201 0

Prediction of plastic anisotropy of sheet metals based on experimental indentation supported by artificial neural networks

Title
Prediction of plastic anisotropy of sheet metals based on experimental indentation supported by artificial neural networks
Author
Jiaping Xia
Alternative Author(s)
하가평
Advisor(s)
윤종헌
Issue Date
2022. 8
Publisher
한양대학교
Degree
Master
Abstract
This paper mainly proposes two kinds of artificial neural network (ANN) models for predicting plastic anisotropy properties of sheet metal using the spherical indentation test, which minimizes measurement time and costs, and simplifies the process. The proposed ANN models for predicting anisotropic properties can replace the traditional complex dimensionless analysis. Moreover, the models are not limited to predicting yield strength anisotropy but also further accurately predict the Lankford coefficient in different directions of materials. To obtain a large dataset for training the ANN, we newly construct an FE spherical indentation model which is suitable for sheet metal in consideration of actual compliance. The constructed FE model is utilized to simulate with one thousand elastoplastic parameter conditions in pure and alloyed engineering metals. We suggest the specific variables of the residual indentation mark which includes the height and length in different directions as input parameters, also with the indentation load-depth curve. Furthermore, according to whether the material properties of the rolling direction of the material are used as input parameters, two different prediction models for anisotropic properties are developed. Experimental validations have been conducted with three different sheet alloys, TRIP1180 steel, zinc alloy, and aluminum alloy 6063-T6, comparing the proposed ANN model and the uniaxial tensile test. In addition, machine vision was used to efficiently analyze the residual indentation marks and automatically measure the indentation profiles in different directions. The proposed ANN model exhibits remarkable performance in the prediction of the flow curves and Lankford coefficient of different directions.|본 연구에서는 측정 시간과 비용을 최소화하고 측정 절차를 단순화 시킨 구형 압입자를 이용한 압입 실험을 이용하여 판재의 소성 이방성 특성을 예측하기 위한 인공 신경망 모델을 제안합니다. 실제 판재에 대한 압입 실험 환경을 유한 요소 모 델로 구현하기 위해 실험 장치의 강성 및 감쇠 효과를 고려하였습니다. 유한 요소 모델을 이용하여 순수 금속 및 합금 범주의 1000개의 매개 변수 조건을 설정하여 인공 신경망 훈련시키기 위한 데이터를 구성하였습니다. 압입 하중에 따른 압입 깊이 곡선과 압입 흔적을 나타내는 변수들을 입력 매개 변수로 설정하였습니다. 각 방향의 높이와 길이를 포함하는 잔류 압흔의 형상은 재료의 이방성 특성을 분 석하는데 이용됩니다. 제안된 인공 신경망 모델과 단축 인장 시험의 결과를 비교 하기 위해 TRIP1180, Al 6063-T6, Zn 합금을 이용하였습니다. 또한 압흔의 효 율적인 분석을 위해 머신 비전을 이용하였으며, 머신 비전을 이용하여 방향별 압 흔에 대한 형상을 측정했습니다. 본 연구에서 제안하는 인공 신경망 모델은 다른 방향의 응력-변형도 곡선과 Lankford 계수의 예측에서 높은 성능을 보입니다.
URI
http://hanyang.dcollection.net/common/orgView/200000626754https://repository.hanyang.ac.kr/handle/20.500.11754/174929
Appears in Collections:
GRADUATE SCHOOL[S](대학원) > MECHANICAL DESIGN ENGINEERING(기계설계공학과) > Theses (Master)
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE