transmission direction, thereby forming multiple energy-absorbing partitions with different compressive yield strengths. Furthermore, the outer yield energy-absorbing layer and the middle platform energy-absorbing layer, as well as the middle platform energy-absorbing layer and the inner flexible isolation layer, are geometrically transitioned through rounded corners.

[0124] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A cloud model-based shipboard voyage data recorder guard storage buffer structure design method, characterized in that, The design method for the protective storage buffer structure of the shipborne navigation data recorder based on the cloud model includes the following steps: The original time-domain signal of the multimodal impact load in the environment where the shipborne navigation data recorder is located is collected, and the peak acceleration, pulse width and energy density are extracted from the original time-domain signal as the characteristic parameter sequence of the impact load. Based on the aforementioned feature parameter sequence, a three-parameter positive cloud model is used to model the uncertainty of shipborne impact loads, resulting in a cloud model that characterizes the random distribution of impact loads. The cloud model is expressed by three feature quantities: expected value, entropy value, and hyperentropy value. Based on the cloud droplet clustering results, the impact loads are divided into multiple impact levels. Based on multiple impact levels, a nested three-level multi-layer buffer structure is designed. The buffer structure consists of an outer yield energy-absorbing layer, a middle platform energy-absorbing layer, and an inner flexible isolation layer from the outside to the inside. Different impact levels are matched with different levels of the buffer structure to achieve response sensitivity matching, so that each level has a differentiated load triggering threshold and energy absorption characteristics. Based on the buffer structure, a mechanical model of each layer of material is established, and the cloud model characteristics and impact level are used as inputs to construct a mapping relationship model from the cloud distribution of impact load to the mechanical response of the buffer structure, so as to predict the stress, strain and energy absorption of the buffer structure under different impact loads.

2. The design method for the protective storage buffer structure of a shipborne navigation data recorder based on a cloud model according to claim 1, characterized in that, Based on the aforementioned feature parameter sequence, a three-parameter positive cloud model is used to model the uncertainty of shipborne impact loads, resulting in a cloud model characterizing the random distribution of impact loads. This cloud model is expressed using three feature quantities: expected value, entropy value, and hyperentropy value. Furthermore, based on cloud droplet clustering results, the impact loads are divided into multiple impact levels, specifically including: The original sequence of impact loads is slid through a fixed sampling window to extract local extreme points to form a feature sequence. The mean of the feature sequence is calculated as the expected value, and the entropy and hyperentropy values ​​are generated based on the three-parameter positive cloud model algorithm. Based on the expected value, entropy value, and hyperentropy value, a cloud droplet sequence that follows a cloud distribution is generated using a two-level random number engine, and the membership degree of each feature parameter is calculated using an exponential function. The cloud droplets are then clustered based on the membership degree. The stability of the cloud model is verified by analyzing the parameter drift trajectory and discrete consistency of the cloud droplet sequence. Based on the cloud droplet clustering results, the impact load is divided into multiple impact levels corresponding to different cloud droplet aggregation areas, and a unique number is assigned to each impact level.

3. The design method for the protective storage buffer structure of a shipborne navigation data recorder based on a cloud model according to claim 1, characterized in that, Based on multiple impact levels, a nested three-level multi-layer buffer structure is designed. The buffer structure, from the outside in, consists of an outer yield energy-absorbing layer, a middle plateau energy-absorbing layer, and an inner flexible isolation layer. Different impact levels are matched with different levels of the buffer structure to ensure that each level has differentiated load triggering thresholds and energy absorption characteristics. Specifically, this includes: The primary energy-absorbing zone is constructed using easily yieldable energy-absorbing material, and its design parameters are matched with the low-level impact loads to absorb the initial peak energy of the impact load. A honeycomb or corrugated partitioned structure with varying thickness is adopted to form multiple secondary energy absorption zones with different yield plateaus. Each partition is matched with a medium-level impact load to perform graded load absorption and path redistribution. A flexible isolation zone is constructed using low-modulus buffer material and is positioned close to the core storage body being protected. It is matched with the high-level or residual impact loads to isolate residual loads and control contact stress. The outer yield energy-absorbing layer, the middle platform energy-absorbing layer, and the inner flexible isolation layer are nested together, and geometric transitions are made between each layer to limit stress concentration.

4. The cloud model based shipboard voyage data recorder guard storage buffer structure design method of claim 1, wherein, The mechanical model of each layer of material is established based on the buffer structure, and the cloud model characteristics and impact level are used as inputs to construct a mapping relationship model from the cloud distribution of impact load to the mechanical response of the buffer structure, so as to predict the stress, strain and energy absorption of the buffer structure under different impact loads. Specifically, this includes: The mechanical parameters of each layer of material were obtained through quasi-static compression tests. Segmented stress-strain constitutive models were established for the different mechanical behaviors of the outer yield energy-absorbing layer, the middle plateau energy-absorbing layer and the inner flexible isolation layer. The segmented mechanical model is assigned to the corresponding buffer layer, and a global finite element model of the three-level multi-layer buffer structure is established, and the contact relationship between the layers is defined. The obtained impact levels and their corresponding cloud model characteristics are bound together to form a load input matrix, and then mapped to the corresponding element set of the finite element model according to the time series. Based on the linear weight superposition method, a response mapping function is constructed with cloud model feature quantities and impact level as inputs and peak stress of each layer of the buffer structure as output. The weight coefficients of the response mapping function are iteratively corrected according to the error between the output of the finite element simulation model and the mapping prediction result until the mapping accuracy meets the requirements.

5. The cloud model based shipboard voyage data recorder guard storage buffer structure design method of claim 4, wherein, After establishing a mechanical model of each layer of material based on the buffer structure, and using cloud model characteristics and impact level as inputs to construct a mapping relationship model from the cloud distribution of impact load to the mechanical response of the buffer structure, in order to predict the stress, strain, and energy absorption of the buffer structure under different impact loads, the cloud model-based design method for the protective storage buffer structure of a shipborne navigation data recorder further includes: Based on the mapping relationship model, the thickness and geometric configuration parameters of the outer yield energy-absorbing layer, the middle platform energy-absorbing layer and the inner flexible isolation layer are optimized, so that the optimized buffer structure can achieve a controllable triggering sequence and stable energy distribution under different impact levels. The optimized buffer structure is subjected to multi-level impact tests, and test response data is collected. A protection performance rating is generated based on the test response data. The cloud model parameters and / or buffer structure parameters are iteratively corrected according to the protection performance rating until the protection requirements are met.

6. The cloud model based shipboard voyage data recorder guard storage buffer structure design method of claim 5, wherein, Based on the mapping relationship model, the thickness and geometric configuration parameters of the outer yield energy-absorbing layer, the middle plateau energy-absorbing layer, and the inner flexible isolation layer are optimized, specifically including: Using the maximization of the total energy absorption of the buffer structure and the minimization of the peak acceleration transferred to the core memory as the multi-objective optimization function, the thickness of the outer yield energy-absorbing layer, the pore size and wall thickness of the cell cells of the middle platform energy-absorbing layer, the thickness of the inner flexible isolation layer, and the density and modulus parameters of each layer material are set as the design variables to be optimized. Based on the installation space limitations of the shipborne navigation data recorder, geometric constraints are set for the overall dimensions of the buffer structure; based on the mechanical properties of each layer of material, performance constraints are set for the yield stress and compaction strain in the segmented stress-strain model. By using the constructed response mapping function as a surrogate model, and combining it with the defined objective optimization function, design variables, and applied constraints, the optimal combination of buffer structure parameters that satisfies the objective optimization function is obtained.

7. The cloud model based shipboard voyage data recorder guard storage buffer structure design method of claim 5, wherein, The step of iteratively correcting the cloud model parameters and / or buffer structure parameters based on the protection performance rating until the protection requirements are met specifically includes: If the protection performance rating is lower than the preset protection performance index, the test response data is first compared with the predicted data of the response mapping function, and the weighting coefficient of the response mapping function is corrected. Based on the revised response mapping function, the parameters of the buffer structure are re-optimized to generate new buffer structure parameters. The second protection performance rating is obtained based on the new buffer structure parameters. If the second protection performance rating is lower than the preset protection performance index, the parameters of the cloud model are adjusted and the impact level is reclassified until the protection performance rating reaches the preset protection performance index.

8. A cloud-based protective storage buffer structure for a shipborne navigation data recorder, wherein the buffer structure is disposed between the recorder housing and the core storage unit, and is designed using the design method described in any one of claims 1-7, characterized in that... The buffer structure is a nested three-level multi-layer energy-absorbing structure, which includes, in sequence along the load transfer direction: an outer yield energy-absorbing layer, made of easily yieldable energy-absorbing material, used to absorb the initial peak energy of the impact load; a middle plateau energy-absorbing layer, adjacent to the inner side of the outer yield energy-absorbing layer, adopting an energy-absorbing configuration with a gradually varying thickness and multiple yield plateaus, used to perform graded absorption of the impact load and energy path redistribution; and an inner flexible isolation layer, adjacent to the inner side of the middle plateau energy-absorbing layer and close to the core memory, made of low-modulus buffer material, used to flexibly isolate residual loads and control the contact stress acting on the core memory.

9. The cloud model based shipboard voyage data recorder guard storage buffer structure of claim 8, wherein, The middle platform energy-absorbing layer is specifically a honeycomb partitioned structure. The cell pore size and cell wall thickness of the honeycomb structure vary in gradient on the cross section parallel to the direction of impact load transmission, thereby forming multiple energy-absorbing partitions with different compressive yield strengths. Furthermore, the outer yield energy-absorbing layer and the middle platform energy-absorbing layer, as well as the middle platform energy-absorbing layer and the inner flexible isolation layer, are geometrically transitioned through rounded corners.

Images