You logged in as
Abstract
Fused Deposition Modeling (FDM) is one of the most widely used Additive Manufacturing technologies that extrude a melted plastic filament through a heated nozzle in order to build final physical models layer-by-layer. In this research, a case study is presented in order to optimize process performance of a low cost FDM 3D printer. Taguchi method was first employed for the experimental procedure design and nine test parts were built according to L9 orthogonal array. The examined process parameters were the deposition angle, layer thickness, and infill ratio each one having three levels. Infill pattern was constant to honeycomb selection. Fabrication time of ABS (Acrylonitrile-Butadiene-Styrene) 3D printed specimens was measured during experiments and analyzed by using Analysis of Means (ANOM) and Analysis of Variance (ANOVA) techniques. Shape accuracy was measured by considering the parts' dimensions in X, Y and Z axes and expressed as the overall error for control. Regression models were developed to use them as objective functions for a group of multi-objective optimization algorithms. Multi-objective Greywolf (MOGWO), multi-objective antlion (MOALO), multi-verse (MOMVO) and multi-objective dragonfly (MODA) algorithms where implemented to simultaneously optimize the bi-objective FDM optimization problem. To evaluate the algorithms and judge superiority with reference to the non-dominated solution sets obtained the hypervolume (area) indicator was adopted. It was verified that algorithms perform differently to the problem formulated for optimizing the FDM process.
Key Words
Fused Deposition Modeling; Additive Manufacturing; ABS; printing time; shape accuracy; process optimization
Address
Nikolaos A. Fountas and Nikolaos M. Vaxevanidis:Laboratory of Manufacturing Processes & Machine Tools (LMProMaT), Department of Mechanical Engineering Educators, School of Pedagogical and Technological Education (ASPETE),
Amarousion GR 151 22, Greece
John D. Kechagias and Aris C. Tsiolikas:Laboratory for Manufacturing Processes and Machine Tools, General Department,University of Thessaly, Gaiopolis Larissa, GR 41500, Greece
Abstract
Today, Gene selection in microarray data is one of the most challenging subjects in the fields of medicine and machine learning. Due to the large number of features and small number of samples in microarray datasets, choosing the desirable genes in these data is a difficult task. Among several methods which have been proposed for gene (feature) selection, ensemble and hybrid methods have attracted more attentions. The purpose of this paper is to find an optimal structure for hybrid-ensemble gene selection method that, by selecting the least number of the genes, yields the desired classification accuracy. For this purpose, the genetic algorithm is used as one of the most popular evolutionary optimization methods to accomplish an optimal hybrid-ensemble feature selection method. The performance of the proposed method is widely tested on 18 microarray datasets, and it is compared to those of the 10 well-known gene selection methods in terms of classification error rates and Gmean. Experimental results demonstrate that the obtained optimal method is considerably superior to the other competing methods over different evaluation methods and datasets.
Key Words
gene selection; high-dimensional data; hybrid methods; metaheuristic; filter methods; ensemble methods
Address
Amirreza Rouhi:Intelligent Data Processing Laboratory (IDPL), Department of Electrical Engineering, Shahid Bahonar University of Kerman, Kerman, Iran/ Department of Electronics and Information, Politecnico di Milano, Italy
Hossein Nezamabadi-pour:Intelligent Data Processing Laboratory (IDPL), Department of Electrical Engineering, Shahid Bahonar University of Kerman, Kerman, Iran
Abstract
Determination of compressive strength of concrete at early testing age is vital in many civil engineering applications. The strength at 7 or 14 days allows engineers to have confidence in the target strength and make a decision in case of unsuspected situations. In this study, the possibility to estimate the early compressive strength of concrete by a machine learning algorithm, namely the support vector regression (SVR), was investigated. To this aim, a database containing 324 data points was gathered from the available literature and use to develop the ML model. For the assessment of the accuracy, common statistical measurements, such as the Pearson correlation coefficient (R) and root mean square error (RMSE) were used. The results showed that the SVR model could successfully model the early compressive strength of concrete with R=0.92386 and RMSE=5.5089 MPa. The sensitivity analysis on the factors exhibiting a positive or negative effect on the early strength of concrete was conducted. The cement content was shown to have the most influential effect on the early development of concrete compressive strength.
Key Words
early compressive strength; concrete; support vector regression; machine learning
Address
Hai-Bang Ly and Binh Thai Pham:University of Transport Technology, Hanoi 100000, Vietnam
Abstract
This study introduced a new predictive approach for estimating the bearing capacity of driven piles. To this end, the required data based on literature such as hammer strikes, soil properties, geometry of the pile, and friction angle between pile and soil were gathered as a suitable database. Then, three predictive models i.e., gene expression programming (GEP), radial basis function type neural networks (RBFNN) and multivariate nonlinear regression (MVNR) were applied and developed for pile bearing capacity prediction. After proposing new models, their performance indices i.e., root mean square error (RMSE) and coefficient of determination (R2) were calculated and compared to each other in order to select the best one among them. The obtained results indicated that the RBFNN model is able to provide higher performance prediction level in comparison with other predictive techniques. In terms of R2, results of 0.9976, 0.9466 and 0.831 were obtained for RBFNN, GEP and MVNR models respectively, which confirmed that, the developed RBFNN model could be selected as a new model in piling technology. Definitely, other researchers and engineers can utilize the procedure and results of this study in order to get better design of driven piles.
Key Words
pile bearing capacity; radial basis function type artificial neural networks (RBFNN); multivariate nonlinear regression (MVNR); gene expression programming (GEP)
Address
Hooman Harandizadeh:Department of Civil Engineering. Faculty of Engineering. Shahid Bahonar University of Kerman.
Pajoohesh Sq. Imam Khomeni Highway. P.O. Box 76169133. Kerman, Iran
Danial Jahed Armaghani:Department of Civil Engineering. Faculty of Engineering. University of Malaya.
50603. Kuala Lumpur, Malaysia
Vahid Toufigh:Faculty of Civil and Surveying Engineering. Graduate University of Advanced Technology. Kerman, Iran
Abstract
Blasting is known as the most common approach for fragmenting rock in open-pit mines. Nevertheless, its side effects are not insignificant, for example, fly rock, ground vibration, dust, toxic by-products, air over-pressure, and back-break. These effects considerably alter the circumambient environment, particularly when pressure is higher than usual. This study proposed and compared four artificial intelligence models for predicting blast-induced air over-pressure, namely multi-layer perceptron (MLP), Random Forest (RF), isotonic regression (IR), and M5-Rules. The air over-pressure was selected as the output variable based on the input variables, i.e., stemming length (T), explosive charge per delay (W), burden (B), monitoring distance (R), and spacing (S). Several statistical performance indices, including coefficient of determination (R
Key Words
blast-induced air over-pressure; artificial intelligence techniques; earth science; quarry mine; soft computing
Address
Hoang Nguyen and Xuan-Nam Bui: Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology,18 Vien Street, Duc Thang Ward, Bac Tu Liem District, Hanoi, Vietnam/ Center for Mining, Electro-Mechanical research, Hanoi University of Mining and Geology,18 Vien st., Duc Thang ward, Bac Tu Liem dist., Hanoi, Vietnam
Panagiotis G. Asteris: Computational Mechanics Laboratory, School of Pedagogical and Technological Education, 14121 Heraklion, Athens, Greece
Quang Hieu-Tran and Phonepaserth Sukhanouvong: Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology,18 Vien Street, Duc Thang Ward, Bac Tu Liem District, Hanoi, Vietnam
Danial Jahed Armaghani: Department of Civil Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia
Masoud Monjezi: Department of Mining, Faculty of Engineering, Tarbiat Modares University, Tehran 14115-143, Iran
Manoj Khandelwal: School of Science, Engineering and Information Technology, Federation University Australia, Ballarat, Australia