APPLICATION OF ARTIFICIAL NEURAL NETWORK (ANN) TO CALCULATE 3D CONCRETE PRINTER LARGE-SIZED

Abstract

The article presents the application of artificial neural network (ANN) to
calculate machine frame structure in 3D concrete large-sized gate in the
construction industry. 3D model of machine was designed on Inventor software.
The finite element model (FEM) of the frame was built in the NX Nastran
environment to evaluate the error of different axes caused by the elastic
deformation of the structure. In order to apply ANN in computational
configuration, the authors build a program to automatically calculate and analyze
the structure of the moving component (X and Z axis) of the machine. The
program to calculate vibrational frequency and elastic deformation of the
structure is built based on ANSYS APDL connected with MATLAB software. Based
on this auto-computation program, the ANN network training dataset was
generated. The research results have important significance in the design and
development and are the premise for calculating and choosing the optimal
machine structure option in large-size 3D, applying optimization algorithms.