KR-102964187-B1 - AI(Artificial Intelligence)-based secondary steel structure 3D modeling automation system and method

KR102964187B1KR 102964187 B1KR102964187 B1KR 102964187B1KR-102964187-B1

Abstract

The AI-based auxiliary steel frame 3D modeling automation method according to the present invention comprises: a step in which a main steel frame search unit searches for main steel frames surrounding a pipe support selected by a user; a step in which a dataset generation unit sets a three-dimensional space containing the main steel frames searched by the main steel frame search unit centered on the pipe support, divides the space into N1 × N2 × N3 , and generates a dataset containing information for each volume; a step in which a data collection unit collects data checking the type and installation type of the pipe support, whether the main steel frame is included, and whether there is interference for each volume divided by the dataset generation unit; a step in which a data input unit processes the data collected by the data collection unit and inputs it into a machine learning library; a step in which an auxiliary steel frame assembly selection unit selects an auxiliary steel frame assembly to be installed on the main steel frame from the machine learning library; and a step in which a dimension assignment unit calculates the location and shape information of the main steel frame to be installed with the selected auxiliary steel frame assembly, the distance between reference points, and the dimensions of the pipe support, and automatically assigns the dimensions of each member. and includes a step of automatically generating a 3D model of an auxiliary steel frame assembly based on the dimensional data of each assigned member by a 3D model generation unit.

Inventors

이동훈
김영민
강신현
이준호
윤지훈
이재구

Assignees

성화산업 주식회사

Dates

Publication Date: 20260513
Application Date: 20250526
Priority Date: 20250417

Claims (14)

A cast steel search unit that searches for cast steel frames around a pipe support selected by the user; A dataset generation unit that sets a three-dimensional space including the cast steel frame searched by the cast steel frame search unit centered on the pipe support, divides the space into N1 × N2 × N3 , and generates a dataset including information for each volume; A data collection unit that collects data checking the type and installation type of the pipe support, whether cast steel is included, and whether there is interference for each volume classified by the above dataset generation unit; A data input unit that processes data collected by the above data collection unit and inputs it into a machine learning library; An auxiliary steel frame assembly selection unit that selects an auxiliary steel frame assembly to be installed on the main steel frame from the above machine learning library; A dimension assignment unit that automatically assigns dimensions to each member by calculating the location and shape information of the main steel frame to be installed with the selected auxiliary steel frame assembly, the distance between reference points, and the dimensions of the pipe support; and It includes a 3D model generation unit that automatically generates a 3D model of an auxiliary steel frame assembly based on the dimensional data of each of the above-mentioned assigned members, and An AI - based auxiliary steel frame 3D modeling automation system characterized by the above dataset generation unit setting a 3D space including a main steel frame and dividing the space into N1 × N2 × N3 using internal logic.
In paragraph 1, An AI-based auxiliary steel frame 3D modeling automation system characterized by the above-mentioned main steel frame search unit searching for main steel frames around a pipe support, wherein the search unit searches for structural type objects including steel frames and panels around the pipe support.
In paragraph 2, The above-described main steel frame search unit is characterized by further searching for objects where interference is expected between the main steel frame and the pipe support, in an AI-based auxiliary steel frame 3D modeling automation system.
delete
In paragraph 1, An AI-based auxiliary steel frame 3D modeling automation system characterized by the above-mentioned data input unit processing data and inputting it into a machine learning library by processing the data into a predefined format and inputting it as a text-format parameter.
In paragraph 1, An AI-based auxiliary steel frame 3D modeling automation system characterized in that, when the auxiliary steel frame assembly selection unit selects an auxiliary steel frame assembly to be installed on the main steel frame from the machine learning library, if the correlation between each type of auxiliary steel frame assembly output from the machine learning library is 0.5 or higher, the corresponding type is automatically selected, and if there are two or more types with similar correlations, the type selected by the user among the top two types is selected.
In paragraph 1, An AI-based auxiliary steel frame 3D modeling automation system characterized by the above-mentioned dimension allocation unit automatically allocating dimensions to each member by calculating the position and shape information of the main steel frame, the distance between reference points, and the dimensions of the pipe support, wherein the user is configured to directly input dimensions in special cases, including cases where a gap between the pipe support and the auxiliary steel frame is required or where a safety margin for the installation of the auxiliary steel frame is required.
a) A step in which a cast steel search unit searches for cast steel frames around a pipe support selected by a user; b) A step in which a dataset generation unit establishes a three-dimensional space containing the cast steel frame searched by the cast steel frame search unit centered on the pipe support, divides the space into N1 × N2 × N3 , and generates a dataset containing information for each volume; c) A step in which a data collection unit collects data checking the type and installation type of the pipe support, whether cast steel is included, and whether there is interference for each volume classified by the dataset generation unit; d) A step in which a data input unit processes data collected by the data collection unit and inputs it into a machine learning library; e) A step in which an auxiliary steel frame assembly selection unit selects an auxiliary steel frame assembly to be installed on the main steel frame from the machine learning library; f) a step in which a dimension allocation unit calculates the location and shape information of the main steel frame where the selected auxiliary steel frame assembly is to be installed, the distance between reference points, and the dimensions of the pipe support, and automatically assigns the dimensions of each member; and g) The 3D model generation unit includes the step of automatically generating a 3D model of the auxiliary steel frame assembly based on the dimensional data of each assigned member, and AI-based auxiliary steel frame 3D modeling automation method, characterized in that, in step b) above, the dataset generation unit sets a three-dimensional space including a main steel frame and divides the space into N1 × N2 × N3 using internal logic.
In paragraph 8, An AI-based auxiliary steel frame 3D modeling automation method characterized in that, in step a) above, the main steel frame search unit searches for main steel frames around the pipe support, and searches for structural type objects including steel frames and panels around the pipe support.
In Paragraph 9, The above-described main steel frame search unit further searches for objects where interference is expected between the main steel frame and the pipe support, an AI-based auxiliary steel frame 3D modeling automation method.
delete
In paragraph 8, An AI-based auxiliary steel frame 3D modeling automation method characterized in that, in step d) above, the data input unit processes the data and inputs it into a machine learning library by processing the data into a predefined format and inputting it as a text-format parameter.
In paragraph 8, AI-based auxiliary steel frame 3D modeling automation method, characterized in that, in step e) above, when the auxiliary steel frame assembly selection unit selects an auxiliary steel frame assembly to be installed on the main steel frame from the machine learning library, if the correlation between each type of auxiliary steel frame assembly output from the machine learning library is 0.5 or higher, the corresponding type is automatically selected, and if there are two or more types with similar correlations, the type selected by the user among the top two types is selected.
In paragraph 8, AI-based auxiliary steel frame 3D modeling automation method, characterized in that in step f) above, the dimension allocation unit calculates the position and shape information of the main steel frame, the distance between reference points, and the dimensions of the pipe support to automatically assign dimensions to each member, and in special cases including when a gap between the pipe support and the auxiliary steel frame is required or when a safety margin for the installation of the auxiliary steel frame is required, the dimensions are directly entered by the user.

Description

AI-based secondary steel structure 3D modeling automation system and method The present invention relates to an AI (Artificial Intelligence)-based auxiliary steel frame 3D modeling automation system, and more specifically, to an AI-based auxiliary steel frame 3D modeling automation system and method that can significantly reduce repetitive manual work by engineers and reduce design man-hours by training a machine learning model with data accumulated from past plant construction projects and using the machine learning model to automatically recommend the shape of an auxiliary steel frame assembly and perform 3D modeling. Pipe supports are auxiliary structures connected to plant structures such as primary steel and walls to properly support piping systems (pipes, cable trays, ducts, etc.) and maintain the function of the equipment, and may include brackets, supports, hangers, etc. Primary steel refers to steel members that form the basic framework supporting major loads to ensure the overall structural stability of the plant during construction. This primary steel is also referred to as "existing steel" or "main steel." Secondary steel is a steel structure installed on the main steel frame to perform additional functions. The existing secondary steel modeling process can be divided into manual modeling and assembly-based modeling. Manual modeling involves the user directly specifying the location of steel structures and modeling them one by one. While it has the advantage of increasing efficiency as the designer's skill level increases, it has the disadvantages of varying design quality depending on the designer and a high likelihood of dimensional errors and interference caused by human error. Assembly-based modeling, developed to complement manual modeling, is a modeling technique built on standardized data of combinations of repeatedly used steel members. While it offers the advantage of designing complex assemblies with a simple click, it has the disadvantage that the complexity of surrounding geometry variables results in an infinite number of possibilities, and the sheer volume of data to consider and the number of possible assembly applications make it difficult to apply appropriate steel assemblies. Conventional technologies as described above have the problem that design quality varies depending on the skill level of the designer, and that design assets accumulated as tacit knowledge are difficult to transfer as know-how. In addition, there is a risk of excessive material expenditure due to over-design caused by the application of inappropriate assemblies, and there is a high possibility of human error, such as interference between equipment and dimensional errors, due to manual work. Furthermore, there is a problem of rising labor costs as it becomes difficult to secure design personnel due to the decline in the domestic population. Meanwhile, Korean Registered Patent Publication No. 10-2734667 (Patent Document 1) discloses a "system for position selection and modeling of pipe supports." The system for position selection and modeling of pipe supports according to this invention builds a database of maximum spacing of pipe supports for pipe specifications to prevent the excessive use of unnecessary pipe supports, and automates the selection of the optimal location where pipe supports can be installed from the modeling information of the steel structure where the pipes are installed. Although this has the effect of maximizing the spacing between pipe supports while minimizing auxiliary steel frames by utilizing existing steel structures for position selection, this is a technology related to the position selection and modeling of pipe supports and has the problem that it is difficult to apply to the 3D modeling of auxiliary steel frames. FIG. 1 is a schematic diagram showing the configuration of an AI-based auxiliary steel frame 3D modeling automation system according to the present invention. FIG. 2 is a flowchart illustrating the execution process of an AI-based auxiliary steel frame 3D modeling automation method according to the present invention. Figure 3 is a diagram showing an overview of the main steel frame around the pipe support where the auxiliary steel frame assembly will be installed. Figure 4 is a diagram showing an overview of dividing a three-dimensional space containing a cast steel frame centered on a pipe support into N1 × N2 × N3 . Figure 5 is a diagram showing an overview of selecting an auxiliary steel frame assembly to be installed on a main steel frame in a machine learning library. Figure 6 is a diagram showing an overview of automatically assigning the dimensions of each member by calculating the location and shape information of the main steel frame to be installed, the distance between reference points, and the dimensions of the pipe supports. Terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary mea