Welding deformation simulation system and method in steel structure construction process

Abstract

The invention relates to a welding deformation simulation system and method in a steel structure construction process. The system comprises a data preprocessing system, a feature analysis system and a deformation process simulation system. Wherein the data preprocessing system comprises a data acquisition module and a data cleaning module; the feature analysis system comprises a geometric feature recognition module and a material and mechanical feature extraction module. The deformation process simulation system comprises a parameter fine tuning module, a multi-source influence analysis module and a process deformation analysis module. According to the system, a basic model is pre-trained based on welding deformation of a steel structure component with a typical section, key parameters are adjusted to ensure that the model can accurately simulate deformation conditions under various load combinations, and geometric, material and mechanical characteristics of a steel structure with an atypical section are converted into equivalent parameters which can be understood by the pre-trained model; therefore, the prediction capability of the model on the atypical section is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of building construction, and in particular relates to a welding deformation simulation system and method in a steel structure construction process. Background Art

[0002] During the construction of steel structures, welding deformation is a complex problem. It is affected by many factors such as welding process parameters, material properties, environmental conditions, and welding sequence, which makes the deformation behavior during welding difficult to predict and control. At present, the simulation of welding deformation of typical steel structures mainly uses traditional finite element analysis methods such as thermoelastic-plastic method and inherent strain method or neural network model to predict welding deformation. However, the traditional finite element analysis method has high calculation cost and requires continuous improvement of strain database. It does not consider the different degrees of deformation of components during steel structure manufacturing, transportation, and on-site installation, and cannot guarantee the accuracy and reliability of finite element analysis.

[0003] Therefore, how to provide a welding deformation simulation system and method for the steel structure construction process is a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0004] In view of the problem that the traditional prediction of welding deformation has low accuracy and reliability, the present invention proposes a welding deformation simulation system and method for the steel structure construction process. In the steel structure construction process, welding deformation is a complex problem. It is affected by many factors such as welding process parameters, material properties, environmental conditions and welding sequence, which makes the deformation behavior in the welding process difficult to predict and control. At present, the simulation of welding deformation of typical steel structures mainly adopts traditional finite element analysis methods such as thermoelastic-plastic method and inherent strain method or neural network model to predict welding deformation. However, the traditional finite element analysis method has high calculation cost and needs to continuously improve the strain database. It does not take into account the different degrees of deformation of components during the steel structure manufacturing, transportation and on-site installation process, and cannot guarantee the accuracy and reliability of finite element analysis.

[0005] In order to solve the above technical problems, the present invention includes the following technical solutions:

[0006] A welding deformation simulation system for a steel structure construction process, comprising:

[0007] A data preprocessing system, wherein the data preprocessing system collects welding deformation data of typical cross-section steel structure components at different construction stages;

[0008] A feature analysis system, wherein the feature analysis system analyzes the features of the atypical cross-section steel structure and extracts geometric features, material features, and mechanical features of the atypical cross-section steel structure;

[0009] The deformation process simulation system integrates the prediction results of the model into various construction management processes to simulate the different degrees of deformation produced during the production, transportation and on-site installation of steel structure components.

[0010] Furthermore, the data preprocessing system includes two modules: a data acquisition module and a data cleaning module. The data collected by the data acquisition module include: welding process parameters, material properties, environmental conditions, welding sequence and results of welding deformation; the data cleaning module preprocesses the data, including normalization processing to eliminate the influence of dimensions, and cleaning the data to remove outliers and noise.

[0011] Furthermore, the feature analysis system includes two modules: a geometric feature recognition module and a material and mechanical feature extraction module. The geometric feature recognition module maps the features of the atypical cross-section to a feature space that can be understood by the pre-trained model, and decomposes it into a combination of several simple geometric shapes, calculates equivalent geometric parameters, and uses these equivalent parameters as input features of the pre-trained model; the material and mechanical feature extraction module is used to integrate the extracted material properties and mechanical features into a set of feature vectors, which will be used in the training and prediction process of the model.

[0012] Furthermore, the deformation process simulation system includes three modules: a parameter fine-tuning module, a multi-source impact analysis module and a process deformation analysis module. The parameter fine-tuning module selects a machine learning model to pre-train the collected data; the multi-source impact analysis module considers various influencing factors in the welding process to process multi-source data and improve prediction accuracy; the process deformation analysis module applies the prediction results of the model to the actual construction process, and simulates the welding deformation that may occur in the steel structure components during processing, transportation and on-site installation.

[0013] Furthermore, the parameter fine-tuning module pre-trains the collected data, including: during the pre-training process, adjusting the parameters of the model so that the model can accurately simulate the deformation of the component under various load combinations, and the parameters include material constitutive relationship parameters, unit type selection, and meshing parameters in the finite element model.

[0014] Furthermore, the multi-source impact analysis module considers multiple influencing factors in the welding process, and the multiple influencing factors include material anisotropy, welding sequence and environmental conditions.

[0015] The present invention also provides a method for simulating welding deformation during steel structure construction, the method comprising:

[0016] Step S1, using the welding deformation simulation system of the steel structure construction process;

[0017] Step S2, completing the welding deformation data of typical cross-section steel structure components at different construction stages through the data acquisition module of the data preprocessing system, and preprocessing the data through the data cleaning module;

[0018] Step S3, decomposing the features of the atypical cross section into a combination of geometric shapes through the geometric feature recognition module of the feature analysis system, calculating the equivalent moment of inertia or equivalent area, and using it as the input feature of the pre-training model; integrating the yield strength, elastic modulus, thermal expansion coefficient, elasticity, plasticity and fracture characteristics into a set of feature vectors through the material and mechanical feature extraction module;

[0019] Step S4: Pre-train the collected data by selecting a machine learning model through the deformation process simulation system, adjust the model parameters so that the model can accurately simulate the deformation of the components under various load combinations, and use random forests or gradient boosting machines to process multi-source data and improve prediction accuracy, and apply the prediction results of the model to the actual construction process, so as to simulate the welding deformation that may occur in the steel structure components during processing, transportation and on-site installation.

[0020] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0021] The present invention provides a welding deformation simulation system and method for a steel structure construction process, the system comprising a data preprocessing system, a feature analysis system and a deformation process simulation system. The data preprocessing system comprises two modules, a data acquisition module and a data cleaning module; the feature analysis system comprises two modules, a geometric feature recognition module and a material and mechanical feature extraction module; the deformation process simulation system comprises three modules, a parameter fine-tuning module, a multi-source influence analysis module and a process deformation analysis module. The welding deformation simulation system for a steel structure component construction process taking into account multi-source influence pre-trains a basic model based on the welding deformation of a steel structure component with a typical cross section, and adjusts key parameters, such as material constitutive relations, unit types and mesh divisions, to ensure that the model can accurately simulate deformation conditions under a variety of load combinations. The geometric, material and mechanical characteristics of atypical cross-section steel structures are converted into equivalent parameters that can be understood by the pre-trained model, such as equivalent moment of inertia and area, thereby improving the model's prediction ability for atypical cross sections.

[0022] Compared with the traditional method, the beneficial effects of the present invention are mainly as follows:

[0023] (1) Generalization ability: Mapping features such as atypical cross-sections and materials into a feature space that can be understood by the pre-trained model enhances the model’s generalization ability for different types of steel structures.

[0024] (2) Comprehensive consideration of multi-source influencing factors: Comprehensive consideration of various influencing factors in the welding process, which is often difficult to achieve in traditional simulation methods, thereby improving the reliability of simulation results.

[0025] (3) Integration of construction process: The prediction results of the model are integrated into the specific construction process so that the simulation results can be directly applied to each actual construction process, thus improving the construction efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a system architecture diagram for simulating welding deformation during steel structure construction in one embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following is a further detailed description of a welding deformation simulation system and method for steel structure construction provided by the present invention in conjunction with the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description.

[0028] Embodiment 1

[0029] Combine the following Figure 1 , a welding deformation simulation system for the steel structure construction process of the present invention is described in detail.

[0030] Please refer to Figure 1 , a welding deformation simulation system for the steel structure construction process, including: a data preprocessing system, a feature analysis system and a deformation process simulation system. Among them, the data preprocessing system includes two modules: a data acquisition module and a data cleaning module; the feature analysis system includes two modules: a geometric feature recognition module and a material and mechanical feature extraction module; the deformation process simulation system includes three modules: a parameter fine-tuning module, a multi-source influence analysis module and a process deformation analysis module. The welding deformation simulation system for the steel structure component construction process considering multi-source influence pre-trains the basic model based on the welding deformation of the steel structure component with a typical cross-section, and adjusts key parameters such as material constitutive relations, unit types and mesh divisions to ensure that the model can accurately simulate the deformation under various load combinations. And the geometric, material and mechanical characteristics of the atypical cross-section steel structure are converted into equivalent parameters that the pre-trained model can understand, such as equivalent moment of inertia and area, so as to improve the model's prediction ability for atypical cross-sections.

[0031] In this embodiment, more preferably, the data preprocessing system is mainly responsible for collecting welding deformation data of typical cross-section steel structure components at different construction stages; the feature analysis system is mainly responsible for analyzing the characteristics of atypical cross-section steel structures, and extracting their geometric characteristics, material characteristics and mechanical characteristics; the deformation process simulation system is mainly responsible for considering multi-source influencing factors, integrating the prediction results of the model into various construction management processes, and simulating the different degrees of deformation generated during the production, transportation, and on-site installation of steel structure components.

[0032] In this embodiment, more preferably, the data preprocessing system is composed of two modules: a data acquisition module and a data cleaning module. The data acquisition module collects welding deformation data of typical cross-section steel structure components at different construction stages. These data should include welding process parameters, material properties, environmental conditions, welding sequence, and welding deformation results. The data cleaning module preprocesses the data, including normalization to eliminate the influence of dimensions, and cleaning the data to remove outliers and noise.

[0033] In this embodiment, more preferably, the feature analysis system is composed of two modules: a geometric feature recognition module and a material and mechanical feature extraction module. The geometric feature recognition module maps the features of the atypical cross-section to a feature space that can be understood by the pre-trained model, decomposes it into a combination of several simple geometric shapes, calculates equivalent geometric parameters (such as equivalent moment of inertia, equivalent area, etc.), and uses these equivalent parameters as input features of the pre-trained model. The material and mechanical feature extraction module is used to integrate the extracted material properties (such as yield strength, elastic modulus, thermal expansion coefficient, etc.) and mechanical characteristics (including elasticity, plasticity, fracture characteristics, etc.) into a set of feature vectors, which will be used in the training and prediction process of the model.

[0034] In this embodiment, more preferably, the deformation process simulation system is composed of three modules: a parameter fine-tuning module, a multi-source impact analysis module, and a process deformation analysis module. The parameter fine-tuning module selects a suitable machine learning model to pre-train the collected data. During the pre-training process, the parameters of the model, such as the material constitutive relationship parameters, unit type selection, meshing parameters, etc. in the finite element model, are adjusted so that the model can accurately simulate the deformation of the component under various load combinations. The multi-source impact analysis module considers various influencing factors in the welding process, such as the anisotropy of the material, the welding sequence, the environmental conditions, etc., and uses machine learning algorithms, such as random forests or gradient boosting machines (GBM), to process multi-source data and improve the prediction accuracy. The process deformation analysis module applies the prediction results of the model to the actual construction process and simulates the welding deformation that may occur in the steel structure components during processing, transportation and on-site installation.

[0035] Please continue to refer to Figure 1The present invention also provides a method for simulating welding deformation during steel structure construction, the method comprising:

[0036] Step S1, using the welding deformation simulation system of the steel structure construction process;

[0037] Step S2, completing the welding deformation data of typical cross-section steel structure components at different construction stages through the data acquisition module of the data preprocessing system, and preprocessing the data through the data cleaning module;

[0038] Step S3, decomposing the features of the atypical cross section into a combination of geometric shapes through the geometric feature recognition module of the feature analysis system, calculating the equivalent moment of inertia or equivalent area, and using it as the input feature of the pre-training model; integrating the yield strength, elastic modulus, thermal expansion coefficient, elasticity, plasticity and fracture characteristics into a set of feature vectors through the material and mechanical feature extraction module;

[0039] Step S4: Pre-train the collected data by selecting a machine learning model through the deformation process simulation system, adjust the model parameters so that the model can accurately simulate the deformation of the components under various load combinations, and use random forests or gradient boosting machines to process multi-source data and improve prediction accuracy, and apply the prediction results of the model to the actual construction process, so as to simulate the welding deformation that may occur in the steel structure components during processing, transportation and on-site installation.

[0040] The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. The above embodiments only express several embodiments of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A welding deformation simulation system for steel structure construction process, characterized in that: include: A data preprocessing system, wherein the data preprocessing system collects welding deformation data of typical cross-section steel structure components at different construction stages; A feature analysis system, wherein the feature analysis system analyzes the features of the atypical cross-section steel structure and extracts geometric features, material features, and mechanical features of the atypical cross-section steel structure; The deformation process simulation system integrates the prediction results of the model into various construction management processes to simulate the different degrees of deformation produced during the production, transportation and on-site installation of steel structure components.

2. The welding deformation simulation system for steel structure construction process according to claim 1 is characterized in that: The data preprocessing system includes two modules: a data acquisition module and a data cleaning module. The data collected by the data acquisition module include: welding process parameters, material properties, environmental conditions, welding sequence and welding deformation results; the data cleaning module preprocesses the data, including normalization processing to eliminate the influence of dimensions, and cleaning the data to remove outliers and noise.

3. The welding deformation simulation system for steel structure construction process according to claim 1 is characterized in that: The feature analysis system includes two modules: a geometric feature recognition module and a material and mechanical feature extraction module. The geometric feature recognition module maps the features of the atypical cross section to a feature space that can be understood by the pre-trained model, decomposes it into a combination of several simple geometric shapes, calculates equivalent geometric parameters, and uses these equivalent parameters as input features of the pre-trained model; The material and mechanical feature extraction module is used to integrate the extracted material properties and mechanical features into a set of feature vectors, which will be used in the model training and prediction process.

4. The welding deformation simulation system for steel structure construction process according to claim 1, characterized in that: The deformation process simulation system includes three modules: a parameter fine-tuning module, a multi-source impact analysis module and a process deformation analysis module. The parameter fine-tuning module selects a machine learning model to pre-train the collected data; the multi-source impact analysis module considers various influencing factors in the welding process to process multi-source data and improve prediction accuracy; the process deformation analysis module applies the prediction results of the model to the actual construction process and simulates the welding deformation that may occur in the steel structure components during processing, transportation and on-site installation.

5. The welding deformation simulation system for steel structure construction process according to claim 4 is characterized in that: The parameter fine-tuning module pre-trains the collected data, including: during the pre-training process, adjusting the parameters of the model so that the model can accurately simulate the deformation of the component under various load combinations, and the parameters include material constitutive relationship parameters, unit type selection, and meshing parameters in the finite element model.

6. The welding deformation simulation system for steel structure construction process according to claim 4 is characterized in that: The multi-source impact analysis module considers multiple influencing factors in the welding process, including material anisotropy, welding sequence, and environmental conditions.

7. A method for simulating welding deformation during steel structure construction, characterized in that: include: Step S1, using the welding deformation simulation system for the steel structure construction process according to any one of claims 1 to 6; Step S2, completing the welding deformation data of typical cross-section steel structure components at different construction stages through the data acquisition module of the data preprocessing system, and preprocessing the data through the data cleaning module; Step S3, decomposing the features of the atypical cross section into a combination of geometric shapes through a geometric feature recognition module of a feature analysis system, calculating the equivalent moment of inertia or equivalent area, and using it as an input feature of a pre-trained model; The yield strength, elastic modulus, thermal expansion coefficient, elasticity, plasticity and fracture characteristics are integrated into a set of feature vectors through the material and mechanical feature extraction module; Step S4: Pre-train the collected data by selecting a machine learning model through the deformation process simulation system, adjust the model parameters so that the model can accurately simulate the deformation of the components under various load combinations, and use random forests or gradient boosting machines to process multi-source data and improve prediction accuracy, and apply the prediction results of the model to the actual construction process, so as to simulate the welding deformation that may occur in the steel structure components during processing, transportation and on-site installation.

Images