Measurement of environmental aspect of 3-D printing process using soft computing methods
Introduction
Mass and customized production of complex shaped products contributes to higher economic growth. During production, the need for energy and materials grows exponentially, which is not considered as environmentally sustainable. As a fairly new concept and advanced technology for the mass customization of products, 3-D printing could be quite beneficial to environmental sustainability. It is because the technology uses the polymer PLA (Polyactic acid) as a material, which is biodegradable, and saves fuel and reduces waste when fabricating prototypes. In addition, the technology can be located near to industries and fabricate raw material itself, resulting in reduction of transport costs and carbon emission.
Among the 3-D Printing processes (Fused deposition modeling (FDM), selective laser sintering (SLS), selective laser melting and stereolithography), SLS deploys a laser beam to selectively fuse powder into a designed solid object layer by layer [1]. Prior studies found that the features of SLS fabricated components including porosity strength, density, shrinkage ratio, etc., have high dependence on parameters such as the material and powder properties and other machine specifications such as laser power, scan speed and scan spacing. For attaining higher environmental sustainability across the globe, this process should be increasingly adopted as a manufacturing procedure. However, due to its high production cost and increased power consumption, the technology is not yet being widely adopted.
It is important to choose the optimum input parameter values (including laser power) since the quality of the components fabricated from the SLS process can be improved resulting in higher productivity and improved environmental performance [2], [3]. In this context, it is important to understand the process behavior. Due to complexity of the process, it is difficult to understand the nature of the effect of these parameters on the component characteristics. To the best of authors’ knowledge, this problem is not just associated with SLS process but also with other major 3-D printing processes such as FDM and stereolithography. This problem of understanding the effect of process parameters has indeed shifted the focus and motivated researchers towards pursuing the investigations on the modeling of additive manufacturing processes [4], [5], [6], [7].
In the perspective of modeling additive manufacturing processes, Garg et al. [2], recently conducted a survey on applications of empirical modeling methods in various processes such as FDM and SLS. It is worth noting that extensive studies were focused on formulating the models for the characteristics of density and shrinkage ratio of the SLS fabricated parts [8], [9], [10], [11], [12], [13], [14]. Only a few studies related to the porosity characteristics were conducted. The models, based on the physics behind the process, can be formulated. However, it may be a difficult task because the SLS process is dynamic and complex in nature due to the occurrence of multiple phenomena, such as transmission and absorption of energy, heating of the powder bed, sintering and cooling of the components [3].
It would be useful to develop a scientific approach for formulating models based on available data since the hidden principles behind the process on using these models could be understood [15], [16], [17], [18], [19], [20], [21], [22]. From the literature, it was found that the novel advanced optimization methods are proposed by hybridizing differential evolution algorithm with receptor editing property of immune system [23], [24], [25], artificial bee colony algorithm with Taguchi’s method [26], [27], differential algorithm with Taguchi’s method [28], cuckoo search algorithm (CS) [29] and immune algorithm with hill climbing local search algorithm [30], [31] for optimization of properties of materials. To meet this objective, several well-known soft computing methods, such as genetic programming (GP), artificial neural networks (ANN), and support vector regression (SVR), can be applied to formulate the relationship between the output and input process parameters of the RP processes. One objective of the present work (Fig. 1) is to explore the ability of these soft computing methods in the prediction of the open porosity of an SLS fabricated prototype. Experiments on SLS are conducted with the measurement of open porosity of the fabricated prototype based on the three input variables (the layer thickness, the laser power and the laser scan speed). The methods are applied on the data obtained from the experiments and its performance is compared using the statistical metrics.
Section snippets
Experimental set-up
The experimental materials being tested is a mixture of ASTM F1185-88 standard hydroxyapatite powder (P218R, Plasma Biotal Ltd, UK) and a standard SLS polymer powder, Polyamide-12 (PA-12) (Duraformw, 3D Systems, Herts, UK). Details of operating conditions of SLS process are given in Savalani et al. [32]. Within the same ratio which was mentioned in Bonfield et al. [33] on HAPEX® , given material was compounded in a Betol BTS40L twin screw extruder (Betol, Luton, UK) to manufacture the HA–PA
Genetic programming
The framework of the GP (Fig. 2) is based on the principle of genetic algorithms. In GP, genes are evolved and every gene is considered as a model. The procedure of GP is discussed in four steps as follows:
Evaluation of models against the 3-D printing experimental data
In this study, the parameter settings of the three soft computing methods, namely GP, SVR and ANN, are adjusted using a trial-and-error approach.
In addition, previous applications of GP used for modeling various complex industrial processes are studied for the settings of GP parameters [38], [39], [40], [41], [42]. The software GPTIPS [43], [44] with a single gene as a setting is used to perform GP in the prediction of open porosity (Fig. 3). Population size of 300, generations of 1000,
Robustness of the laser power based open porosity model by sensitivity analysis
After evaluating the performance of the proposed models, this section attempts to validate the robustness of the best model, i.e. GP, by performing the sensitivity analysis (SA) about the mean. The main reason behind performing the analysis is to perform a check whether the information obtained from the models matches with those from the actual understanding about the experimental studies [32]. Description of the performed sensitivity analysis is given as follows:
The SA percentage of the output
Conclusions
In view of improving the environmental performance of the SLS process, the present work proposed three soft computing methods, namely GP, SVR and ANN, in formulating the models. The formulation procedure of laser power-based-open porosity models using the three methods is discussed. The results show that the GP model outperforms the other two models. The sensitivity and parametric analysis validates the robustness of the laser power based GP model by unveiling relationships between open
Acknowledgments
The study was supported by the Singapore MPLP project, Nanyang Technological University Ref. M4061473 and Research Grant Ref. M060030008.
References (49)
- et al.
Improving accuracy through shrinkage modelling by using Taguchi method in selective laser sintering
Int. J. Mach. Tools Manuf.
(2007) An experimental design approach to selective laser sintering of low carbon steel
J. Mater. Process. Technol.
(2003)- et al.
Stepwise approach for the evolution of generalized genetic programming model in prediction of surface finish of the turning process
Adv. Eng. Softw.
(2014) - et al.
A framework of neural networks based consensus control for multiple robotic manipulators
Neurocomputing
(2014) - et al.
Synchronized control with neuro-agents for leader–follower based multiple robotic manipulators
Neurocomputing
(2014) A comparative study of population-based optimization algorithms for turning operations
Inf. Sci.
(2012)Comparison of evolutionary based optimization algorithms for structural design optimization
Eng. Appl. Artif. Intell.
(2013)A new hybrid differential evolution algorithm for the selection of optimal machining parameters in milling operations
Appl. Soft Comput.
(2013)A new hybrid artificial bee colony algorithm for robust optimal design and manufacturing
Appl. Soft Comput.
(2013)Optimization of cutting parameters in multi-pass turning using artificial bee colony-based approach
Inf. Sci.
(2013)
Hybrid Taguchi-differential evolution algorithm for optimization of multi-pass turning operations
Appl. Soft Comput.
An effective hybrid immune-hill climbing optimization approach for solving design and manufacturing optimization problems in industry
J. Mater. Process. Technol.
Interfaces in analogue biomaterials
Acta Mater.
A study on shrinkage compensation of the SLS process by using the Taguchi method
Int. J. Mach. Tools Manuf
Combined CI-MD approach in formulation of engineering moduli of single layer graphene sheet
Simul. Model. Pract. Theory
State-of-the-art in empirical modelling of rapid prototyping processes
Rapid Prototyp. J.
Density prediction of crystalline polymer sintered components at various powder bed temperatures
Rapid Prototyp. J.
Model of the selective laser sintering of bisphenol – a polycarbonate
Ind. Eng. Chem. Res.
Numerical prediction of temperature and density distributions in selective laser sintering processes
Rapid Prototyp. J.
DOE based three-dimensional finite element analysis for predicting density of a laser-sintered component
Rapid Prototyp. J.
Density prediction of selective laser sintering components based on support vector regression
Acta Phys. Sin.
Cited by (29)
Residual stress computation in direct metal deposition using integrated artificial neural networks and finite element analysis
2024, Materials Today CommunicationsModeling spatial variations in co-axial melt pool monitoring signals in laser powder bed fusion
2023, Journal of Manufacturing ProcessesData-driven modeling of process, structure and property in additive manufacturing: A review and future directions
2022, Journal of Manufacturing ProcessesDirected energy deposition processes and process design by artificial intelligence
2022, Advances in Metal Additive ManufacturingMachine learning in predicting mechanical behavior of additively manufactured parts
2021, Journal of Materials Research and TechnologyCitation Excerpt :The obtained results confirmed that porosity prediction was accurate under any parameter setting. In [102,103], layer thickness and laser power were considered as inputs of a ML-based system in order to predict porosity and achieve desired porosity by changing the process parameters in SLS printing technique. Specifically, in [102] multi-gene genetic programming approach was used to formulate the model.
Machine learning in additive manufacturing: State-of-the-art and perspectives
2020, Additive ManufacturingCitation Excerpt :Moreover, open porosity is required in some cases, such as auxetic structures for energy absorption and porous structures for medical implants. In SLS processing of PLA material, SVM and MLP techniques were explored to predict the open porosity [80]. Multi-gene genetic programming (MGGP) is an evolutionary approach that can automatically evolve the model structure and its coefficient while one disadvantage is that its generalization ability is limited.