A steel pipe concrete special-shaped column joint rigidity prediction and optimization system

Abstract

The application relates to the technical field of building structure engineering, and particularly discloses a steel pipe concrete special-shaped column node stiffness prediction and optimization system, which comprises the following steps: obtaining the geometric size, material performance and structural details of a node to be optimized, establishing a three-dimensional entity simulation copy, extracting an initial stiffness value from a bending moment-rotation angle curve obtained by solving based on applied boundary conditions and displacement loads, and constructing a mapping relationship copy between structural parameters and stiffness based on the initial stiffness values under multiple different structural parameters; taking the mapping relationship copy as an evaluation basis, automatically searching for an optimal structural parameter combination and outputting a search result in a preset feasible region with the maximum stiffness or the minimum material consumption as a target; and the application realizes rapid prediction of node stiffness and automatic optimization of structural parameters.

Description

Technical Field

[0001] This invention relates to the field of building structure engineering technology, specifically to a system for predicting and optimizing the stiffness of irregular-shaped steel-concrete composite column joints. Background Technology

[0002] Steel-concrete composite irregular column structures have been increasingly widely used in high-rise residential and public buildings due to their advantages such as flexible cross-sectional shapes, high load-bearing capacity, and convenient building layout. As key components in a structure that transmit internal forces and coordinate deformation, the rotational stiffness of joints directly affects the overall structure's lateral displacement response, internal force distribution, and even seismic performance.

[0003] Existing technologies suffer from the problem that the stress mechanism of steel-concrete composite irregular column joints is complex, making it difficult to balance the accuracy of stiffness calculation and optimization efficiency. Specifically, when performing multi-objective optimization within the feasible domain of structural parameters, traditional high-precision simulation methods cannot be directly embedded into optimization algorithms that require a large number of iterative evaluations due to the excessive time required for a single calculation. As a result, the design results often remain at the local optimal solution based on experience or limited examples rather than the global optimal solution. Summary of the Invention

[0004] The purpose of this invention is to provide a system for predicting and optimizing the stiffness of irregular-shaped steel-concrete composite column joints, so as to solve the problems mentioned above.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A system for predicting and optimizing the stiffness of irregular-shaped steel-concrete composite column joints includes:

[0007] The node construction parameter acquisition module is used to acquire the construction parameters of the steel-concrete composite irregular column node to be optimized. The construction parameters include: geometric dimension information, material performance parameters and construction detail features.

[0008] The node stress simulation replica construction module establishes a three-dimensional solid simulation replica that reflects the actual stress characteristics of the node based on the structural parameters of the steel-concrete composite irregular column node to be optimized.

[0009] The node stiffness numerical extraction module applies preset boundary conditions and load modes to the three-dimensional solid simulation copy, obtains the moment-rotation response curve of the node by solving, and extracts the initial stiffness value characterizing the node's ability to resist rotational deformation from the moment-rotation response curve.

[0010] The stiffness mapping relationship construction module extracts the initial stiffness values ​​from the nodal stiffness values ​​under multiple different construction parameters, and constructs a copy of the mapping relationship between the construction parameters and the initial stiffness values ​​through interpolation or fitting.

[0011] The optimized parameter search module uses the mapping relationship copy as the evaluation basis. Within the preset feasible domain of construction parameters, it aims to maximize the initial stiffness of the node or minimize the amount of material used in the node, while satisfying the preset bearing capacity and construction constraints. It automatically searches for the optimal combination of node construction parameters.

[0012] The optimal construction parameter output module is used to output the optimal combination of node construction parameters obtained by the optimization parameter search module.

[0013] As a further aspect of the present invention: the construction process of the three-dimensional solid simulation copy is as follows:

[0014] Based on the geometric dimensions, three-dimensional solids of the steel pipe and the core concrete were created respectively, and an initial interference was set between the inner wall of the steel pipe and the outer surface of the concrete to simulate the clamping effect of the steel pipe on the concrete.

[0015] The actual stress-strain curve of steel in the material performance parameters is discretized into multiple linear data points, and the uniaxial compressive stress-strain curve of concrete is corrected according to the constrained constitutive model considering the confining pressure effect, and then assigned to the corresponding entities in the form of data.

[0016] Based on the location and size of the weld in the structural details, a weld entity is established at the corresponding part of the steel pipe and given the same material properties as the base material but with a higher yield strength;

[0017] The geometric entity is meshed using eight-node hexahedral elements, and local seed points are set in the weld seam and node core area to refine the mesh, ultimately generating a 3D solid simulation copy.

[0018] As a further aspect of the present invention: the modification of the uniaxial compressive stress-strain curve of concrete according to a constrained constitutive model considering the confining pressure effect specifically includes:

[0019] Extract the constraint stress values ​​of the steel pipe on the core concrete in each direction from the three-dimensional solid simulation copy, and calculate their average value as the constant confining pressure value.

[0020] Based on a constant confining pressure, the peak stress point of the uniaxial compressive stress-strain curve of concrete is raised, and the strain value corresponding to the peak stress point is shifted in the direction of increasing strain.

[0021] Based on the magnitude of the constant confining pressure, the absolute value of the slope of the descending segment of the uniaxial compressive stress-strain curve of concrete is reduced, and the end of the descending segment of the stress-strain curve is extended towards the direction of greater strain.

[0022] The new stress-strain data points generated after peak stress elevation, peak strain shift, and slope adjustment of the descending segment are used to replace the corresponding positions on the original uniaxial compressive stress-strain curve of concrete, forming a corrected stress-strain relationship that takes into account the confining pressure effect.

[0023] As a further aspect of the present invention: the calculation process of the initial stiffness value is as follows:

[0024] Rigid coupling surfaces are established at the top and bottom faces of the column in the 3D solid simulation replica. Fixed constraints are applied to the top face of the column, and axial pressure pointing towards the column body is applied to the bottom face of the column to simulate the actual stress state.

[0025] A rigid loading surface is established on the free end face of the beam in the 3D solid simulation copy, and a monotonically increasing displacement load perpendicular to the beam axis is applied to the loading surface until the node reaches its ultimate bearing capacity.

[0026] During the monotonically increasing displacement load application process, the reaction force value of the beam end loading surface and the relative rotation displacement of the beam and column at the corresponding time are recorded under each incremental step. The reaction force value is multiplied by the beam length to convert it into a bending moment value, and the relative rotation displacement is converted into a rotation value to generate a bending moment-rotation data point set.

[0027] Linear regression fitting was performed on the data points in the initial loading stage of the moment-rotation data point set, and the slope value of the fitted straight line was taken as the initial stiffness value characterizing the node's ability to resist rotational deformation.

[0028] As a further aspect of the present invention: the establishment of a rigid loading surface on the free end face of the beam in the three-dimensional solid simulation copy specifically includes:

[0029] Set the loading rate of the displacement load so that the loading rate gradually decreases according to the preset attenuation coefficient as the cumulative displacement of the loading surface increases, so as to avoid node instability due to excessive loading.

[0030] After applying displacement in each incremental step, the reaction force change on the loading surface and the energy balance residual in the solution process are monitored simultaneously. Only when the reaction force increment and the energy balance residual are both lower than their respective preset thresholds is the current incremental step determined to have converged and the next step is initiated.

[0031] When the reaction force on the loading surface is detected to show a continuous downward trend and the decrease exceeds the preset limit, it is determined that the node has reached its ultimate bearing capacity, and the application of displacement load is immediately terminated.

[0032] As a further aspect of the present invention: the process of constructing the mapping relationship copy is as follows:

[0033] Multiple different construction parameters and their corresponding initial stiffness values ​​are grouped according to the construction parameter type. Within each group, a discrete data point set is generated with the initial stiffness value as the ordinate and the construction parameter as the abscissa.

[0034] For each set of discrete data points, calculate the rate of change of the slope of the line segment between each adjacent data point, and mark the interval where the rate of change of the slope exceeds the preset abrupt change threshold as the parameter sensitive area;

[0035] In the parameter-sensitive region, cubic spline interpolation is used to insert new interpolation points between adjacent data points, while in the non-sensitive region, linear interpolation is used to supplement interpolation points.

[0036] All existing data points and newly added interpolation points are aggregated to construct a copy of the continuous and smooth mapping relationship between the construction parameters and the initial stiffness values.

[0037] As a further aspect of the present invention: the process of obtaining the optimal combination of node construction parameters specifically includes:

[0038] Multiple initial combinations of construction parameters are randomly generated within the feasible domain of the construction parameters, and each initial combination is assigned a current evaluation value;

[0039] All initial construction parameter combinations are sorted according to their current evaluation values, and the top-scoring combinations are selected as the preferred combinations based on the preset optimization objective.

[0040] Random perturbations are applied to the construction parameters of the preferred combination to generate new construction parameter combinations, and the new combinations are merged with the remaining initial combinations to form a new generation of parameter combination population;

[0041] Repeat the process of sorting, selecting and applying random perturbation until the variation of the optimal current evaluation value for multiple consecutive generations is less than the preset fluctuation limit. Then, output the construction parameter combination corresponding to this time as the optimal node construction parameter combination.

[0042] As a further aspect of the present invention: the process of constructing the new generation parameter combination population is as follows:

[0043] S1, calculate the adaptive perturbation amplitude coefficient based on the current iteration number, so that the adaptive perturbation amplitude coefficient gradually decays to a preset minimum value as the iteration number increases;

[0044] S2, for each dimension construction parameter in each preferred combination, multiply the current value of the construction parameter by an adaptive perturbation amplitude coefficient and by a perturbation direction factor randomly generated in the range of negative one to positive one to obtain the perturbation offset of the corresponding dimension parameter.

[0045] S3, add the current value of each parameter to the disturbance offset to obtain the initial new parameter value, and determine whether the corresponding initial new parameter value falls within the preset feasible region of construction parameters. If it exceeds the feasible region, adjust it to the nearest feasible region boundary value.

[0046] S4. Summarize the new parameter values ​​after adjusting all dimensional parameters to form a new construction parameter combination. Repeat the process from S1 to S3 for each preferred combination to generate a specified number of new combinations. Then merge all the new combinations with the remaining unselected initial combinations to form a new generation of parameter combination population.

[0047] The beneficial effects of this invention are:

[0048] (1) By establishing a three-dimensional solid simulation copy and extracting the initial stiffness value, this invention avoids the high cost and long cycle of repeatedly making test pieces in traditional physical experiments; on this basis, a copy of the mapping relationship between construction parameters and stiffness is constructed, so that designers can quickly evaluate the nodal stiffness under any combination of parameters without repeatedly calling the finite element software, thereby improving the efficiency of the multi-scheme comparison stage.

[0049] (2) The present invention uses an iterative search method with adaptive perturbation within the feasible domain of structural parameters to find the optimal combination of parameters, which can take into account both the maximization of node stiffness and the minimization of material usage, while satisfying the bearing capacity and structural constraints. Compared with the traditional design method that relies on empirical formulas or local trial calculations, the results obtained are closer to the global optimal solution, effectively avoiding insufficient safety redundancy or material waste caused by improper parameter selection. Attached Figure Description

[0050] The invention will now be further described with reference to the accompanying drawings.

[0051] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Please see Figure 1 As shown, this invention is a system for predicting and optimizing the stiffness of irregular-shaped steel-concrete composite column joints, comprising:

[0054] The node construction parameter acquisition module is used to obtain the construction parameters of the steel-concrete composite irregular column node to be optimized. These construction parameters include: geometric dimension information, material performance parameters, and detailed construction features, specifically including:

[0055] The node construction parameter acquisition module operates on the irregular steel-concrete composite column node to be optimized. This node consists of a steel pipe and core concrete poured inside the steel pipe, and its cross-section is L-shaped, T-shaped, or cross-shaped. This module is used to acquire the construction parameters necessary for subsequent simulation analysis of the node. The specific process is as follows:

[0056] The first step is to obtain the geometric dimensions of the node. This information includes the overall external dimensions of the steel pipe, namely the length, width, and thickness of each leg of the irregular column; the thickness of the steel pipe wall; the cross-sectional dimensions of the core concrete; and the cross-sectional height, width, and length of the beam connected to the node. This geometric dimension information can be obtained directly by reading the original construction drawing design documents of the node, or, when analyzing existing structural nodes, by measuring the actual components on-site using a laser rangefinder and vernier calipers.

[0057] The second step is to obtain the material performance parameters of the node. These parameters include the steel grade used to manufacture the steel pipe and its corresponding elastic modulus, yield strength, tensile strength, and Poisson's ratio; the strength grade of the core concrete used for pouring and its corresponding standard value of axial compressive strength and elastic modulus; and the strength grade of the welds used to connect the various components of the node. These material performance parameters are obtained by consulting the corresponding performance data in the national standard material database, based on the material standards specified in the construction drawing design documents, or by performing uniaxial tensile and compression tests on specimens taken from the same batch of materials.

[0058] The third step is to obtain the structural details of the node. These details include the splicing method of the steel pipes at the junctions of the various limbs (whether cold bending or steel plate welding is used); the interface treatment method between the core concrete and the inner wall of the steel pipe (whether it is poured in natural contact or pre-applied with a bonding medium); whether stiffening plates are installed in the core area of ​​the node, and if so, their specific location, thickness, and opening details; and the weld type at the connection between the beam and the steel pipe column wall (whether fillet weld or grooved full penetration weld is used), and the corresponding weld leg dimensions. These structural details are obtained directly by consulting the node's construction drawings and welding procedure qualification report.

[0059] In the nodal stress simulation replica construction module, based on the structural parameters of the steel-concrete composite irregular column node to be optimized, a three-dimensional solid simulation replica reflecting the actual stress characteristics of the node is established, specifically including:

[0060] Based on the aforementioned collected construction parameters, a three-dimensional solid simulation copy that can reflect the mechanical behavior of the node to be optimized under actual stress conditions is established. The specific construction process is carried out in the following steps.

[0061] The first step is to create three-dimensional solids of the steel pipe and the core concrete based on the aforementioned geometric dimensions. Specifically, the steps are as follows: First, based on the wall thickness and outer contour dimensions of each limb of the steel pipe, a three-dimensional geometry of the steel pipe is created using a solid extrusion method. Second, based on the clear dimensions of the inner cavity of the steel pipe, a three-dimensional geometry of the core concrete is created. After completing the geometric modeling, an initial interference fit is set between the inner wall of the steel pipe and the outer surface of the concrete. This initial interference fit is set to five-thousandths of the radial thickness of the steel pipe, to simulate the radial compression, i.e., the clamping effect, caused by the shrinkage of the steel pipe on the core concrete during actual stress.

[0062] The second step involves assigning the material performance parameters to the aforementioned geometric entities in the form of discrete data points. Specifically, for steel, the actual stress-strain curve obtained from uniaxial tensile tests is used to select 20 data points at equal intervals according to strain increment control, forming a multilinear data point set, which is then assigned to the steel pipe and weld entities. Secondly, for concrete, the uniaxial compressive stress-strain curve is modified according to a constrained constitutive model considering confining pressure effects, resulting in a modified stress-strain data point set, which is then assigned to the core concrete entity.

[0063] The third step is to establish the weld entity based on the weld location and size in the structural details. Specifically, firstly, based on the obtained weld type and weld leg size, independent three-dimensional weld entities are established at the connection between the steel pipe and the beam flange, as well as at the splicing points of the various steel pipe segments, using scanning or stretching methods. Secondly, the yield strength of the weld material is obtained by increasing the yield strength corresponding to the steel grade in the material performance parameters by 15%, and this value, along with other elastic parameters, is assigned to the weld entity to reflect the characteristic that the strength of the weld is higher than that of the base material.

[0064] The fourth step involves meshing the aforementioned geometric entities, including the steel pipes, core concrete, and welds. Specifically, all geometric entities are meshed using eight-node hexahedral elements, with an initial element side length of 20 mm. Next, local seed points are set in the pre-marked node core area and the weld entity and its surrounding 10 mm region. The spacing between these seed points is reduced to 5 mm to refine the mesh in this area, ultimately generating a 3D solid simulation copy for subsequent numerical solutions.

[0065] The process of modifying the uniaxial compressive stress-strain curve of concrete according to the constrained constitutive model that takes into account the confining pressure effect is carried out in the following steps.

[0066] The first step is to apply a reference axial pressure equivalent to 20% of the material's standard strength to a 3D solid simulation copy that has completed the initial mesh generation but has not yet undergone constitutive replacement. The pre-solution is then performed to extract the compressive stress values ​​of all nodes on the core concrete surface in the direction perpendicular to the surface. The arithmetic mean of these compressive stress values ​​is calculated, and this mean is used as the constant confining pressure value of the node under this stress state.

[0067] The second step is to correct the peak stress point of the uniaxial compressive stress-strain curve of concrete based on the constant confining pressure value. Specifically, the steps are as follows: First, consult the national standard for the axial compressive strength of concrete of this strength grade and use this standard value as the original peak stress. Second, calculate the increase in peak stress according to the ratio of 8% for each MPa constant confining pressure value, and add the original peak stress to the increase to obtain the corrected peak stress. Third, multiply the strain value corresponding to the original peak stress point by an amplification factor, which is set to 1.0 plus 0.005 multiplied by the constant confining pressure value, to obtain the corrected peak strain.

[0068] The third step is to adjust the descending segment of the uniaxial compressive stress-strain curve of concrete based on the constant confining pressure value. Specifically, the steps are as follows: First, obtain the point corresponding to 85% of the peak stress on the descending segment of the original curve as a reference point. The absolute value of the slope of the line connecting this reference point and the peak stress point is taken as the original descending segment slope. Second, set a slope reduction factor, which is equal to 1.0 divided by 1.0 plus 0.1 multiplied by the constant confining pressure value. Multiply the original descending segment slope by this slope reduction factor to obtain the corrected descending segment slope. Third, multiply the strain value corresponding to the end of the descending segment of the original curve by an extension factor, which is set to 1.5 plus 0.1 multiplied by the constant confining pressure value, to obtain the corrected end strain value. This strain value is used as the termination point of the corrected curve.

[0069] The fourth step is to generate corrected concrete stress-strain data. Specifically, first, using the corrected peak stress and peak strain as control points, and following the corrected descending slope, new descending segment data points are generated by extending from the peak strain point towards the corrected final strain point. Second, keeping the original ascending segment data points unchanged, the end of the ascending segment is smoothly connected to the new peak stress point. Finally, the original ascending segment data points, the corrected peak stress-strain points, and the newly generated descending segment data points are summarized, and these are used to replace the corresponding data on the original uniaxial compressive stress-strain curve of the concrete, forming a complete set of corrected stress-strain data points used to characterize the core concrete entity.

[0070] In the node stiffness numerical extraction module, preset boundary conditions and load modes are applied to the 3D solid simulation copy. The moment-rotation response curves of the nodes are obtained by solving, and the initial stiffness values ​​characterizing the nodes' ability to resist rotational deformation are extracted from the moment-rotation response curves. Specifically, this includes:

[0071] The process of extracting the node stiffness values ​​is used to obtain initial stiffness values ​​characterizing the node's ability to resist rotational deformation from the constructed 3D solid simulation copy. This process is performed according to the following steps.

[0072] The first step is to establish rigid coupling surfaces on the top and bottom end faces of the column in the 3D solid simulation replica. Specifically, motion coupling constraints are established on all nodes of the top end face, creating a rigid connection between each node and a reference point located at the centroid of the end face; similarly, rigid coupling surfaces are established on the bottom end face. Then, fixed constraints are applied to the reference point on the top end face, restricting its translational and rotational degrees of freedom in three directions; an axial pressure is applied to the reference point on the bottom end face, pointing towards the column. This pressure is 20% of the design value of the column's axial force under the design load, simulating the initial axial load borne by the column in actual stress conditions.

[0073] The second step involves establishing a rigid loading surface on the free end face of the beam in the 3D solid simulation replica. Specifically, motion coupling constraints are established on all nodes of the beam end face, rigidly connecting them to a reference point located at the centroid of the beam end face. Then, a monotonically increasing displacement load perpendicular to the beam axis is applied to this reference point. The direction of the displacement load is set to cause the beam to bend outwards towards the column, and the displacement load is continuously applied until the node reaches its ultimate bearing capacity. The specific method of applying this displacement load will be executed in subsequent detailed steps.

[0074] The third step involves recording the response data for each incremental step during the application of the monotonically increasing displacement load. Specifically, after each incremental step, the reaction force value of the reference point on the rigid loading surface at the beam end is extracted from the solution results. Simultaneously, the vertical relative displacement of this reference point relative to the column top reference point is extracted and recorded as the beam end vertical displacement. The beam end vertical displacement is then multiplied by the calculated length of the beam (i.e., the vertical distance from the beam end loading surface to the column side surface) to obtain the beam-column relative rotation angle value corresponding to that incremental step, in radians. The beam end reaction force value is multiplied by the calculated length of the beam to obtain the beam end bending moment value corresponding to that incremental step. The bending moment and rotation angle values ​​calculated for each incremental step are stored as a data point in a data point set.

[0075] The fourth step is to extract the initial stiffness value from the moment-rotation data set. Specifically, this involves selecting data points from the origin up to 30% of the node's ultimate moment as the initial linear segment. A linear regression method is then used to fit a straight line to these data points, and the slope of this line is taken as the initial stiffness value characterizing the node's resistance to rotational deformation. The slope is calculated using the following formula:

[0076] ;

[0077] in, This is the initial stiffness value, in kilonewton-meters per radian. This represents the total number of data points in the initial linear segment. For the first The angle value of each data point, in radians; For the first The bending moment values ​​for each data point are expressed in kilonewton-meters.

[0078] The specific process of applying a monotonically increasing displacement load on the rigid loading surface at the beam end is as follows: In order to prevent the solution from diverging or the nodes from becoming unstable due to excessively rapid loading, and to make the loading process closer to the actual quasi-static stress, the process of applying the displacement load is controlled by the following detailed steps.

[0079] The first step is to set the loading rate of the displacement load. The loading rate is defined as the incremental displacement value applied in each increment step. This loading rate is then gradually reduced according to a preset attenuation coefficient as the cumulative displacement of the loading surface increases. Specifically, the initial loading rate is first set. The value is one millimeter per second; let the current cumulative displacement be... Units are in millimeters; preset attenuation coefficient The value is set to 0.02 per millimeter. Therefore, the loading rate for the current increment step is... Calculate using the following formula:

[0080] ;

[0081] in, It is a natural constant; The initial loading rate is 1 millimeter per second; The attenuation coefficient is 0.02 per millimeter. This represents the cumulative displacement applied so far, in millimeters. This formula ensures that the loading rate decreases exponentially with increasing displacement. When the cumulative displacement reaches 50 mm, the loading rate decreases to approximately 37% of the initial value, thus allowing for smaller increments to ensure solution accuracy as the ultimate bearing capacity approaches.

[0082] The second step involves simultaneously monitoring the change in reaction force on the loading surface and the energy balance residual during the solution process after applying displacement in each incremental step. Specifically, after convergence of each incremental step, the reaction force value at the end of that incremental step is extracted. and the reaction force value at the end of the previous increment step Calculate the absolute value of the reaction force increment. Simultaneously, the energy balance residual for this incremental step is extracted from the solver output; that is, the absolute value of the difference between external force work and internal force work. The reaction force increment threshold is set to 5% of the current reaction force value, and the energy balance residual threshold is set to 1% of the current total external force work. The current incremental step is considered convergent only if both the absolute value of the reaction force increment and the energy balance residual are below their respective thresholds, allowing the next incremental calculation to proceed. Otherwise, the displacement increment of the current incremental step is halved and re-solved until the convergence condition is met or the incremental step size is less than a preset minimum value (one-thousandth of the initial step size), at which point the analysis terminates.

[0083] The third step involves determining that the node has reached its ultimate bearing capacity when a continuous downward trend in the reaction force on the loading surface is detected, and the decrease exceeds a preset limit. Specifically, the reaction force value for each incremental step is recorded. When the reaction force value for three consecutive incremental steps is less than the value of the previous incremental step, the reaction force is considered to have entered a decreasing phase. The reaction force value at the starting point of this decreasing phase is taken as the peak reaction force. Subsequently, when the reaction force value in subsequent incremental steps decreases to below 85% of the peak reaction force, the node is determined to have lost its ability to continue bearing load, the application of displacement load is immediately terminated, and the state at this point is recorded as the ultimate bearing capacity state. This method accurately captures the ultimate bearing capacity of the node, avoiding simulation result distortion due to overloading.

[0084] In the stiffness mapping relationship construction module, based on the initial stiffness values ​​extracted from the nodal stiffness values ​​under multiple different construction parameters, a copy of the mapping relationship between the construction parameters and the initial stiffness values ​​is constructed through interpolation or fitting. Specifically, this includes:

[0085] The process of constructing the stiffness mapping relationship involves establishing a continuous mapping copy that reflects the correspondence between construction parameters and initial stiffness based on the initial stiffness values ​​calculated under multiple different construction parameters. This copy will serve as the evaluation basis for subsequent parameter searches. This process is performed according to the following steps.

[0086] The first step involves grouping multiple different structural parameters and their corresponding initial stiffness values ​​according to their structural parameter types and generating discrete data point sets. Specifically, all structural parameters involved in the analysis are first divided into three categories based on their physical properties: the first category consists of geometric dimensional parameters, including steel pipe wall thickness, length and thickness of each limb, and stiffening plate thickness; the second category consists of material property parameters, including steel yield strength and concrete axial compressive strength; and the third category consists of structural detail parameters, including weld dimensions and initial interference values. Then, for each category of structural parameters, discrete data points are generated in a two-dimensional coordinate system, with the specific value of that parameter as the x-axis and the corresponding initial stiffness value as the y-axis. Each group of data points constitutes an independent discrete data point set.

[0087] The second step involves calculating the slope change rate of the line segments between adjacent data points for each set of discrete data points, and marking the intervals containing data points whose slope change rate exceeds a preset abrupt change threshold as parameter-sensitive regions. Specifically, for a set of discrete points containing m data points, sort them by their x-coordinate values ​​from smallest to largest, and connect adjacent data points sequentially to form a total of m-1 line segments. For the a-th line segment, its slope value is equal to the difference between the y-coordinates of its two endpoints divided by the difference between their x-coordinates. Starting from the second line segment, calculate the absolute value of the difference between the slope of the current line segment and the slope of the previous line segment, then divide this absolute value by the absolute value of the slope of the previous line segment. The result is defined as the slope change rate at the a-th interval. The preset abrupt change threshold is set to 50%. Starting from the second data point and iterating backwards, if the slope change rate at the a-th interval exceeds 50%, the x-coordinate interval between the a-th data point and the (a+1)-th data point is marked as a parameter-sensitive region. Simultaneously, the slope change rates of the intervals between the first and second data points, and between the second-to-last and last data points, are compared only with their adjacent single intervals and judged and marked according to the aforementioned threshold. All intervals that do not meet the condition of a slope change rate exceeding 50% are marked as non-sensitive regions.

[0088] The third step involves using cubic spline interpolation within the parameter-sensitive region and linear interpolation within the non-sensitive region to supplement new interpolation points. Specifically, for each marked parameter-sensitive region, the left and right endpoints are the original data point A and B, respectively. Take the adjacent original data point to the left (denoted as point L) and the adjacent original data point to the right (denoted as point R), along with points A and B, for a total of four points, as control points for spline interpolation. Based on the x and y coordinates of these four points, construct a cubic polynomial curve passing through points L, A, B, and R, requiring the curve to have continuous first derivatives at points A and B. On this curve, starting from the x-coordinate of point A and ending at the x-coordinate of point B, insert new interpolation points at equal intervals, with the insertion interval set to 1 / 5 of the minimum interval between the original data points. Secondly, for each marked non-sensitive region, the left and right endpoints are the original data point C and the original data point D, respectively. Connect points C and D directly with a straight line segment. On this straight line segment, insert new interpolation points at equal intervals from the x-coordinate of point C to the x-coordinate of point D, using the same insertion spacing as the sensitive area.

[0089] The fourth step involves summarizing all existing data points and newly added interpolation points to construct a copy of the continuous and smooth mapping relationship between construction parameters and initial stiffness values. Specifically, this involves: first, reordering all original data points in the discrete data point set corresponding to each type of construction parameter, along with all new interpolation points generated through interpolation within each interval of that parameter type, according to their ascending x-coordinate values, and merging them into a complete data point column. Any two adjacent points in this data point column are continuous and equidistant along the x-coordinate direction. This data point column is then stored in a table, with each row containing a construction parameter value and its corresponding initial stiffness value, serving as the direct basis for querying the initial stiffness value corresponding to any construction parameter value during subsequent parameter searches.

[0090] In the parameter optimization search module, using the mapping relationship copy as the evaluation basis, within the preset feasible domain of construction parameters, the optimization objective is to maximize the initial stiffness of the node or minimize the amount of material used in the node, while satisfying the preset bearing capacity and construction constraints. The module automatically searches for the optimal combination of node construction parameters, specifically including:

[0091] The optimization parameter search process is used to automatically search for the optimal combination of node construction parameters within the preset feasible domain of construction parameters, based on the mapping relationship copy as the evaluation basis. This process is performed according to the following steps.

[0092] The first step involves randomly generating multiple initial structural parameter combinations within the feasible domain of the structural parameters, and assigning a current evaluation value to each initial combination. Specifically, the feasible domain is defined based on the allowable value ranges of various structural parameters. For example, the wall thickness of the steel pipe ranges from 6 mm to 20 mm, and the yield strength of the steel ranges from 235 MPa to 420 MPa. Within the value range of each parameter dimension, a specified number of initial structural parameter combinations, such as 100, are generated using a uniformly distributed random number generation method. Then, for each initial combination, if the optimization objective is to maximize the initial stiffness of the node, the mapping relationship copy is queried based on the structural parameters of the combination, and the queried initial stiffness value is used as the current evaluation value of the combination. If the optimization objective is to minimize the material usage of the node, the total volume of the steel pipe and concrete is calculated based on the geometric dimensions of the combination, and then the total volume is multiplied by their respective densities and summed to obtain the total weight. This total weight value is used as the current evaluation value, and in this case, a smaller material usage indicates a better evaluation value.

[0093] The second step involves sorting all initial structural parameter combinations according to their current evaluation values, and selecting the top-scoring combinations as preferred combinations based on the preset optimization objective. Specifically, if the optimization objective is to maximize the initial stiffness of the nodes, the combinations are sorted in descending order of their current evaluation values; if the optimization objective is to minimize the material usage of the nodes, the combinations are sorted in ascending order of their current evaluation values. After sorting, the top 20% of the combinations are selected as preferred combinations. For example, if there are 100 combinations in total, the first 20 are selected as preferred combinations, and the remaining 80 combinations are marked as the remaining initial combinations.

[0094] The third step involves applying random perturbations to the construction parameters of the preferred combinations to generate new combinations of construction parameters, and then merging these new combinations with the remaining initial combinations to form a new generation of parameter combination population. The specific implementation of this step will be carried out in the subsequent refinement process.

[0095] The fourth step involves repeating the sorting, selection, and random perturbation process described above until the variation range of the optimal current evaluation value across multiple generations is less than a preset fluctuation limit. The corresponding construction parameter combination at this point is then output as the optimal node construction parameter combination. Specifically, after each generation of search, the optimal current evaluation value among all combinations in that generation is recorded. Starting from the 5th generation, after each generation's calculation, the optimal evaluation value of the current generation is compared with the optimal evaluation value of the previous 5 generations. The absolute value of the difference is calculated and divided by the former to obtain the variation range within 50 generations. The preset fluctuation limit is set to 1%. When the variation range obtained from 50 consecutive generations is less than 1%, the search is considered converged, the iteration process is terminated, and the construction parameter combination corresponding to the optimal evaluation value in the last generation is extracted and output as the optimal node construction parameter combination.

[0096] The process of constructing a new generation of parameter combination populations: The process of constructing a new generation of parameter combination populations, namely, the specific operation of applying random perturbation to the preferred combination and merging it with the remaining combination, is carried out according to the following steps.

[0097] The first step is to calculate the adaptive perturbation amplitude coefficient based on the current iteration count, so that this coefficient gradually decays to a preset minimum value as the number of iterations increases. Specifically, let the total number of iterations performed so far be t, the preset maximum number of iterations be 200, the preset initial perturbation amplitude coefficient be 1.0, and the preset minimum perturbation amplitude coefficient be 0.1. The adaptive perturbation amplitude coefficient corresponding to the current iteration count is calculated as follows: divide the current iteration count t by the maximum number of iterations 200 to obtain a proportional value; subtract this proportional value from 1.0 to obtain the decay factor; multiply the initial perturbation amplitude coefficient 1.0 by this decay factor, and then compare it with the minimum perturbation amplitude coefficient 0.1, taking the larger value as the adaptive perturbation amplitude coefficient for the current generation. This calculation method allows the perturbation amplitude to gradually decrease from 1.0 to 0.1 as iterations progress, allowing for larger perturbations in the early stages of the search to explore a wide region, and using smaller perturbations in the later stages of the search to finely approximate the optimal solution.

[0098] The second step involves taking the current value of each construction parameter in each preferred combination as a baseline, multiplying it by an adaptive perturbation amplitude coefficient, and then multiplying it by a perturbation direction factor randomly generated within the range of -1 to +1 to obtain the perturbation offset of that dimension parameter. Specifically, for a given preferred combination, let it contain p dimensions of construction parameters, for example, the first dimension being the steel pipe wall thickness and the second dimension being the steel yield strength. For the j-th dimension parameter, firstly, a random number uniformly distributed within the range of -1 to +1 is generated as the perturbation direction factor. Then, the current value of this dimension parameter is multiplied by the adaptive perturbation amplitude coefficient calculated in the first step, and then multiplied by the random perturbation direction factor. The product is used as the perturbation offset of that dimension parameter. The perturbation offset can be positive or negative; a positive value indicates an increase based on the current value, and a negative value indicates a decrease based on the current value.

[0099] The third step involves adding the current value of each dimension parameter to the disturbance offset to obtain a preliminary new parameter value. Then, it is determined whether this preliminary new parameter value falls within the preset feasible region of the construction parameters. If it exceeds the feasible region, it is adjusted to the nearest feasible region boundary value. Specifically, for the j-th dimension parameter, its current value is added to the disturbance offset calculated in the second step to obtain a preliminary new parameter value. Then, the feasible region range of this dimension parameter is checked; for example, the steel pipe wall thickness range is 6 mm to 20 mm. If the preliminary new parameter value is less than the lower limit of this range (6 mm), the adjusted new parameter value is set to 6 mm; if the preliminary new parameter value is greater than the upper limit of this range (20 mm), the adjusted new parameter value is set to 20 mm; if the preliminary new parameter value is between 6 mm and 20 mm, it is directly taken as the adjusted new parameter value.

[0100] The fourth step involves summarizing the adjusted parameter values ​​for all dimensions to form a new set of construction parameters. Steps one through three are then repeated for each preferred combination to generate a specified number of new combinations. All new combinations are then merged with the remaining unselected initial combinations to form a new generation of parameter combination population. Specifically, the number of new combinations to be generated for each preferred combination is set, for example, 3 new combinations per preferred combination. For the 20 preferred combinations selected in the first step, a total of 60 new combinations are generated. These 60 new combinations are merged with the remaining 80 unselected initial combinations from the second step, resulting in a total of 140 combinations. From these 140 combinations, 100 are randomly selected as the new generation of parameter combination population for the next round of sorting and selection. This method maintains a constant population size while ensuring the inheritance of superior genes generated by the preferred combinations.

[0101] The optimal construction parameter output module is used to output the optimal combination of node construction parameters obtained by the optimization parameter search module, specifically including:

[0102] The first step is to read the combination of the optimal evaluation values ​​corresponding to the final iteration convergence from the optimization parameter search process, and extract all the construction parameters contained in the combination, specifically including the steel pipe wall thickness value, steel yield strength value, concrete strength grade value, stiffening plate thickness value, and weld size value.

[0103] The second step is to organize the extracted structural parameters into a data list according to the preset output format. Each row in the list corresponds to a parameter name and its value. The value retains two significant figures after the decimal point. For example, the wall thickness of the steel pipe is recorded as 8.50 mm and the yield strength of the steel is recorded as 345.00 MPa.

[0104] The third step is to output the completed data list as a text file to the specified storage path and display the list synchronously in the operation interface for direct use in subsequent construction drawing design or simulation verification.

[0105] The working principle of this invention is as follows: By collecting node construction parameters and establishing a three-dimensional solid simulation copy, the moment-rotation response curve of the node is obtained after applying boundary conditions and displacement loads, and the initial stiffness value is extracted from it; then, based on the initial stiffness values ​​under multiple different construction parameters, a mapping relationship copy between construction parameters and initial stiffness values ​​is constructed; finally, using this mapping relationship copy as the evaluation basis, within the preset feasible domain of construction parameters, the optimal combination of node construction parameters that satisfies the bearing capacity and construction constraints is automatically obtained and output through an iterative search method with the optimization objective of maximizing the initial stiffness of the node or minimizing the amount of material used in the node.

[0106] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A system for predicting and optimizing the stiffness of irregular-shaped steel-concrete composite column joints, characterized in that, The application relates to a steel tube concrete special-shaped column joint optimization method, which comprises the following steps: a node construction parameter acquisition module is used for acquiring construction parameters of a steel tube concrete special-shaped column joint to be optimized, wherein the construction parameters comprise geometric size information, material performance parameters and construction detail characteristics; a node stress simulation copy construction module is used for constructing a three-dimensional entity simulation copy reflecting actual stress characteristics of the node to be optimized based on the construction parameters of the steel tube concrete special-shaped column joint to be optimized; a node stiffness value extraction module is used for applying preset boundary conditions and load modes to the three-dimensional entity simulation copy, obtaining a moment-rotation response curve of the node by solving, and extracting an initial stiffness value representing the rotation deformation resistance of the node from the moment-rotation response curve; a stiffness mapping relationship construction module is used for constructing a mapping relationship copy between the construction parameters and the initial stiffness values by interpolation or fitting based on the initial stiffness values output by the node stiffness value extraction under multiple different construction parameters; an optimization parameter search module is used for searching optimal node construction parameter combinations in a preset construction parameter feasible region based on the mapping relationship copy, taking the maximum initial stiffness of the node or the minimum material consumption of the node as the optimization target, and meeting preset bearing capacity and construction constraint conditions; an optimal construction parameter output module is used for outputting the optimal node construction parameter combinations obtained by the optimization parameter search module.

2. The system for predicting and optimizing the joint stiffness of a concrete-filled steel tubular (CFST) special-shaped column according to claim 1, wherein The construction process of the three-dimensional entity simulation copy is as follows: three-dimensional entities of the steel tube and the core concrete are respectively established according to the geometric size information, and an initial interference amount is arranged between the inner wall of the steel tube and the outer surface of the concrete to simulate the tight clamping effect of the steel tube on the concrete; the real stress-strain curve of the steel material in the material performance parameters is discretized into multiple linear data points, and the uniaxial compression stress-strain curve of the concrete is modified according to the constraint constitutive relation considering the confining pressure effect, and then the data is respectively assigned to the corresponding entity; weld entities are established at the corresponding parts of the steel tube according to the position and size of the weld in the construction detail characteristics, and the weld entities are assigned with the same material performance as the base material but higher yield strength; the geometric entity is meshed by adopting eight-node hexahedral elements, and local seed points are arranged at the weld and the node core area to densify the mesh, and finally the three-dimensional entity simulation copy is generated.

3. The system for predicting and optimizing the joint stiffness of a concrete-filled steel tubular (CFST) special-shaped column according to claim 2, characterized in that, The modified uniaxial compression stress-strain curve of the concrete according to the constraint constitutive relation considering the confining pressure effect comprises the following steps: the constraint stress values generated by the steel tube on the core concrete in each direction are extracted, and the average value is calculated as a constant confining pressure value; the peak stress point of the uniaxial compression stress-strain curve of the concrete is lifted according to the constant confining pressure value, and the strain value corresponding to the peak stress point is translated to the strain increasing direction; the absolute value of the slope of the descending segment of the uniaxial compression stress-strain curve of the concrete is reduced according to the size of the constant confining pressure value, and the descending segment end is extended to the strain larger direction; the new stress-strain data points generated after the peak stress lifting, the peak strain translation and the slope adjustment of the descending segment are used to replace the corresponding positions on the original uniaxial compression stress-strain curve of the concrete, and the modified stress-strain relationship considering the confining pressure effect is formed.

4. The system for predicting and optimizing the joint stiffness of a concrete-filled steel tubular (CFST) special-shaped column according to claim 1, wherein, The calculation process of the initial stiffness value is as follows: Rigid coupling surfaces are established at the top and bottom faces of the column in the 3D solid simulation replica. Fixed constraints are applied to the top face of the column, and axial pressure pointing towards the column body is applied to the bottom face of the column to simulate the actual stress state. A rigid loading surface is established on the free end face of the beam in the 3D solid simulation copy, and a monotonically increasing displacement load perpendicular to the beam axis is applied to the loading surface until the node reaches its ultimate bearing capacity. During the monotonically increasing displacement load application process, the reaction force value of the beam end loading surface and the relative rotation displacement of the beam and column at the corresponding time are recorded under each incremental step. The reaction force value is multiplied by the beam length to convert it into a bending moment value, and the relative rotation displacement is converted into a rotation value to generate a bending moment-rotation data point set. Linear regression fitting was performed on the data points in the initial loading stage of the moment-rotation data point set, and the slope value of the fitted straight line was taken as the initial stiffness value characterizing the node's ability to resist rotational deformation.

5. The system for predicting and optimizing the joint stiffness of a concrete-filled steel tubular (CFST) special-shaped column according to claim 4, wherein, The establishment of a rigid loading surface on the free end face of the beam in the three-dimensional solid simulation copy specifically includes: Set the loading rate of the displacement load so that the loading rate gradually decreases according to the preset attenuation coefficient as the cumulative displacement of the loading surface increases, so as to avoid node instability due to excessive loading. After applying displacement in each incremental step, the reaction force change on the loading surface and the energy balance residual in the solution process are monitored simultaneously. Only when the reaction force increment and the energy balance residual are both lower than their respective preset thresholds is the current incremental step determined to have converged and the next step is initiated. When the reaction force on the loading surface is detected to show a continuous downward trend and the decrease exceeds the preset limit, it is determined that the node has reached its ultimate bearing capacity, and the application of displacement load is immediately terminated.

6. The system for predicting and optimizing the joint stiffness of a concrete-filled steel tubular (CFST) special-shaped column according to claim 1, wherein The process of constructing the mapping relationship replica is as follows: Multiple different construction parameters and their corresponding initial stiffness values ​​are grouped according to the construction parameter type. Within each group, a discrete data point set is generated with the initial stiffness value as the ordinate and the construction parameter as the abscissa. For each set of discrete data points, calculate the rate of change of the slope of the line segment between each adjacent data point, and mark the interval where the rate of change of the slope exceeds the preset abrupt change threshold as the parameter sensitive area; In the parameter-sensitive region, cubic spline interpolation is used to insert new interpolation points between adjacent data points, while in the non-sensitive region, linear interpolation is used to supplement interpolation points. All existing data points and newly added interpolation points are aggregated to construct a copy of the continuous and smooth mapping relationship between the construction parameters and the initial stiffness values.

7. The system for predicting and optimizing the joint stiffness of a concrete-filled steel tubular (CFST) special-shaped column according to claim 1, wherein, The optimal combination of node construction parameters is derived by means of: Multiple initial combinations of construction parameters are randomly generated within the feasible domain of the construction parameters, and each initial combination is assigned a current evaluation value; All initial construction parameter combinations are sorted according to their current evaluation values, and the top-scoring combinations are selected as the preferred combinations based on the preset optimization objective. Random perturbations are applied to the construction parameters of the preferred combination to generate new construction parameter combinations, and the new combinations are merged with the remaining initial combinations to form a new generation of parameter combination population; Repeat the process of sorting, selecting and applying random perturbation until the variation of the optimal current evaluation value for multiple consecutive generations is less than the preset fluctuation limit. Then, output the construction parameter combination corresponding to this time as the optimal node construction parameter combination.

8. The system for predicting and optimizing the joint stiffness of a concrete-filled steel tubular (CFST) special-shaped column according to claim 7, wherein, The formation process of the new generation of parameter combination population is as follows: S1, calculate the adaptive perturbation amplitude coefficient based on the current iteration number, so that the adaptive perturbation amplitude coefficient gradually decays to a preset minimum value as the iteration number increases; S2, for each dimension construction parameter in each preferred combination, multiply the current value of the construction parameter by an adaptive perturbation amplitude coefficient and by a perturbation direction factor randomly generated in the range of negative one to positive one to obtain the perturbation offset of the corresponding dimension parameter. S3, add the current value of each parameter to the disturbance offset to obtain the initial new parameter value, and determine whether the corresponding initial new parameter value falls within the preset feasible region of construction parameters. If it exceeds the feasible region, adjust it to the nearest feasible region boundary value. S4. Summarize the new parameter values ​​after adjusting all dimensional parameters to form a new construction parameter combination. Repeat the process from S1 to S3 for each preferred combination to generate a specified number of new combinations. Then merge all the new combinations with the remaining unselected initial combinations to form a new generation of parameter combination population.

Images