0
Research Papers

In-Process Monitoring of Selective Laser Melting: Spatial Detection of Defects Via Image Data Analysis

[+] Author and Article Information
Marco Grasso

Dipartimento di Meccanica,
Politecnico di Milano,
Via La Masa 1,
Milan 20156, Italy
e-mail: marcoluigi.grasso@polimi.it

Vittorio Laguzza

Dipartimento di Meccanica,
Politecnico di Milano,
Via La Masa 1,
Milan 20156, Italy
e-mail: vittorio.laguzza@polimi.it

Quirico Semeraro

Dipartimento di Meccanica,
Politecnico di Milano,
Via La Masa 1,
Milan 20156, Italy
e-mail: quirico.semeraro@polimi.it

Bianca Maria Colosimo

Dipartimento di Meccanica,
Politecnico di Milano,
Via La Masa 1,
Milan 20156, Italy
e-mail: biancamaria.colosimo@polimi.it

1Corresponding author.

Manuscript received December 9, 2015; final manuscript received August 28, 2016; published online November 10, 2016. Assoc. Editor: Dragan Djurdjanovic.

J. Manuf. Sci. Eng 139(5), 051001 (Nov 10, 2016) (16 pages) Paper No: MANU-15-1647; doi: 10.1115/1.4034715 History: Received December 09, 2015; Revised August 28, 2016

Selective laser melting (SLM) has been attracting a growing interest in different industrial sectors as it is one of the key technologies for metal additive manufacturing (AM). Despite the relevant improvements made by the SLM technology in the recent years, process capability is still a major issue for its industrial breakthrough. As a matter of fact, different kinds of defect may originate during the layerwise process. In some cases, they propagate from one layer to the following ones leading to a job failure. In other cases, they are hardly visible and detectable by inspecting the final part, as they can affect the internal structure or structural features that are difficult to measure. This implies the need for in-process monitoring methods able to rapidly detect and locate defect onsets during the process itself. Different authors have been investigating machine sensorization architectures, but the development of statistical monitoring techniques is still in a very preliminary phase. This paper proposes a method for the detection and spatial identification of defects during the layerwise process by using a machine vision system in the visible range. A statistical descriptor based on principal component analysis (PCA) applied to image data is presented, which is suitable to identify defective areas of a layer. The use of image k-means clustering analysis is then proposed for automated defect detection. A real case study in SLM including both simple and complicated geometries is discussed to demonstrate the performances of the method.

Copyright © 2017 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

Examples of defective parts produced via SLM: local defects in (a) complicated geometries, (b) contours of solid parts, (c) lattice structures, and (d) part–support interfaces

Grahic Jump Location
Fig. 2

Three monitoring scales for SLM processes

Grahic Jump Location
Fig. 3

Experimental setup with the high-speed camera outside the build chamber: side view picture (left panel) and schematic representation of the same view (central panel)

Grahic Jump Location
Fig. 4

Left panel: bottom view of the CAD model, where three triangular features resulting from part slicing are highlighted (referred, respectively, as triangles 1–3) and right panel: bottom view of the manufactured part showing the effects of local overheating in overhang acute corners

Grahic Jump Location
Fig. 5

Triangular features resulting from slicing the CAD model of the part

Grahic Jump Location
Fig. 6

Example of laser scanning path for triangle 1 (left panel) and detail of corner C (right panel)

Grahic Jump Location
Fig. 7

Example of an image stream that consists of J frames of size M×N pixels

Grahic Jump Location
Fig. 8

Traditional (left) and proposed (right) unfolding approaches from ℝJ×M×N to ℝJ×p and to ℝp×J, respectively, where p=M×N

Grahic Jump Location
Fig. 9

Spatial distribution of T2(X,Y) for one slice of the cylindrical shape (left panel) and clusters identified using the k-means approach (central panel; black: background and gray: normal melting) and corresponding D(k) statistics (right panel)

Grahic Jump Location
Fig. 10

Examples of three intensity profiles for pixels belonging to corner A (top panel), corner B (central panel), and corner C (bottom panel)—triangle 1

Grahic Jump Location
Fig. 11

Two-dimensional (top panels) and three-dimensional (bottom panels) representations of the spatial distribution of T2(X,Y) for triangle 1 (left panels), triangle 2 (central panels), and triangle 3 (right panels); the color map ranges from dark blue (lower values) to bright yellow (larger values); see figure online for color

Grahic Jump Location
Fig. 12

Example of intensity profiles corresponding to pixels that exhibit a spike of the T2(X,Y) indicator identified as spike I and spike II (triangle 1)

Grahic Jump Location
Fig. 13

Clusters identified using the k-means approach in the out-of-control examples (top panels; black: background cluster, gray: normal melting cluster, and red: defect cluster) and corresponding D(k) statistics (bottom panels; vertical red-dashed line corresponds to the selected number k of clusters); see figure online for color

Grahic Jump Location
Fig. 14

k-means clustering of the T2(X,Y) statistic iteratively updated for a growing number of frames for one slice of the cylindrical shape (black: background and gray: normal melting)

Grahic Jump Location
Fig. 15

k-means clustering of the T2(X,Y) statistic iteratively updated for a growing number of frames for triangle 1 (black: background, gray: normal melting, and red: defect); see figure online for color

Grahic Jump Location
Fig. 16

k-means clustering of the T2(X,Y) statistic iteratively updated for a growing number of frames for triangle 2 (black: background, gray: normal melting, and red: defect); see figure online for color

Grahic Jump Location
Fig. 17

k-means clustering of the T2(X,Y) statistic iteratively updated for a growing number of frames for triangle 3 (black: background, gray: normal melting, and red: defect); see figure online for color

Grahic Jump Location
Fig. 18

Results of competitor approach based on average pixel intensities for triangle 1 (black: background, gray: normal melting)

Grahic Jump Location
Fig. 19

Results of competitor approach based on average pixel intensities for triangle 2 (black: background, gray: normal melting, and red: defect); see figure online for color

Grahic Jump Location
Fig. 20

Results of competitor approach based on average pixel intensities for triangle 3 (black: background, gray: normal melting)

Grahic Jump Location
Fig. 21

T2(X,Y) spatial distribution for different pixel arranging methods: row-wise (top-left panel) and column-wise (top-right panel), row-wise arrangement with randomized order within the row (bottom-left panel) and complete order randomization (bottom-right panel)—triangle 1

Grahic Jump Location
Fig. 22

Mean T2(X,Y) spatial distribution (left panel) and the corresponding standard deviation (right panel) for an ensemble of 1000 realizations—triangle 1

Grahic Jump Location
Fig. 23

Loadings of retained PCs estimated by analysis of the entire image stream for triangle 1 (dotted lines mark the time window during which the overheating effects of corner C were visible)

Grahic Jump Location
Fig. 24

Scores of retained PCs estimated by analysis of the entire image stream for triangle 1

Grahic Jump Location
Fig. 25

Three-dimensional representation of the SPE(X,Y) statistic triangle 1; the color map (see figure online for color) ranges from dark blue (lower values) to bright yellow (larger values)

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In