Visualization Support For Virtual Redesi
Visualization Support For Virtual Redesi
Visualization Support For Virtual Redesi
com
Abstract
Rapidly changing products and market demand call for manufacturing systems to be continuously adapted and developed. The
process of modifying manufacturing systems requires large amounts of planning involving contributions from personnel across an
organization. These people need a shared understanding of the future system, including but not limited to its design, functions, and
expected performance. One common representation in the virtual manufacturing system domain are 2D CAD layouts. Typical
problems with such traditional 2D models are that only experts understand the content fully. For increased understanding, 3D CAD
models could bridge the gap between different areas of expertise. However, creating 3D models representing the complete system is
traditionally time-consuming, resulting in oversimplified models or limited to parts of the system. Furthermore, such models
normally contain uncertainty about building-related geometries that could incur costly mistakes if used as basis for decisions, e.g.
realizing during installation of a machine that roof-
f beams interfere with the planned placement. This paper evaluates what type of
problems can be solved with better visualization support, e.g. issues concerning workshop-layout, production flow, workplace
design, etc. The evaluation is based on two case studies at different manufacturing sites during ongoing system redesign processes.
The case studies implemented visualization using a combination of CAD models and 3D laser scanned as-built data of the current
.
Bringing this concept into the early phases of manufacturing system redesign has the potential to facilitate the creation of a shared
understanding of the future system within cross-functional project teams.
© 2013 The
© 2013 TheAuthors.
Authors.Published
Publishedbyby Elsevier
Elsevier B.V.
B.V.
Selection and/or
Selection and peer-review
peer-review underunder responsibility
responsibility of Professor
of Professor PedrodoFilipe
Pedro Filipe CarmodoCunha
Carmo Cunha
Keywords: Virtual manufacturing; Production systems; Visualization; 3D laser scanning; Point clouds
2.2. Spatial data collection be done to various degrees depending on the target
application and processing performance. To be able to
One key to solving many industrial engineering use the data for purposes beyond visualization of static
problems is accurate measurements of spatial data. 3D backgrounds, manual intervention is necessary. Typical
imaging is a field within measurement science dealing operations performed in a factory environment are object
with capture of three-dimensional spatial data. There are based selection and grouping of a subset of points, an
many different technologies for that purpose, see e.g. Bi example of such a model is presented in Fig 2. This
and Wang [6], or Sansoni et al. [13] for an overview. A could e.g. be used to separate a machine body from the
phase based 3D laser scanner has been the technology data and save it as a stand-alone point cloud file. Most
used to capture spatial data in the study presented in this software designed to work with point cloud data also
paper. The scanner has a near 360 degrees field of view have the capability of rendering and/or editing CAD data
and typically captures tens of millions of data points in a in parallel, which is useful for a number of applications.
few minutes. While the technology is rather new, there
are ongoing efforts to standardize testing and assessment
of equipment and the data formats for storing the
measurements [14]. The technology is being used across
a number of different fields, e.g. heritage documentation,
forensics, and tunnel mapping [13], [15].
A brief description of the workflow pertaining to data
capture using a 3D laser scanner is [16]:
Prepare Scanning - Plan the scanning and reference
objects positions to ensure that all necessary data can
be captured. The line of sight from the scanner to the
objects of interest has to be considered as well as the
line of sight to the reference objects. At least three
corresponding reference objects have to be visible in
two separate scans if they are to be combined
successfully.
Perform Scanning - Position the scanner on the
planned positions and execute the data capture. For
good results, it is important that the scanned
environment remains motionless throughout the Fig. 2. Example of a point cloud model representing a robot
scanning process. The measurement process duration
per scan position is about 5 minutes using a resolution 3. Method
that is suitable for larger indoor areas.
Process Scan Data - Register the scans to align and This paper is based on case studies at two
combine them into one data set and clean the data manufacturing sites producing high-end engine
from any unwanted artifacts. The alignment is done components to the aerospace industry. The current
using the reference objects mentioned. Examples of manufacturing systems in the two factories are mainly
artifacts are sensor noise and partially captured organized as functional layouts, but the company aims to
moving objects. change the systems towards a more product-oriented
The result is a large dataset containing measurement layout. 3D scan data has been used in both studies in
points described by x, y, and z coordinates and an RGB- parallel with the normal work method for visualization
color code (also normal direction and reflection consisting of basic 2D CAD layouts and for some
intensity). A section of a factory 50 by 50 meters could specific parts 3D CAD models. The point cloud models
typically consist of 500 million points, and generate a were used to support the visualization of the current
file size of approximately 4 gigabytes. There are many situation as well as future changes. The research goal
vendor specific data formats, which could potentially with the two studies was to evaluate potential benefits
lead to reduced data interchangeability and transparency. and issues with the scanning method.
However, standardized formats have recently been The case studies were performed in three steps as
brought forth as a remedy to this, see ASTM E2807 [14]. follow:
The data is simply a large amount of points in space, 1. 3D scanning of the factory
which has coined the commonly used name point cloud. 2. Pre-work of the scan data
The point clouds can be made sparser by filtering away a 3. Work meeting and discussion
percentage of the points. This reduces data size and can
422 Erik Lindskog et al. / Procedia CIRP 7 (2013) 419 – 424
The scanning process was performed as described in organization. With assistance from one of the authors the
section 2.2. The first case study was used as a feasibility project members were able to walk around in the model
study and experiences were taken into consideration and make measurements of interest. For example
when performing the second study. The two case studies distances between pillars and walls as well as other
are presented in this section in the same sequence as they building geometries.
were performed.
3.2. Case study B
3.1. Case study A
The second case study was performed in a factory
The section of the factory studied in this case was with production running at full speed. The work content
undergoing a major reconstruction. Most of the old in this study can be separated into two steps. The
machines were moved to another location and replaced primary goal of the case was to evaluate a proposed
with new machines to create a new manufacturing layout considering a new machine and its attached
system. The goal with the study was to evaluate if the equipment before start of installation. To free up room
new layout proposal would fit and be suitable for the for the new installation an existing machine was moved
existing factory building. to another location.
To cover the selected part of the factory the scanning Two people carried out the 3D scanning process in
process required 13 scan positions. The complete part of one workday during full production but limited to the
the factory for the new manufacturing system was section of the factory affected by the reconstruction. To
captured with a focus on building geometries. Static cover enough detail regarding the machines and other
reference points were mounted in the building to enable equipment 22 scan positions were required.
synchronization with future complementary scans. Most The first step in the modification of the point cloud
of the old machines and equipment had been moved model was to select the points representing the existing
when the scanning was performed which resulted in a new
more or less empty facility with reservation for some location. The main body of the machine was placed onto
construction-work equipment and material. an existing machine foundation, resulting in that
Based on the 3D scan data a point cloud model was attached equipment overlapped with the material-
created. In the model construction material and handling aisle. Points representing the equipment were
equipment was removed to get a clean model of the moved around the machine to evaluate new locations.
factory building. Fig 3 shows a section of the point cloud In step two a hybrid model was created where 3D
model before any modifications had been made. The CAD models of the new machine and necessary
point cloud model was synchronized to the factory equipment were imported and placed according to the
coordinate system and compared with the 2D CAD proposed 2D CAD layout solution. A section of the
layout to find possible differences. To visualize the hybrid model is presented in Fig 4.
outcome from the scanning process, other than the actual
point cloud model, a fly-through movie of the factory
was created.
well as parts of the point cloud to create alternative it would not have been possible to store enough trollies
solutions. Different solutions were studied from several with material and products close to the machine. With
perspectives and measurements were taken to verify the this new information at hand, the group discussed
model. alternative solutions, which were then tested in the
hybrid model. 3D CAD objects and point cloud objects
4. Result and discussion were translated in the model to positions suggested by
the group, ultimately resulting in a new layout proposal.
Project members in the two case studies were Measurements were made in the hybrid model to
impressed by the visualization possibility point cloud validate the model to the real world and to support
models holds and found possible areas of application for further discussion topics. Without the visual support
their specific interest. Those thoughts and main from the hybrid model and the fact that experts from
reflections are presented and discussed in this section. different parts of the organization were gathered at the
As mentioned in section 3, case study A was used as same discussion, the installation would have been made
a feasibility study to evaluate the potential of the based on the first layout proposal, which could have
technology. This study did show that the technology was resulted in problems during and/or after the installation.
well accepted in the project group. Presenting the result A known layout issue described by the project group
to the group did have a positive outcome, especially for was to get the proposed solution accepted by the
persons working with layout planning and machine operators. Traditionally, a 2D layout is presented to the
acquisition. During the discussion some typical operators, most often receiving approval. But after the
problems regarding the layout planning came up. Even if new machine or workstation had been installed and the
the project meetings normally are held in close contact operators were fully able understand the changes, they
to the factory section the group did find it very useful to were not always so approving of it, at times this has
have a digital copy of the actual factory building. They resulted in costly rebuilding activities. Therefore, an
would then have the possibility to quite easily and in a operator was invited at the work meeting and came with
short time study the model or take measurements when valuable insights based on experiences from the
certain issues were discussed. Normally they would operation phase.
either go out onto the factory floor to see for themselves The method of presenting and working with the
during the meeting or postpone the task to the next hybrid model could also be discussed. In these cases a
meeting. Issues solved by such information are for projector beam was used with a connected laptop, from
example the height between the floor and ceiling or where the model was navigated and modified. The
other parts that could interfere with the planned machine authors believe that it may be more suitable to use some
location. This information would not have been possible sort of interactive solution instead. A reflection from the
to acquire from the 2D layout, which they normally use case studies was to use an environment that encourages
for most of the planning. creativity. Currently the tools used in the case studies
The first modification in case study B was to move require an expert to be used to its full potential.
the existing machine to the new location. By moving the However, it could be discussed whether this should be
point cloud in the model representing the machine it was necessary or not. If the tool and model enable interactive
possible to visualize the machine at the new location and use by everyone it may improve the outcome of the
evaluate different installation alternatives. Comparing discussion even further.
the model with the real installation, the result of the Another interesting improvement could be to couple
installation was very close to what was visualized in the this scanning method with other visualization
model. A problem with moving specific parts of the technologies such as augmented reality and virtual
point cloud is to define all points belonging, in this case, reality. The common factor and overarching goal
to the machine. remains; improving the ability to provide a good
Before the work meeting regarding the installation of understanding of the future situation for all users.
the new machine the project group had already decided
upon a layout solution to be used. This layout was 5. Conclusions
mainly based on a 2D CAD drawing with some support
from simplified 3D CAD models. Based on that layout a The initial studies presented in this paper have shown
hybrid model was created by importing the 3D CAD that a realistic virtual model of the factory with high
objects into the point cloud model including all level of details and accuracy can be used as high-level
surrounding support equipment. All persons present at visual support when redesigning manufacturing systems.
the meeting realized by analyzing the hybrid model that Using 3D laser scanning to create point cloud models
the machine and supporting equipment require more holds the capability to provide this realistic visualization
space than was available. With the first proposed layout and enable a better understanding throughout an
424 Erik Lindskog et al. / Procedia CIRP 7 (2013) 419 – 424
organization. Visualizing the manufacturing system so Production Initiative and the Production Area of
that people with different prior knowledge and interest Advance at Chalmers University of Technology. The
can understand it and create the same mental model will support is gratefully acknowledged.
enable the model to be more efficiently used in
discussions and as a decision support tool. The hybrid
model used in the second case study did to some extent References
enable the possibility of
[1] Schuh G, Aghassi S, Orilski S, Schubert J, Bambach M,
to discuss the proposed solution based on a common Freudenberg R, Hinke C, Schiffer M. Technology roadmapping
view of the manufacturing system. for the production in high-wage countries. Production
To achieve good results during such discussions it is Engineering 2011;5:463-473.
[2] Cohen MA, Eliashberg J, Ho T. New product development: The
important to gather a cross-functional team of people performance and time-to-market tradeoff. Management Science
with different background and knowledge. The elevated 1996;42:173-186.
form of visualization that has been proposed in this [3] Dangelmaier W, Fischer M, Gausemeier J, Grafe M, Matysczok C,
paper holds the potential to be especially useful for Mueck B. Virtual and augmented reality support for discrete
manufacturing system simulation. Computers in Industry
including operational personnel in the discussions. 2005;56:371-383.
During the discussion everyone should, based on the [4] Becker MC, Salvatore P, Zirpoli F. The impact of virtual
virtual representation, have the possibility to present simulation tools on problem-solving and new product
development organization. Research Policy 2005;34:1305-1321.
their view of the solution and highlight possible [5] Wiendahl HP, Harms T, Fiebig C. Virtual factory design - a new
problems in their respective working areas. This would tool for a co-operative planning approach. International Journal
make it possible to prevent mistakes early in the of Computer Integrated Manufacturing 2003;16:535-540.
[6] Bi ZM, Wang L. Advances in 3D data acquisition and processing
planning process that may not have been found for industrial applications. Robotics and Computer-Integrated
otherwise. Manufacturing 2010;26:403-413.
The solution may not be limited to point cloud [7] Liker JK. The Toyota way: 14 management principles from the
. Blacklick: McGraw-Hill
models, but the focus should be headed towards an Professional Publishing; 2004.
effective method for creating virtual representations of [8] Bicheno J, Holweg M. The Lean Toolbox: The Essential Guide to
the future factories with realistic visualization and high Lean Transformation. Buckingham: PICSIE Books; 2009.
level of detail in a short time. [9] Vallhagen J, Stahre J, Johansson B. Visual Production - strategic
manufacturing system development tools for aerospace industry.
In ISABE 2011, Gothenburg.
6. Future work and visions [10] Kuehn W. Digital factory: integration of simulation enhancing the
product and production process towards operative control and
optimization. International Journal of Simulation 2006;7:27-39.
Further user studies with experts from different areas [11] Iqbal M, Hashmi MSJ. Design and analysis of a virtual factory
of the organization will be performed as the next step of layout. Journal of Materials Processing Technology
this research. Based on experiences and demands from 2001;118:403-410.
[12] Menck N, Yang X, Weidig C, Winkes P, Lauer C, Hagen H,
users in those studies a method will be proposed. This Hamann B, Aurich JC. Collaborative Factory Planning in Virtual
method should include a structured approach for analysis Reality, In 45th CIRP Conference on Manufacturing Systems
and synthesis work in cross-functional teams to fully 2012, Athens.
utilize the potential of high-level visualization support. [13] Sansoni G, Trebeschi M, Docchio F. State-of-The-Art and
Applications of 3D Imaging Sensors in Industry, Cultural
There are different factors that need to be covered in this Heritage, Medicine, and Criminal Investigation. Sensors
approach such as the overall material flow, material 2009;9:568-601.
handling within the workstation, tasks for replacing [14] ASTM standard E2807: Standard Specification for 3D Imaging
Data Exchange, Version 1.0. ASTM International; 2011.
tools, maintenance etc. [15] Hu H, Fernandez-Steeger TM. 3D Modeling using LiDAR data
The future goal and vision is that the method is and its geological and geotechnical applications. In 18th
developed into a support tool used in all suitable projects International Conference on Geoinformatics 2010, Beijing.
[16] Lindskog E, Berglund J, Vallhagen J, Berlin R, Johansson B.
working with developing manufacturing systems. An Combining Point Cloud Technologies with Discrete Event
envisioned scenario is to use the tool throughout the Simulation. In Proceedings of the 2012 Winter Simulation
whole project by users from a wide scope of areas and to Conference 2012, Berlin.
use the 3D scan data to create the main factory
representation.
Acknowledgements