Multi-class strong nonlinear curve prediction method and system of complex superstructure

Abstract

The invention discloses a multi-class strong nonlinear curve prediction method and system for a complex superstructure, and relates to the technical field of superstructures. Aiming at the problems of complex mechanical behaviors, strong nonlinear characteristics, multiple types, too high calculation cost and the like in the current complex superstructure mechanical property curve prediction, the method greatly improves the prediction accuracy of the multiple types of strong nonlinear curves by accurately describing the curve characteristics; and moreover, the problem that multiple curve features are mixed is effectively avoided, construction of the prediction model can be completed only by a small amount of data, and the cost of data generation and neural network training is reduced. The method provides a new thought for predicting the multi-class strong nonlinear mechanical property curve of the complex superstructure, and has a good application prospect in the field of complex superstructure research.

Description

Background Art

[0002] In recent years, the research on mechanical superstructures has achieved remarkable results, realizing a series of unconventional and counterintuitive mechanical properties that are rare or even non-existent in nature. By rationally designing the size parameters, superstructures can exhibit programmable mechanical responses. For example, by rationally designing the unit cell shape, the stiffness, Poisson's ratio, and thermal expansion coefficient of the superstructure can be programmed to span from negative values ​​to positive values. Based on a simple geometric parameterized microstructure layout design method, superstructures can achieve customized design of mechanical responses to meet different functional requirements under actual working conditions. Therefore, it is an important task to study the influence of the structural size parameters of complex metamaterials on mechanical properties in order to accurately predict the mechanical properties.

[0003] Through the reasonable arrangement of units, negative stiffness superstructures can exhibit sudden jump behavior in which the force changes suddenly in an instant, which has been proven to help further improve the buffering performance of superstructures. However, the sudden jump behavior makes the mechanical performance curve non-smooth. This strong nonlinear mutation feature greatly increases the difficulty of accurate prediction. Secondly, the mechanical properties of complex superstructures are affected by the geometric parameters of the structure itself. When there are many geometric parameters, the high-dimensional design space will increase the difficulty of accurate prediction. Therefore, the strong nonlinearity of the mechanical response of complex superstructures and the high-dimensional design space greatly increase the complexity of performance prediction, which makes the prediction of superstructure mechanical curves a very challenging task.

[0004] Machine learning can be used as a method for processing data curves. It can derive logic and rules from the provided data set and predict new data based on the obtained logic and rules. In order to improve the prediction accuracy, the most commonly used method is to increase the sample size of the training set, but this will greatly increase the computational cost of generating the data set. Therefore, an efficient and feasible method is needed to accurately predict superstructure multi-category curves, especially non-smooth curves with strong nonlinear characteristics. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for predicting multi-category strongly nonlinear curves of complex superstructures in response to the above-mentioned deficiencies in the prior art, thereby improving the accuracy while saving time, cost and computing resources, and solving the technical problems that traditional machine learning algorithms require a large number of simulations and experiments, a large number of training sets, and a lot of time and cost.

[0006] The present invention adopts the following technical solutions: A method for predicting multi-category strong nonlinear curves of complex superstructures comprises the following steps: S1. According to the influencing factors of the mechanical properties of complex superstructures, the geometric parameters of the complex superstructures are selected as the size parameters, and different structures are obtained by changing the size parameters. Quasi-static compression simulation is performed to obtain a correlation database between the structural size parameters and the mechanical response, and the curves and size parameters are normalized to obtain a curve data set; S2, using the curve monotonicity and smoothness to classify the curve features of the curve data set obtained in step S1, to obtain a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve; S3, according to the monotonically increasing positive stiffness curve, the smooth negative stiffness curve and the non-smooth sudden negative stiffness curve obtained in step S2, training a neural network in combination with a machine learning algorithm, and constructing a curve type prediction model; S4, based on the initial data set obtained in step S1, partition and reconstruct the curve with strong nonlinearity, that is, the non-smooth sudden jump negative stiffness curve obtained by the curve prediction model in step S3, by interpolation calculation, and concentrate the similar features of the curve in a specific area, that is, some specific sections of the curve, to facilitate feature prediction; S5. For the curve reconstructed in step S4, neural networks are trained respectively in combination with machine learning algorithms to construct a curve feature prediction model. The curve feature prediction model uses size parameters as input parameters and outputs a predicted quasi-static compression mechanical property curve.

[0007] Preferably, according to the influencing factors of the mechanical properties of the complex superstructure, 9 geometric parameters of the complex superstructure are selected as size parameters, and the value ranges of the 9 parameters are as follows:

[0008] in, is the thickness of the thin rod of AB rod and the thin rod of C rod, is the boundary length of the AB rod and the boundary length of the C rod, is the thickness of the AB rod and the thickness of the C rod, is the length of C rod L2, is the angle between the AB rods, is the C-bar angle, the specific parameters are as follows Figure 3 shown.

[0009] Preferably, the curve feature classification of the curve data set using the curve monotonicity and smoothness is specifically as follows: When all first-order derivatives are greater than 0, it is classified as a first-class curve; When there is a derivative less than 0, the curve is considered smooth when it does not exceed the set threshold and is classified as the second type of curve; On the contrary, it is a non-smooth jump behavior with sudden force change, which is classified as the third type of curve.

[0010] Preferably, the monotonically increasing positive stiffness curve, the smooth negative stiffness curve and the non-smooth sudden negative stiffness curve are specifically:

[0011] in, F and u are the normalized force and displacement, Indicates the threshold used to determine whether the curve is smooth.

[0012] Preferably, the curve type prediction model is constructed as follows: The size parameters are taken as input, the monotonically increasing positive stiffness curve, the smooth negative stiffness curve and the non-smooth sudden negative stiffness curve are marked as 1-3 respectively, and the category number is taken as output. The LSTM neural network is used for classification training to obtain the curve type prediction model.

[0013] Preferably, the forget gate of the curve type prediction model:

[0014] in, is the sigmoid function, is the weight matrix of the forget gate, is the hidden state at the previous time step, is the input of the current time step, is the bias term; Input Gate:

[0015]

[0016] in, is the activation value of the input gate, is a candidate memory cell, and is the weight matrix of the input gate and the candidate memory cell, and is the bias term; Update memory unit:

[0017] in, is the memory cell of the current time step, is the memory unit of the previous time step; Output Gate:

[0018]

[0019] in, is the activation value of the output gate, is the hidden state at the current time step, is the weight matrix of the output gate, is the bias term.

[0020] Preferably, the partition reconstruction is specifically as follows: For the positive stiffness curve, machine learning classification training is directly performed. For the smooth negative stiffness curve, when reconstructing the curve, the curve is divided into three regions, including two positive stiffness regions and one negative stiffness region; For the curve with a sharp change in negative stiffness, the curve is reconstructed into 5 regions, including 2 smooth segments and 1 rapidly decreasing non-smooth segment, plus 2 positive stiffness regions, a total of 5 regions.

[0021] Preferably, an interpolation calculation method is used to make the same partition of the same category of curves contain the same number of feature points, and obtain a mechanical response vector that describes the curve characteristics and is composed of the feature points on each partition and the points between different partitions.

[0022] Preferably, an error evaluation mechanism of the minimum Euclidean distance is established to verify the constructed curve feature prediction model, the force-displacement curve is discretized into 10,000 evaluation points, the Euclidean distance from each prediction point to the 10,000 evaluation points is calculated, and the minimum distance is extracted to evaluate the deviation of the prediction point from the reference curve.

[0023] In a second aspect, an embodiment of the present invention provides a multi-category strong nonlinear curve prediction system for a complex superstructure, comprising The data module selects the geometric parameters of the complex superstructure as the size parameters according to the influencing factors of the mechanical properties of the complex superstructure, changes the size parameters to obtain different structures, performs quasi-static compression simulation to obtain the correlation database between the structural size parameters and the mechanical response, and normalizes the curves and size parameters to obtain the curve data set; The classification module uses the monotonicity and smoothness of the curve to classify the curve features of the curve data set, and obtains a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve; A construction module is used to train a neural network based on a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve, and to construct a curve type prediction model by combining a machine learning algorithm; The reconstruction module reconstructs the strongly nonlinear curve by partitioning it according to the obtained initial data set through interpolation calculation, and concentrates the similar features of the curve in a specific area, that is, some specific sections of the curve, to facilitate feature prediction; The prediction module trains neural networks for the reconstructed curves in combination with machine learning algorithms to build a curve feature prediction model. The curve feature prediction model uses dimensional parameters as input parameters and outputs a predicted quasi-static compression mechanical property curve.

[0024] In a third aspect, a computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for predicting multi-category strongly nonlinear curves of complex superstructures when executing the computer program.

[0025] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for predicting multi-category strong nonlinear curves of complex superstructures.

[0026] Compared with the prior art, the present invention has at least the following beneficial effects: A method for predicting multi-category strong nonlinear curves of complex superstructures, which pre-processes the data curve set by curve classification and curve reconstruction, can greatly reduce the data of the training set while ensuring the prediction accuracy, and significantly reduce the experimental and simulation calculation costs. The prediction method for multi-category nonlinear curves of complex superstructures based on machine learning can accurately predict the mechanical properties of complex superstructures with an accuracy rate of over 95%, providing an algorithmic basis for accurate and efficient prediction of the mechanical properties of complex superstructures, and has good social and economic benefits.

[0027] Furthermore, nine geometric parameters of the complex superstructure are selected as dimensional parameters, aiming to deeply study the influence of these parameters on the mechanical properties of the superstructure and realize precise control of the mechanical response through reasonable geometric design. The selection of nine geometric parameters as dimensional parameters can comprehensively cover the geometric characteristics of the superstructure and ensure that the model can fully consider the influence of various design variables. By adjusting these nine key geometric parameters, the optimal design of the superstructure can be achieved, so that its mechanical properties meet the requirements of specific applications.

[0028] Furthermore, the curve features of the curve data set are classified using the curve monotonicity and smoothness, with the aim of efficiently and accurately identifying curves of different categories, thereby improving the model's prediction accuracy for the mechanical properties of complex superstructures. By analyzing the monotonicity of the curve, the curves can be effectively divided into two categories: positive stiffness curves and negative stiffness curves; smoothness classification can help identify areas of sudden changes in the curve, and further subdivide the negative stiffness curves into smooth and non-smooth categories. This classification method has important guiding significance for subsequent model training and prediction.

[0029] Furthermore, the nonlinear curves are initially classified to improve the adaptability and prediction accuracy of strongly nonlinear and non-smooth curves.

[0030] Furthermore, complex nonlinear mechanical property curves are partitioned so that similar features can be concentrated in specific areas, thereby simplifying the prediction process of complex curves, improving prediction accuracy, reducing training data requirements, and being applicable to a variety of neural network models.

[0031] Furthermore, the minimum Euclidean distance error evaluation mechanism is adopted. When dealing with the prediction of multi-category nonlinear mechanical property curves of complex superstructures, this method greatly improves the adaptability of the model to non-smooth and strongly nonlinear curve characteristics through discrete sampling points. It can be understood that the beneficial effects of the second aspect mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here.

[0032] In summary, the present invention greatly improves the prediction accuracy of multi-category strong nonlinear curves by accurately describing the curve characteristics; and effectively avoids the problem of mixing multiple curve characteristics, and only requires a small amount of data to complete the construction of the prediction model, reducing the cost of data generation and neural network training; it provides a new idea for the prediction of multi-category strong nonlinear mechanical property curves of complex superstructures, and has good application prospects in the field of complex superstructure research.

[0033] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings to be used in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 is a flow chart of the method of the present invention; Figure 2 A neural network flow chart for establishing a curve feature prediction model according to the present invention; Figure 3 A schematic diagram of the complex superstructure and its geometric parameters studied by the present invention; Figure 4 It is a schematic diagram of the curve classification and curve reconstruction method of the present invention; Figure 5 It is the error evaluation method, the principle diagram of minimum Euclidean distance; Figure 6A comparison chart of the prediction accuracy of different neural networks under feature-driven and non-feature-driven conditions; Figure 7 A schematic diagram of a computer device provided by an embodiment of the present invention; Figure 8 The present invention is a block diagram of an electronic device provided according to an embodiment. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] In the description of the present invention, it should be understood that the terms “include” and “comprises” indicate the presence of described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0038] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0039] It should be further understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present invention generally indicates that the associated objects are in an "or" relationship.

[0040] It should be understood that, although the terms first, second, third, etc. may be used to describe preset ranges, etc. in the embodiments of the present invention, these preset ranges should not be limited to these terms. These terms are only used to distinguish preset ranges from each other. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0041] The word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0042] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0043] The present invention provides a method for predicting multi-category strong nonlinear curves of complex superstructures. Based on curve preprocessing technologies such as classification and reconstruction, the method realizes fast and accurate prediction of multi-category strong nonlinear curves. The method does not need to design a structure with specific parameters or perform simulation. The mechanical properties of the complex superstructure can be directly predicted by inputting its characteristic parameters.

[0044] See also Figure 1 and Figure 2 The present invention provides a method for predicting multi-category strong nonlinear curves of complex superstructures, comprising the following steps: S1. Data generation: According to the factors affecting the mechanical properties of complex superstructures, 9 geometric parameters of complex superstructures are selected as dimensional parameters. Within a reasonable range, the sizes of various parameters are changed, and quasi-static compression simulation is performed in the commercial finite element software Abaqus to obtain the stress-strain data curve set of the structure. Data preprocessing is performed on the data curve set and dimensional parameters; See also Figure 3, study the geometric structure of the complex superstructure shown in the figure, and determine its 12 geometric parameters. The complex superstructure is composed of three rods A, B, and C, and the 12 geometric parameters are: AB rod angle α, C rod angle β, AB rod thin rod thickness t1, AB rod thickness δ1, C rod thin rod thickness t2, C rod thickness δ2, AB rod boundary length l1, C rod boundary length l2, AB rod angle length l3, AB rod angle thickness t3, AB rod length L1, C rod length L2; then, reasonably change 9 of the 12 geometric parameters, use the Py code based on Python secondary development, perform finite element simulation experiments in Abaqus software, obtain stress-strain curves under different parameter modeling, complete the construction of the initial database, and perform batch simulation of 1000 sets of data in this embodiment.

[0045] The value ranges of the 9 parameters are as follows:

[0046] In order to improve the efficiency of simulation, Abaqus was secondary developed based on Python, and Py files were used for finite element simulation to obtain curve data sets.

[0047] The initial data set is randomly divided into training set and test set according to 8:2, and all size parameters are normalized to reduce the influence of different size parameters on the results due to different dimensions and units, and eliminate the adverse effects caused by singular sample data.

[0048] The normalization formula is as follows:

[0049] in, X is the normalized result, X i is the original size parameter value, X min The minimum value of a size parameter range. X max The maximum value in the range.

[0050] S2. Curve classification: According to the different characteristics of the curves, the curve features of the obtained data set are classified using the curve monotony and smoothness, and three different types of curves are obtained: monotonically increasing positive stiffness curves, smooth negative stiffness curves, and non-smooth sudden negative stiffness curves; For the classification of curves, when all first-order derivatives are greater than 0, that is, the curve is monotonically increasing and exhibits positive stiffness characteristics, it is classified as a first-class curve; When there is a derivative less than 0, that is, there is a decreasing interval, which shows a negative stiffness characteristic. On this basis, the size of the slope is compared. When it does not exceed the set threshold, the curve is considered smooth and classified as the second type of curve; Otherwise, it is considered that a non-smooth jump behavior of sudden force change has occurred, which is classified as the third type of curve.

[0051] See also Figure 4 , calculate the slope of each point on the curve and the difference between the two adjacent points of the curve, and divide the curve into three different categories according to the monotonicity and smoothness characteristics: monotonically increasing positive stiffness curve, smooth negative stiffness curve and non-smooth sudden negative stiffness curve:

[0052] in, F and u are the normalized force and displacement, Indicates the threshold used to determine whether the curve is smooth.

[0053] S3. Establish a curve type prediction model: Based on the curve classification results and combined with machine learning algorithms, a curve type prediction model is constructed; After the curve classification, the size parameter is used as input, and the three types of curves are marked as 1-3 respectively. The 9 designable size parameters are used as input and the category number is used as output. The LSTM neural network is used for classification training, which contains 100 neurons and a total of 1500 iterations to build a prediction model of the curve type. LSTM, long short-term memory network, is a special type of recurrent neural network (RNN) composed of three gates (gating units): forget gate, input gate and output gate. The following is the basic formula of LSTM: Forget Gate:

[0054] in, is the sigmoid function, is the weight matrix of the forget gate, is the hidden state at the previous time step, is the input of the current time step, is the bias term.

[0055] Input Gate:

[0056]

[0057] in, is the activation value of the input gate, is a candidate memory cell, and is the weight matrix of the input gate and the candidate memory cell, and is the bias term.

[0058] Update memory unit:

[0059] in, is the memory cell of the current time step, is the memory cell of the previous time step.

[0060] Output Gate:

[0061]

[0062] in, is the activation value of the output gate, is the hidden state at the current time step, is the weight matrix of the output gate, is the bias term.

[0063] S4, curve reconstruction: Based on the obtained initial data set, the curve that is difficult to predict is reconstructed through interpolation calculation, and the similar features of the curve are concentrated in a specific area, that is, some specific sections of the curve, to facilitate feature prediction; According to the curve type prediction model in step S3, the initial database is divided into three small databases of different curve types; then the curves containing negative stiffness are partitioned by calculating the slope. For the second type of weak nonlinear curves, they are divided into three partitions of positive stiffness, negative stiffness, and positive stiffness according to the positive and negative slopes; for the third type of strong nonlinear curves, firstly, the curves are divided into three partitions of positive stiffness, negative stiffness, and positive stiffness according to the positive and negative slopes. In the negative stiffness partition, the curves are divided into three partitions of weak negative stiffness, strong negative stiffness, and weak negative stiffness according to whether the slope exceeds the threshold, for a total of five partitions.

[0064] When the obtained support reaction force data curve has a negative stiffness interval, because the positive stiffness interval has fewer features and is easy to predict, while the negative stiffness interval has complex features, the curve changes quickly and is difficult to predict, so through curve reconstruction, the positive and negative stiffness areas and negative stiffness areas of the curve are separated, and similar features are concentrated in specific areas. This can not only improve the prediction accuracy, but also reduce the number of training sets, making training easier. Figure 4 As shown, the specific steps are as follows: For the positive stiffness curve, machine learning classification training can be directly performed due to its simple curve characteristics. For the non-monotonic curve of negative stiffness, there are mainly two types, smooth negative stiffness curve and non-smooth sudden negative stiffness curve. For the smooth negative stiffness curve, when reconstructing the curve, the curve is divided into three regions, namely, two positive stiffness regions and one negative stiffness region.

[0065] For the curve with a sharp change in negative stiffness, the curve is reconstructed into 5 regions. Because the slope of the negative stiffness region changes sharply and the local characteristics at both ends of the negative stiffness region are obvious, the negative stiffness part is further subdivided according to the continuity characteristics, including 2 smooth segments and 1 rapidly decreasing non-smooth segment, plus 2 positive stiffness regions, a total of 5 regions.

[0066] Afterwards, the interpolation calculation method is used to make the same partition of the same category of curves contain the same number of feature points, and the mechanical response vector describing the curve characteristics is obtained by the feature points on each partition and the points between different partitions. After the curve is reconstructed, the sampling points in the negative stiffness area are increased, which improves the above-mentioned curve preprocessing technology based on curve classification, curve reconstruction, etc., and realizes the fast and accurate prediction of strong nonlinear curves. This method is called feature-driven method.

[0067] S5. Establish a curve feature prediction model: Based on the response vector that describes the curve characteristics, multiple shallow and lightweight neural network trainings are performed on different types of curves to construct a curve feature prediction model; an error evaluation mechanism of the minimum Euclidean distance is established for the curve feature prediction model.

[0068] With the dimensional parameters as input and the mechanical response vector as output, multiple shallow and lightweight LSTM neural networks are trained for the classified and reconstructed curves. The training results are then tested using the test set, and finally a curve feature prediction model is established. When testing the training results, the minimum Euclidean distance is used as the error evaluation method. The force-displacement curve is discretized into 10,000 evaluation points, and the Euclidean distance from each prediction point to the 10,000 evaluation points is calculated. The minimum distance is extracted to evaluate the deviation between the prediction point and the reference curve. The formula for MED (minimum Euclidean distance) is as follows:

[0069] Where i represents the i-th sampling point, j represents the j-th evaluation point, ( , ) is the coordinate of the prediction point, ( , ) is the reference point coordinate, m is the number of evaluation points, and n is the number of sampling points.

[0070] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as systems, methods or program products. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, which may be collectively referred to herein as "circuits", "modules" or "platforms".

[0071] In yet another embodiment of the present invention, a multi-category strong nonlinear curve prediction system for a complex superstructure is provided. The system can be used to implement the multi-category strong nonlinear curve prediction method for a complex superstructure. Specifically, the multi-category strong nonlinear curve prediction system for a complex superstructure includes a data module, a classification module, a construction module, a reconstruction module and a prediction module.

[0072] Among them, the data module selects the geometric parameters of the complex superstructure as the size parameters according to the influencing factors of the mechanical properties of the complex superstructure, changes the size parameters to obtain different structures, performs quasi-static compression simulation to obtain the correlation database between the structural size parameters and the mechanical response, and normalizes the curves and size parameters to obtain the curve data set; The classification module uses the monotonicity and smoothness of the curve to classify the curve features of the curve data set, and obtains a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve; A construction module is used to train a neural network based on a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve, and to construct a curve type prediction model by combining a machine learning algorithm; The reconstruction module reconstructs the strongly nonlinear curve by partitioning it according to the obtained initial data set through interpolation calculation, and concentrates the similar features of the curve in a specific area, that is, some specific sections of the curve, to facilitate feature prediction; The prediction module trains neural networks for the reconstructed curves in combination with machine learning algorithms to build a curve feature prediction model. The curve feature prediction model uses dimensional parameters as input parameters and outputs a predicted quasi-static compression mechanical property curve.

[0073] In another embodiment of the present invention, a terminal device is provided, the terminal device includes a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions to implement corresponding method processes or corresponding functions; the processor described in the embodiment of the present invention can be used for the operation of a multi-category strong nonlinear curve prediction method for complex superstructures, including: According to the influencing factors of the mechanical properties of complex superstructures, the geometric parameters of the complex superstructure are selected as dimensional parameters, and different structures are obtained by changing the dimensional parameters. Quasi-static compression simulation is performed to obtain a database of associations between structural dimensional parameters and mechanical responses, and the curves and dimensional parameters are normalized to obtain a curve data set; the curve monotonicity and smoothness are used to classify the curve features of the obtained curve data set to obtain a monotonically increasing positive stiffness curve, a smooth negative stiffness curve and a non-smooth sudden negative stiffness curve; based on the obtained monotonically increasing positive stiffness curve, smooth negative stiffness curve and non-smooth sudden negative stiffness curve, a neural network is trained in combination with a machine learning algorithm to construct a curve type prediction model; based on the obtained initial data set, the curve with strong nonlinearity, that is, the non-smooth sudden negative stiffness curve obtained by the curve prediction model, is partitioned and reconstructed through interpolation calculation, and the similar features of the curve are concentrated in specific areas, that is, certain specific sections of the curve, to facilitate feature prediction; for the reconstructed curves, the neural networks are trained separately in combination with the machine learning algorithm to construct a curve feature prediction model, which takes the dimensional parameters as input parameters and outputs a predicted quasi-static compression mechanical property curve.

[0074] In another embodiment of the present invention, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a terminal device for storing programs and data. It is understandable that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and the extended storage medium supported by the terminal device, and can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that more specific examples (non-exhaustive list) of the computer-readable storage medium here include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0075] Computer readable storage media also include data signals propagated in baseband or as part of a carrier wave, which carry readable program codes. Such propagated data signals can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable storage medium can also be any readable medium other than a readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program code contained on the readable storage medium can be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0076] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0077] The processor may load and execute one or more instructions stored in a computer-readable storage medium to implement the corresponding steps of the method for predicting multi-category strong nonlinear curves of complex superstructures in the above embodiment; the processor may load and execute the following steps: According to the influencing factors of the mechanical properties of complex superstructures, the geometric parameters of the complex superstructure are selected as dimensional parameters, and different structures are obtained by changing the dimensional parameters. Quasi-static compression simulation is performed to obtain a database of associations between structural dimensional parameters and mechanical responses, and the curves and dimensional parameters are normalized to obtain a curve data set; the curve monotonicity and smoothness are used to classify the curve features of the obtained curve data set to obtain a monotonically increasing positive stiffness curve, a smooth negative stiffness curve and a non-smooth sudden negative stiffness curve; based on the obtained monotonically increasing positive stiffness curve, smooth negative stiffness curve and non-smooth sudden negative stiffness curve, a neural network is trained in combination with a machine learning algorithm to construct a curve type prediction model; based on the obtained initial data set, the curve with strong nonlinearity, that is, the non-smooth sudden negative stiffness curve obtained by the curve prediction model, is partitioned and reconstructed through interpolation calculation, and the similar features of the curve are concentrated in specific areas, that is, certain specific sections of the curve, to facilitate feature prediction; for the reconstructed curves, the neural networks are trained separately in combination with the machine learning algorithm to construct a curve feature prediction model, which takes the dimensional parameters as input parameters and outputs a predicted quasi-static compression mechanical property curve.

[0078] See also Figure 7 , the terminal device is a computer device, and the computer device 60 of this embodiment includes: a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When the computer program 63 is executed by the processor 61, the method for predicting multiple categories of strongly nonlinear curves of a complex superstructure in the embodiment is implemented. To avoid repetition, it is not described one by one here. Alternatively, when the computer program 63 is executed by the processor 61, the functions of each model / unit in the system for predicting multiple categories of strongly nonlinear curves of a complex superstructure in the embodiment are implemented. To avoid repetition, it is not described one by one here.

[0079] The computer device 60 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will appreciate that Figure 7 This is only an example of the computer device 60 and does not constitute a limitation of the computer device 60. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the computer device may also include input and output devices, network access devices, buses, etc.

[0080] The processor 61 may be a central processing unit (CPU), or other general-purpose processors, central processing units, graphics processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, data processing logic based on quantum computing, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0081] The memory 62 may be an internal storage unit of the computer device 60, such as a hard disk or memory of the computer device 60. The memory 62 may also be an external storage device of the computer device 60, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the computer device 60.

[0082] Furthermore, the memory 62 may include both an internal storage unit of the computer device 60 and an external storage device. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 may also be used to temporarily store data that has been output or is to be output.

[0083] Any reference to a memory, database or other medium used in the embodiments provided in the present application may include at least one of a non-volatile and a volatile memory. Non-volatile memory may include a read-only memory (ROM), a magnetic tape, a floppy disk, a flash memory, an optical memory, a high-density embedded non-volatile memory, a resistive random access memory (ReRAM), a magnetic random access memory (MRAM), a ferroelectric random access memory (FRAM), a phase change memory (PCM), a graphene memory, etc. Volatile memory may include a random access memory (RAM) or an external cache memory, etc. As an illustration and not limitation, RAM may be in various forms, such as a static random access memory (SRAM) or a dynamic random access memory (DRAM), etc.

[0084] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0085] See also Figure 8 The terminal device 600 is an electronic device, and the electronic device is in the form of a general computing device. The components of the electronic device may include but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0086] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 performs the steps according to various exemplary embodiments of the present invention described in the above method section of this specification. For example, the processing unit 610 can perform the following steps: Figure 1 Follow the steps shown in .

[0087] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0088] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0089] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0090] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

[0091] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention described and shown in the drawings here can usually be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0092] See also Figure 5 ,This figure reflects the advantage of using the minimum Euclidean distance method when dealing with non-smooth curves with large slope changes.

[0093] Horizontal axis: "Strain" indicates the degree of deformation of a material when subjected to an external force, usually expressed as a percentage or a specific value. In this chart, the range of strain is from -0.1 (compressive strain) to 1.0 (tensile strain), increasing in steps of 0.2.

[0094] Vertical axis: "Absolute deviation (MPa)" indicates the difference between the actual measured value and a reference value or predicted value at a specific strain, in MPa (megapascals). This indicator reflects the accuracy and reliability of the measurement results.

[0095] Reference curve: Starting from the lower left corner, as the strain increases (from negative to positive values, i.e. from compression to tension), the absolute deviation decreases gradually. At higher tensile strains, the material's performance is more stable, or the measurement system is more accurate under these conditions.

[0096] Prediction curve: The trend is opposite to that of the reference curve. It starts from the upper left corner, and as the strain increases, the absolute deviation gradually decreases, but it is generally below the reference curve. This means that the predicted value is closer to the actual value than the reference value in most cases, or the prediction model can more accurately reflect the relationship between strain and absolute deviation than the reference standard in some aspects.

[0097] The small box on the right side of the chart is labeled "Absolute Deviation (MPa)" and "Minimum Euclidean Distance". This is a quantitative description of the difference between the curves in the chart or between the curves and an ideal state. The minimum Euclidean distance is a common method to measure the distance between two points in space. It is used here to evaluate the closeness or difference between the predicted curve and the reference curve.

[0098] If the prediction accuracy does not meet the set standard, continue to increase the training set for training until the set condition is met. In addition, due to the use of curve classification and curve reconstruction preprocessing methods, the curve features are concentrated, and BP neural network and RBF neural network can also be used for training, and the accuracy is significantly improved, indicating that the feature-driven machine learning method in the present invention is applicable to any neural network.

[0099] See also Figure 6 , compares the minimum Euclidean distance between feature-driven machine learning algorithms and traditional algorithms, where the X-axis represents the type of neural network, from left to right: BP, LSTM, and RBF.

[0100] The Y-axis represents the scale of the minimum Euclidean distance, which is divided into two main parts. The bottom part is (×10 -2 ) is the scale, indicating a smaller distance value; the top is (×10²) as the scale, indicating a larger distance value. This dual-scale design allows the chart to simultaneously display data changes at different orders of magnitude.

[0101] BP neural network: On the smaller scale of the Y axis (×10 -2 ) area, the position of the data points of the BP network is relatively low, indicating that at a smaller scale (i.e., when the minimum Euclidean distance is small), the output performance of the BP network is better. However, in the area with a larger Y-axis scale (×10²), the position of the data points of the BP network rises significantly, indicating that at a larger scale (i.e., when the minimum Euclidean distance is large), the performance of the BP network is poor.

[0102] LSTM neural network: The LSTM network maintains a relatively low and stable data point position at each scale of the Y-axis, which means that the LSTM network performs well at different scales and has good stability and robustness.

[0103] RBF neural network: Similar to BP network, RBF network is more efficient at smaller scales (×10 -2 ) also shows good performance, with lower data point locations. However, at a larger scale (×10²), the performance of the RBF network decreases, but the decrease is smaller than that of the BP network, and the performance is still better than that of the BP network.

[0104] By comparing the minimum Euclidean distance values ​​of the three neural networks, BP, LSTM, and RBF, at different scales, the performance differences in dealing with specific problems are shown. Among them, the LSTM network stands out for its excellent stability and robustness, while the BP and RBF networks perform well at smaller scales but have limitations when dealing with larger scale data.

[0105] In summary, the present invention provides a method and system for predicting multi-category strong nonlinear curves of complex superstructures, which greatly improves the prediction accuracy of multi-category strong nonlinear curves by accurately describing the characteristics of the curves; and effectively avoids the problem of mixing multiple curve characteristics, and only requires a small amount of data to complete the construction of the prediction model, thereby reducing the cost of data generation and neural network training; it provides a new idea for the prediction of multi-category strong nonlinear mechanical property curves of complex superstructures, and has good application prospects in the field of complex superstructure research.

[0106] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0107] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0108] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0109] In the embodiments provided by the present invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0110] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0111] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0112] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0113] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0116] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A method for predicting multi-class strongly nonlinear curves of complex superstructures, characterized in that: The following steps are involved: S1. According to the influencing factors of the mechanical properties of complex superstructures, the geometric parameters of the complex superstructures are selected as the size parameters, and different structures are obtained by changing the size parameters. Quasi-static compression simulation is performed to obtain a correlation database between the structural size parameters and the mechanical response, and the curves and size parameters are normalized to obtain a curve data set; S2, using the curve monotonicity and smoothness to classify the curve features of the curve data set obtained in step S1, to obtain a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve; S3, according to the monotonically increasing positive stiffness curve, the smooth negative stiffness curve and the non-smooth sudden negative stiffness curve obtained in step S2, training a neural network in combination with a machine learning algorithm, and constructing a curve type prediction model; S4, based on the initial data set obtained in step S1, partition and reconstruct the non-smooth sudden negative stiffness curve obtained by the curve prediction model in step S3 through interpolation calculation, and concentrate similar features of the curve in a specific area; S5. For the curve reconstructed in step S4, neural networks are trained respectively in combination with machine learning algorithms to construct a curve feature prediction model. The curve feature prediction model uses size parameters as input parameters and outputs a predicted quasi-static compression mechanical property curve.

2. The method for predicting multi-class strong nonlinear curves of complex superstructures according to claim 1, characterized in that: According to the influencing factors of the mechanical properties of complex superstructures, 9 geometric parameters of complex superstructures are selected as size parameters. The value ranges of the 9 parameters are as follows: in, is the thickness of the thin rod of AB rod and the thin rod of C rod, is the boundary length of the AB rod and the boundary length of the C rod, is the thickness of the AB rod and the thickness of the C rod, is the length of C rod L2, is the angle between the AB rods, is the C-bar angle.

3. The multi-category strong nonlinear curve prediction method for complex superstructures according to claim 1, characterized in that: The curve feature classification of the curve data set using the curve monotonicity and smoothness is as follows: When all first-order derivatives are greater than 0, it is classified as a first-class curve; When there is a derivative less than 0, the curve is considered smooth when it does not exceed the set threshold and is classified as the second type of curve; On the contrary, it is a non-smooth jump behavior with sudden force change, which is classified as the third type of curve.

4. The method for predicting multi-class strong nonlinear curves of complex superstructures according to claim 3, characterized in that: The monotonically increasing positive stiffness curve, the smooth negative stiffness curve and the non-smooth sudden negative stiffness curve are as follows: in, F and u are the normalized force and displacement, Indicates the threshold used to determine whether the curve is smooth.

5. The method for predicting multi-class strong nonlinear curves of complex superstructures according to claim 1, characterized in that: The specific steps of constructing the curve type prediction model are as follows: The size parameters are taken as input, the monotonically increasing positive stiffness curve, the smooth negative stiffness curve and the non-smooth sudden negative stiffness curve are marked as 1-3 respectively, and the category number is taken as output. The LSTM neural network is used for classification training to obtain the curve type prediction model.

6. The method for predicting multi-category strong nonlinear curves of complex superstructures according to claim 5, characterized in that: Forget gate of curve type prediction model: in, is the sigmoid function, is the weight matrix of the forget gate, is the hidden state at the previous time step, is the input of the current time step, is the bias term; Input Gate: in, is the activation value of the input gate, is a candidate memory cell, and is the weight matrix of the input gate and the candidate memory cell, and is the bias term; Update memory unit: in, is the memory cell of the current time step, is the memory unit of the previous time step; Output Gate: in, is the activation value of the output gate, is the hidden state at the current time step, is the weight matrix of the output gate, is the bias term.

7. The method for predicting multi-category strong nonlinear curves of complex superstructures according to claim 1, characterized in that: Partition reconstruction is as follows: For the positive stiffness curve, machine learning classification training is directly performed. For the smooth negative stiffness curve, when reconstructing the curve, the curve is divided into three regions, including two positive stiffness regions and one negative stiffness region; For the curve with a sharp change in negative stiffness, the curve is reconstructed into 5 regions, including 2 smooth segments and 1 rapidly decreasing non-smooth segment, plus 2 positive stiffness regions, a total of 5 regions.

8. The method for predicting multi-category strong nonlinear curves of complex superstructures according to claim 7, characterized in that: By using the interpolation calculation method, the same partition of the same category of curves contains the same number of characteristic points, and the mechanical response vector describing the characteristics of the curve is obtained, which is composed of the characteristic points on each partition and the points between different partitions.

9. The method for predicting multi-class strong nonlinear curves of complex superstructures according to claim 1, characterized in that: An error evaluation mechanism of minimum Euclidean distance was established to verify the constructed curve feature prediction model. The force-displacement curve was discretized into 10,000 evaluation points, the Euclidean distance from each prediction point to the 10,000 evaluation points was calculated, and the minimum distance was extracted to evaluate the deviation between the prediction point and the reference curve.

10. A multi-category strong nonlinear curve prediction system for complex superstructures, characterized in that: include: The data module selects the geometric parameters of the complex superstructure as the size parameters according to the influencing factors of the mechanical properties of the complex superstructure, changes the size parameters to obtain different structures, performs quasi-static compression simulation to obtain the correlation database between the structural size parameters and the mechanical response, and normalizes the curves and size parameters to obtain the curve data set; The classification module uses the monotonicity and smoothness of the curve to classify the curve features of the curve data set, and obtains a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve; A construction module is used to train a neural network based on a monotonically increasing positive stiffness curve, a smooth negative stiffness curve, and a non-smooth sudden negative stiffness curve, and to construct a curve type prediction model by combining a machine learning algorithm; The reconstruction module reconstructs the strongly nonlinear curve by partitioning it according to the obtained initial data set through interpolation calculation, and concentrates the similar features of the curve in a specific area, that is, some specific sections of the curve, to facilitate feature prediction; The prediction module trains neural networks for the reconstructed curves in combination with machine learning algorithms to build a curve feature prediction model. The curve feature prediction model uses dimensional parameters as input parameters and outputs a predicted quasi-static compression mechanical property curve.

Images