Fig 1 - uploaded by Ehsan Malekipour
Content may be subject to copyright.
A schematic diagram of direct metal laser sintering (DMLS) process [6]

A schematic diagram of direct metal laser sintering (DMLS) process [6]

Source publication
Article
Full-text available
Powder bed fusion (PBF) process is a metal additive manufacturing process, which can build parts with any complexity from a wide range of metallic materials. Research in the PBF process predominantly focuses on the impact of a few parameters on the ultimate properties of the printed part. The lack of a systematic approach to optimizing the process...

Contexts in source publication

Context 1
... each layer. Selective laser melting (SLM) and selective laser sintering (SLS) are the main two PBF processes. Unlike the SLM process, where the powder is completely melted down to form a homogeneous part, the SLS process partially melts the material (sinter the powder) layerby-layer at the molecular level [5]. The schematic diagram in Fig. 1 shows the overall process of the PBF process [6]. The 3D printer machine consists of a supply station for the metal powder and a sintering/melting unit. A laser selectively sinters/melts the powder with respect to the layer geometry along a prescribed pattern. After sintering/melting of a layer, the powder dispenser platform moves ...
Context 2
... MATLAB image processing to measure the porosity of each sample in two steps (Fig. 3). First, MATLAB creates black and white (BW) images from the micrographs. In these images, the black pixels represent the porosity and the white pixels represent the solid. In this step, the threshold level is adjusted by comparing the pore size in the SEM image (Fig. 10.a) with the image generated by MATLAB (Fig. 10.b) to increase the accuracy of the method [41]. Then, we calculate the ratio of the number of black pixels to the total pixels for the horizontal BW micrographs and the vertical ones separately. The overall average ratio for both magnifications on the BW images of horizontal cross-section and ...
Context 3
... of each sample in two steps (Fig. 3). First, MATLAB creates black and white (BW) images from the micrographs. In these images, the black pixels represent the porosity and the white pixels represent the solid. In this step, the threshold level is adjusted by comparing the pore size in the SEM image (Fig. 10.a) with the image generated by MATLAB (Fig. 10.b) to increase the accuracy of the method [41]. Then, we calculate the ratio of the number of black pixels to the total pixels for the horizontal BW micrographs and the vertical ones separately. The overall average ratio for both magnifications on the BW images of horizontal cross-section and vertical crosssections represent the porosity ...
Context 4
... applied to the samples generate three different types of porosity according to low, medium, or high value of volumetric energy density (VED). Low volumetric energy density leads to incomplete melting of the powder particles and formation of irregular pores due to lack of fusion such as sample 3 (the low VED with LP 100 W and SS 900 mm/s) (Fig. 11.a). While, the exertion of high volumetric energy vaporizes the material and hence, it leads to the formation of circular gas pores such as sample 13 (the highest VED with LP 175 W and SS 700 mm/s) (Fig. 11.b). These circles can be a cross-section for a keyhole porosity. Samples with medium VED such as sample 16 (medium VED with LP 175 W ...
Context 5
... of the powder particles and formation of irregular pores due to lack of fusion such as sample 3 (the low VED with LP 100 W and SS 900 mm/s) (Fig. 11.a). While, the exertion of high volumetric energy vaporizes the material and hence, it leads to the formation of circular gas pores such as sample 13 (the highest VED with LP 175 W and SS 700 mm/s) (Fig. 11.b). These circles can be a cross-section for a keyhole porosity. Samples with medium VED such as sample 16 (medium VED with LP 175 W and SS 1000 mm/s) possess microscale holes with a nearly uniform distribution throughout the cross-section, which is an evidence in better mechanical properties compared with the other types [42]. Table 4 ...
Context 6
... vertical cross sections, which lead to different porosity percentage in each cross section. Vertical cross sections illustrate the less frequent but bigger size porosity usually progressing through layers. Whereas, horizontal cross sections illustrate the widespread porosity in different size ranges, which scatters throughout the entire section (Fig. 12). The average porosity will be used in the future phase of this framework. The porosities in horizontal and vertical cross-sections for the samples according to the applied VED is shown in Fig.12. As Fig. 13 shows, the energy density alters between 55 and 138 J/mm 3 . This energy density creates a part with the density between 95.52% ...
Context 7
... average porosity will be used in the future phase of this framework. The porosities in horizontal and vertical cross-sections for the samples according to the applied VED is shown in Fig.12. As Fig. 13 shows, the energy density alters between 55 and 138 J/mm 3 . ...
Context 8
... porosities in horizontal and vertical cross-sections for the samples according to the applied VED is shown in Fig.12. As Fig. 13 shows, the energy density alters between 55 and 138 J/mm 3 . This energy density creates a part with the density between 95.52% and 99.31% with a maximum of 99.31%. ...
Context 9
... (BP) algorithm to develop a multiple input/output optimization models. The inputs include the process parameters (laser power, scan speed, hatch space, and beam diameter) while the outputs include ultimate properties of a fabricated part (densification, mechanical properties, surface roughness, dimensional accuracy, and fabrication time) (see Fig. 14). Development of this multi-input/output ANN function will eventually lead to an intelligent system capable of controlling multiple parameters simultaneously. The framework presented in this work employs two sets of experiments to first, narrow down the process parameters (the inputs) to their optimized ranges and second, study how a ...
Context 10
... system will go towards a comprehensive control system in the PBF process within two steps. First, we develop a NN intelligent system, which makes users/manufacturers capable of selecting the process parameters according with the required/desired ultimate properties of a fabricated part (the current project) (Fig. 15). Second, we integrate this system with an online monitoring and control (OMC) system to fabricate nearly flawless parts with the desired ultimate qualities (the long-term objective). Plentiful research nowadays has focused on the development of OMC systems [43][44][45][46][47][48] to avoid/diminish the defects and abnormalities ...
Context 11
... work for post-processing. Optimization of process parameters by using an ANN model in this project integrated with an OMC system can considerably improve the mechanical properties and surface quality of fabricated parts, increase the repeatability, reduce fabrication time, and significantly decrease the need for the post-processing operations. Fig. 15. The algorithm of selection process parameters [57] ...

Similar publications

Article
Full-text available
Tungsten inert gas (TIG) welding of thin stainless steel tubes is an important manufacturing process with many applications, in particular for the closure of sheathed thermocouples tubes and thermocouple wells. During the welding process, the microstructure, properties as well as the surface roughness is directly influenced by the setup of the weld...

Citations

... The optimum parameters play a crucial role in enhancing the quality of the building part and its mechanical properties. Most of the researchers utilized different ML techniques to optimize process parameters [44][45][46]. ...
Article
Full-text available
Machine learning (ML) is one of the artificial intelligence tools which uses past data to learn the relationship between input and output and helps to predict future trends. Powder bed fusion additive manufacturing (PBF-AM) is extensively used for a wide range of applications in the industry. The AM process establishment for new material is a crucial task with trial-and-error approaches. In this work, ML techniques have been applied for the prediction of the density of PBF-AM. Density is the most vital property in evaluating the overall quality of the AM building part. The ML techniques, namely, artificial neural network (ANN), K-nearest neighbor (KNN), support vector machine (SVM), and linear regression (LR), are used to develop a model for predicting the density of the stainless steel (SS) 316L build part. These four methods are validated using R-squared values and different error functions to compare the predicted result. The ANN and SVM model performed well with the R-square value of 0.95 and 0.923, respectively, for the density prediction. The ML models would be beneficial for the prediction of the process parameters. Further, the developed ML model would also be helpful for the future application of ML in additive manufacturing .
... They developed process maps for 316L and 17-4 PH stainless steel using that model. Marrey et al. (2019), Mehrpouya et al. (2019), and Nikam et al. (2020) developed ANN models to optimize process parameters. Marrey et al. (2019) proposed an ANN framework to find optimum process parameters for desired mechanical properties in LBPF fabricated 316L stainless steel. ...
... Marrey et al. (2019), Mehrpouya et al. (2019), and Nikam et al. (2020) developed ANN models to optimize process parameters. Marrey et al. (2019) proposed an ANN framework to find optimum process parameters for desired mechanical properties in LBPF fabricated 316L stainless steel. Mehrpouya et al. (2019) developed an MLP neural networks-based model to predict optimal process parameters for additively manufactured NiTi shape memory alloys. ...
Article
Full-text available
Additive manufacturing (AM) offers the advantage of producing complex parts more efficiently and in a lesser production cycle time as compared to conventional subtractive manufacturing processes. It also provides higher flexibility for diverse applications by facilitating the use of a variety of materials and different processing technologies. With the exceptional growth of computing capability, researchers are extensively using machine learning (ML) techniques to control the performance of every phase of AM processes, such as design, process parameters modeling, process monitoring and control, quality inspection, and validation. Also, ML methods have made it possible to develop cybermanufacturing for AM systems and thus revolutionized Industry 4.0. This paper presents the state-of-the-art applications of ML in solving numerous problems related to AM processes. We give an overview of the research trends in this domain through a systematic literature review of relevant journal articles and conference papers. We summarize recent development and existing challenges to point out the direction of future research scope. This paper can provide AM researchers and practitioners with the latest information consequential for further development.
... Zuowei et al. [17] used ANN in order to predict the geometric deviations for fused deposition modelling, classifying the process images as well. Marrey et al. [18] presented a framework for the PBF process parameters optimization by mapping the ultimate quality of a fabricated part into the process parameters and material properties. The effects of laser power and scan speed on the microstructure and porosity of SS 316L fabricated parts were investigated in their first phase studies. ...
Article
Full-text available
Directed energy deposition (DED) is an additive manufacturing process used in manufacturing free form geometries, repair applications, coating and surface modification, and fabrication of functionally graded materials. It is a process in which focused thermal energy is used to fuse materials by melting. Thermal effects can cause distortions and defects on the parts during the DED process, therefore they should be evaluated and taken into account during the manufacturing of products. Melting pool control and DED bead geometries should be defined properly as well. In this work, an Artificial Neural Network model has been applied considering the DED process parameters in order to predict the geometrical patterns and create a local reinforced product as a hybrid manufacturing technology. Although lots of studies are available on topology optimization for manufacturing methods such as casting, extrusion, and powder bed fusion, topology optimization for the DED process is not widely taken into consideration to predict the design geometrical patterns. DOE RSM and ANN approaches were applied in this study to predict convenient dimensions, topology based geometrical patterns of local stiffeners and heat source power optimizing the energy, total mass, and peak force results of the hybrid part. A single bead track deposition is simulated in terms of validation of the numerical heat source model, and cross-sections of the beads are analysed. A cross-member structure is manufactured using the DED device and the structure is correlated under the three point bending physical conditions on test bench. It has been investigated that locally reinforced cross beam has much more energy absorption and peak force values than plain model. The results showed that the proposed NN-GA is a promising approach to generate the topology based geometrical patterns and process parameters which can be used to create a local reinforced product as hybrid manufacturing technologies.
... The downward-facing surface condition can be optimized by minimizing supports using different methods such as numerical simulations, statistical analyses of LPBF build parameter, [4] topologically optimized supports [212] and artificial neural networks. [213] Contactless supports were explored by researchers via building a thin blade parallel to the down-facing surface. [212,214] The key point behind this design is to facilitate efficient heat transfer from the melt pool to the thin blade through the powder bed. ...
Article
The design freedom offered by additive manufacturing (AM) enables the fabrication of components with internal surfaces that are challenging to access post-manufacture. This is of concern, as the surface condition can markedly deteriorate fatigue performance. Additionally, the adaptation of surface finishing methods for AM components with topologically optimised designs can be a costly practice. It is therefore desirable to consider deploying AM parts with no or minimal surface processing for targeted applications. This requires an in-depth understanding of the formation of various types of AM surfaces, including the variation in surface condition and controlling factors, and their influence on mechanical performance. The last few years have seen significant research advances in these aspects. Ti-6Al-4V is the most extensively studied alloy for AM. The research data available now allows an informative treatment of this topic for both practical applications and future research. Using laser powder bed fusion (LPBF) of Ti-6Al-4V as a model AM−alloy system, this article examines (i) the characteristics of various types of LPBF surfaces including horizontal, vertical, inclined, upward, downward, internal isolated, and slotted surfaces; (ii) the design features and LPBF variables that affect the surface topography; (iii) the capabilities of existing post-AM surface processing methods; and (iv) the influence of AM surface topography on mechanical properties by focusing on the fatigue performance. On this basis, design considerations are recommended for AM of consistent surfaces, and priority surface-related research issues are identified. The purpose is to establish an essential knowledge base for improved commercial designs for LPBF for suitable dynamically loaded applications, with no or minimal surface processing. While centring on LPBF of Ti-6Al-4V, the insights derived are expected to be applicable to other AM processes or metallic materials.
... Direct metal laser sintering process(Marrey 2019). For a color version of this figure, seewww.iste.co.uk/kumar/materials.zip ...
Chapter
This chapter explains an appraisal of the methodologies and suitable materials used in additive manufacturing (AM) for advancing the aerospace industry with the performances of topological and material optimization, design integration and part consolidation that promotes profitable quantity and quality products. It discusses several 3D printing methods and the use of different materials and their applications in the aerospace industry. The prominent process of AM is classified into the following seven types, namely laser beam melting, electron beam melting, selective laser melting, direct metal laser sintering, laser metal fusion, and direct metal deposition. The materials play a predominant role in AM, which are considered for engineered operational applications. AM contributes its foremost capability to serving the aerospace industry, because this industry faced a great challenge to reduce the weight and size of components in order to avoid dangers and promote safety and durability.
... For the MAM process monitoring, the NN first takes the original measurement data or its features as the input, adjusts the structure and parameters of the network through learning algorithms, and iteratively optimizes the networks to map the build quality against the measurements. The NN has shown a strong advantage in defect localization and prediction [98][99][100][101][102][103][104][105][106][107][108][109][110][111]. In selective laser sintering (SLS), the part density is one of the most important indications of build quality. ...
... The selection of different ML methods are based on the types of input features and defects. In general, the NN is could be applied in most cases of quality monitoring among all the ML methods [98][99][100][101][102][103][104][105][106][107][108][109][110]. In the case of small samples and high-dimensional features, the SVM approaches have many advantages, but they are limited in multiple defect classifications [114], and insensitive to subtle state variations [111,115]. ...
Article
Full-text available
The metal-based additive manufacturing (MAM) processes have great potential in wide industrial applications, for their capabilities in building dense metal parts with complex geometry and internal characteristics. However, various defects in the MAM process greatly affect the precision, mechanical properties and repeatability of final parts. These defects limit its application as a reliable manufacturing process, especially in the aerospace and medical industries where high quality and reliability are essential. MAM process monitoring provides a technical basis for avoiding and eliminating defects to improve the build quality. Based on of the nature of the MAM build defects, this article conducts a thorough investigation of monitoring methods, and proposes a machine learning (ML) framework for process condition monitoring. According to the structure of ML models, they are divided into shallow ML-based and deep learning-based methods. The state-of-the-art ML monitoring approaches, as well as the advantages and disadvantages of their algorithmic implementations, are discussed. Finally, the prospects of ML based process monitoring researches are summarized and advised.
... We used two different magnifications --60X and 300X (100µm and 10µm scale, respectively) -for each micrograph and then employed MATLAB image processing to measure the porosity of each sample in three steps (Figure 7). First, we generated bi-color black and white images from each micrograph; second, the threshold level was adjusted by comparing the pore size between the SEM image and the MATLAB-generated image to increase the method accuracy [36]; finally, we obtained the porosity percentage by calculating and averaging the ratio of black parts (pores) to the white parts in the micrographs related to each cross-section [37]. By analyzing the horizontal cross-section, we revealed three different types of porosity -low, medium, high --according to the level of VED. ...
Preprint
Powder bed fusion (PBF) process is a metal additive manufacturing process which can build parts with any complexity from a wide range of metallic materials. PBF process research has predominantly focused on the impact of only a few parameters on product properties due to the lack of a systematic approach for optimizing a large set of process parameters simultaneously. The pivotal challenges regarding this process require a quantitative approach for mapping the material properties and process parameters onto the ultimate quality; this will then enable the optimization of those parameters. In this study, we propose a two-phase framework for optimizing the process parameters and developing a predictive model for 316L stainless steel material. We also discuss the correlation between process parameters -- i.e., laser specifications -- and mechanical properties and how to achieve parts with high density (> 98%) as well as better ultimate mechanical properties. In this paper, we introduce and test an innovative approach for developing AM predictive models, with a relatively low error percentage of 10.236% that are used to optimize process parameters in accordance with user or manufacturer requirements. These models use support vector regression, random forest regression, and neural network techniques. It is shown that the intelligent selection of process parameters using these models can achieve an optimized density of up to 99.31% with uniform microstructure, which improves hardness, impact strength, and other mechanical properties.
... We used two different magnifications --60X and 300X (100µm and 10µm scale, respectively) -for each micrograph and then employed MATLAB image processing to measure the porosity of each sample in three steps (Figure 7). First, we generated bi-color black and white images from each micrograph; second, the threshold level was adjusted by comparing the pore size between the SEM image and the MATLAB-generated image to increase the method accuracy [36]; finally, we obtained the porosity percentage by calculating and averaging the ratio of black parts (pores) to the white parts in the micrographs related to each cross-section [37]. By analyzing the horizontal cross-section, we revealed three different types of porosity -low, medium, high --according to the level of VED. ...
Article
Full-text available
Powder bed fusion (PBF) process is a metal additive manufacturing process which can build parts with any complexity from a wide range of metallic materials. PBF process research has predominantly focused on the impact of only a few parameters on product properties due to the lack of a systematic approach for optimizing a large set of process parameters simultaneously. The pivotal challenges regarding this process require a quantitative approach for mapping the material properties and process parameters onto the ultimate quality; this will then enable the optimization of those parameters. In this study, we propose a two-phase framework for optimizing the process parameters and developing a predictive model for 316L stainless steel material. We also discuss the correlation between process parameters-i.e., laser specifications-and mechanical properties and how to achieve parts with high density (> 98%) as well as better ultimate mechanical properties. In this paper, we introduce and test an innovative approach for developing AM predictive models, with a relatively low error percentage of 10.236% that are used to optimize process parameters in accordance with user or manufacturer requirements. These models use support vector regression, random forest regression, and neural network techniques. It is shown that the intelligent selection of process parameters using these models can achieve an optimized density of up to 99.31% with uniform microstructure, which improves hardness, impact strength, and other mechanical properties.
... For the MAM process monitoring, the NN first takes the original measurement data or its features as the input, adjusts the structure and parameters of the network through learning algorithms, and iteratively optimizes the networks to map the build quality against the measurements. The NN has shown a strong advantage in defect localization and prediction [98][99][100][101][102][103][104][105][106][107][108][109][110][111]. In selective laser sintering (SLS), the part density is one of the most important indications of build quality. ...
... The selection of different ML methods are based on the types of input features and defects. In general, the NN is could be applied in most cases of quality monitoring among all the ML methods [98][99][100][101][102][103][104][105][106][107][108][109][110]. In the case of small samples and high-dimensional features, the SVM approaches have many advantages, but they are limited in multiple defect classifications [114], and insensitive to subtle state variations [111,115]. ...
Preprint
Full-text available
1-The metal-based additive manufacturing (MAM) processes have great potential in wide industrial applications, for their capabilities in building dense metal parts with complex geometry and internal characteristics. However, various defects in the MAM process greatly affect the precision, mechanical properties and repeatability of final parts. These defects limit its application as a reliable manufacturing process, especially in the aerospace and medical industries where high quality and reliability are essential. MAM process monitoring provides a technical basis for avoiding and eliminating defects to improve the build quality. Based on of the nature of the MAM build defects, this study conducts a thorough investigation of monitoring methods, and proposes a machine learning (ML) framework for process condition monitoring. According to the structure of ML models, they are divided into shallow ML based and deep learning based methods. The state-of-the-art ML monitoring approaches, as well as the advantages and disadvantages of their algorithmic implementations, are discussed. Finally, the prospects of ML based process monitoring researches are summarized and advised.
... Khorasani et al. [23] included energy density in their investigation on the effect of different L-PBF process parameters on part density, hardness, tensile strength and surface quality and their results showed correlation between hardness, density and meltpool formations. King et al. [24] concluded that the energy density of the laser beam controls the transition from a conduction-mode to keyhole-mode melting in the L-PBF process, while some advocate the usage of energy density due to its ability to capture track width evolution and the normalized melt pool depth [22,[25][26][27][28]. ...
Article
Full-text available
Down-facing surfaces are one of the most challenging features in metal parts produced by laser powder bed fusion (LPBF). A combination of reasons, primary of which are residual stresses and overheating cause these features to have the worst surface finish and dimensional accuracy of all LPBF surfaces. In order to examine this phenomenon, a Design of Experiments (DoE) study is conducted for three different inclination angles, namely 45°, 35° and 25° and for two different layer thicknesses of 60 µm and 90 µm. The results from the DoE are used to establish quadratic regression equations that can be used to predict the quality marks of surface roughness and the relative dimensional error.This fundamental investigation helps to explain the reasons for the major defects in down-facing surfaces of parts produced with Ti-6AL-4V material, namely the dross formation and attempts to improve the predictability of quality within the region. Further to the establishment of the quadratic equations, a discussion is conducted on the thermomechanical processes involved in the mechanism of dross formation and explanations are given on the reasons behind the observed physical phenomena. The trend of the propagation of (Root Mean Square) RMS Surface roughness (Sq) and the relative dimensional error with respect to the Volumetric Energy Density (VED) is discussed in detail. The respective quadratic equations are then tested by a second round of validation prints, and the results confirm the feasibility of the developed quadratic models to accurately predict process outcomes especially when operating near the suggested optimal printing zones. The high roughness of low VED printing is attributed to the formation of ‘inverse mushroom’ structures, and the low roughness of high VED surface is attributed to the formation of large flat regions formed as adjacent meltpools that can fuse together at various locations.