Dynamic optimization-based real-time simulation method and system for net rack assembly process

Abstract

The invention provides a dynamic optimization-based real-time simulation method and a dynamic optimization-based real-time simulation system for a net rack assembly process, and the method comprises the steps: constructing a parameterized digital twinborn body with a mechanical response capability by obtaining net rack component parameters, material attributes, a connection relationship and construction monitoring data; carrying out assembly process simulation based on real-time data assimilation, and predicting structural response; assembly sequence optimization is realized through deformation energy evaluation and precision constraint; and self-adaptive configuration of the support system is carried out according to a statics equilibrium criterion. Real-time prediction of mechanical response in the whole net rack splicing process, active anti-deviation optimization of a splicing sequence and self-adaptive precise cooperation of temporary support configuration are achieved, so that the construction process is promoted to be converted from static experience driving to dynamic data intelligent decision making, and the splicing precision, the construction safety and the resource utilization efficiency are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent construction, in particular to a net rack assembly process real-time simulation method and system based on dynamic optimization. BACKGROUND

[0002] As the core form of modern large public buildings, the assembly construction quality of large-span space net rack structure is directly related to the safety and durability of the whole project. Because such structures will experience a series of transient systems with continuous changes in geometric shape and mechanical state during assembly, the internal forces and deformations are constantly redistributed, which is extremely easy to cause installation deviation out of limit or even instability risk due to improper assembly sequence or insufficient temporary support, therefore, the precise control of the assembly process is put forward with extremely high requirements.

[0003] In the prior art, the assembly construction of the net rack mainly relies on static three-dimensional design drawings based on building information model, pre-prepared assembly sequence scheme and temporary support arrangement based on engineering experience. Before construction, although finite element software is used for off-line simulation of specific working conditions to evaluate the final forming state or the mechanical properties of a few key steps, the determination of the assembly sequence and the configuration of the temporary support mostly depend on the experience judgment of the engineers, and both of them as key decision elements are usually designed and optimized separately.

[0004] However, the above method has significant technical problems, including that the construction process is unpredictable, the off-line simulation cannot dynamically simulate the mechanical response of all transient structures in the assembly path, and there is a safety blind area; the assembly sequence and temporary support configuration depend on experience and have poor coordination, which may cause non-optimal sequence and excessive deviation, and the support may be too conservative or insufficient; the simulation model is disconnected from the physical construction, and cannot be corrected by real-time monitoring data, which makes the simulation prediction value gradually deviate from the real state, and finally leads to the control of installation precision and process stability being lagging and passive. SUMMARY

[0005] In view of the above actual situation, the present application proposes a net rack assembly process real-time simulation method and system based on dynamic optimization to solve the technical problems of unpredictable assembly process, assembly sequence and temporary support configuration depending on experience and poor coordination, and installation precision and process stability control lagging and passive caused by the disconnection of the simulation model from the physical construction in the prior art.

[0006] A net rack assembly process real-time simulation method based on dynamic optimization, the method comprising the following steps:

[0007] S1, obtaining to-be-processed data, the to-be-processed data including target installation precision, a geometric parameter set of a space truss member, a material mechanics attribute parameter set, a structural connection relationship data set, an initial assembly rule parameter set, a real-time construction monitoring data stream, and a construction constraint parameter set; the target installation precision is an installation deviation tolerance value required in a space truss assembly process; the material mechanics attribute parameter set is a mechanics characteristic parameter set including an elastic modulus and a density; the structural connection relationship data set is a data set defining a connection type and a degree of freedom constraint between members; the initial assembly rule parameter set is a parameter set defining an assembly sequence priority level and a grouping strategy based on design specifications and engineering experience; the real-time construction monitoring data stream is a field construction dynamic data; and the construction constraint parameter set is a construction limitation parameter set including an operation space and equipment capacity;

[0008] S2, performing digital twin body construction processing on the geometric parameter set of the space truss member, the material mechanics attribute parameter set, and the structural connection relationship data set, so as to obtain a parameterized digital twin body; the digital twin body construction processing is to establish a parameterized geometric model by solving an equation set of node geometric constraints, and to generate a digital entity with mechanics response capability in combination with material attribute mapping;

[0009] S3, performing assembly process simulation processing on the parameterized digital twin body, the initial assembly rule parameter set, and the real-time construction monitoring data stream, so as to obtain structure response prediction data; the assembly process simulation processing is response prediction based on dynamic simulation of sequentially activated members in combination with assimilation analysis of the real-time construction monitoring data stream;

[0010] S4, performing assembly sequence optimization processing on the structure response prediction data and the target installation precision, so as to obtain deviation-resistant assembly sequence data; the assembly sequence optimization processing is sequence optimization under precision constraint based on path integral deformation energy evaluation;

[0011] S5, performing support system configuration processing on the deviation-resistant assembly sequence data, the parameterized digital twin body, and the construction constraint parameter set, so as to obtain adaptive support configuration data; the support system configuration processing is transient system inconsistency analysis and constraint satisfaction solving process based on statics equilibrium criterion.

[0012] Further, the S2 step includes the following substeps:

[0013] S201, performing geometric topology reconstruction on the geometric parameter set of the space truss member and the structural connection relationship data set, so as to obtain a parameterized geometric model; the geometric topology reconstruction is an equation set solving algorithm based on geometric constraints between connection nodes;

[0014] S202, performing a mechanical model assignment on the parameterized geometric model and the material mechanics attribute parameter set, so as to obtain a parameterized digital twin, wherein the mechanical model assignment is to map the material attributes to the geometric model to generate a digital entity with mechanical response capability.

[0015] Further, the S3 step includes the following sub-steps:

[0016] S301, performing an assembly process simulation based on the parameterized digital twin and an initial assembly rule parameter set, so as to obtain theoretical structure response data, wherein the assembly process simulation is to generate an assembly sequence according to the initial rules and simulate the assembly steps by sequentially activating the components;

[0017] S302, performing a data assimilation analysis combining the real-time construction monitoring data stream and the theoretical structure response data, so as to obtain structure response prediction data, wherein the data assimilation analysis is to fuse the measured data and the simulation data by a Kalman filtering algorithm to correct the prediction results.

[0018] Further, the S4 step includes the following sub-steps:

[0019] S401, calculating the cumulative deformation energy of different assembly sequences by a deformation energy evaluation algorithm based on the structure response prediction data, so as to obtain a sequence deformation energy evaluation set, wherein the deformation energy evaluation algorithm is a calculation process of path integration of structure strain energy in the assembly process based on the principle of minimum potential energy;

[0020] S402, performing sequence screening on the sequence deformation energy evaluation set in combination with the target installation accuracy, so as to obtain anti-deviation assembly sequence data, wherein the sequence screening is to select an assembly sequence that meets the accuracy requirement and has the minimum cumulative deformation energy.

[0021] Further, the S5 step includes the following sub-steps:

[0022] S501, performing critical state stability calculation based on the anti-deviation assembly sequence data and the parameterized digital twin, so as to obtain critical support configuration data, wherein the critical state stability calculation is a calculation process of determining the minimum number and position of supports required to maintain the stability of the system in the assembly path by performing a rank deficiency analysis on each transient system in the assembly path according to the statics equilibrium criterion;

[0023] S502, performing support scheme generation combining the construction constraint parameter set and the critical support configuration data, so as to obtain adaptive support configuration data, wherein the support scheme generation is processed by constraint satisfaction solving based on a binary decision diagram.

[0024] Further, the geometric topology reconstruction in the S201 is based on the equation set solving algorithm of the geometric constraints between the connected nodes, including node geometric constraint definition processing and constraint equation set solving processing; the mechanical model assignment in the S202 is to map material properties to the geometric model to generate a digital entity with mechanical response capability, including element stiffness matrix assembly processing and mass matrix generation processing.

[0025] Further, the assembly process simulation in the S301 is to simulate the assembly steps by generating an assembly sequence according to initial rules and sequentially activating components, including assembly sequence generation processing and sequential activation simulation processing; the data assimilation analysis in the S302 is to correct the prediction results by fusing measured data and simulation data through Kalman filtering algorithm, including state prediction update processing and measurement update correction processing.

[0026] Further, the deformation energy evaluation algorithm in the S401 is a calculation process based on the principle of minimum potential energy to path-integrate the structural strain energy in the assembly process, including transient strain energy calculation processing and path-integral accumulation processing; the sequence screening in the S402 is to select the assembly sequence that meets the accuracy requirements and has the minimum cumulative deformation energy, including accuracy compliance checking processing and minimum deformation energy sequence selection processing.

[0027] Further, the critical state stability calculation in the S501 is a calculation process that determines the minimum number and position of supports required to maintain system stability by performing rank deficiency analysis on each transient system in the assembly path according to the statics equilibrium criterion, including transient system equilibrium matrix construction processing and system rank deficiency analysis processing; the support scheme generation in the S502 is processed by constraint satisfaction solving based on a binary decision diagram, including constraint condition encoding processing and binary decision diagram solving processing.

[0028] In addition, the application also discloses a real-time simulation system for the assembly process of a net rack based on dynamic optimization, characterized in that the system comprises:

[0029] An acquisition unit is configured to acquire to-be-processed data, the to-be-processed data including target installation precision, a geometric parameter set of a space truss member, a material mechanics attribute parameter set, a structural connection relationship data set, an initial assembly rule parameter set, real-time construction monitoring data flow, and a construction constraint parameter set; the target installation precision is an installation deviation tolerance value required in a space truss assembly process; the material mechanics attribute parameter set is a mechanics characteristic parameter set including an elastic modulus and a density; the structural connection relationship data set is a data set defining a connection type and a degree of freedom constraint between members; the initial assembly rule parameter set is a parameter set defining an assembly sequence priority level and a grouping strategy based on design specifications and engineering experience; the real-time construction monitoring data flow is on-site construction dynamic data; and the construction constraint parameter set is a construction limitation parameter set including an operation space and equipment capacity.

[0030] A digital twin construction unit is configured to perform digital twin body construction processing on the geometric parameter set of the space truss member, the material mechanics attribute parameter set, and the structural connection relationship data set, so as to obtain a parameterized digital twin body; the digital twin body construction processing is to establish a parameterized geometric model by solving an equation set of node geometric constraints, and to generate a digital entity with mechanics response capability in combination with material attribute mapping;

[0031] A simulation analysis unit is configured to perform assembly process simulation processing on the parameterized digital twin body, the initial assembly rule parameter set, and the real-time construction monitoring data flow, so as to obtain structural response prediction data; the assembly process simulation processing is response prediction based on dynamic simulation of sequentially activated members in combination with assimilation analysis of real-time construction monitoring data flow.

[0032] A sequence optimization unit is configured to perform assembly sequence optimization processing on the structural response prediction data and the target installation precision, so as to obtain anti-deviation assembly sequence data; the assembly sequence optimization processing is sequence optimization based on path integral deformation energy evaluation and precision constraint.

[0033] A configuration unit is configured to perform support system configuration processing on the anti-deviation assembly sequence data, the parameterized digital twin body, and the construction constraint parameter set, so as to obtain adaptive support configuration data; the support system configuration processing is transient system inconsistency analysis and constraint satisfaction solving process based on statics equilibrium criterion.

[0034] The method and system for real-time simulation of a space truss assembly process based on digital twin proposed in the application realize real-time and accurate prediction of mechanics response of the whole space truss assembly process, active anti-deviation optimization of an assembly sequence, and adaptive and accurate cooperation of temporary support configuration, thereby achieving a paradigm shift from static experience driving to dynamic data intelligent decision-making in a construction process, and significantly improving assembly precision, construction safety, and resource utilization efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 A method flowchart of a dynamic optimization-based grid assembly process real-time simulation method is provided for the present application.

[0036] Figure 2 A dynamic optimization-based grid assembly process real-time simulation system structure diagram is provided for the present application. DETAILED DESCRIPTION

[0037] The simulation technical route in the embodiments of the present application will be described clearly and completely in combination with the drawings of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0038] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail in combination with the drawings and specific embodiments.

[0039] The features and performances of the present application will be further described in detail in combination with the embodiments. Please refer to the drawings and specific embodiments. Figure 1 As shown in the drawings, a dynamic optimization-based grid assembly process real-time simulation method includes the following steps:

[0040] S1, obtaining the to-be-processed data, the to-be-processed data including a target installation precision, a grid member geometric parameter set, a material mechanics attribute parameter set, a structure connection relationship data set, an initial assembly rule parameter set, a real-time construction monitoring data stream, and a construction constraint parameter set; the target installation precision is an installation deviation tolerance value required in the grid assembly process; the grid member geometric parameter set is a parameter set describing the size and shape of the grid member; the material mechanics attribute parameter set is a mechanics characteristic parameter set including the elastic modulus and the density; the structure connection relationship data set is a data set defining the connection type and the degree of freedom constraint between members; the initial assembly rule parameter set is a parameter set of the assembly sequence priority level and grouping strategy defined based on the design specification and engineering experience; the real-time construction monitoring data stream is the on-site construction dynamic data collected through the Internet of Things sensor; and the construction constraint parameter set is a construction limitation parameter set including the operation space and the equipment capacity;

[0041] In some embodiments, the data to be processed constitutes multi-source information input required for simulation and optimization, which fully covers all elements from structural design, material properties, construction logic to site monitoring and constraint conditions, ensuring that the subsequent processed digital twin has high fidelity and real-time response capability. The target installation precision is the installation deviation tolerance value required during the assembly process of the space truss, which is directly derived from the engineering design specification. Specifically, it specifies the allowable deviation threshold of the node spatial coordinates and the angle deviation limit of the member axis. This precision index will serve as the quality target for the assembly sequence optimization process and the termination criterion for iteration.

[0042] In the present embodiment, the set of space truss member geometric parameters is a parameter set describing the size and shape of the space truss members, specifically including the length, cross-sectional area, and moment of inertia of each rod about different principal axes, as well as the initial design coordinates (X, Y, Z) of all nodes. It should be noted that this data set is the geometric basis for geometric topology reconstruction to generate a parameterized geometric model, and its data format is strictly consistent with the entity information in the building information model or computer-aided design drawings.

[0043] In some embodiments, the set of material mechanical property parameters includes the elastic modulus E and the density p. It should be noted that the elastic modulus and the density dominate different mechanical behaviors of the structure in finite element analysis. The elastic modulus is a core parameter for calculating the structural stiffness matrix K, which defines the linear relationship between stress s and strain e of the material within the elastic range, i.e., s = Ee. This formula is used to calculate the internal force and elastic deformation of the member under load. The density is used to calculate the structural mass matrix M, which, together with the volume V of the member, determines the gravity F g and inertial characteristics, which are related by: g F = pV g where g is the gravitational acceleration vector. It should be noted that in the assembly process simulation, the self-weight of the member is one of the main loads causing structural deformation and internal force during the construction stage, so the density and elastic modulus are two basic material properties that are indispensable for constructing a calculable structural model.

[0044] In the present embodiment, the set of structural connection relationship data is a data set defining the connection type and degree of freedom constraint between members; specifically, this data set is stored in the form of a node and element topology relationship table and a connection attribute matrix. The connection type is defined as hinged or rigid, with hinged releasing all bending moment degrees of freedom and only transmitting axial force and shear force, while rigid coupling all force and moment degrees of freedom. It is used to determine the assembly rules and boundary conditions of the overall structural stiffness matrix K, and is the fundamental basis for determining the geometric invariance of the structural system (i.e., whether it is stable).

[0045] In some implementations, the initial assembly rule parameter set is a set of parameters based on the assembly sequence priority and grouping strategy defined by design specifications and engineering experience. Further, this set is specifically encoded as a set of priority rules, such as prioritizing the assembly of support units and following a symmetrical path extending from the core area to the periphery. These rules, by influencing the activation order of components in the assembly process simulation, directly determine the initial calculation path of the theoretical structural response data, providing an engineeringly reasonable and feasible starting point for subsequent data assimilation and sequence optimization.

[0046] In this embodiment, the real-time construction monitoring data stream is dynamic on-site construction data collected by IoT sensors. Specifically, this data stream includes node displacement monitoring data and component strain monitoring data. The node displacement monitoring data is the three-dimensional coordinate change ΔP of key nodes of the grid structure during assembly, obtained through a navigation satellite system receiver or total station, used to characterize the geometric configuration deviation of the structure. The component strain monitoring data is the micro-strain value ε of the grid structure components under construction loads, collected by strain gauges. measured This is used to reflect the internal force state of the structural members. It should be noted that in the subsequent S302 data assimilation analysis step, the nodal displacement monitoring data is used as the observation vector y. k The position component in the vector y, wherein the strain monitoring data of the component is used as the observation vector y. k The strain components in the data, together with the strain components, constitute the input of the Kalman filter algorithm, used to correct the state vector x in the theoretical structural response data in real time. k (Including displacement and strain), thereby ensuring that the predicted structural response data is consistent with the dynamic behavior of the physical entity.

[0047] In some implementations, the construction constraint parameter set is a set of construction limitation parameters including operating space and equipment capacity; this set quantifies the actual constraints on site into explicit model constraints. Specifically, it includes the working radius R and lifting capacity of the tower crane. The lifting restraints constituted The restricted operating areas defined by the spatial envelope of the construction equipment, and the upper limits on the location and number of temporary support frames. It should be noted that these constraints will be formalized as a set of inequality constraints in the support system configuration process, and will directly participate in the search and determination of feasible solutions, ensuring that the generated adaptive support configuration data has strict field feasibility.

[0048] S2, performing digital twin construction processing on the set of net structure member geometric parameters, the set of material mechanics attribute parameters, and the set of structure connection relationship data, so as to obtain a parameterized digital twin, wherein the digital twin construction processing is to establish a parameterized geometric model by solving an equation group of node geometric constraints, and to generate a digital entity with mechanics response capability in combination with material attribute mapping;

[0049] Specifically, the step includes the following sub-steps:

[0050] S201, performing geometric topology reconstruction on the set of net structure member geometric parameters and the set of structure connection relationship data, so as to obtain a parameterized geometric model, wherein the geometric topology reconstruction is based on an equation group solving algorithm of geometric constraints between connected nodes;

[0051] In some embodiments, the equation group solving algorithm of geometric constraints between connected nodes includes node geometric constraint definition processing and constraint equation group solving processing. The node geometric constraint definition processing is a mathematical definition process of converting member connection into distance constraints and direction constraints between nodes based on the node and element topology relationship and the connection attribute matrix in the set of structure connection relationship data, wherein the distance constraints are directly derived from the member length parameters in the set of net structure member geometric parameters, and the direction constraints indirectly affect the degree of freedom allocation in the subsequent mechanics model through the connection type. The constraint equation group solving processing is a calculation process of solving the geometric constraint equation group by using a numerical iteration algorithm to obtain node coordinate parameters that satisfy all constraint conditions, which ensures that the geometric model is consistent with the design intent. Figure 1 .

[0052] In some embodiments, the node geometric constraint definition processing specifically combines the member length l ij and the initial design coordinates (X i , Y i , Z i ) of the nodes in the set of net structure member geometric parameters with the node and element topology relationship table in the set of structure connection relationship data; specifically, for each member of the connected nodes i and j, a distance constraint equation is defined as follows: wherein l ij is the design length of the member, and the equation forces the Euclidean distance between the nodes i and j to be equal to the member length. Further, the processing also includes definition of a direction constraint, which is indirectly embodied by the connection type flag recorded in the connection attribute matrix in the set of structure connection relationship data; specifically, at the geometric topology level, the direction constraint represents the potential limitation requirement on the relative rotation degree of freedom between nodes, and the connection type (hinged or rigid) defined in the set of structure connection relationship data does not directly change the geometric constraint equation at the geometric topology reconstruction stage, but records the connection type flag t in the connection attribute matrix.ij , where t ij =0 indicates hinged connection, t ij =1 indicates a rigid connection; it should be noted that the connection types (hinged or rigid) defined in the structural connection relationship dataset inject mechanical behavior into the geometric model through the concept of directional constraints during the geometric topology reconstruction stage. The connection type flag is embedded in the parametric geometric model, providing input for element type selection, degree-of-freedom allocation, and stiffness matrix assembly in the subsequent mechanical model assignment steps. Preferably, the cross-sectional area A in the geometric parameter set of the network structure components... ij and the moment of inertia I of the cross section about different principal axes xx,ij with I yy,ij Although it does not directly participate in solving geometric constraints, it is associated and stored as a member property parameter. In the geometric model, it is bound to the connection topology information as an additional parameter, ensuring the complete integration of geometric and mechanical properties.

[0053] In this embodiment, the constraint equations are solved using the Newton-Raphson iterative algorithm; furthermore, this algorithm uses the node coordinate vector X = [X1, Y1, Z1, X1, Y1, Z2, ..., X...]. n ,Y n Z n ] T As an unknown variable, the residual vector R(X) of the constraint equation system is defined as the difference between the actual distance and the design length, specifically in the form: The solution process iteratively updates the node coordinates X. (k+1) =X (k) -J -1 R(X (k) The process is performed, where J is the Jacobian matrix of the residual vector R with respect to the node coordinate vector X, and k represents the iteration number. It should be noted that the elements of the Jacobian matrix are derived from the partial derivatives. Composition, where d ij This is the calculated distance between nodes i and j in the current iteration. The iteration continues until the norm of the residual vector ||R(X)|| is less than the preset tolerance, indicating that the geometric constraints have been satisfied.

[0054] In some implementations, the parametric geometric model is the output of a geometric topology reconstruction; specifically, the model represents the geometry of the grid structure in parametric form, and its model is defined by a node coordinate matrix P and a connection topology matrix C. The node coordinate matrix P is an n×3 matrix, where n is the number of nodes, and each row stores the three-dimensional coordinates (X, Y, C) of a node. i ,Y i Z i), which is the final value after the solving process of the constraint equations. The connection topology matrix C is a m x 4 matrix, where m is the number of members, each row stores the index of a pair of connected nodes (i, j), the connection type flag t ij and the associated cross-sectional area A ij and the inertia parameter. It is noted that the connection type flag t ij directly determines the assembly rule and boundary conditions of the global structure stiffness matrix K in the subsequent steps; for example, in the mechanical model assignment step, for a hinged connection (t ij = 0) element, it is regarded as a truss element, only transferring axial force, its element stiffness matrix only depends on the cross-sectional area A ij and the material elastic modulus, the degree of freedom constraint is released for the bending moment; for a rigid connection (t ij = 1) element, it is regarded as a beam element, coupling axial force, shear force and bending moment, its element stiffness matrix also depends on the cross-sectional inertia moment, the degree of freedom constraint is fully coupled. Further, the parameterized geometric model not only captures the geometric configuration of the space truss, but also integrates the connection type and cross-sectional properties through the connection topology matrix, providing a seamless transition for the mechanical model assignment of step S202, ensuring that the digital twin can accurately simulate the mechanical behavior of hinged or rigid connections, thereby supporting stability analysis and deformation prediction in the assembly process simulation.

[0055] S202, the parameterized geometric model and the material mechanics attribute parameter set are subjected to mechanical model assignment, thereby obtaining a parameterized digital twin, the mechanical model assignment is to map the material properties to the geometric model to generate a digital entity with mechanical response capability;

[0056] In some embodiments, the mechanical model assignment is to map the material properties to the geometric model to generate a digital entity with mechanical response capability includes element stiffness matrix assembly processing and mass matrix generation processing. The element stiffness matrix assembly processing is a calculation process of constructing the element stiffness matrix of each member and assembling it into a global structure stiffness matrix based on the connection topology and cross-sectional geometric properties in the parameterized geometric model, combined with the elastic modulus in the material mechanics attribute parameter set. The mass matrix generation processing is a calculation process of constructing a lumped mass matrix or a consistent mass matrix to represent the inertia characteristics of the structure based on the member geometric information in the parameterized geometric model and the material density.

[0057] In some embodiments, the element stiffness matrix assembly processing takes the connection topology matrix C and the node coordinate matrix P in the parameterized geometric model as input. Specifically, for each member element recorded in the connection topology matrix C, the corresponding element mechanical model is selected according to its connection type flag t ij . When t ij= 0, the element is modeled as a spatial truss element, only subjected to axial forces, whose element stiffness matrix is represented in local coordinate system as: where E is the elastic modulus, A ij is the cross-sectional area of the bar, and l ij is the length of the bar. When t ij = 1, the element is modeled as a spatial beam element, subjected to axial forces, shear forces, bending moments and torsional moments, whose element stiffness matrix is a symmetric matrix containing 12 degrees of freedom, whose elements explicitly depend on the elastic modulus E, the cross-sectional area A ij and the cross-sectional inertia I xx,ij and I yy,ij . It is important to note that all element stiffness matrices need to be transformed from local to global coordinate system by means of a coordinate transformation matrix T, forming the global element stiffness matrix K e = T T k e T. Further, all global element stiffness matrices K e are assembled into the global structure stiffness matrix K according to their nodal degrees of freedom numbering, which is the core of the subsequent simulation process for static or dynamic analysis.

[0058] In the present embodiment, the mass matrix generation process is also based on the parametric geometric model and the material density p. Preferably, a lumped mass matrix is employed, which is a diagonal matrix. For each bar element, its mass m e = pA ij l ij is equally distributed to its two connected nodes. For the spatial truss element, each node gets a mass of m e / 2 and sets a mass value on its corresponding three translational degrees of freedom. For the spatial beam element, each node also gets a mass of m e / 2 and sets a mass value on its corresponding three translational and three rotational degrees of freedom, where the calculation of the rotational inertia is based on a simplified model of the element mass and geometric dimensions. The element lumped mass matrix is then assembled into the global lumped mass matrix M. It is important to note that the global mass matrix is crucial when dynamic effects are considered or modal analysis is performed, which together with the global stiffness matrix defines the basic dynamic properties of the structure.

[0059] In some embodiments, the parameterized digital twin is the output of the mechanical model assigning step, which is a computable structural model with mechanical response capability. The mathematical model of the parameterized digital twin is defined by a four-tuple M = <P, C, K, M>, where P is the node coordinate matrix, which stores the real-time geometric configuration of the structure; C is the connection topology matrix, which stores the connection relationship, type and cross-section attribute of the members; K is the global stiffness matrix, which represents the stiffness characteristics of the structure; and M is the global mass matrix, which represents the inertia characteristics of the structure. It should be noted that the parameterization characteristic is reflected in that the node coordinate matrix P can be updated according to the real-time construction monitoring data stream, while the global stiffness matrix K and the global mass matrix M can dynamically change with the activation or deactivation of the members in the assembly process, thereby realizing the synchronous mapping and real-time simulation of the physical entity construction process.

[0060] S3, performing assembly process simulation processing on the parameterized digital twin, the initial assembly rule parameter set and the real-time construction monitoring data stream, thereby obtaining structure response prediction data, the assembly process simulation processing being response prediction based on dynamic simulation of sequentially activated members combined with assimilation analysis of the real-time construction monitoring data stream;

[0061] Specifically, this step includes the following sub-steps:

[0062] S301, performing assembly process simulation based on the parameterized digital twin and the initial assembly rule parameter set, thereby obtaining theoretical structure response data, the assembly process simulation being a method of simulating assembly steps by generating an assembly sequence according to initial rules and sequentially activating members;

[0063] In some embodiments, the assembly process simulation, which is a method of simulating assembly steps by generating an assembly sequence according to initial rules and sequentially activating members, includes assembly sequence generation processing and sequential activation simulation processing. The assembly sequence generation processing is a calculation process of sorting all members in the parameterized digital twin into an installation order that meets the engineering logic based on the priority rules and grouping strategies encoded in the initial assembly rule parameter set. The sequential activation simulation processing is a numerical simulation process of simulating the mechanical behavior of the members during the gradual installation by dynamically modifying the system matrix of the parameterized digital twin according to the generated assembly sequence, and calculating the structure response quantity at each step.

[0064] In some embodiments, the assembly sequence generation processing takes the initial assembly rule parameter set and the connection topology matrix of the parameterized digital twin as input. Specifically, the processing uses the node connection relationship in C and the member attribute (such as whether it is a support member), combined with the "expanding from the core area to the periphery" strategy defined in the rules, to construct an ordered assembly sequence S = [e1, e2,..., en], where n is the total number of members in the structure. mIt should be noted that although the node coordinate matrix P does not directly participate in the sorting calculation in this process, it is the geometric basis for defining the spatial relationship between the core area and the surrounding space. The spatial strategy in the rules depends on the node spatial distribution defined by P.

[0065] In this embodiment, the sequential activation simulation processing is performed using P, K, and M. This processing takes the assembly sequence S and the parameterized digital twin M = <P, C, K, M> as input, and simulates the assembly process by cyclically executing activation steps. At each simulation step k (corresponding to assembling the k-th component), the processing first activates the sequence unit e. k The nodal coordinate matrix P serves as an invariant geometric reference in this step, used for locating all elements, calculating member lengths, and subsequently calculating nodal displacements U. (k) The reference coordinate system. The processing then extracts the corresponding sub-matrices of the activated elements from the global stiffness matrix and global mass matrix of the parameterized digital twin based on the currently active set of elements, assembling them to form the reduced global stiffness matrix K of the current transient structure. (k) and reducing the overall quality matrix M (k) . Specifically, K (k) and M (k) It is a submatrix directly derived from K and M, which is achieved by introducing a Boolean activation state vector a. (k) To achieve this, in which This indicates that unit i has been activated at step k; otherwise, it is 0. The processing is then based on the current activation state vector a. (k) Assemble the reduced overall stiffness matrix K of the current transient structure (k) and reducing the overall quality matrix M (k) K (k) and M (k) Only by all activated units (i.e. The structure is assembled from the element stiffness matrix and element mass matrix of the unit. Furthermore, for this transient structure, its static equilibrium equations are solved to calculate the load under self-weight. Nodal displacement U under action (k) : Among them, self-weight load The mass is composed of the weights of the activated units, whose mass information is directly derived from the relevant elements on the diagonal of the overall mass matrix M. It should be noted that at each step k, the processing is also based on the solved displacement U. (k) Based on the constitutive relation of the elements, calculate the internal forces N of the activated elements. (k) Stress σ (k)These calculations also rely on the unit stiffness properties derived from K. Specifically, the calculation of internal forces and stresses of activated units based on the derived displacements is a standard post-processing procedure following the finite element theory. Specifically, for each activated unit e, its internal force calculation relies on its stiffness properties and its deformation. The nodal displacement vector of unit e in the global coordinate system is extracted This vector is a subset of the global displacement vector U (k) , containing the displacement degrees of freedom of all nodes of the unit. For a spatial truss unit, each node has 3 translational degrees of freedom; for a spatial beam unit, each node has 6 degrees of freedom (3 translational, 3 rotational). The nodal displacement is then converted to the local coordinate system of the unit using the unit coordinate transformation matrix T defined in step S202, resulting in the local coordinate system displacement In the local coordinate system, the calculation of internal forces (including axial force, shear force, bending moment, etc.) of the unit becomes straightforward. The nodal force vector of the unit in the local coordinate system is obtained by where k e is the stiffness matrix of the unit in the local coordinate system (as defined in S202 or ). The elements in the vector directly correspond to the nodal internal forces of the unit. For a truss unit, its axial force N is simply the relevant element in ; for a beam unit, contains the axial force, two shear forces, and the bending moment. Finally, the unit stress σ (k) is calculated based on the unit internal forces. For a truss unit, its axial stress σ is calculated from the axial force N and the cross-sectional area A: For a beam unit, its bending normal stress σ b is calculated from the bending moment M and the cross-sectional moment of inertia I, for example where y is the distance to the neutral axis, and the total stress is the sum of the axial stress and the bending normal stress.

[0066] In some embodiments, the theoretical structural response data is a result set output from the sequential activation simulation process, which fully records the theoretical mechanical behavior of the entire assembly process. This data is represented by an ordered set of time-step response quantities , where k represents the sequence number of the assembly simulation step, which is a counter starting from 1, recording the number of assembly operations, and m represents the total number of component units in the parameterized digital twin, i.e., the total length of the assembly sequence S. It should be noted that each response quantity U theory , N (k) , σ (k) in D (k)All of these are calculated by taking P, K, and M in a specific activation state k within the parameterized digital twin. Therefore, the theoretical structural response data is essentially the dynamic behavior output of the parameterized digital twin under a given assembly sequence S. It provides benchmark simulation data for subsequent S302 data assimilation analysis, allowing for comparison and fusion with real-time monitoring data.

[0067] S302, combine the real-time construction monitoring data stream and theoretical structural response data to perform data assimilation analysis, thereby obtaining structural response prediction data. The data assimilation analysis is to use the Kalman filter algorithm to fuse measured data and simulation data to correct the prediction results.

[0068] In some implementations, the data assimilation analysis involves fusing measured and simulated data using a Kalman filter algorithm to correct prediction results. This includes state prediction update processing and measurement update correction processing. The state prediction update processing is a mathematical process that uses theoretical structural response data and its uncertainty information to predict the mean and covariance of the structural state at the next moment. The measurement update correction processing is a computational process that weights and fuses real-time construction monitoring data streams with state prediction values ​​to obtain the optimal state estimate.

[0069] In some implementations, the state prediction update process is based on the theoretical structural response data generated in step S301. Specifically, the theoretical state of assembly simulation step k is defined as the state vector x. k This vector contains the displacements U of all nodes at that moment. (k) The theoretical structural response data D theory The predicted value x for the state is provided. k|k-1 It should be noted that the dynamic evolution of the state vector is described by a linear stochastic state-space model: x k =F k x k-1 +w k , where F k It is the state transition matrix, which is composed of the mechanical properties of the parameterized digital twin (such as K). (k) and M (k) After discretization, the derivation revealed the physical evolution law of the structural state from step k-1 to step k; w k It is process noise, with a mean of zero and a covariance matrix of Q. k The Gaussian distribution is used to quantify the uncertainty of the model forecast. The processing simultaneously forecasts the state covariance matrix P. k|k-1： The covariance matrix characterizes the predicted state value x. k|k-1 The degree of uncertainty.

[0070] In the present embodiment, the measurement update correction process takes the real-time construction monitoring data stream as the observation input. The observation vector y k is constructed from the nodal displacement monitoring data ΔP and the member strain monitoring data ε measured . The relationship between the observation and state values is defined by the observation equation: y k = H k x k + v k , where H k is the observation matrix that maps the state vector x k (containing displacements) to the observation vector y k , and v k is the observation noise, which is assumed to be Gaussian with zero mean and covariance matrix R k to quantify the uncertainty in the sensor measurements. In some embodiments, the calculation of the Kalman gain K k is essentially a process of determining the optimal weight between the theoretical prediction and the measured data. Specifically, the Kalman gain is equal to the current state prediction uncertainty P k|k-1 mapped by the observation matrix H k , multiplied by the inverse of the overall uncertainty (the mapped prediction uncertainty plus the observation noise uncertainty R k ). This calculation indicates that when the prediction uncertainty is much larger than the observation uncertainty, the Kalman gain tends to 1, and the correction process will trust the measured data more; conversely, when the observation uncertainty is larger, the Kalman gain tends to 0, and the correction process will trust the theoretical prediction more.

[0071] In the present embodiment, the state vector is optimally corrected using the calculated Kalman gain K k . The correction is achieved by multiplying the Kalman gain K k with the prediction residual (i.e., the difference between the measured value y k and the predicted observation value H k x k|k-1 ), generating a correction term, and adding this correction term to the original state prediction value x k|k-1 to obtain the optimal state estimate after data assimilation. This step incorporates the new information contained in the measured data into the predicted state at the optimal scale. Further, after the state vector is corrected, the uncertainty of the state needs to be evaluated and updated. The update of the covariance matrix is achieved by subtracting the product of the Kalman gain K k and the observation matrix H k from the identity matrix, and multiplying the result with the original state prediction covariance matrix P k|k-1 . This operation reflects the principle that after the measured data is incorporated, the uncertainty of the system state is reduced, and the updated covariance matrix P k|kIt quantitatively reflects the corrected optimal state estimate x. k|k The new, reduced confidence level.

[0072] In some implementations, the structural response prediction data is the output of a data assimilation analysis step; it is a more accurate state estimation sequence that integrates theoretical simulations and field measurements. This data uses a modified state vector sequence. This is used to represent the data. It should be noted that this is different from the theoretical structural response data D. theory In comparison, D prediction Each state vector x in k|k All of them include correction information based on actual monitoring data, and their estimation error is determined by the covariance matrix P. k|k Quantitative description. Preferably, the structural response prediction data provides input for the deformation energy assessment algorithm in the subsequent S401 step because its displacement field U (k) (included in x) k|k The path integral of the cumulative deformation energy calculated based on this more accurately reflects the actual deformation state of the physical entity, thus making the path integral of the cumulative deformation energy more representative of the mechanical properties of the actual assembly process.

[0073] S4, perform assembly sequence optimization processing on the structural response prediction data and target installation accuracy to obtain anti-deviation assembly sequence data. The assembly sequence optimization processing is based on the deformation energy assessment of path integral and sequence optimization under accuracy constraints.

[0074] Specifically, this step includes the following sub-steps:

[0075] S401, Based on the structural response prediction data, the cumulative deformation energy of different assembly sequences is calculated using a deformation energy evaluation algorithm to obtain a sequence deformation energy evaluation set. The deformation energy evaluation algorithm is a process of calculating the structural strain energy during assembly based on the principle of minimum potential energy.

[0076] In some implementations, the deformation energy assessment algorithm, based on the principle of minimum potential energy, calculates the structural strain energy during assembly through path integration. This process includes transient strain energy calculation and path integral accumulation. The transient strain energy calculation is a physical quantity calculation process that, for each transient structural state in the assembly path, calculates the total structural strain energy in that state based on its displacement field and internal force field. The path integral accumulation process is a weighted summation of the strain energies of all transient states throughout the entire assembly sequence, representing the integral process of the overall deformation energy accumulation caused by the assembly sequence.

[0077] In some implementations, the transient strain energy calculation process takes the structural response prediction data output from step S302 as input. Specifically, for the k-th transient state in the assembly sequence, from the state vector x...k|k extracted from the data assimilation corrected nodal displacement vector U (k) According to the principle of elasticity, the total strain energy V (k) of the transient structure is composed of the sum of strain energy contributions from all activated components. For linear elastic materials, the strain energy of each component is calculated through its nodal displacement and element stiffness matrix. In the present embodiment, the transient strain energy calculation process calculates the total elastic strain energy stored in the transient structure for each particular state in the assembly sequence. The strain energy of each component is calculated through its nodal displacement and element stiffness matrix. Specifically, for the kth assembly state, the total strain energy V (k) of the transient structure is calculated based on the principle of minimum potential energy. The calculation uses the data assimilation corrected nodal displacement vector U (k) and the global stiffness matrix K (k) of the current transient structure. The core of the calculation is to take the transpose of the displacement vector U (k) , multiply it with the global stiffness matrix K (k) , multiply the result with the displacement vector U (k) itself, and multiply the obtained result by one half. It is noted that the global stiffness matrix characterizes the global stiffness properties of the structure provided by the current set of activated components, while the displacement vector characterizes the deformation pattern of the structure under the load. This step quantifies the total elastic deformation energy accumulated in the structure in order to reach the current deformation pattern. The energy is a scalar, whose value is always positive, and the larger the value is, the higher the deformation level or the more significant the internal force state of the transient structure.

[0078] In some embodiments, the path integral accumulation process evaluates the cumulative mechanical effects induced by a complete assembly sequence S j . Instead of simply adding the strain energies of the states, the process treats the discrete assembly steps as samples of a continuous process, numerically integrates the transient strain energy sequence, and accurately characterizes the continuous accumulation of deformation energy along the assembly path. Specifically, the accumulated deformation energy E total (S j ) is calculated by the following path integral formula: where s is the continuous assembly progress variable, V(s) is the strain energy function along the assembly path, and m is the total number of component elements in the parameterized digital twin. Since the assembly process is carried out in discrete steps, the integral is solved using numerical methods. Preferably, the composite trapezoidal rule is used for discrete integral approximation: where Δs is the integral step, which takes value 1 in the case of uniform assembly step. It is noted that this integral method can more accurately reflect the cumulative effect of strain energy on the entire assembly path than simple summation, because it considers the linear change of strain energy between adjacent states, rather than regarding it as a step function. Further, the calculation process needs to be repeated for a plurality of different candidate assembly sequences S j (where j = 1, 2,..., n represents the index of different sequences) repeatedly. It is noted that an optimal assembly sequence should make the deformation energy path integral experienced by the structure in the entire assembly process (path) minimum, which conforms to the physical law that the structure tends to a low-energy stable state as guided by the principle of minimum potential energy.

[0079] In some embodiments, the sequence deformation energy evaluation set is the output of the deformation energy evaluation algorithm step, which provides a quantitative decision basis for subsequent sequence screening. The set is represented by a set containing all candidate sequences and their corresponding cumulative deformation energies , where n is the total number of candidate assembly sequences evaluated. The sequence deformation energy evaluation set ε establishes a clear mapping relationship between the assembly sequence S j and a mechanical performance indicator E total (S j ). Preferably, the smaller the cumulative deformation energy E total (S j ), the smaller the forced deformation of the structure caused by the assembly path, and the lower the residual internal force in the component, thereby theoretically leading to higher installation accuracy and better structure forming state, which lays a direct numerical comparison basis for selecting the anti-deviation assembly sequence that meets the target installation accuracy in the S402 step.

[0080] S402, sequence screening of the sequence deformation energy evaluation set is performed in combination with the target installation accuracy, so as to obtain anti-deviation assembly sequence data, and the sequence screening is to select an assembly sequence that meets the accuracy requirement and has the smallest cumulative deformation energy;

[0081] In some embodiments, the sequence screening that selects an assembly sequence that meets the accuracy requirement and has the smallest cumulative deformation energy includes accuracy compliance verification processing and minimum deformation energy sequence selection processing. The accuracy compliance verification processing is a logical judgment process of pre-evaluating the installation deviation of all candidate sequences in the sequence deformation energy evaluation set based on the target installation accuracy indicator, and screening a feasible sequence set that meets the accuracy constraint. The minimum deformation energy sequence selection processing is a decision process of selecting an assembly sequence with the smallest cumulative deformation energy from the feasible sequence set that passes the accuracy verification as the optimal solution.

[0082] In some embodiments, the accuracy compliance checking process takes as input the target installation accuracy defined in step S1 and the sequence deformation energy evaluation set generated in step S401. Specifically, the target installation accuracy defines the allowable deviation threshold AL max and the angular deviation limit Aq of the member axis max . For each candidate sequence S j in the sequence deformation energy evaluation set, the process computes its maximum nodal displacement ||U prediction || (m) and the angular change of the key member by querying the nodal displacement vector U (m) in the structure response prediction data D ∞ generated in step S302 at its corresponding final assembly state (i.e. k = m). The accuracy compliance condition is defined by the following inequality set: ||U (m) || ∞ ≤ AL max , max |q i - q i,0 | ≤ Aq max , where q i and q i,0 denote the final azimuth angle and the initial design azimuth angle of member i respectively. It is noted that only the candidate sequence S j that satisfies all the accuracy inequality constraints simultaneously is retained, constituting the feasible sequence set F.

[0083] In the present embodiment, the minimum deformation energy sequence selection process takes as input the feasible sequence set F and its corresponding cumulative deformation energy data. The process seeks and selects the sequence that makes the cumulative deformation energy E total achieve the minimum value among all the feasible sequences that satisfy the accuracy requirement as the optimal solution. This selection process is formulated as the following optimization problem: where S * is the sought anti-deviation assembly sequence. It is noted that this process guarantees that the selected assembly sequence not only satisfies the final installation accuracy requirement, but also causes the minimum structural forced deformation and internal cumulative strain energy in the entire assembly path, which conforms to the optimal control idea guided by the principle of minimum potential energy, and reduces the generation of assembly process deviation from the mechanical essence.

[0084] In some embodiments, the anti-deviation assembly sequence data is the output of the sequence screening step, which provides the core assembly scheme for the subsequent support system configuration. This data is represented by a two-tuple (S * , E total (S * )), where S * is the selected optimal assembly sequence, and E total (Sj ) is the minimum cumulative deformation energy value corresponding to the sequence. It should be noted that the anti-bias assembly sequence data S * is the optimal solution obtained after digital twin simulation, data assimilation correction, deformation energy evaluation and multi-objective screening, which has the ability to resist assembly bias in theory. Based on the sequence, the stability of each transient system in the assembly path is analyzed, and the minimum support requirement matched therewith is derived, forming a complete and coordinated intelligent construction scheme from assembly sequence to support configuration.

[0085] S5, support system configuration processing is performed on the anti-bias assembly sequence data, parameterized digital twin and construction constraint parameter set, so as to obtain adaptive support configuration data, and the support system configuration processing is a transient system rank defect analysis and constraint satisfaction solving process based on statics equilibrium criterion.

[0086] Specifically, this step includes the following sub-steps:

[0087] S501, critical state stability calculation is performed based on the anti-bias assembly sequence data and the parameterized digital twin, so as to obtain critical support configuration data, and the critical state stability calculation is a calculation process of performing rank defect analysis on each transient system in the assembly path according to the statics equilibrium criterion, to determine the minimum number and position of supports required to maintain the stability of the system;

[0088] In some embodiments, the critical state stability calculation is a calculation process of performing rank defect analysis on each transient system in the assembly path according to the statics equilibrium criterion, to determine the minimum number and position of supports required to maintain the stability of the system, including transient system balance matrix construction processing and system rank defect analysis processing. The transient system balance matrix construction processing is an algebraic modeling process of establishing the coefficient matrix corresponding to the force balance equation for each assembly transient state defined by the anti-bias assembly sequence data. The system rank defect analysis processing is a calculation process of identifying the geometric variability of the system and determining the minimum constraint requirement by analyzing the rank and degree of freedom relationship of the balance matrix.

[0089] In some embodiments, the transient system balance matrix construction processing takes the anti-bias assembly sequence data and the parameterized digital twin obtained in the S402 step as input. Specifically, for the kth transient state in the assembly sequence, the substructure composed of all activated components in this state is extracted. According to the topological relationship C (k) and the node coordinates P (k) , the static force balance equation is established: B (k) t (k) = f (k) , where t (k) is the internal force vector of the component, f (k) is the node load vector, and B(k) i.e. the equilibrium matrix. In the present embodiment, the parametric digital twin M = <P, C, K, M> is the basis of the mechanical analysis for the critical state stability computation, the member internal force vector t (k) and the nodal load vector f (k) are directly derived from the mechanical state computation of the parametric digital twin, where the nodal load vector f (k) is computed by the product of the global mass matrix M in the parametric digital twin and the gravity acceleration vector g, i.e. f (k) = M (k) g, while the member internal force vector t (k) is obtained by multiplying the global stiffness matrix K (k) in the parametric digital twin with the nodal displacement vector U (k) provided by the structural response prediction data, i.e. t (k) = K (k) U (k) , thus ensuring that all the mechanical quantities required for the equilibrium matrix construction are consistent with the real-time state of the digital twin. It is noted that the equilibrium matrix B (k) has a dimension of d n x d m , where d n is the total number of degrees of freedom of the transient system, and d m is the number of activated members. Each element of the equilibrium matrix is determined by the nodal coordinates and member direction cosines, which specifically reflect the contribution of each member internal force to the nodal equilibrium.

[0090] In the present embodiment, the system rank deficiency analysis process is based on the linear algebra analysis of the equilibrium matrix B (k) . The process first computes the rank r (k) = rank(B (k) ) of the equilibrium matrix, and then computes the rank deficiency d (k) of the system: d (k) = d n - r (k) , where d n is the total number of degrees of freedom of the system. It is noted that the rank deficiency d (k) directly represents the number of geometric variable degrees of freedom of the system. When d (k) > 0, it indicates that the transient system has d (k)An individual mechanism displacement mode requires the same number of independent constraints to be applied to make it a geometrically invariant system. Further, by analyzing the zero space basis vectors of the equilibrium matrix, the specific mechanism displacement mode form and its corresponding nodal displacement direction can be identified, thereby determining the type and spatial location of the degrees of freedom that need to be constrained. It is worth mentioning that the core of the system deficiency analysis process based on linear algebra analysis of the equilibrium matrix is to determine the static characteristics of the transient structure system by calculating the rank of the matrix, where the rank of the matrix rank(B (k) ) represents the number of independent equations in the system of equilibrium equations, and the total degrees of freedom d n of the system represents the number of independent variables required to describe the spatial configuration of the system; when the number of independent equations is less than the number of degrees of freedom, i.e., d n -r (k) > 0, it indicates that there are d (k) unconstrained, possible rigid body displacement or mechanism motion modes in the system, i.e., geometrically variable; the purpose of this process is to accurately identify the number and spatial direction of these unstable degrees of freedom, and the key information obtained, the rank deficiency d (k) , and the mechanism displacement mode ， obtained by analyzing the zero space of the equilibrium matrix, directly indicates the minimum number and specific orientation of external support constraints that need to be applied to eliminate geometric variability, thereby providing a basis for determining the minimum support requirement.

[0091] In some embodiments, the critical support configuration data is the output of the critical state stability calculation step, which defines the minimum support requirement to ensure the stability of the entire assembly process. This data is represented by a support requirement set S support = {(k, N k , D k )} k∈K , where k is the assembly step number, K is the index set of all unstable transient states, N k is the node number set that needs to be supported in the kth transient state, and D k is the degree of freedom direction set that needs to be constrained for the corresponding node. It should be noted that each node support is equivalent to applying the corresponding constraint, eliminating one or more mechanism degrees of freedom. Preferably, the critical support configuration data ensures that each transient state from the start of assembly to the formation of the structure is geometrically stable, providing the theoretical minimum support requirement for the support scheme generation of the subsequent S502 step, so that the support system configured meets the stability requirement and avoids resource waste caused by excessive support.

[0092] At S502, support scheme generation is performed based on the construction constraint parameter set and the critical support configuration data, to obtain adaptive support configuration data. The support scheme generation is processed by constraint satisfaction solving based on a binary decision diagram.

[0093] In some embodiments, the support scheme generation processed by constraint satisfaction solving based on a binary decision diagram includes constraint condition encoding processing and binary decision diagram solving processing. The constraint condition encoding processing is a formal description process of uniformly transforming spatial restrictions and resource restrictions defined in the construction constraint parameter set and minimum support requirements defined in the critical support configuration data into a Boolean logic expression form. The binary decision diagram solving processing is a combination optimization process of constructing a binary decision diagram data structure based on the encoded Boolean logic expression, and solving a feasible support arrangement scheme that satisfies all constraint conditions by a graph traversal algorithm.

[0094] In this step, the mechanical requirements for structural stability are unified with the actual limitations of the construction site into a framework for solving. The minimum support requirements (i.e. the degrees of freedom and positions that must be constrained) determined by the critical support configuration data are translated into a set of Boolean logic assertions that must be true, while the physical limitations defined by the construction constraint parameter set, such as the operation space forbidden zone and the upper limit of the support resource quantity, are translated into another set of Boolean logic constraints. These constraints are collectively coded into a Boolean function, each satisfying solution (output true) of which represents a feasible support arrangement scheme that meets the stability requirements and does not violate any site conditions. By constructing a binary decision diagram of this Boolean function, a standardized data structure, the graph theory properties of which are used to efficiently traverse all possible combinations of variables (i.e. all possible support arrangement schemes) and quickly locate the feasible solutions that meet all the constraints, and even select the optimal solution with the least number of supports or the lowest comprehensive cost. The adaptive support configuration data enables the precise, dynamic and optimized configuration of the support system, providing a support scheme that precisely matches the optimal assembly sequence and is allocated on demand. This scheme ensures that each transient structure in the assembly path is geometrically stable at the technical level, fundamentally eliminating the risk of instability and minimizing the amount of temporary support material, installation and removal time, and related mechanical shifts, thereby reducing construction costs. At the same time, the scheme fully complies with the crane operating range, spatial obstacles and other real conditions of the site, and has the ability to be implemented directly on the ground. The generated instruction set not only specifies which node positions in three-dimensional space need to be arranged with supports, but also clearly defines the specific directions that each support needs to be constrained (such as only constraining Z-direction displacement, or simultaneously constraining X, Y-direction translation), and the dynamic use plan of these supports in the assembly sequence, i.e. when each set of supports should be installed and removed after the structure has the ability to be self-stable. Thus, the grid assembly process is organized as an organic process of coordinated evolution of the support system and the main structure assembly, with efficient turnover of resources, ensuring the highest safety and precision standards while maximizing construction efficiency.

[0095] In some embodiments, the constraint condition coding process takes the construction constraint parameter set defined in step S1 and the critical support configuration data output in step S501 as inputs. Specifically, a Boolean decision variable x i is defined for each potential support installation position, where x i = 1 indicates that a support is placed at this position, and x i = 0 indicates that no support is placed. The critical support configuration data is coded as a set of hard constraints that must be met: for each unstable transient state k ∈ K, the required support node set N k and the constraint direction set D kare transformed into logical "and" and "or" relations, ensuring that all necessary degrees of freedom constraints are satisfied in the state. Meanwhile, the set of construction constraints parameters are encoded as restrictions: the operation space forbidden zones are transformed into assertions that the corresponding decision variables must be 0, the support number upper limit are transformed into global cardinality constraints The crane working range constraints are then transformed into logical implication relations between the decision variables of the specific regions.

[0096] In the present embodiment, the binary decision diagram solving process receives the set of encoded Boolean logic expressions as input. A binary decision diagram is a compact, canonical representation of a Boolean function, which represents all variable assignment combinations and their corresponding function outputs (true / false) through a directed acyclic graph. The process first combines all the Boolean constraints (including stability requirements and construction restrictions) into an overall Boolean function F(x) through logical "and" operations: F(x) = C stability ∧ C exclusion ∧ C resource ∧..., where C stability represents the stability constraints encoded from the critical support configuration data, C exclusion represents the operation forbidden zone constraints, C resource represents the resource upper limit constraints. The process then constructs a binary decision diagram representing F(x). It is noted that due to the canonical nature of binary decision diagrams, each path from the root node to the terminal node 1 (representing true) corresponds to a feasible support arrangement x that satisfies all the constraints. Further, the process finds an optimal or satisfactory solution among all feasible solutions through an efficient graph traversal algorithm. Preferably, the optimization objective can be defined as minimizing the total support number∑ i x i or minimizing the total cost of support arrangement, which is achieved by evaluating the objective function values corresponding to different paths during the traversal process.

[0097] In some embodiments, the adaptive support configuration data is the output of the support scheme generation step, which represents the optimal support arrangement scheme that guarantees the stability of the entire assembly process under the premise of satisfying all construction constraints. This data is represented by a triple A = <X * , L * , T * >, where X * is a Boolean vector representing the decision of whether to place a support at each potential support location (1 for placement, 0 for no placement); L * records the spatial coordinates of all selected support locations and the required degrees of freedom directions; and T * defines the timing strategy for support usage, i.e., the placement sequence S *The timing of activation and dismantling is carefully controlled to ensure that it always covers all critical unstable transient states. It should be noted that the adaptive support configuration data A is an optimized assembly sequence combined with a perfectly matched dynamic support scheme, forming a complete intelligent operation plan that can directly guide on-site construction, ensuring consistency in accuracy, stability, and construction feasibility during the space frame assembly process.

[0098] It is worth mentioning that, after obtaining feasible solutions by traversing the binary decision graph, the binary decision graph solving process further selects an optimal solution from all feasible solutions. The optimal solution is directly represented as a specific Boolean decision variable assignment combination X that makes the overall Boolean function F(x) true. * =(x1,x2,...,x n The vector X * This is the optimal deployment decision. It should be noted that X... * This is the direct output of the binary decision graph solution process. It records which potential support locations are selected under the premise of satisfying all stability and construction constraints. In this embodiment, the optimal layout decision X is obtained. * Subsequently, the support scheme generation process performs data mapping to generate L. * and T * The position and orientation data L * The generation is achieved by X * The decision variable x takes a value of 1. i The coordinates are mapped back to their corresponding physical locations, and this coordinate information is pre-stored in the available support point library defined by the construction constraint parameter set. Simultaneously, for each selected location, the required degree of freedom direction D is determined by the critical support configuration data S. support Specifically, the decision involves integrating all the degree-of-freedom constraints on the node at that location under all transient states k, and taking their union to form a complete set of constraint directions. The time-series policy data T... * The generation is based on the critical support configuration data S support The process records the correspondence between each transient state k and the support requirement; the process assigns an activation opportunity to each selected support (i.e., the assembly step k when the support is first required). start ) and a removal timing (i.e., the assembly step k where the support is no longer needed in all subsequent assembly steps). end This forms a dynamic support usage plan. In some implementations, the adaptive support configuration data is an integration of the above data processing procedures. It should be noted that X... * L is derived from the optimization results of solving the binary decision graph. * By analyzing X * Perform physical mapping and combine S supportmechanical requirements in the process of assembling the grid structure, T * is generated by completely analyzing the time sequence of S support .

[0099] Based on the description of the above-mentioned dynamic optimization-based grid structure assembly process real-time simulation method embodiment, the application also discloses a dynamic optimization-based grid structure assembly process real-time simulation system. The dynamic optimization-based grid structure assembly process real-time simulation system can be a computer program (including program code) running the above-mentioned dynamic optimization-based grid structure assembly process real-time simulation method. Please refer to FIG. 10, which shows that the dynamic optimization-based grid structure assembly process real-time simulation system can run the following units: Figure 2

[0100] The acquisition unit 110 is configured to acquire to-be-processed data, wherein the to-be-processed data includes a target installation precision, a grid structure member geometric parameter set, a material mechanical property parameter set, a structure connection relationship data set, an initial assembly rule parameter set, a real-time construction monitoring data stream, and a construction constraint parameter set. The target installation precision is an installation deviation tolerance value required in the grid structure assembly process. The material mechanical property parameter set is a mechanical property parameter set including an elastic modulus and a density. The structure connection relationship data set is a data set defining the connection type and the degree of freedom constraint between members. The initial assembly rule parameter set is a parameter set of assembly sequence priority and grouping strategy defined based on design specifications and engineering experience. The real-time construction monitoring data stream is on-site construction dynamic data. The construction constraint parameter set is a construction limitation parameter set including an operation space and equipment capacity.

[0101] The digital twin construction unit 120 is configured to perform digital twin body construction processing on the grid structure member geometric parameter set, the material mechanical property parameter set, and the structure connection relationship data set, so as to obtain a parameterized digital twin body. The digital twin body construction processing is to establish a parameterized geometric model by solving an equation group of node geometric constraints, and to generate a digital entity with mechanical response capability in combination with material property mapping.

[0102] The simulation analysis unit 130 is configured to perform assembly process simulation processing on the parameterized digital twin body, the initial assembly rule parameter set, and the real-time construction monitoring data stream, so as to obtain structure response prediction data. The assembly process simulation processing is response prediction based on dynamic simulation of sequentially activated members in combination with assimilation analysis of the real-time construction monitoring data stream.

[0103] The sequence optimization unit 140 is configured to perform assembly sequence optimization processing on the structure response prediction data and the target installation precision, so as to obtain anti-deviation assembly sequence data. The assembly sequence optimization processing is sequence optimization based on path integral deformation energy evaluation and precision constraint.

[0104] ​The configuration unit 150 is configured to perform support system configuration processing on the anti-deviation assembly sequence data, the parameterized digital twin, and the construction constraint parameter set, so as to obtain adaptive support configuration data, wherein the support system configuration processing is based on a transient system rank deficiency analysis and a constraint satisfaction solving process according to a statics balance criterion.

[0105] The above merely illustrates the preferred embodiments of the present application, and it should be understood that the present application is not limited to the forms disclosed herein, and should not be considered as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concepts described herein, by the above teachings or related art or knowledge. Any modification and change made by those skilled in the art without departing from the spirit and scope of the present application shall fall within the protection scope of the claims of the present application.

Claims

1. A real-time simulation method for the assembly process of a space frame based on dynamic optimization, characterized in that, The method includes the following steps: S1. Obtain the data to be processed, which includes target installation accuracy, set of geometric parameters for grid structure components, set of material mechanical properties parameters, set of structural connection relationship data, set of initial assembly rule parameters, real-time construction monitoring data stream, and set of construction constraint parameters. The target installation accuracy is the required installation deviation tolerance value during grid assembly. The set of material mechanical properties parameters is a set of mechanical characteristic parameters including elastic modulus and density. The set of structural connection relationship data is a set of data defining the connection type and degree of freedom constraints between components. The set of initial assembly rule parameters is a set of parameters based on the assembly sequence priority level and grouping strategy defined by design specifications and engineering experience. The real-time construction monitoring data stream is dynamic data of on-site construction. The set of construction constraint parameters is a set of construction restriction parameters including operating space and equipment capacity. S2, perform digital twin construction processing on the set of geometric parameters of the network components, the set of material mechanical property parameters, and the set of structural connection relationships to obtain a parameterized digital twin. The digital twin construction process establishes a parameterized geometric model by solving the equations of the node geometric constraints and generates a digital entity with mechanical response capability by combining material property mapping. S3, perform assembly process simulation processing on the parameterized digital twin, the initial assembly rule parameter set and the real-time construction monitoring data stream to obtain structural response prediction data. The assembly process simulation processing is a response prediction based on the dynamic simulation of sequentially activated components combined with the assimilation analysis of the real-time construction monitoring data stream. S4, perform assembly sequence optimization processing on the structural response prediction data and target installation accuracy to obtain anti-deviation assembly sequence data. The assembly sequence optimization processing is based on the deformation energy assessment of path integral and sequence optimization under accuracy constraints. S5, the anti-deviation assembly sequence data, parameterized digital twin and construction constraint parameter set are processed for support system configuration to obtain adaptive support configuration data. The support system configuration processing is a transient system rank deficiency analysis and constraint satisfaction solution process based on static equilibrium criteria.

2. The real-time simulation method for the assembly process of a space frame based on dynamic optimization according to claim 1, characterized in that, Step S2 includes the following sub-steps: S201, Perform geometric topology reconstruction on the set of geometric parameters of the network components and the set of structural connection relationships to obtain a parametric geometric model. The geometric topology reconstruction is based on an algorithm for solving a system of equations that solves the geometric constraints between the connection nodes. S202, apply a mechanical model to the parametric geometric model and the set of material mechanical property parameters to obtain a parametric digital twin. The application of the mechanical model involves mapping material properties to the geometric model to generate a digital entity with mechanical response capabilities.

3. The real-time simulation method for the assembly process of a space frame based on dynamic optimization according to claim 1, characterized in that, Step S3 includes the following sub-steps: S301, based on the parameterized digital twin and the initial assembly rule parameter set, the assembly process is simulated to obtain theoretical structural response data. The assembly process simulation is to generate an assembly sequence according to the initial rules and use the method of sequentially activating components to simulate the assembly steps. S302, combining the real-time construction monitoring data stream and theoretical structural response data to perform data assimilation analysis, thereby obtaining structural response prediction data. The data assimilation analysis is performed by fusing measured data and simulation data using the Kalman filter algorithm to correct the prediction results.

4. A real-time simulation method for the assembly process of a space frame based on dynamic optimization according to any one of claims 1-3, characterized in that, Step S4 includes the following sub-steps: S401, based on the structural response prediction data, the cumulative deformation energy of different assembly sequences is calculated by the deformation energy evaluation algorithm to obtain the sequence deformation energy evaluation set. The deformation energy evaluation algorithm is a process of calculating the structural strain energy during the assembly process by means of path integral based on the principle of minimum potential energy. S402, the sequence deformation energy evaluation set is screened in combination with the target installation accuracy to obtain anti-deviation assembly sequence data. The sequence screening is to select the assembly sequence that meets the accuracy requirements and has the smallest cumulative deformation energy.

5. The real-time simulation method for the assembly process of a space frame based on dynamic optimization according to claim 4, characterized in that, Step S5 includes the following sub-steps: S501, based on the anti-deviation assembly sequence data and parameterized digital twin, critical state stability calculation is performed to obtain critical support configuration data. The critical state stability calculation is a calculation process that performs rank deficiency analysis on each transient system in the assembly path according to the static equilibrium criterion to determine the minimum number and location of supports required to maintain system stability. S502, combining the construction constraint parameter set and critical support configuration data, a support scheme is generated to obtain adaptive support configuration data. The support scheme generation is processed by solving the constraint satisfaction based on a binary decision graph.

6. The real-time simulation method for the assembly process of a space frame based on dynamic optimization according to claim 2, characterized in that, The geometric topology reconstruction in S201 is based on the equation solving algorithm of geometric constraints between connecting nodes, including node geometric constraint definition processing and constraint equation solving processing; the mechanical model assignment in S202 is to map material properties to the geometric model to generate a digital entity with mechanical response capability, including element stiffness matrix assembly processing and mass matrix generation processing.

7. The real-time simulation method for the assembly process of a space frame based on dynamic optimization according to claim 3, characterized in that, The assembly process simulation in S301 is to generate an assembly sequence based on initial rules and use the method of sequential activation of components to simulate the assembly steps, including assembly sequence generation processing and sequential activation simulation processing. The data assimilation analysis in S302 is to fuse measured data and simulation data using the Kalman filter algorithm to correct the prediction results, including state prediction update processing and measurement update correction processing.

8. The real-time simulation method for the assembly process of a space frame based on dynamic optimization according to claim 4, characterized in that, The deformation energy assessment algorithm in S401 is based on the principle of minimum potential energy and calculates the structural strain energy during the assembly process by performing path integral calculation, including transient strain energy calculation and path integral accumulation. The sequence screening in S402 is to select the assembly sequence that meets the accuracy requirements and has the minimum cumulative deformation energy, including accuracy compliance verification processing and minimum deformation energy sequence selection processing.

9. The real-time simulation method for the assembly process of a space frame based on dynamic optimization according to claim 5, characterized in that, The critical state stability calculation in S501 is based on the static equilibrium criterion to perform rank deficiency analysis on each transient system in the assembly path to determine the minimum number and location of supports required to maintain system stability. The calculation process includes the construction of the transient system equilibrium matrix and the rank deficiency analysis of the system. The support scheme generation in S502 is processed by the constraint satisfaction solution based on the binary decision graph, including constraint condition encoding and binary decision graph solving.

10. A real-time simulation system for the assembly process of a space frame based on dynamic optimization, characterized in that, The system includes: The acquisition unit is used to acquire data to be processed, which includes target installation accuracy, geometric parameter set of grid structure components, material mechanical property parameter set, structural connection relationship dataset, initial assembly rule parameter set, real-time construction monitoring data stream, and construction constraint parameter set. The target installation accuracy is the required installation deviation tolerance value during grid assembly. The material mechanical property parameter set is a set of mechanical property parameters including elastic modulus and density. The structural connection relationship dataset is a data set defining the connection type and degree of freedom constraints between components. The initial assembly rule parameter set is a set of parameters based on the assembly sequence priority level and grouping strategy defined by design specifications and engineering experience. The real-time construction monitoring data stream is dynamic data of on-site construction. The construction constraint parameter set is a set of construction restriction parameters including operating space and equipment capacity. The digital twin construction unit is used to perform digital twin construction processing on the set of geometric parameters of the network components, the set of material mechanical property parameters, and the set of structural connection relationships to obtain a parameterized digital twin. The digital twin construction process establishes a parameterized geometric model by solving the equations of the node geometric constraints and generates a digital entity with mechanical response capability by combining material property mapping. The simulation analysis unit is used to perform assembly process simulation processing on the parameterized digital twin, the initial assembly rule parameter set and the real-time construction monitoring data stream, so as to obtain structural response prediction data. The assembly process simulation processing is based on response prediction by dynamic simulation of sequentially activated components combined with assimilation analysis of real-time construction monitoring data stream. The sequence optimization unit is used to perform assembly sequence optimization processing on the structural response prediction data and target installation accuracy to obtain anti-deviation assembly sequence data. The assembly sequence optimization processing is based on the deformation energy assessment of path integral and sequence optimization under accuracy constraints. The configuration unit is used to perform support system configuration processing on the anti-deviation assembly sequence data, parameterized digital twin, and construction constraint parameter set to obtain adaptive support configuration data. The support system configuration processing is a transient system rank deficiency analysis and constraint satisfaction solution process based on the static equilibrium criterion.

Images