Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending
碩士 === 國立高雄第一科技大學 === 機械與自動化工程所 === 95 === All sheet metal forming operation incorporate some bending; often it is the major feature. After forming, some elastic springback occurs and considerable residual stresses may result. Accurate prediction and controlling of springback is essential for the de...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2007
|
Online Access: | http://ndltd.ncl.edu.tw/handle/61609839669014215059 |
id |
ndltd-TW-095NKIT5689022 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-095NKIT56890222016-05-20T04:18:04Z http://ndltd.ncl.edu.tw/handle/61609839669014215059 Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending 應用適應性模糊系統於U型彎回彈預測 Wen-cheng Hsu 許文政 碩士 國立高雄第一科技大學 機械與自動化工程所 95 All sheet metal forming operation incorporate some bending; often it is the major feature. After forming, some elastic springback occurs and considerable residual stresses may result. Accurate prediction and controlling of springback is essential for the design of sheet metal forming tools. In this paper, we used the finite element software Dynaform 5.1 to collect the springback data of U-bending. Three main sheet metal forming parameters are the radius of punch(Rp), the radius of die corner(Rd) and punch-die clearance(C) with four levels respectively. A predictive model for springback is developed by using Adaptive Neuro-Fuzzy Inference System (ANFIS). We can carry out the construction of predicting model to provide the available reference for die designer. Besides, parameters influencing springback have been evaluated quantitatively using the Taguchi method. We obtained the training data(64 sets) and the checking data(27 sets) by finite element analysis, respectively. The training datas are applied to construct the predicting model for the springback of U-bending. In addition, the check data are employed to check the accuracy for the predicting model. At the same time, in the ANFIS, the types of membership function for the antecedent part will affect the accuracy of the predicting model. Therefore, four different membership functions are triangle function, trapezoid function, generalized bell function and Gauss function were used during the process of ANFIS in this study, so that we can select the better membership function of the antecedent part for the predicting model. Comparing the predicted values with the analysis results, we can find that the Gauss membership function can get the error least. The L΄16 Taguchi analysis results showed that the contribution of three parameters: the radius of punch(Rp) is 59.32%, the radius of die corner(Rd) is 1.94%, and punch-die clearance(C) is 37.12%. The main influencing parameter of springback is the radius of punch. Bor-tsuen Lin 林栢村 2007 學位論文 ; thesis 93 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立高雄第一科技大學 === 機械與自動化工程所 === 95 === All sheet metal forming operation incorporate some bending; often it is the major feature. After forming, some elastic springback occurs and considerable residual stresses may result. Accurate prediction and controlling of springback is essential for the design of sheet metal forming tools.
In this paper, we used the finite element software Dynaform 5.1 to collect the springback data of U-bending. Three main sheet metal forming parameters are the radius of punch(Rp), the radius of die corner(Rd) and punch-die clearance(C) with four levels respectively. A predictive model for springback is developed by using Adaptive Neuro-Fuzzy Inference System (ANFIS). We can carry out the construction of predicting model to provide the available reference for die designer. Besides, parameters influencing springback have been evaluated quantitatively using the Taguchi method.
We obtained the training data(64 sets) and the checking data(27 sets) by finite element analysis, respectively. The training datas are applied to construct the predicting model for the springback of U-bending. In addition, the check data are employed to check the accuracy for the predicting model. At the same time, in the ANFIS, the types of membership function for the antecedent part will affect the accuracy of the predicting model. Therefore, four different membership functions are triangle function, trapezoid function, generalized bell function and Gauss function were used during the process of ANFIS in this study, so that we can select the better membership function of the antecedent part for the predicting model. Comparing the predicted values with the analysis results, we can find that the Gauss membership function can get the error least. The L΄16 Taguchi analysis results showed that the contribution of three parameters: the radius of punch(Rp) is 59.32%, the radius of die corner(Rd) is 1.94%, and punch-die clearance(C) is 37.12%. The main influencing parameter of springback is the radius of punch.
|
author2 |
Bor-tsuen Lin |
author_facet |
Bor-tsuen Lin Wen-cheng Hsu 許文政 |
author |
Wen-cheng Hsu 許文政 |
spellingShingle |
Wen-cheng Hsu 許文政 Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending |
author_sort |
Wen-cheng Hsu |
title |
Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending |
title_short |
Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending |
title_full |
Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending |
title_fullStr |
Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending |
title_full_unstemmed |
Application of Adaptive Neuro-Fuzzy Inference System on Predicting Springback of U-Bending |
title_sort |
application of adaptive neuro-fuzzy inference system on predicting springback of u-bending |
publishDate |
2007 |
url |
http://ndltd.ncl.edu.tw/handle/61609839669014215059 |
work_keys_str_mv |
AT wenchenghsu applicationofadaptiveneurofuzzyinferencesystemonpredictingspringbackofubending AT xǔwénzhèng applicationofadaptiveneurofuzzyinferencesystemonpredictingspringbackofubending AT wenchenghsu yīngyòngshìyīngxìngmóhúxìtǒngyúuxíngwānhuídànyùcè AT xǔwénzhèng yīngyòngshìyīngxìngmóhúxìtǒngyúuxíngwānhuídànyùcè |
_version_ |
1718273635523756032 |