A method and system for flexible design and optimization of a building machine structure suitable for multiple working conditions

Abstract

The application belongs to the technical field of flexible design, and discloses a building machine structure flexible design and optimization method and system suitable for multiple working conditions, which adopts a machine learning agent model to predict complex truss system structure response, sets multiple reinforcement measures to realize structure resistance improvement of the building machine, and realizes flexible design of the building machine by optimizing design of loads in stages respectively. Based on a parametric design method, a generative design framework and a flexible design method of introducing reinforcement measures, a scheme set containing four structures is obtained in a real engineering case, the maximum stress ratio of the scheme structure under the corresponding working condition is kept below 1.0, and the safety of the structure is ensured. Compared with the robustness scheme under the extreme wind working condition, the optimal design scheme obtained by the application reduces the steel consumption by 12.45% under the construction working condition, and reduces the steel consumption by an average of 7.02% under each level of wind working condition.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of flexible design, and particularly relates to a flexible design and optimization method and system for a building machine structure suitable for multiple working conditions. BACKGROUND

[0002] With the improvement of national economy and urbanization level, the population density of cities increases, and the demand for high-rise buildings steadily increases. At present, the frame-core tube system is generally used for high-rise building structures, and the formwork construction technology plays a crucial role in the construction of high-rise building core tubes. The formwork can be divided into sliding formwork system, climbing formwork system, etc. However, the core tube construction has large workload, complex process, high precision requirement, limited operation space, high construction risk, and various construction equipment layout is dense, the working areas are mutually crossed and restricted, which greatly increases the difficulty of core tube construction. Therefore, designing a safe, convenient and efficient engineering machine is the key to the development of the construction industry.

[0003] With the development of building technology and industrialization, the air building machine (ABM) realizes the industrialization and intelligent construction of high-rise building construction. The use of the building machine simplifies the assembly and movement of the formwork frame, reduces repetitive work, and speeds up the construction speed. It provides a stable working environment, improves construction safety, and at the same time, saves materials through repeated turnover of the formwork, and constructs a digital information management system. The air building machine has been successfully applied to multiple projects. Lightweight, standardization, modularization, intelligentization and industrialization are the development direction of future ABM.

[0004] With the development of computer technology, some computer-aided design techniques are applied to the design process of the building machine, such as parametric design. Parametric design is an algorithm that uses parameters to represent the characteristics of a system, including numerical and geometric relationships of the structure, and can quickly generate the structure according to the change of the parameters. In the traditional design method, the designer needs to manually remodel to adjust the finite element model, resulting in a large amount of repetitive work in the modeling analysis cycle. Parametric design is based on structural modularization and incorporates algorithms into the design process. The designer not only establishes the spatial position of the model, but also combines the position relationship, structural characteristics and connection method into the preset parameters in the model. The completed model can be easily converted according to the designer's expectations. By changing the variable parameters, the structure can be regenerated, thereby reducing the workload of repetitive modeling and realizing the rapid iteration of the model.

[0005] Designers usually consider multiple objectives such as stress, displacement, frequency, mass, etc. when designing, and the responses of these objectives have a complex relationship with the structure itself. Some objectives often conflict with each other and cannot be achieved simultaneously, so it is impossible to obtain a perfect structure that optimizes all objectives. At this point, multi-objective optimization algorithms (MOEA) can help designers. Multi-objective optimization algorithms find the Pareto front by iterating multiple samples in the solution space, which is a set of objectives that achieve optimality without degrading other objectives. However, current multi-objective optimization algorithms usually require more than 10,000 iterations to converge, which is difficult to achieve for large and complex structures with high computational cost. With the continuous upgrading of computer algorithms and performance, proxy models built by machine learning (ML) have become an effective technical route to solve this problem. Proxy models are response functions that use data-driven algorithms to fit complex real-world models. Machine learning models use self-learning mechanisms to analyze the relationship between the input and output of training data and obtain high-precision proxy models. Since finite element calculations are time-consuming, using proxy models to predict finite element models can greatly reduce the time cost of calculations and make it possible to use optimization algorithms that require a lot of computational cost. At the same time, proxy models have generalization capabilities and can use fewer training samples to predict the response of the structure within a larger parameter variation range. Thus, an optimization design framework is formed, realizing a method for fast generation and optimization of optimal design schemes for general structures.

[0006] Past ABM designs have been primarily based on robustness considerations, and the resulting designs are static, meaning that the design must account for the construction conditions, jacking conditions, and potential extreme working conditions that the structure might encounter. However, during the typical 2-3 year period of ABM use, the probability of experiencing extreme conditions is relatively low. As a result, a significant amount of material is used to withstand extreme conditions, resulting in a waste of resources during the majority of the structure's usage time. This approach is inconsistent with the trend towards lightweight ABM design and presents a challenge in balancing material conservation with addressing the demands of extreme and normal construction conditions. In contrast to the concept of robustness, flexibility refers to the ability of a system or design to adapt and meet changing demands, emphasizing the ability of a system to be modified to accommodate new conditions or requirements, allowing the system to track changes in demand and remain effective in different environments. The introduction of flexibility aims to improve resource utilization, adapt to changing demands, and is applied in areas such as manufacturing systems, supply chains, software development, etc. The concept of flexibility is consistent with the goal of lightweight design in ABMs, as it allows for efficient use of resources while adapting to different demands and conditions.

[0007] Through the above analysis, the problems and defects of the prior art are: how to integrate the flexible design concept into the design of the ABM steel platform to achieve resource conservation and cost reduction. SUMMARY

[0008] In view of the problems of the prior art, the application provides a flexible design and optimization method and system for a building machine structure suitable for multiple working conditions, which overcomes the deficiencies of the prior art in the design of the building machine by considering environmental load changes, structural resistance redundancy and the like, and combining a parametric design method and a generative design method to realize the rapid generation and design of the building machine structure.

[0009] The application is implemented as follows: a flexible design and optimization method for a building machine structure suitable for multiple working conditions, which uses a machine learning agent model to predict the structural response of a complex truss system, sets multiple reinforcement measures to improve the structural resistance of the building machine, and realizes the flexible design of the building machine by optimizing the design of the load in stages.

[0010] Further, the flexible design and optimization method for a building machine structure suitable for multiple working conditions specifically comprises:

[0011] S1. Parametric design and response calculation of the building machine in the air: the model establishment process is performed in Grasshopper; first, the candidate points of the support columns and the size of the steel platform are preset according to the design drawing of the core tube shear wall, and a plurality of main truss layouts are preset accordingly; then, a parametric building machine model is established, the main structural components of the building machine are created based on the module design, and are connected with each other according to the module rules, wherein the auxiliary structure is simplified and equivalent to the load applied to the main structure; through parametric programming statements, the software can set the position of the Bailey piece and combine it into a steel platform, set the position of the support column and connect it with the steel platform, and set the reinforcement measures;

[0012] S2. Multi-objective optimization design based on artificial intelligence: the optimization design method combining the agent model and the multi-objective optimization algorithm is used for multi-objective optimization; a plurality of combination schemes of independent variables are generated by using the filling method, are substituted into the finite element software for calculation, and the steel consumption, the maximum stress and the maximum displacement response of the scheme are obtained, which together constitute the data set for training the agent model; the machine learning algorithm is used to train and test the stress, displacement and steel consumption respectively, and three agent models are obtained; through the multiple iterations of the population of the multi-objective optimization algorithm in the variable space of the parameters of the plurality of reinforcement measures, a series of non-dominated solution sets on the Pareto surface are obtained for the designer to select; the schemes in the solution set cannot obtain a scheme with a better target without deteriorating other targets, which represent the optimal schemes under different target weights, and the designer can select according to the requirements and preferences.

[0013] S3. Flexible design with multiple reinforcement measures: the structure realizes the force flexibility through various reinforcement measures, the initial structure may not meet the force requirements under all conditions, but the force is improved by adding reinforcement measures when necessary, allowing the structure to dynamically adjust and become relatively optimal under each working condition, and the scheme set is composed of design solutions generated for each time point; the flexible design adopts a progressive design mode, which gradually increases various reinforcement measures on the same basic structural framework to adapt to different working conditions by setting constraints in the optimization design process.

[0014] Further, S1 further comprises: performing elastic linear analysis by SAP2000 to obtain the response of the structure; in the elastic analysis, the response under each working mode is calculated independently, and the total result is obtained by linear combination, which reduces the time cost of calculation; the selected responses are the maximum stress ratio, the maximum point displacement and the steel consumption of the structure, which represent the resistance, stiffness and cost of the structure, and are also the optimization objectives in the optimization design.

[0015] Further, S2 specifically comprises: the proxy model uses the Deep Forest algorithm, which includes multi-granularity scanning and cascading forest structure; multi-granularity scanning is used to extract features from input data at multiple scales or resolutions, including applying different size sliding windows to input data and extracting feature vectors from each window, using different window sizes allows the algorithm to capture different levels of details in the input data; smaller window sizes capture local information and fine-grained patterns, while larger window sizes capture more global information and larger structures; the cascading forest structure is composed of multiple levels, each level contains a group of random forests, and the forests receive input feature vectors extracted from multi-granularity scanning for training; in the decision tree, each leaf node is assigned a feature to be split until only one type of instance is on the last leaf node, and the final prediction is calculated based on the maximum value in the average vector of the final level of forest, in order to reduce the risk of overfitting, the class vector generated by each forest is generated through k-fold cross-validation, and then averaged to generate the final class vector as the enhanced features for the next level in the cascade; compared with deep neural networks, deep forests show obvious advantages, including significantly reducing the number of hyperparameters, and showing considerable robustness in the performance of various hyperparameter settings, which makes it suitable for fitting requirements with strong nonlinearity;

[0016] The multi-objective optimization algorithm uses the AGE-MOEA algorithm, which inherits the overall framework of NSGA-II and uses a survival score combining diversity and proximity instead of crowded distance; in the first step, the algorithm generates an initial set P of N random solutions, then generates new solutions, offspring Q, using crossover and mutation; the new population including the current population and the offspring population is non-dominant sorted to obtain the non-dominant front F; after normalizing the non-dominant front, a new population is generated, and a survival score considering diversity and proximity is assigned to the first non-dominant front; the new population of M solutions is selected from the non-dominant front in order until the solution exceeds M; then, the remaining solutions are selected from the front F in descending order of survival score.

[0017] Further, S3 specifically comprises the following steps in the flexible design process adopted by the application: first, a plurality of basic structural frameworks are predefined, from which the optimal basic structure under construction conditions is determined as the blueprint for subsequent design; second, the range of wind speed that the structure can withstand is determined from the construction condition to the maximum extreme condition, and then divided into smaller intervals by design points; the design points are equidistantly arranged or rearranged according to the design results; third, the structural response under extreme conditions is analyzed to find the weak areas in the structure, and several local strengthening measures are proposed for specific areas; fourth, for each design point, the optimal structure is generated by multi-objective optimization based on the selected basic structure; the parameters of multi-objective optimization include the setting parameters of various strengthening measures, and the objectives are the maximum stress, maximum displacement and steel consumption of the structure; in the next level of optimization design, the strengthening measures set in the previous level should be retained.

[0018] Another object of the present application is to provide an adaptive multi-condition building machine structure flexible design and optimization system for realizing the adaptive multi-condition building machine structure flexible design and optimization method, comprising:

[0019] The parametric design and response calculation module is used for parametric design and response calculation, and the model establishment process is carried out in Grasshopper; first, the support column candidate points and steel platform size are preset according to the core tube shear wall design drawing, and a plurality of main truss layouts are preset accordingly; then, the parametric building machine model is established, the main structural components of the building machine are respectively created based on the module design, and are connected with each other according to the module rule, wherein the accessory structure is simplified and equivalent to the load applied to the main structure; through parametric programming statements, the software can set the position of the Bailey piece and combine it into a steel platform, set the position of the support column and connect it with the steel platform, and set the strengthening measures;

[0020] Multi-objective optimization design module: a multi-objective optimization design method combining a surrogate model and a multi-objective optimization algorithm is used for multi-objective optimization; a filling method is used to generate a plurality of combination schemes of independent variables, which are substituted into a finite element software to obtain the steel consumption, maximum stress and maximum displacement response of the scheme, which together constitute a data set for training of the surrogate model; a machine learning algorithm is used to train and test the data of stress, displacement and steel consumption respectively to obtain three surrogate models; a plurality of iterations of a population of a multi-objective optimization algorithm are used to search in a variable space of a plurality of strengthening measure setting parameters to obtain a series of non-dominated solution sets on a Pareto surface for designers to select; the schemes in the solution set cannot obtain a scheme with a better target without degrading other targets, representing the optimal scheme under different target weights, and the designer can select according to requirements and preferences.

[0021] Flexible design module: used for flexible design, the structure realizes force flexibility through various strengthening measures, the initial structure may not meet the force requirements under all conditions, but the force can be improved by adding strengthening measures when necessary, allowing the structure to dynamically adjust and become relatively optimal under each working condition, and the scheme set is composed of design solutions generated for each time point; the flexible design adopts a progressive design mode, which gradually increases various strengthening measures on the same basic structure framework to adapt to different working conditions by setting constraints in the optimization process.

[0022] Another object of the present application is to provide a computer device comprising a memory and a processor, the memory storing a computer program, the computer program being executed by the processor to make the processor execute the steps of the flexible design and optimization method for the building machine structure adapted to multiple working conditions.

[0023] Another object of the present application is to provide a computer readable storage medium storing a computer program, the computer program being executed by the processor to make the processor execute the steps of the flexible design and optimization method for the building machine structure adapted to multiple working conditions.

[0024] Another object of the present application is to provide an information data processing terminal for realizing the flexible design and optimization system for the building machine structure adapted to multiple working conditions.

[0025] In combination with the above technical solutions and solved technical problems, the technical solution to be protected by the present application has the following advantages and positive effects:

[0026] Firstly, the present application proposes a flexible design and optimization technology for the building machine structure adapted to multiple working conditions, which realizes the rapid generation of the building machine structure scheme set under the large wind working condition by introducing a plurality of strengthening measures through a structure parameterization design method, a multi-objective optimization algorithm and a hierarchical design method.

[0027] The application is based on a parameterized design method, a generative design framework and a flexible design method using various reinforcement measures. In a real engineering case, a set of four structural schemes is obtained, and the maximum stress ratio of the scheme structure under corresponding working conditions is kept below 1.0, ensuring the safety of the structure. Compared with the robustness scheme under extreme wind conditions, the optimal design scheme obtained by the application reduces the steel consumption by 12.45% under construction conditions and by an average of 7.02% under various wind conditions. The results show that the flexible design method used in the application effectively improves the structural design efficiency of the aerial building machine and reduces the construction cost.

[0028] Second, the parameterized design method abstracts the structural components as parameters, simplifying the modeling and design work to adjusting parameters and finding the optimal parameter combination, improving the modeling speed, simplifying the design process, and greatly improving the design and optimization efficiency. The finite element proxy model of the building machine model is constructed using artificial intelligence technology to predict the response of the model, saving a lot of calculation time. The multi-objective optimization algorithm is used to form a Pareto solution set, providing the designer with design schemes that meet different design preferences and obtaining a more economical building machine design scheme. The load analysis design method generates structures corresponding to various extreme conditions, enabling the building machine to adopt the most economical scheme under different extreme conditions. The flexible design scheme using various reinforcement measures realizes the flexibility of the building machine's resistance at a relatively low cost, enabling it to change its structure to meet the requirements under different extreme conditions. The gradual design scheme reduces the difference between schemes and reduces the workload of structural modification.

[0029] Third, the expected benefits and commercial value of the technical solution of the application after transformation are: using the parameterized modeling method, the model is quickly generated, the workload of model modification is reduced, and the design efficiency is improved; using the proxy model to simulate finite element simulation greatly reduces the calculation time of repeated simulation and improves the work efficiency of building machine design; the multi-objective optimization algorithm is used for structural performance optimization, and schemes that meet the different design preferences of the designer are provided, improving the performance of the building machine design structure; a series of building machine design schemes are generated to provide scheme guidance for building machine designers and meet the resistance requirements of building machines under different working conditions; the resistance flexibility of the building machine is realized through various local reinforcement measures, the wind resistance of the building machine is gradually improved, the generated different structures can meet the relative optimum under different working conditions, the steel consumption of the building machine is reduced, the design and construction costs of the building machine are reduced, and the benefits of the engineering project are further improved.

[0030] Fourth, the technical problems solved by the application and the significant technical progress obtained mainly reflect in the following aspects:

[0031] 1. Solution to technical problems:

[0032] Adaptive challenge: Traditional formwork structure design often targets specific working conditions, lacking adaptability and flexibility in multiple working conditions. This invention introduces a flexible design concept, achieving dynamic adjustment and optimization of formwork in different working conditions, solving the problem of traditional design difficulty in adapting to changing construction environments.

[0033] Optimization efficiency problem: Traditional structural optimization process usually relies on a large number of tests and finite element analysis, which is time-consuming and labor-intensive. This invention uses a machine learning surrogate model to predict the structural response of complex truss systems, greatly accelerating the optimization process and improving design efficiency.

[0034] Structural resistance improvement: To address the problem of insufficient structural resistance of formwork in complex working conditions, this invention sets up various strengthening measures to significantly improve structural resistance, enhancing the stability and safety of formwork.

[0035] 2. Technical progress and innovation:

[0036] Parametric design and response calculation: This invention uses parametric design methods to quickly adjust and optimize design parameters through Grasshopper software modeling. At the same time, through response calculation, the structural performance under different design schemes can be accurately evaluated, providing strong support for optimization design.

[0037] Multi-objective optimization design based on artificial intelligence: Combined with surrogate models and multi-objective optimization algorithms, this invention achieves a balance and optimization between multiple objectives such as steel consumption, maximum stress, maximum displacement, etc. This design method not only improves design quality but also makes the design scheme more in line with actual needs.

[0038] Flexible design and progressive strengthening measures: By adopting a flexible design philosophy, this invention allows structures to increase resistance through additional strengthening measures when necessary, achieving dynamic adjustment and performance optimization of structures. This progressive design pattern enables formwork to better adapt to the needs of different construction stages.

[0039] This invention solves the technical problems of existing technology, achieving significant technical progress and innovation in formwork structure design. These advances not only improve design efficiency and quality but also enhance the adaptability and stability of formwork, making a positive contribution to the development of the construction field. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings needed to be used in the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0041] Figure 1 is a flexible design and optimization method flowchart of a building machine structure suitable for multiple working conditions provided by the embodiment of the present application;

[0042] Figure 2 is a flexible design framework schematic diagram provided by the embodiment of the present application;

[0043] Figure 3 is a flexible design framework schematic diagram of a building machine provided by the embodiment of the present application;

[0044] Figure 4 is a schematic diagram of an aerial building machine physical model provided by the embodiment of the present application;

[0045] Figure 5 is a schematic diagram of a basic structure framework provided by the embodiment of the present application; wherein, (a) structure one; (b) structure two; (c) structure three; (d) structure four; (e) structure five; (f) structure six; (g) structure seven; (h) structure eight;

[0046] Figure 6 is a schematic diagram of a basic structure selection provided by the embodiment of the present application;

[0047] Figure 7 is a schematic diagram of a strengthening measure structure provided by the embodiment of the present application;

[0048] Figure 8 is a schematic diagram of a linear regression result of a true value and a predicted value provided by the embodiment of the present application; wherein, (a) stress ratio; (b) displacement; (c) steel consumption;

[0049] Figure 9 is a schematic diagram of a multi-objective optimization Pareto frontier provided by the embodiment of the present application;

[0050] Figure 10 is a schematic diagram of a structure response value provided by the embodiment of the present application;

[0051] Figure 11 is a schematic diagram of a flexible design response provided by the embodiment of the present application; wherein, (a) stress ratio; (b) displacement;

[0052] Figure 12 is a schematic diagram of a building machine structure flexible design and optimization system structure suitable for multiple working conditions provided by the embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0054] In view of the problems in the prior art, the application provides a flexible design and optimization method and system for a building machine structure suitable for multiple working conditions.

[0055] The application provides a flexible design and optimization method for a building machine structure suitable for multiple working conditions, adopts a machine learning agent model to predict the structural response of a complex truss system, sets multiple reinforcement measures to improve the structural resistance of the building machine, and realizes flexible design of the building machine by optimizing design of loads in stages.

[0056] As shown in the drawings, the application specifically comprises the following steps. Figure 1

[0057] S1. Parameterized design and response calculation of an aerial building machine

[0058] The model establishment process of the application is performed in Grasshopper. First, candidate points of support columns and sizes of steel platforms are preset according to the design drawings of a core tube shear wall, and a plurality of main truss layouts are preset according to the candidate points and the sizes. Then, a parameterized building machine model is established, main structural components of the building machine are respectively created based on module design, and the main structural components are connected with each other according to the module rules, wherein the accessory structures are simplified and equivalent to loads applied to the main structure. Through parameterized programming statements, the software can set the positions of the Bailey pieces and combine the Bailey pieces into the steel platforms, set the positions of the support columns and connect the support columns with the steel platforms, and set the reinforcement measures.

[0059] The application adopts SAP2000 to perform elastic linear analysis and obtain the response of the structure. In the elastic analysis, the response in each working mode can be independently calculated, the total result is obtained through linear combination, and the time cost of calculation is reduced. The responses selected in the application are the maximum stress ratio, the maximum point displacement and the steel consumption of the structure, which represent the resistance, stiffness and cost of the structure and are also the optimization targets in the optimization design. Swallow is a plug-in supporting data transmission from Grasshopper to SAP2000. It allows definition of cross sections, member loads and conversion from a geometric model to a SAP2000 model. However, some functions are not implemented in this plug-in, including constraints and elastic supports, and the definition thereof is set by calling the API interface of SAP2000 by means of the GH_CPython plug-in. GH_CPython provides a component for users to implement Python code inside Grasshopper. The use of these software and algorithms realizes automation of the processes of model establishment, data transmission, finite element analysis and data acquisition.

[0060] S2. Multi-objective optimization design based on artificial intelligence

[0061] ​The present invention employs an optimization design method that combines a surrogate model with a multi-objective optimization algorithm. The surrogate model uses a Deep Forest algorithm, which includes a multi-grained scanning and a cascading forest structure. Multi-grained scanning is used to extract features from input data at multiple scales or resolutions. It involves applying different sized sliding windows to the input data and extracting a feature vector from each window. Using different window sizes allows the algorithm to capture details at different levels in the input data. Smaller window sizes capture local information and fine-grained patterns, while larger window sizes capture more global information and larger structures. Multi-grained scanning enhances the algorithm's ability to capture and utilize information at different scales or resolutions in the input data, thereby improving task performance. The cascading forest structure consists of multiple levels, each containing a set of random forests. The forests receive the input feature vectors extracted from the multi-grained scanning for training. In the decision trees, each leaf node is assigned a feature to split on until there is only one type of instance on the last leaf node. The final prediction is the maximum value in the average vector computed from the final level of forests. To reduce the risk of overfitting, the class vectors generated by each forest are averaged to generate a final class vector as the enhanced feature for the next level in the cascade, generated by k-fold cross-validation. Compared to deep neural networks, Deep Forest exhibits clear advantages, including a significantly reduced number of hyperparameters and a considerable robustness in performance across various hyperparameter settings, which makes it suitable for fitting responses with strong nonlinearity.

[0062] The multi-objective optimization algorithm uses the AGE-MOEA algorithm, which inherits the overall framework of NSGA-II and uses a viability score that combines diversity and proximity instead of crowding distance. In the first step, the algorithm generates an initial set P of N random solutions. Then, it generates new solutions, offspring Q, using crossover and mutation. The new population, including the current population and the offspring population, is non-dominant sorted to obtain the non-dominant front F. After normalizing the non-dominant front, a new population is generated, and a viability score that considers diversity and proximity is assigned to the first non-dominant front. A new population of M solutions is selected from the non-dominant front in order, until the solutions exceed M. Then, it selects the remaining solutions from the front F in descending order of viability score.

[0063] S3. Flexible design method with multiple reinforcement measures

[0064] In the robust structure design of ABM, the structure is required to consider all design conditions, and the design result should meet the resistance requirements at each time point without changing the structure. Generally, the final design result is the envelope of the optimal structure that meets all constraints under different conditions, which achieves global relative optimality, but may not be optimal at most time points, resulting in waste of resistance of the structure during most of the use period. In the present invention, the structure realizes resistance flexibility through various strengthening measures. The initial structure may not be able to meet the resistance requirements under all conditions, but the resistance can be improved by adding strengthening measures when necessary. In the proposed flexible design method, various working conditions that the structure may experience are also considered. However, this design method is not intended to find a single structure that can meet all working condition requirements, but instead proposes a series of structures that allow the structure to dynamically adjust and become relatively optimal under each working condition, and the solution set is composed of design solutions generated for each time point. In order to reduce the differences between the solutions, improve the convenience and feasibility of structure conversion, and reduce the complexity of the design, the flexible design adopts a progressive design mode, which gradually increases various strengthening measures on the same basic structure framework by setting constraints in the optimization design process.

[0065] In the flexible design process adopted in the present invention, the following steps are taken: First, several basic structure frameworks are predefined, from which the optimal basic structure under construction conditions is determined as the blueprint for subsequent design. Second, the range of wind speeds that the structure may withstand is determined, from the construction condition to the maximum extreme condition, and then divided into smaller intervals through several design points. The design points can be set equidistantly or reset according to the design results. Third, the structural response under extreme conditions is analyzed, and the weak areas in the structure are found out, and several local strengthening measures are proposed for specific areas. Fourth, for each design point, the optimal structure is generated based on the selected basic structure through multi-objective optimization. The parameters of multi-objective optimization include the setting parameters of various strengthening measures, and the objectives are the maximum stress, maximum displacement, and steel consumption of the structure. In the next level of optimization design, the strengthening measures set in the previous level should be retained. The strengthening measures, as a kind of progressive constraint, realize the gradual increase of resistance and facilitate the conversion between structures. The solution set formed by the flexible design is composed of these structures. The workflow of this method is summarized as Figure 2 shown.

[0066] Referring to Figure 3 shown, the present embodiment provides a working condition grading air building machine structure flexible design and optimization technology, and the case is the application of a lightweight air building machine for a high-rise building construction. The main steps include:

[0067] Step 1, selection of the basic structure of the building machine. The selection of the basic structure frame is crucial because it determines the overall performance and maximum load-bearing capacity of the structure. Since the building machine is in the construction condition most of the time, the selection of the basic structure is based on the construction condition. The building machine is constructed as shown in Figure 4 Due to the condition that the support columns need to be fixed on the shear wall of the core tube, 8 basic structure frames are proposed as shown in Figure 5 On the basis of these basic structure frames, transverse supports are added, and the number of transverse supports set on each side varies from 0 to 3. The stress ratio, displacement and steel consumption of the structure under construction conditions are calculated, and the results are shown in Figure 6 The optimal result with the minimum value of the three target values is highlighted by the star symbol. The number of structure types in this case is about 200, and it is feasible to calculate all structure types. If more parameters need to be considered in the actual situation, it is better to use multi-objective optimization for structure selection. The structure represented by the optimization result is designed under the construction condition, and the basic structure frame is shown in Figure 5 (f).

[0068] Step 2, construction of training data for building machine response surrogate model. In this invention, the deep forest model is selected as the surrogate model to predict the response, which is a simplification of finite element calculation. The surrogate model can significantly reduce the calculation cost of each calculation, and maintain high accuracy using a large number of input samples. In addition, it can predict the response of the structure in a larger range based on a limited number of samples. The surrogate model used in this invention considers the changes of the load applied to the structure and the implementation of multiple strengthening measures, which enables the surrogate model to predict the response of the structure with different strengthening measures under various wind loads. In order to reduce the number of parameters, improve the calculation efficiency and improve the fitting accuracy, the design of the Bailey structure and the change of the cross-sectional size are not considered in this invention, and the numerical values refer to the actual engineering model parameters. The selected parameters of the surrogate model are shown in Table 1, including load variables and 5 kinds of strengthening measure variables, and the construction of the strengthening measures is shown in Figure 7 The parameter x1 represents the basic structure frame selected in the previous step. The finite element calculation result can be obtained by inputting the generated parameter combination into the parametric design process. The training data set contains the combination of input parameters and target responses. The generated about 4000 samples are divided into training set and test set, the proportion is 0.8:0.2. The training result is shown in Figure 8 The R 2 index of stress ratio, displacement and steel consumption is 0.9532, 0.9663 and 0.9590 respectively, and the detailed result index is shown in Table 2. This indicates that the surrogate model can accurately fit the finite element model and meet the required prediction accuracy.

[0069] Table 1. Model parameters

[0070]

[0071] Table 2. Model performance evaluation

[0072]

[0073] Step 3, flexible design of the building machine using working condition classification. The design wind speed from 20 to 40 m / s is divided into several design points, and each design point is optimized and designed, and the Pareto front of the optimization result is as shown in Figure 9 The points on the Pareto front can be returned to the finite element model to obtain more accurate response values and more accurate corresponding intervals. In extreme conditions, the structure needs to meet a stress ratio of less than 1.0 to avoid structural failure. In addition, people tend to choose a design with lower steel consumption to obtain higher economic benefits. The final design is selected according to the design points marked with a star symbol in Figure 9 , for example, the star symbol indicates the design scheme selected when designing at this point.

[0074] The results show that the present application can select a suitable design scheme set from the Pareto front, as shown in Figure 10 In this case, a scheme set containing 4 structures is obtained, and the specific design parameters are shown in Table 3. As shown in Figure 11 Through finite element analysis, the maximum stress ratio of each scheme of the structure under the corresponding working condition is maintained below 1.0, ensuring the safety of the structure. By comparing with the design results of the traditional design method, the optimal design scheme obtained by the present application reduces the steel consumption by 12.45% under the construction working condition, and reduces the steel consumption by an average of 7.02% under each level of wind working condition. The optimization degree of each structure is shown in Table 4, and the results show that the best design scheme has a significant optimization improvement effect compared with the traditional design.

[0075] Table 3. Design scheme strengthening measure arrangement

[0076]

[0077] Table 4. Optimization degree of flexible design results

[0078]

[0079] As shown in Figure 12 , the flexible design and optimization system of the building machine structure adapted to multiple working conditions provided by the embodiment of the present application comprises:

[0080] Parametric design and response calculation module: used for parametric design and response calculation, the model building process is carried out in Grasshopper; first, the candidate points of support columns and the size of steel platforms are preset according to the core tube shear wall design drawing, and a plurality of main truss layouts are preset accordingly; then, a parameterized building machine model is established, the main structural components of the building machine are respectively created based on the module design, and are connected with each other according to the module rule, wherein the auxiliary structure is simplified and equivalent to load applied to the main structure; through parameterized programming statements, the software can set the position of the Bailey piece and combine it into a steel platform, set the position of the support column and connect it with the steel platform, and set the strengthening measures;

[0081] Multi-objective optimization design module: used for multi-objective optimization design by using the optimization design method combining the proxy model and the multi-objective optimization algorithm; a fast and flexible filling method based on clustering is used to generate about 4000 combination schemes of strengthening measure settings, which are substituted into SAP2000 for finite element calculation to obtain the steel consumption, the maximum stress and the maximum displacement response, which together constitute the data set for training of the proxy model algorithm; the deepforest algorithm is used for data training of stress, displacement and steel consumption respectively to obtain three proxy models; the AGE-MOEA multi-objective optimization algorithm is used for optimization iteration to search the variable space of the strengthening measure setting parameters to obtain a series of non-dominated solution sets on the Pareto surface for designers to select, and the scheme with the least steel consumption is usually selected under the condition that the stress and displacement response meet certain requirements.

[0082] Flexible design module: used for flexible design, the structure realizes resistance flexibility through various strengthening measures, the initial structure may not meet the resistance requirements under all conditions, but the resistance can be improved by adding strengthening measures when necessary, allowing the structure to dynamically adjust and become relatively optimal under each working condition, and the scheme set is composed of design solutions generated for each time point; the flexible design adopts a progressive design mode, and various strengthening measures are gradually increased on the same basic structural framework to adapt to different working conditions by setting constraints in the optimization design process.

[0083] The application embodiment of the present application provides a computer device, which comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the flexible design and optimization method of the building machine structure suitable for multiple working conditions.

[0084] The application embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the flexible design and optimization method of the building machine structure suitable for multiple working conditions.

[0085] The application embodiment of the present application provides an information data processing terminal, which is used for realizing a flexible design and optimization system of a building machine structure adapting to multiple working conditions.

[0086] The following are two specific embodiments for showing the practical application of the flexible design and optimization method of the building machine structure adapting to multiple working conditions:

[0087] Embodiment one:

[0088] In the construction of a certain high-rise building, the flexible design and optimization method of the building machine structure adapting to multiple working conditions is adopted. First, the parametric design and response calculation of the aerial building machine are carried out in the Grasshopper software. According to the design drawings, the candidate points of the support column and the size of the steel platform are preset, and a plurality of main truss layouts are preset accordingly. Subsequently, a parametric building machine model is established, and the main structure components, including the steel platform and the support column, are created based on the module design and connected with each other according to the module rules. The auxiliary structure is simplified and applied to the main structure in the form of equivalent load.

[0089] Next, the multi-objective optimization design method based on artificial intelligence is adopted. A plurality of independent variable combination schemes are generated by using the filling method, and are substituted into the finite element software for calculation, so as to obtain the steel consumption, the maximum stress and the maximum displacement response of each scheme, which together constitute the data set for training the surrogate model. The stress, displacement and steel consumption are trained and tested by using the machine learning algorithm, and the corresponding surrogate model is obtained. Then, the multi-objective optimization algorithm is used to perform multiple iterative searches in the variable space of the strengthening measure setting parameters, and a series of non-dominated solution sets located on the Pareto surface are obtained. The designer selects the optimal scheme from the solution set according to the engineering requirements and preferences.

[0090] In the flexible design stage, a plurality of strengthening measures are adopted to improve the structural resistance of the building machine. According to different working condition requirements, the strengthening measures are gradually increased, such as adding temporary supports and adjusting the connection mode of the steel platform. This progressive design mode enables the building machine to dynamically adjust under different working conditions and maintain relatively optimal performance.

[0091] Embodiment two:

[0092] In another large-scale building project, the flexible design and optimization method of the building machine structure adapting to multiple working conditions is also applied. In the parametric design and response calculation stage, a more complex building machine model is created in Grasshopper according to the actual situation and design requirements of the project. The model considers a plurality of main truss layouts and support column configurations to adapt to the requirements of different construction stages.

[0093] In the multi-objective optimization phase, more advanced machine learning algorithms and surrogate model techniques were used. Through extensive finite element calculations and analysis, more accurate surrogate models were obtained, and a more comprehensive set of non-dominated solutions was obtained through multi-objective optimization algorithms. Designers selected the optimal reinforcement configuration scheme from the solution set according to the specific requirements and constraints of the project.

[0094] In the flexible design phase, a variety of reinforcement combination schemes were designed for different working conditions and load conditions that may occur in the project. These schemes can be dynamically adjusted when needed to improve the structural resistance and stability of the building machine. Through the application of the progressive design mode, the building machine can maintain excellent performance and safety throughout the construction process.

[0095] These two examples demonstrate the application and effect of the flexible design and optimization method of the building machine structure in different construction projects. Through the combination of parametric design, multi-objective optimization and flexible design, the structural performance and adaptability of the building machine can be effectively improved to meet the construction requirements under different working conditions.

[0096] It should be noted that the embodiments of the present application can be realized by hardware, software or a combination of software and hardware. The hardware part can be realized by special logic; the software part can be stored in the memory and executed by the appropriate instruction execution system, such as microprocessor or special designed hardware. Those skilled in the art can understand that the above devices and methods can be realized by computer executable instructions and / or included in processor control code, such as carrier medium, such as magnetic disk, CD or DVD-ROM, programmable memory, such as read-only memory (firmware), or data carrier, such as optical or electronic signal carrier. The device and its modules of the present application can be realized by hardware circuit, such as ultra-large scale integrated circuit or gate array, semiconductor, such as logic chip, transistor, etc., or programmable hardware device, such as field programmable gate array, programmable logic device, etc., or by software executed by various types of processors, or by a combination of the above hardware circuit and software, such as firmware.

[0097] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any modification, equivalent replacement and improvement within the technical range disclosed by the present application, which is within the spirit and principle of the present application, should be covered within the protection scope of the present application.​​​​

Claims

1. A method for flexible design and optimization of a building machine structure adapted to multiple working conditions, characterized in that, A machine learning surrogate model is used to predict the structural response of a complex truss system, and various strengthening measures are combined to improve the structural resistance of the building machine. By classifying the loads that the building machine will bear and conducting independent optimization design for each level of load, the flexible design and optimization of the building machine structure are realized, so that the building machine can adjust its structure under different working conditions to achieve the relatively optimal performance. The method comprises the following steps: S1. Parametric design and response calculation of the building machine in the air, wherein: In the Grasshopper software environment, the core tube shear wall design drawing is parameterized, the support column position and steel platform size are preset; A plurality of main truss layout schemes are preset, and the main structure component model of the building machine is established based on modular design; The position of the Bailey piece is automatically set by parameterization programming, and the Bailey piece is combined into a steel platform, the support column and the steel platform are connected, and the strengthening measures are set; S2. Multi-objective optimization design based on artificial intelligence, wherein: The optimization design is carried out by combining the surrogate model and the multi-objective optimization algorithm to generate a plurality of design schemes; The steel consumption, maximum stress and maximum displacement response of each design scheme are calculated by using the finite element software, and the training data set of the surrogate model is constructed; The machine learning algorithm is applied to train different response data to obtain the surrogate model of stress, displacement and steel consumption; The Pareto optimal solution set is found in the variable space of the setting parameters of the various strengthening measures through the iterative search of the multi-objective optimization algorithm, which is selected by the designer according to the demand and preference; S3. Flexible design using various strengthening measures, wherein: The structure realizes flexible resistance through the strengthening measures, allowing dynamic adjustment of structural resistance under different working conditions; An incremental design mode is adopted, and the strengthening measures are gradually increased on the basis of the same basic structure frame to adapt to different working conditions by setting constraints in the optimization process; S3 specifically includes the following steps in the flexible design process: first, a basic structure frame is predefined, from which the optimal basic structure under construction conditions is determined as a blueprint for subsequent design; second, the range of wind speeds that the structure may withstand is determined, from the construction conditions to the maximum extreme conditions, and then divided into smaller intervals by design points; the design points are equidistantly arranged or rearranged according to the design results; third, the structural response under extreme conditions is analyzed to find the weak areas in the structure, and several local strengthening measures are proposed for the specific areas; fourth, for each design point, the optimal structure is generated by multi-objective optimization based on the selected basic structure; the parameters of the multi-objective optimization include the setting parameters of various strengthening measures, and the targets are the maximum stress, maximum displacement and steel consumption of the structure; in the next level of optimization design, the strengthening measures set in the previous level should be retained.

2. The method of claim 1, wherein, The multi-objective optimization design step further comprises: Various machine learning algorithms are used to train the surrogate model, including but not limited to support vector machines, neural networks, and random forest algorithms; In the optimization algorithm, a constraint handling strategy is adopted to meet the requirements of structural safety and engineering practice; The cross-validation method is used to evaluate the prediction performance of the surrogate model, ensuring that the resulting model has good generalization ability and prediction accuracy. The fitness function of the multi-objective optimization algorithm is set to balance the relationship between each design objective and update the population during the iteration process to improve the diversity of solutions and optimization efficiency. The generated Pareto optimal solution set can reflect the impact of different strengthening measures on the performance of the structure and the performance trade-off under different working conditions, providing comprehensive design scheme selection for decision-makers.

3. The method of claim 1, wherein the method further comprises: determining a plurality of design parameters of the building structure; and determining a plurality of design parameters of the building structure based on the plurality of design parameters of the building structure and the plurality of design parameters of the building structure. The flexible design and optimization method of the building machine structure suitable for multiple working conditions includes: Parametric design and response calculation of the aerial building machine: The model establishment process is carried out in Grasshopper. First, the candidate points of support columns and the size of steel platforms are preset according to the core tube shear wall design drawings, and then several main truss layouts are preset. Then, the parametric building machine model is established, and the main structural components of the building machine are created based on module design and connected with each other according to the module rules. The auxiliary structure is simplified and equivalent to the load applied to the main structure. Through parametric programming statements, the software can set the position of the steel platform and combine it into a steel platform, set the position of the support column and connect it with the steel platform, and set the strengthening measures. Multi-objective optimization design based on artificial intelligence: The optimization design method combines the surrogate model with the multi-objective optimization algorithm for multi-objective optimization. The filling method is used to generate several combinations of independent variables, which are input into the finite element software for calculation to obtain the steel consumption, maximum stress, and maximum displacement response of the scheme, which together constitute the data set for training the surrogate model. Machine learning algorithms are used to train and test the stress, displacement, and steel consumption data to obtain three surrogate models. Through multiple iterations of the multi-objective optimization algorithm population in the variable space of the multiple strengthening measure setting parameters, a series of non-dominated solution sets on the Pareto surface are obtained for designers to choose from. The schemes in the solution set cannot obtain a better scheme for a certain objective without deteriorating other objectives, representing the optimal scheme under different objective weights. Designers can choose according to their requirements and preferences. Flexible design using multiple strengthening measures: The structure realizes the flexibility of resistance through various strengthening measures. The initial structure may not meet the resistance requirements under all conditions, but the resistance can be improved by adding strengthening measures when necessary, allowing the structure to dynamically adjust and become relatively optimal under each working condition. The scheme set is composed of design solutions generated for each time point. The flexible design adopts a progressive design mode, that is, by gradually increasing various strengthening measures on the same basic structural framework to adapt to different working conditions. SAP2000 is used for elastic linear analysis to obtain the response of the structure. In the elastic analysis, the response under each working mode is calculated independently, and the total result is obtained by linear combination, reducing the time cost of calculation. The selected responses are the maximum stress ratio, maximum point displacement, and steel consumption, which represent the resistance, stiffness, and cost of the structure and are also the optimization objectives in the optimization design.

4. The method of claim 1, wherein the method further comprises: determining a plurality of design parameters of the building structure; and determining a plurality of design parameters of the building structure based on the plurality of design parameters of the building structure and the plurality of design parameters of the building structure. S2 specifically includes: the agent model uses a deep forest algorithm, which includes multi-granularity scanning and a cascaded forest structure; multi-granularity scanning is used to extract features from input data at multiple scales or resolutions, including applying different sized sliding windows to the input data and extracting feature vectors from each window, using different window sizes allows the algorithm to capture different levels of detail in the input data; smaller window sizes capture local information and fine-grained patterns, while larger window sizes capture more global information and larger structures; the cascaded forest structure consists of multiple levels, each level contains a set of random forests, the forests receive input feature vectors extracted from multi-granularity scanning for training; in the decision tree, each leaf node is assigned a feature to split until only one type of instance is on the last leaf node, the final prediction is calculated based on the maximum value in the average vector of the final level of forest, in order to reduce the risk of overfitting, the class vector generated by each forest is generated by k-fold cross-validation, then averaged to generate the final class vector as the enhanced feature of the next level in the cascade; compared with deep neural networks, deep forests show obvious advantages, including significantly reducing the number of hyperparameters, and showing considerable robustness in performance under various hyperparameter settings, which makes it suitable for fitting requirements with strong nonlinear response; The multi-objective optimization algorithm uses the AGE-MOEA algorithm, which inherits the overall framework of NSGA-II and uses a survival score that combines diversity and proximity instead of crowded distance; in the first step, the algorithm generates an initial set P of N random solutions, then generates new solutions using crossover and mutation, offspring Q; the new population including the current population and the offspring population is non-dominant sorted to obtain the non-dominant front F; after normalizing the non-dominant front, a new population is generated and a survival score that considers diversity and proximity is assigned to the first non-dominant front; the new population of M solutions is selected from the non-dominant front in order until the solution exceeds M; then, the remaining solutions will be selected from the front F in descending order of survival score.

5. A flexible design and optimization system for a multi-condition adaptive building machine structure according to the method of any one of claims 1 to 4, characterized in that, It includes: Parametric design and response calculation module: used for parametric design and response calculation, the model building process is carried out in Grasshopper; first, according to the core tube shear wall design drawing, the support column candidate point and steel platform size are preset, and according to this, several main truss layouts are preset; Then, the parameterized building machine model is established, the main structure components of the building machine are created based on the module design, and they are connected with each other according to the module rule, among which the accessory structure is simplified and equivalent to the load applied to the main structure; through parameterized programming statements, the software can set the position of the steel platform according to the scheme parameters and combine it into a steel platform, set the position of the support column and connect it with the steel platform, and set the strengthening measures; Multi-objective optimization design module: a multi-objective optimization design method combining a surrogate model and a multi-objective optimization algorithm is used for multi-objective optimization; a filling method is used to generate a plurality of combination schemes of independent variables, which are substituted into a finite element software to obtain the steel consumption and maximum stress and maximum displacement response of the scheme, which together constitute a data set for training of the surrogate model; A machine learning algorithm is used to train and test the data of stress, displacement and steel consumption respectively, and three surrogate models are obtained; a series of non-dominated solution sets on the Pareto surface are obtained by multiple iterations of the multi-objective optimization algorithm population in the variable space of the parameters of various strengthening measures, for designers to select; the schemes in the solution set cannot obtain a better scheme for a certain target without degrading other targets, representing the optimal scheme under different target weights, which can be selected by designers according to requirements and preferences; Flexible design module: used for flexible design, the structure realizes resistance flexibility through various strengthening measures, the initial structure may not meet the resistance requirements under all conditions, but the resistance can be improved by adding strengthening measures when necessary, allowing the structure to dynamically adjust and become relatively optimal under each working condition; the scheme set is composed of design solutions generated for each time point; the flexible design adopts a progressive design mode, which gradually increases various strengthening measures to adapt to different working conditions by setting constraints in the optimization design process on the same basic structure framework. 6.A computer device, comprising a memory and a processor, the memory storing a computer program, the computer program being executed by the processor to make the processor execute the steps of the flexible design and optimization method for the structure of the building machine adapted to multiple working conditions according to any one of claims 1-4. 7.A computer readable storage medium, storing a computer program, the computer program being executed by the processor to make the processor execute the steps of the flexible design and optimization method for the structure of the building machine adapted to multiple working conditions according to any one of claims 1-4. 8.An information data processing terminal for implementing the flexible design and optimization system for the structure of the building machine adapted to multiple working conditions according to claim 5.

Images