Active control method and device for cable-strut tension structure based on physical information neural network

Abstract

The present invention relates to an active control method and device of a cable-strut tension structure based on a physical information neural network, the method comprising the following steps: obtaining the actuation of a number of randomly generated units; calculating the tangent stiffness matrix of the structure after applying prestress and the actuation, and when the matrix is ​​a positive definite matrix, using a physical information neural network to fit the displacement of the structure; calculating the unit internal force of the corresponding structure based on the maximum displacement obtained by fitting, judging whether the obtained unit internal force and maximum displacement meet the constraint conditions, and if so, calculating the loss function value, further judging whether the maximum number of iterations has been reached, and guiding the iterative process of the physical information neural network with the loss function value in the next iteration; taking the actuation corresponding to the minimum displacement of the structure fitted in each iteration as the final required result, and realizing active control based on the final required result. Compared with the prior art, the present invention has the advantages of high calculation efficiency and wide application range.

Description

Technical Field

[0001] The present invention relates to the field of intelligent construction of cable structures, and in particular to an active control method and device for a cable-rod tension structure based on a physical information neural network. Background Art

[0002] Active structural control is a technology that can improve structural performance. It uses sensors, controllers and actuators to monitor the dynamic response of the structure in real time, and adjusts the behavior of the structure through active intervention to reduce adverse effects. A sensing structure refers to a structure that can monitor and diagnose the structural system, and has various sensors that can monitor the structural state response; an adaptive structure refers to a system equipped with actuators that can control movable components through instructions to change the structural geometry and force characteristics. Intelligent structures can make good use of the characteristics of both to achieve closed-loop active control, and have learning capabilities, and can achieve self-monitoring, self-diagnosis, and self-adaptation without human intervention.

[0003] For cable-strut tension structures with active control functions, they must go through four states from the construction stage to the use stage: initial state, prestressed state, load state, and regulation state. In the process of active structural control, the key is to determine the control strategy of the structure in the corresponding state according to the control target. During the use of the structure, its shape may need to be adjusted to meet specific needs. To this end, actuators that can adjust the length of the unit can be introduced into the high-sensitivity unit of the cable-strut tension structure. These actuators realize active control of the structure by changing the internal force of the unit, thereby effectively regulating the structure under the action of load.

[0004] In the past, active control of cable-strut tension structures was mostly based on control equation method, inverse iteration method and optimization algorithm. However, these control equation methods are complex to calculate and require high professional quality; the inverse iteration method may not converge; the optimization algorithm is slow and consumes a lot of computing resources. Therefore, it is necessary to develop a new active control method for cable-strut tension structures. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a method and device for active control of cable-strut tension structure based on physical information neural network with high computational efficiency and wide application range.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] An active control method for a cable-strut tension structure based on a physical information neural network comprises the following steps:

[0008] S1. Obtain the actuation of a number of randomly generated units;

[0009] S2, calculating the tangent stiffness matrix of the structure after applying the prestress and the actuation, and determining whether the tangent stiffness matrix is ​​a positive definite matrix, if so, executing step S3, if not, returning to step S1;

[0010] S3, using the physical information neural network to fit the structural displacement, and recording the maximum displacement of this iterative fitting;

[0011] S4, calculating the unit internal force of the corresponding structure based on the maximum displacement value obtained by fitting in step S3, and judging whether the obtained unit internal force and maximum displacement value meet the constraint conditions, if so, executing step S5, if not, returning to step S1;

[0012] S5, based on the maximum displacement obtained by fitting in step S3 and the actual node displacement derived from the tangent stiffness matrix in step S2, calculate the loss function value, and determine whether the maximum number of iterations has been reached. If so, execute step S6, if not, return to step S3, and use the loss function value to guide the iteration process of the physical information neural network;

[0013] S6. Taking the actuation amount corresponding to the minimum displacement of the structure fitted in each iteration as the final required result, and realizing active control of the cable-strut tension structure based on the final required result.

[0014] Furthermore, the calculation step of the tangent stiffness matrix includes:

[0015] S21. Calculate the elastic stiffness matrix K of the structure e ;

[0016] S22. Calculate the geometric stiffness matrix K of the structure p , the geometric stiffness matrix is ​​expressed as the contribution of the prestress of each element;

[0017] S23, based on the elastic stiffness matrix K e and the geometric stiffness matrix K p The sum of the tangent stiffness matrix K of the structure is obtained T .

[0018] Furthermore, the elastic stiffness matrix K e The calculation formula is:

[0019]

[0020] In the formula, a k represents the direction cosines of the nodes at both ends of unit k; b represents the number of units; E k represents the elastic modulus of unit k; A k is the cross-sectional area of ​​unit k; l k is the length of unit k.

[0021] Furthermore, the geometric stiffness matrix K p The calculation formula is:

[0022]

[0023] Where, t k is the axial force of unit k; g k is the direction vector of the geometric stiffness matrix of element k; b represents the number of elements; l k is the length of unit k.

[0024] Furthermore, the constraint conditions include stress conditions, stability conditions and tension-compression relationships, wherein:

[0025] The expression of the stress condition is:

[0026]

[0027] In the formula, σ t , σ c , [σ t ]、[σ c ] represent the cable stress, strut stress, cable material strength and strut material strength respectively;

[0028] The stability condition refers to the stability coefficient satisfying the stability condition of the strut;

[0029] The expression of the tension-compression relationship is:

[0030]

[0031] In the formula, f t and f c They represent the initial tension of the cables and struts of the cable-strut tension structure respectively.

[0032] Furthermore, the calculation formula of the stability coefficient is:

[0033]

[0034] In the formula, is the stability coefficient, f is the design value of the compressive strength of the rod, A represents the cross-sectional area, and N represents the axial force of the brace.

[0035] Furthermore, the calculation process of the loss function value includes:

[0036] S51. The actual node displacement is derived from the tangent stiffness matrix. The calculation formula of the actual node displacement Δ is:

[0037]

[0038] In the formula, K Trepresents the tangent stiffness matrix, F(Act) is the prestress corresponding to the input action;

[0039] S52. Calculate the loss function value ε k , the calculation formula is:

[0040]

[0041] Wherein, Δ(θ) is the maximum displacement value obtained by fitting in step S3.

[0042] The present invention also provides an active control device for a cable-rod tension structure based on a physical information neural network, comprising:

[0043] A random actuation generation module is used to obtain the actuations of a number of randomly generated units;

[0044] A tangent stiffness matrix calculation module is used to calculate the tangent stiffness matrix of the structure after applying the prestress and the actuation, and to determine whether the tangent stiffness matrix is ​​a positive definite matrix. If not, the random actuation generation module is called again;

[0045] The fitting module responds when the tangent stiffness matrix is ​​determined to be a positive definite matrix, and is used to fit the structural displacement using the physical information neural network, and record the maximum displacement of this iterative fitting;

[0046] The constraint condition judgment module calculates the unit internal force of the corresponding structure based on the maximum displacement value obtained by the fitting module, and judges whether the obtained unit internal force and maximum displacement value meet the constraint condition. If not, the random action moment generation module is called again;

[0047] A loss function calculation module, which responds when the constraint condition is met, calculates the loss function value based on the maximum displacement value obtained by the fitting module and the actual node displacement derived from the tangent stiffness matrix of the tangent stiffness matrix calculation module, and determines whether the maximum number of iterations has been reached. If not, the fitting module is re-called, and the iterative process of the physical information neural network is guided by the loss function value;

[0048] The control output module responds when the maximum number of iterations is reached, and is used to take the actuation corresponding to the minimum displacement of the structure fitted in previous iterations as the final required result, and realize active control of the cable-strut tension structure based on the final required result.

[0049] The present invention also provides an active control device for a cable-strut tension structure based on a physical information neural network, comprising one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the active control method for a cable-strut tension structure based on a physical information neural network as described above.

[0050] The present invention also provides a computer-readable storage medium, comprising one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the active control method of the cable-strut tension structure based on the physical information neural network as described above.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. Active control based on physical information neural network (PINN) can not only effectively solve the active control problem of single self-stress mode structure, but also solve the active control problem of multiple self-stress modes, and has a wide range of applications.

[0053] 2. The present invention does not rely on external finite element programs and has high computational efficiency. In the control of each stage, the degree of optimization of displacement regulation can reach 60% to 85%. It has strong regulation capability and is suitable for other cable-strut tension structures with the same physical information control equations, and has strong generalization capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is the overall flow chart of the present invention;

[0055] Figure 2 is a geometric dimension diagram of a cable rod tension structure in an embodiment of the present invention;

[0056] Figure 3 3a is an isometric view and a top view of the cable-rod tension structure in an embodiment of the present invention, wherein (3a) is an isometric view and (3b) is a top view;

[0057] Figure 4 4 is a graph showing the change of the loss function when the structure is actively controlled based on PINN during the iteration process in an embodiment of the present invention, wherein (4a) is a graph showing the change of the loss function in stage 3, and (4b) is a graph showing the change of the loss function in stage 5;

[0058] Figure 5 5a is the control diagram of the inner ring rope displacement control diagram, and 5b is the outer ring rope displacement control diagram. DETAILED DESCRIPTION

[0059] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0060] The present embodiment provides an active control method for a cable-strut tension structure based on a physical information neural network. A tangent stiffness matrix is ​​obtained based on a randomly generated actuation, and a physical information neural network is used to fit the structural displacement. The next iteration of the neural network is guided by a loss function value between the node displacement obtained by fitting and the actual node displacement derived from the tangent stiffness matrix. The actuation corresponding to the minimum displacement of the structure fitted in all iterations is used as the final required result, and active control of the cable-strut tension structure is achieved based on the final required result.

[0061] This embodiment takes the Kiewitt type cable dome, a typical cable-rod tension structure, as an example. The Kiewitt type cable dome model is as follows: Figure 2 and Figure 3 As shown, by selecting the required actuator unit, the active control of the cable-rod tension structure can be automatically calculated and realized. For the specific process, refer to Figure 1 As shown, the following steps are included:

[0062] S1, randomly generate a set of unit motions.

[0063] S2, calculate the tangent stiffness of the structure after applying prestress and actuation. If the tangent stiffness matrix is ​​a positive definite matrix, evaluate the displacement result of the structure; otherwise, return to S1.

[0064] Further, in step S2, the calculation steps of the tangent stiffness matrix of the cable-strut tension structure are as follows:

[0065] S21, calculate the elastic stiffness matrix K of the structure e , the calculation formula is as follows:

[0066]

[0067] In the formula, a k The vector actually represents the direction cosines of the nodes i and j at both ends of unit k; b represents the number of units; E k represents the elastic modulus of unit k; A k is the cross-sectional area of ​​unit k; l k is the length of unit k;

[0068] S22, calculate the structural geometric stiffness matrix K p , the geometric stiffness matrix is ​​expressed as the contribution of the prestress of each unit, the geometric stiffness matrix K p The calculation formula is as follows:

[0069]

[0070] Where, t k is the axial force of the unit; g k is the direction vector of the geometric stiffness matrix of element k;

[0071] S23, the structural elasticity matrix K calculated in the first two steps e and the structural geometric stiffness matrix K p Adding together, we can get the tangent stiffness matrix K of the structure T .

[0072] S3, use the PINN network to fit the structural displacement and record the maximum displacement of this iterative fitting.

[0073] S4, using the node displacement fitted by the neural network to calculate the internal force of the corresponding structural unit, evaluate whether the obtained unit internal force and node displacement meet the constraint conditions; if the constraint conditions are met, proceed to the next step of calculating the loss function; otherwise, return to S1 for loop processing.

[0074] Furthermore, in step S4, the constraints that the unit internal forces and node displacements must satisfy include:

[0075] (1) Stress conditions: The stress of the cables and struts of the lower cable-strut system shall not exceed the material strength. The judgment conditions are as follows:

[0076]

[0077] In the formula, σ t , σ c , [σ t ]、[σ c ] represent the cable stress, strut stress, and material strength of the cables and struts, respectively;

[0078] (2) Stability conditions: by calculating the stability coefficient Determine whether the stability condition of the brace is met; the stability coefficient The calculation formula is as follows:

[0079]

[0080] In the formula, f is the design value of the compressive strength of the rod, A represents the cross-sectional area, and N represents the axial force of the brace;

[0081] (3) Tension-compression relationship: To determine whether the tension-compression relationship is satisfied, the cables of the structure should be subjected to tension and the struts should be subjected to compression during the calculation process. The expression is as follows:

[0082]

[0083] S5, by comparing the node displacement obtained by neural network fitting with the actual node displacement derived from the tangent stiffness matrix, the loss function value is calculated; the loss function is used to guide the neural network to perform the next iteration, and the final prediction result is output when the maximum number of iterations is reached;

[0084] Further, in step S5, the calculation steps of the loss function are as follows:

[0085] S51, the actual node displacement derived from the tangent stiffness matrix. According to the node equilibrium equation of the structure:

[0086] K T Δ=F(Act)

[0087] Since the tangent stiffness matrix K is determined in step S2 T is a positive definite matrix, so K T It must be reversible, then the actual node displacement is:

[0088]

[0089] Among them, the load F(Act) is the prestress corresponding to the input action.

[0090] The above node displacement equation is the physical information control equation. This method is applicable to other cable-strut tension structures with the same physical information control equation.

[0091] S52, calculate the loss function ε k , and its calculation formula is as follows:

[0092]

[0093] Here, Δ(θ) is the maximum displacement value output in step S3.

[0094] S6. Add the node displacement obtained in this iteration to the original node coordinates to form new node coordinates; repeat S1-S5 until the maximum number of iterations is reached; the actuating force corresponding to the minimum displacement of the structure fitted in all iterations is the required result, and active control of the cable-strut tension structure is achieved based on the final required result.

[0095] Figure 4 The loss function of the active structural control based on PINN during the iterative process of the regulation state is shown. It can be seen from the figure that before 40 iterations, the loss function drops rapidly during the iterative optimization process and then tends to be flat. Since this embodiment is intended to control the displacement of the structure, it can be concluded that with the continuous iteration of PINN, the maximum displacement of the structure is significantly optimized.

[0096] Table 1 shows the load conditions of the structure at different stages. Figure 5The control effect of the Kiewitt-type structure is demonstrated, and the active control based on PINN has a good control effect on the structure. In stage 1, the structure is in a zero displacement state; after the design load is applied to the structure in stage 2, the maximum displacement of the structure is close to 50mm. In stage 3, after the actuator applies a force to the actuator unit, the displacements of all nodes are greatly reduced, and the maximum displacement is reduced by more than 85%. In stage 4, the design live load continues to be applied, and the structure produces a downward deformation of about 100mm. After regulation in stage 5, the displacement is also reduced by about 85%. Therefore, the feasibility of the present invention is verified.

[0097] Table 1 Control stage and load condition table

[0098]

[0099] If the above method 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 technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0100] In another embodiment, a cable-strut tension structure active control device based on a physical information neural network is provided, including a random actuation generation module, a tangent stiffness matrix calculation module, a fitting module, a constraint judgment module, a loss function calculation module and a control output module, wherein the random actuation generation module is used to obtain the actuation of a plurality of randomly generated units; the tangent stiffness matrix calculation module is used to calculate the tangent stiffness matrix of the structure after applying prestress and the actuation, and to determine whether the tangent stiffness matrix is ​​a positive definite matrix, and if not, to re-call the random actuation generation module; the fitting module responds when the tangent stiffness matrix is ​​determined to be a positive definite matrix, and is used to fit the structural displacement using the physical information neural network, and record the maximum displacement of this iterative fitting; the constraint judgment module is based on the fitting module The unit internal force of the corresponding structure is calculated based on the maximum displacement value obtained, and it is determined whether the obtained unit internal force and maximum displacement value meet the constraint conditions. If not, the random actuation generation module is called again; the loss function calculation module responds when the constraint conditions are met, and the loss function value is calculated based on the maximum displacement obtained by the fitting module and the actual node displacement derived from the tangent stiffness matrix of the tangent stiffness matrix calculation module to determine whether the maximum number of iterations has been reached. If not, the fitting module is called again, and the iterative process of the physical information neural network is guided by the loss function value; the control output module responds when the maximum number of iterations is reached, and is used to take the actuation corresponding to the minimum displacement of the structure fitted in all iterations as the final required result, and realize active control of the cable-strut tension structure based on the final required result.

[0101] In another embodiment, an active control device for a cable-strut tension structure based on a physical information neural network is also provided, comprising one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the active control method for a cable-strut tension structure based on a physical information neural network as described above.

[0102] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0103] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks 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 produce 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.

[0104] 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 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0105] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. An active control method for a cable-strut tension structure based on a physical information neural network, characterized in that: The following steps are involved: S1. Obtain the actuation of a number of randomly generated units; S2, calculating the tangent stiffness matrix of the structure after applying the prestress and the actuation, and determining whether the tangent stiffness matrix is ​​a positive definite matrix, if so, executing step S3, if not, returning to step S1; S3, using the physical information neural network to fit the structural displacement, and recording the maximum displacement of this iterative fitting; S4, calculating the unit internal force of the corresponding structure based on the maximum displacement value obtained by fitting in step S3, and judging whether the obtained unit internal force and maximum displacement value meet the constraint conditions, if so, executing step S5, if not, returning to step S1; S5, based on the maximum displacement obtained by fitting in step S3 and the actual node displacement derived from the tangent stiffness matrix in step S2, calculate the loss function value, and determine whether the maximum number of iterations has been reached. If so, execute step S6, if not, return to step S3, and use the loss function value to guide the iteration process of the physical information neural network; S6, taking the actuation corresponding to the minimum displacement of the structure fitted in each iteration as the final required result, and realizing active control of the cable-rod tension structure based on the final required result; The calculation steps of the tangent stiffness matrix include: S21. Calculate the elastic stiffness matrix K of the structure e ; S22. Calculate the geometric stiffness matrix K of the structure p , the geometric stiffness matrix is ​​expressed as the contribution of the prestress of each element; S23, based on the elastic stiffness matrix K e and the geometric stiffness matrix K p The sum of the tangent stiffness matrix K of the structure is obtained T ; The elastic stiffness matrix K e The calculation formula is: In the formula, a k represents the direction cosines of the nodes at both ends of unit k; b represents the number of units; E k represents the elastic modulus of unit k; A k is the cross-sectional area of ​​unit k; l k is the length of unit k; The geometric stiffness matrix K p The calculation formula is: Where, t k is the axial force of unit k; g k is the direction vector of the geometric stiffness matrix of element k; b represents the number of elements; l k is the length of unit k; The constraint conditions include stress conditions, stability conditions and tension-compression relationships, among which: The expression of the stress condition is: In the formula, σ t , σ c , [σ t ]、[σ c ] represent the cable stress, strut stress, cable material strength and strut material strength respectively; The stability condition refers to the stability coefficient satisfying the stability condition of the strut; The expression of the tension-compression relationship is: In the formula, f t and f c They represent the initial tension of the cables and struts of the cable-strut tension structure respectively; The calculation process of the loss function value includes: S51. The actual node displacement is derived from the tangent stiffness matrix. The calculation formula of the actual node displacement Δ is: In the formula, K T represents the tangent stiffness matrix, f(Act) is the prestress corresponding to the input action; S52. Calculate the loss function value ε k , the calculation formula is: Wherein, Δ(θ) is the maximum displacement value obtained by fitting in step S3.

2. The active control method of cable-strut tension structure based on physical information neural network according to claim 1 is characterized in that: The calculation formula of the stability coefficient is: In the formula, is the stability coefficient, f is the design value of the compressive strength of the rod, A represents the cross-sectional area, and N represents the axial force of the brace.

3. An active control device for a cable-rod tension structure based on a physical information neural network, characterized in that: include: A random actuation generation module is used to obtain the actuations of a number of randomly generated units; A tangent stiffness matrix calculation module is used to calculate the tangent stiffness matrix of the structure after applying the prestress and the actuation, and to determine whether the tangent stiffness matrix is ​​a positive definite matrix. If not, the random actuation generation module is called again; The fitting module responds when the tangent stiffness matrix is ​​determined to be a positive definite matrix, and is used to fit the structural displacement using the physical information neural network, and record the maximum displacement of this iterative fitting; The constraint condition judgment module calculates the unit internal force of the corresponding structure based on the maximum displacement value obtained by the fitting module, and judges whether the obtained unit internal force and maximum displacement value meet the constraint condition. If not, the random action moment generation module is called again; A loss function calculation module, which responds when the constraint condition is met, calculates the loss function value based on the maximum displacement value obtained by the fitting module and the actual node displacement derived from the tangent stiffness matrix of the tangent stiffness matrix calculation module, and determines whether the maximum number of iterations has been reached. If not, the fitting module is re-called, and the iterative process of the physical information neural network is guided by the loss function value; A control output module responds when the maximum number of iterations is reached, and is used to take the actuation corresponding to the minimum displacement of the structure fitted in all iterations as the final required result, and realize active control of the cable-rod tension structure based on the final required result; The calculation steps of the tangent stiffness matrix include: S21. Calculate the elastic stiffness matrix K of the structure e ; S22. Calculate the geometric stiffness matrix K of the structure p , the geometric stiffness matrix is ​​expressed as the contribution of the prestress of each element; S23, based on the elastic stiffness matrix K e and the geometric stiffness matrix K p The sum of the tangent stiffness matrix K of the structure is obtained T ; The elastic stiffness matrix K e The calculation formula is: In the formula, a k represents the direction cosines of the nodes at both ends of unit k; b represents the number of units; E k represents the elastic modulus of unit k; A k is the cross-sectional area of ​​unit k; l k is the length of unit k; The geometric stiffness matrix K p The calculation formula is: Where, t k is the axial force of unit k; g k is the direction vector of the geometric stiffness matrix of element k; b represents the number of elements; l k is the length of unit k; The constraint conditions include stress conditions, stability conditions and tension-compression relationships, among which: The expression of the stress condition is: In the formula, σ t , σ c , [σ t ]、[σ c ] represent the cable stress, strut stress, cable material strength and strut material strength respectively; The stability condition refers to the stability coefficient satisfying the stability condition of the strut; The expression of the tension-compression relationship is: In the formula, f t and f c They represent the initial tension of the cables and struts of the cable-strut tension structure respectively; The calculation process of the loss function value includes: S51. The actual node displacement is derived from the tangent stiffness matrix. The calculation formula of the actual node displacement Δ is: In the formula, K T represents the tangent stiffness matrix, F(Act) is the prestress corresponding to the input action; S52. Calculate the loss function value ε k , the calculation formula is: Wherein, Δ(θ) is the maximum displacement value obtained by fitting in step S3.

4. An active control device for a cable-rod tension structure based on a physical information neural network, characterized in that: It comprises one or more processors, a memory and one or more programs stored in the memory, wherein the one or more programs include instructions for executing the active control method of the cable-strut tension structure based on the physical information neural network as described in any one of claims 1-2.

5. A computer-readable storage medium, characterized in that: It includes one or more programs for execution by one or more processors of an electronic device, wherein the one or more programs include instructions for executing the active control method of a cable-strut tension structure based on a physical information neural network as described in any one of claims 1-2.

Images