TY - JOUR
T1 - Optimizing the hardness of SLA printed objects by using the neural network and genetic algorithm
AU - Hu, Guang
AU - Cao, Zhi
AU - Hopkins, Michael
AU - Hayes, Conor
AU - Daly, Mark
AU - Zhou, Haiying
AU - Devine, Declan M.
N1 - Publisher Copyright:
© 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
PY - 2019
Y1 - 2019
N2 - In the developing field of manufacturing, 3D printing is rapidly increasing the horizon of what is possible. However, the possibility of implementing new 3D methods of production has brought new challenges for industries, particularly in the case of changing traditional mind-sets about methods of manufacturing. This is due to the traditional and fixed mind-sets of experienced designers and of course owing to the lack of knowledge on 3D printing. In this paper, 3D printing processes were optimized by using a new algorithm; this advanced algorithm is created by combining the characteristics of an artificial neural network (ANN) and a genetic algorithm (GA). Furthermore, the print efficiency and quality of final products can be improved by optimizing 3D printing experimental conditions. In the current study, stereolithography (SLA) was employed as the 3D printing technique. This particular technique is commonly used to fabricate solid objects that are photochemically solidified. Based on previous research results, three main contents of process planning in 3D printing were defined and used as input to build the ANN model to predict the hardness. With orientation ranging from 0 to 90 degrees, ultraviolet post-curing (UV curing) time ranging from 20 to 60 minutes and annealing time from 0 to 4 hours, over 100 samples were tested to create a large sample set. It was observed that the orientation had the most significant impact while UV curing time had the lowest significant impact on the printed object's hardness. In addition, based on the hardness results, the predicted orientation of 0 degrees, UV curing time of 60 minutes and an annealing time of 2.88 hours were the optimum experimental conditions for the final printed object's hardness. From this study, it was concluded the new algorithm could be used to optimize the hardness of printed objects and to provide key information for the improvement of existing 3D printing technology.
AB - In the developing field of manufacturing, 3D printing is rapidly increasing the horizon of what is possible. However, the possibility of implementing new 3D methods of production has brought new challenges for industries, particularly in the case of changing traditional mind-sets about methods of manufacturing. This is due to the traditional and fixed mind-sets of experienced designers and of course owing to the lack of knowledge on 3D printing. In this paper, 3D printing processes were optimized by using a new algorithm; this advanced algorithm is created by combining the characteristics of an artificial neural network (ANN) and a genetic algorithm (GA). Furthermore, the print efficiency and quality of final products can be improved by optimizing 3D printing experimental conditions. In the current study, stereolithography (SLA) was employed as the 3D printing technique. This particular technique is commonly used to fabricate solid objects that are photochemically solidified. Based on previous research results, three main contents of process planning in 3D printing were defined and used as input to build the ANN model to predict the hardness. With orientation ranging from 0 to 90 degrees, ultraviolet post-curing (UV curing) time ranging from 20 to 60 minutes and annealing time from 0 to 4 hours, over 100 samples were tested to create a large sample set. It was observed that the orientation had the most significant impact while UV curing time had the lowest significant impact on the printed object's hardness. In addition, based on the hardness results, the predicted orientation of 0 degrees, UV curing time of 60 minutes and an annealing time of 2.88 hours were the optimum experimental conditions for the final printed object's hardness. From this study, it was concluded the new algorithm could be used to optimize the hardness of printed objects and to provide key information for the improvement of existing 3D printing technology.
UR - http://www.scopus.com/inward/record.url?scp=85083532614&partnerID=8YFLogxK
U2 - 10.1016/j.promfg.2020.01.016
DO - 10.1016/j.promfg.2020.01.016
M3 - Conference article
AN - SCOPUS:85083532614
SN - 2351-9789
VL - 38
SP - 117
EP - 124
JO - Procedia Manufacturing
JF - Procedia Manufacturing
T2 - 29th International Conference on Flexible Automation and Intelligent Manufacturing, FAIM 2019
Y2 - 24 June 2019 through 28 June 2019
ER -
