Multi-modal input deep neural network and frame structure beam column design method and device

Abstract

The invention provides a multi-modal input deep neural network and a frame structure beam column design method and device. The multi-modal input deep neural network comprises a convolutional neural network column arrangement module and a graph neural network beam arrangement module. a convolutional neural network column arrangement module comprises a feature coding fusion network and an image generation network; a feature coding fusion network is used for carrying out feature fusion; an image generation network is used for generating a column arrangement image based on the fusion features; a graph neural network beam arrangement module comprises a column node and side information extraction network and a graph feature generation network; a column node and side information extraction network is used for extracting node information from the column arrangement image and extracting side information in combination with the column arrangement image, the building load partition image information and the building function partition image information; a map feature generation network is used for generating map features based on the node information and the side information. According to the key building image and design information, frame structure beam column design can be quickly completed.

Description

technical field [0001] The invention relates to the technical fields of building structure design and machine learning, in particular to a multimodal input deep neural network, a frame structure beam column design method and device. Background technique [0002] In the design of the frame structure building scheme and the initial design of the structure, in order to ensure the safety and rationality of the final design results, it is necessary to conduct a rapid and reasonable beam-column preliminary analysis based on the building function zoning, building load zoning, and some design information. design. [0003] A good and fast preliminary design scheme of beams and columns can not only assist the update and optimization of architectural schemes, but also assist the later deepening design of structural schemes. However, the current beam-column design method for frame structures that relies on manual experience is time-consuming, and the interactive design efficiency betwe...

AI Technical Summary

Problems solved by technology

However, the current beam-column design method for frame structures that relies on manual experience is time-consuming, and the interactive design efficiency between construction engineers and structural engineers is low; manual design relies on the experience of designers, resulting in differences in the design results of different designers, making it difficult to coordinate and coordinate. Considering the experience of many excellent designers, it is difficult to make full use of the existing design drawing resources, and it is difficult to fully absorb the historical design experience

At the same time, the existing computer-aided frame structure beam-column design method consumes a lot of computing resources, takes a long time, depends on the selection of objective functions, has poor universality, and is difficult to effectively apply the existing mature design results

The existing artificial intelligence-assisted structural design methods cannot carry out the integrated design of beams and columns of frame structures, and cannot take into account the efficient image processing capabilities of convolutional neural networks and the efficient topological relationship processing capabilities of graph neural networks. The image data in the structural beam-column design plan data and the topological relationship between the beam-column are considered comprehensively

Existing methods are time-consuming, inefficient, and difficult to use existing design drawing resources, which is not conducive to iterative changes in design schemes, making it difficult to meet the rapid design requirements in the preliminary design stage of beams and columns of frame structures

Images