You are on page 1of 11

Journal of Manufacturing Systems 31 (2012) 444454

Contents lists available at SciVerse ScienceDirect

Journal of Manufacturing Systems


journal homepage: www.elsevier.com/locate/jmansys

Technical Paper

Flexibility in manufacturing automation: A living lab case study of Norwegian


metalcasting SMEs
Rhythm Suren Wadhwa
Inst. for produksjons-og kvalitetstek, NTNU Valgrinda, Trondheim 7491, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Flexibility can be dened as the ability to respond efciently to the changing demands of the customer and
Received 29 May 2012 is different in SMEs (Small-to-Medium manufacturing Enterprises) than the traditional OEMs (Original
Received in revised form 17 July 2012 Equipment Manufacturers). Costs involved in implementing manufacturing exibility to meet customer
Accepted 18 July 2012
demand are more important in the SMEs, especially that are labor intensive for example metalcasting
Available online 9 August 2012
companies, when located in a high cost country like Norway. Therefore the Norwegian Research Council
initiated the Autocast project to promote SME business cooperation and exibility by automating the
Keywords:
manufacturing processes within the Norwegian casting/foundry SMEs. The project intends to identify and
SME metalcasting
Foundry automation
improve the areas crucial for exible manufacturing to help SMEs stay competitive in global competition.
Manufacturing exibility Literature conrms the difculty in the development of objective measures for exibility, for example,
Operations because of its property of multidimensionality. Nevertheless, an attempt has been made in this paper to
Norway identify the suitability of essential spheres of manufacturing exibility to a foundry set-up accompanied
by two living lab foundry case studies. The paper also completes the set of design rules specically for
foundry automation and assist exibility in manufacturing. Efforts have been made to sustain the existing
exible automation implementation on the basis of poka-yoke principles. The ndings from the living
lab case study assist in proposing the design for exible foundry automation (DFFFA) guidelines for exible
automation in a consortium collaborative environment could allow for better response to customer needs
and support on-the-move interaction collaboration.
2012 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

1. Introduction to Europe, with a huge potential to generate wealth, jobs and a


better quality of life. Manufacturing activity in Europe represents
Globalization has created an environment of similar opportuni- approximately 21% of the EU GDP and provides about 20% of all
ties for manufacturing competitors around the world [14]. It has jobs (more than 30 million) in 25 different industrial sectors, largely
created a world market driven by erce competition among com- dominated by SMEs.
panies that are located in different parts of the world but produce The ability to respond efciently to the changing demands
similar products. Organizations, both large and small, require a set of the customer and is different in SMEs (Small-to-Medium
of recongurable equipment to meet customer demands of one manufacturing Enterprises) than the traditional OEMs (Original
of kind, or small batch quantities of customized products. Client Equipment Manufacturers). Literature shows that the central dis-
demands for small volumes of customizable product leads to a tinction between large and small rms is the greater external
paradigm shift in how effectively a SME (less than 250 employees uncertainty of the environment in which the small rm operates,
and less than D50 million annual turnover: European Commission together with the greater internal consistency of its motivations
2010); would operate to satisfy varying customer demands. and actions [24]. It has been observed that identifying best practices
Flexible automation allows rapid recongurability of the pro- is a tricky process difcult to implement, which is more notice-
duction system in order to manufacture several different products, able when the companies are small- to medium-sized enterprises
achieving high degree of machine utilization, reduction of in- (SMEs). Typically, SMEs have severe resource constraints and lim-
process inventory, as well as decrease in response times to meet ited knowledge of manufacturing automation methodologies.
the changing customer preferences. Automation is the force behind Costs involved in implementing manufacturing exibility via
the rationalization of manufacturing processes to increase compet- automation, to meet customer demand are more important in
itiveness and productivity. Manufacturing is of high importance the SMEs, especially those that are labor intensive for example
metalcasting companies, when located in a high labor cost coun-
tries. Metalcasting companies are particularly in need of help from
E-mail address: rhythm.s.wadhwa@ntnu.no automation technology, because of some of the drivers listed below

0278-6125/$ see front matter 2012 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jmsy.2012.07.008
R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454 445

Casting process is intensive in manual labor. slow response to customer demand. In the present market envi-
The extreme environmental conditions of handling hot molten ronment, foundry companies are looking for exibility to enable
metal, dust, and HSE issues such as injury from handling sharp shorter manufacturing lead times that could meet the ever chang-
hot cast parts. ing demands of the customers. But for this, it is necessary to have
Inspite of the above drivers as mentioned above these SMEs face economic production of which is possible only if the plant automa-
a major deterrent standing in the way of automating the process tion is well managed because the large setup and handling times in
and that is due to high part variability resulting from differ- foundries not only increase the in-process waste but also adds to
ent metal cooling rates. This makes the implementation of an the inability to respond quickly to customer needs.
automation solution especially a challenge. In general, manufacturing automation is considered a problem
for production personnel only, as its true that the shop oor is
In this article we have chosen an action research approach, the place where the manufacturing problems can be seen. Many
which means that we enter a real industrial situation with the aim other areas can have an impact on the feasibility of automation,
of both improving it and creating knowledge. Action research is for example lean technologies such as poka-yoke. The presented
well suited methodology for the living lab, since both approaches work shows the following two step approach to reducing non-value
emphasize interaction between theory and practical situation, adding activities, particularly for foundry automation:
involves multiple stakeholders with distinct roles, and highlight
the importance of constant reection in order to follow up with Technical guidelines for manufacturing automation
the implementation results. Poka-yoke capabilities to sustain the quality of casted parts pro-
This paper provides a comprehensive compilation of general duced
exible manufacturing parameters that could be applicable to
metalcasting SMEs. A limitation of the study is the number of The presented study will show that the role of the automation
SME foundries (nine national foundries). A detailed analysis of designer is indispensible. He plays an important role in inuencing
results based on such small sample of participants is not practi- the exibility in batch production foundry environment, inuenc-
cal, nevertheless this number is representative of the total foundry ing quick changeovers, before they are actually performed. He is
businesses in the country. The paper helps in the development responsible for all the technical aspects of automation since these
of further studies in the addressing the automation needs of the are determined by the concept and setup of the machine. The
foundry industry. awareness of the importance of having foundry automation guide-
lines can help a manufacturing engineer to build an automation
friendly equipment setup. This work has highlighted this approach
2. Industrial context
as design for exible foundry automation (DFFFA).
A set of technical guidelines are given which might prove use-
Literature shows that the central distinction between large and
ful to be a good practice while planning an automation system.
small rms is the greater external uncertainty of the environment
These rules are suggested from a technical point of view in line
in which the small rm operates, together with the greater internal
with Shingos poka-yoke technique. These are aimed to reduce the
consistency of its motivations and actions [24]. It has been observed
enable exibility elements such as quick changeovers and improve
that identifying best practices is a tricky process difcult to imple-
part quality compliance.
ment, which is more noticeable when the companies are small-
In the presented work, we have used a living lab approach, for
to medium-sized enterprises (SMEs). Typically, SMEs have severe
identifying and understanding the different areas for application
resources and limited knowledge of automation methodologies.
of exible automation in a real world context with the consortium
Additionally, the company culture is not prone to change and learn,
lead users. Living labs started to emerge in the beginning of 2000
as observed in many cases.
and the concept has since grown. A precondition in living lab activ-
There have been very limited published studies for SME
ities is that they are used in a real-world context. A living lab could
foundries, and to help them establish tools to promote competitive-
be understood as an environment of publicprivate partnerships in
ness. Ribiero et al. [19] developed a methodology for benchmarking
which the stakeholders get together to stimulate and challenge the
the Portuguese SME metalcasting consortium (eight foundries),
research and development, by actively taking part in the innovation
to promote business cooperation within the industry and enlarge
process. The approach is similar to open source methodologies, e.g.,
available business information. This study only focused on devel-
open innovation, involving stakeholders and crowdsourcing.
oping a performance measurement framework, and identied
The automated living lab test cell was installed directly at one
developing manufacturing exibility as a critical factor for measur-
of the participating foundries. Followed by the described technol-
ing innovation in manufacturing. Spangler et al. published results
ogy demonstration, discussions have resulted in rst ideas and
proposing manufacturing performance measurement tools related
feedback for improvement. In parallel, several other processes and
to the metalcasting industry in the US based on the needs identi-
experimentations were identied and prepared for real life exper-
ed by the Defense Logistics Agency (DLA). Separate questionnaires
imentation. During the above living lab process, constant feedback
developed for the sand casting, die casting and investment casting
for improvement was collected and transformed into a requirement
suppliers are mailed to the 283 foundries listed with the Ameri-
list for the technology providers.
can Foundrymens Society (AFS) foundry database. The researchers
The automation solutions were implemented over a period of
received 39 completed surveys which are used to develop guide-
20 months. The solution implementation in only two of the nine
lines and perform further analysis to identify relationships between
foundries has been currently disclosed due to proprietary nature
variables for the DoD and DLA.
of the information at present. These two case studies demonstrate
the implementation of the needs identied by the consortium. The
3. Research method living labs reference used in the context of this research, is in the
context of industrial implementation to demonstrate the use of the
Improving exibility is vital to casting SMEs success because a research work. This ts well with the dual role we had in this work
time consuming changeover of dies as well as excessive manual from the research university standpoint. Our responsibility was to
handling of parts, is expensive: it could demand long production emphasize the interaction between theory and practice, involve
runs to justify its cost and this in turn leads to excess inventory and many stakeholders with distinct role relevant in the situation and
446 R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454

use the industrial facilities to demonstrate the application of the manufacturing and assembly operations like painting, cutting,
exible manufacturing concept and improve SME capabilities. welding, etc. These applications comprise only a minor part of the
industrial work where robots can be utilized, especially considering
4. Automation in metalcasting the variety of applications and the number of SMEs.

It is becoming increasingly evident over the past few years that 5. Flexible manufacturing
one of the major possible ways to improve the production rate and
reduce labor costs is to automate the process of handling of parts The analysis of the literature regarding the topic of manufactur-
to cope with varying production rates. The above view is consis- ing exibility, in high volume production environments, highlights
tent with the literature and is justied because of the following the presence of four main categories of scientic works. The publi-
observations cations in the rst group deal with the analysis of the interpretation
of manufacturing exibility and their relationships with manufac-
A review of the current manufacturing processes in most SME turing problems. A key contribution in this area was presented
industries reveals that over 65% of the total manufacturing time by Upton, who dened exibility as the ability to change mini-
is spent in manual handling of the material by human operators. mizing penalty with time. The second group of publications deal
It was also found that the cost of manufacturing a product is with classication of existing exibility forms through concep-
roughly between 30% and 40% of total manufacturing cost, which tual frameworks. An early denition of exibility was provided by
increases with increased labor cost. Gupta [10] who credited Mascarenhas as having dened exibil-
ity. The third category focuses on the development of approaches
Fig. 1 shows the evolution of the price of a particular industrial and models, both qualitative and quantitative, to support the sys-
robot in the average and annual wages for a Norwegian indus- tem design while considering the given system exibility forms.
trial worker demonstrated for a 40-year period. As the price of the Koren [14] studied the relationship between the level of embedded
industrial robot has increased over the years, so have the person- system exibility and the system performance. The fourth group
nel wages. The curve shows that an industrial robot is becoming of research aims at systemizing the higher number of exibility
relatively cheaper. In 1976, an industrial robot purchase cost cor- denitions related to the real implementation of various forms of
responded to 8-years of personnel labor for the investment to be exibility in a manufacturing system.
paid back. This number was reduced to four years in 1986, and one The key feature distinguishing exible manufacturing systems
year in the present day scenario. Most modern production tech- from dedicated manufacturing systems is, in fact, their exibil-
niques are in general characterized by such a development, and ity. Yet, there is no agreement to date on the precise denition
this is not just limited to robotics. of exibility. Literature conrms the difculty in the develop-
The robot is advantageous and often necessary in applica- ment of objective measures for exibility, for example, because
tions that are dull, dirty or dangerous for humans. Traditionally, of its property of multidimensionality. Gupta et al. [10] proposed
industrial robots (above 1 million operational installations world- a method to evaluate value of exibility based on time scale
wide, IFR 2009) are used to replace human labor in manufacturing decomposition of short, medium and long term changes; to simply
such that repeatability, quality and accuracy is increased in the problem to measuring the value of exibility. Investments in

Fig. 1. Price development for industrial robot Cincinnati Milacron T 3 relative the cost of an employee executed by Norwegian industrial workers on average.
Extended from Lien [33].
R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454 447

exible manufacturing scenarios have been difcult to justify due


to lack of techniques to quantify exibility. Wadhwa [29] proposed
a methodical concept utilizing real options to evaluate exible
foundry batch manufacturing system. Pyoun et al. [16] suggested a
method for quantifying exibility. separately analyzed exibility in
mass manufacturing systems from a users and manufacturers point
of view by classifying exibility into 11 and 4 elements respec-
tively, applied to a transmission assembly scenario. Although the
work is about analyzing exibility separately from a systems and
users point of view, however one can observe that these views can
vary widely between industries. Foundry industry, for example,
has signicant challenges in the current regulatory and political
environment with developing an economical sustainable business
model. Nevertheless, the approach could prove useful in batch
manufacturing scenarios.
The work by Fernandes et al. [7] investigates the decisions about
equipment choice along with the impact of product mix distribu-
tion between standard and customized; and is based on the product
mix denition of Berry and Cooper. Wang et al. [31] proposed
a scalability planning methodology by introducing a design-for-
scalability method to address the recongurability issue to meet
the changing demand of the customer. They proposed a genetic
algorithm heuristic to solve the line balancing optimization prob-
lem. The main issue in a foundry exible manufacturing system
is however, the issue of time involved in equipment changeover,
which needs to be addressed to enable efcient batch production in
a metalcasting environment [30]. The die changeover setup process Fig. 2. (a) Flexibility forms and (b) importance v/s difculty of implementation of
depends on the technical aspect of the foundry system, organiza- exibility forms.
tion of work content, and to a great degree to the method to perform
the changeover. The setup time affecting the foundry changeover
the internal processes, and uncertainty and changes in the envi-
exibility also depends on the technique of the machine itself,
ronment. Such factors have affected manufacturing companies for
which in-turn depends on the given cast product specications.
a long time, but the inuence of them has escalated during the
As conrmed from an extensive literature review, the exibility
past 20 years as a result of advances in manufacturing technol-
denitions presented could have different interpretations in dif-
ogy, and demand for mass customization. Nevertheless, an attempt
ferent organizational contexts. The following example illustrates
has been made to identify the suitability of essential spheres of
how a promising exibility form can lead to different results. When
manufacturing exibility to a foundry set-up.
using changeover to evaluate the system exibility different com-
Changeover exibility. This is dened as the number of different
panies dened its components of product modularity, volume and
products being produced concurrently over time and is a result of
product mix, differently according to their individual manufactur-
uncertainty as to the length of product life cycles [9]. An alternative
ing environments (sand casting, aluminum die casting, etc.), and
approach suggested by Gustavsson [11] is to calculate the ratio of
associated it differently with different levels of difculties in imple-
the equipment investment relevant for the next product to that of
mentation. This issue was discussed in detail with the alliance
the total investment.
partners. It was observed that the lead company sometimes forces
Fig. 3 shows an aluminum foundry process that was evaluated
the inclusions of particular exibility forms already being used
by breaking the equipment into ve major systems. Those systems
internally in their company. This issue had to be handled with care.
are the core machines, the castlines, the nish lines, the heat treat
On one hand, these forms are tested and familiar. On the other hand,
cells and the pre-machining line. Though each of these areas may
including all options in the proposed exibility implementation
include several machines or sub-systems, they will be evaluated as
proposal, could lead to an excessive number of choices.
a department, based on the input and output information available
In the next step, the number and denition exibility forms
from the plant. Fig. 3 shows the equipment layout and Table 1 shows
was rened. The nine companies scored, using ve point Likert
the machine assignment, where b1, . . ., b6 denotes the in-process
scales, the exibility forms against two dimensions: importance
buffers.
of the exibility form in an SME context, and difculty in imple-
From the structural model, we put the model into the aggre-
menting it in a manufacturing facility. The importance scale ranged
gation machines following the aggregation algorithm for parallel
from not important (1) to highly important (5) and the difculty
machines and serial machines. Putting all the factors together, the
scale ranges from easily implementable to difcult to implement
nal structural model of the entire system was simplied as a 4
(5). This procedure helped in identifying the exibility forms that
machine 3 buffer serial line to conduct throughput and bottleneck
(description in Section 3) relevant to manufacturing SMEs. Fig. 2
shows the results that came out of this procedure.

5.1. Dening exibility in a manufacturing context

In recent years, exibility has attracted signicant attention


from both SME industry and academia due to the external factors
resulting from varying customer demands and increasing competi-
tion. Changing operating conditions are forcing rms to be exible
in terms of handling variations in demand, product variations in Fig. 3. Foundry layout structural model.
448 R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454

Table 1 In foundry automation context, the term machine refers could


Machine number assignments for the system structural model.
refer to assembly robot.
m1 Core machine 1 The machine level provides the basic framework for exibility.
m2 Core machine 2 In other words, machine exibility is necessary for other exibili-
m3 Core machine 3
ties. Machine exibility allows for:
m4 Castline 1
m5 Castline 2
m6 Castline 3 Lower batch sizes and resulting savings in inventory costs
m7 Finish line 1 Higher machine utilizations [12]
m8 Finish line 2 Production of complex parts
m9 Finish line 3
Shorter lead times for new product introductions [5]
m10 Heart treat cell 1
m11 Heart treat cell 2
Better product quality realizations in the face of random varia-
m12 Pre-machining tions in input quality [19,25]
Achieving machine exibility

Electronic and software development might help to create


analysis on the system, acting as an input to quick changeovers for machine exibility in a cost-efcient way. The rst aspect of
batch production parts. machine exibility can be measured by the number of different
Product modularity exibility. The modular product design with operations that a machine can perform without requiring more
component modularity helps to leverage product variations and use than a specied amount of effort [19]. The second aspect of machine
of common components across product families. Modular product exibility may be measured in terms of time and/or cost required
conguration can reduce the variety of components while offering in switching from one operation to another [30].
a wide range of products. Modular exibility helps in reducing the Part handling exibility. Flexibility of a material handling system
lead-time of producing new products and allows combining stan- is its ability to move different part types efciently for proper posi-
dardized modules in a variety of ways to create unique nished tioning and processing through the manufacturing facility it serves.
products. The denition covers loading and unloading of parts, transporting
The components embedded within a certain allowable range them from machine to machine, and eventually storing them under
of variation provide great exibility in composing the nal prod- varying conditions of the manufacturing facility. This denition is
ucts. The standardized interface of components in modular product consistent with the discussion in Stecke and Browne [2] and Chat-
design enables effective coordination of development in mod- terjee et al. [6]. Kusiak and Chatterjee et al. dene the ability of
ular organization structures. In the context of the described the material handling system in terms of physical location of each
living lab setup, the part modularity could assist in developing group of machines. They also dene linkages between each pair of
pre-programmed gripping congurations leading to an 100% auto- groups and between each pair of machines within each group, and
mated factory. the times for every possible move between machines. From these,
Product mix exibility. Product exibility relates to the ability of it is possible to dene the set of all possible material paths that can
the system to cope with changes products to be processed by the be supported in the factory.
system. Generally, the productivity tends to decline with increas- Better layouts, more space, cleaner environment, better
ing job exibility [3]. Most of the emphasis on achieving exibility ergonomics, and the use of autonomous work groups improve
in manufacturing systems has been related to product exibility. material handling exibility in assembly environments [13].
Essentially this can be achieved either at the machine level or at Parallel-assembly arrangements allow an assembler not to be
the system level. At the foundry machine automation level, prod- impeded by a person who might be slower or experiencing a prob-
uct exibility could be achieved by increasing the capabilities of the lem. Changes such as local testing and inspection in the assembly
robot by using ATC (automatic tool changer) to handle multiple part process increase the cycle time. These changes may considerably
families. The robot could be made more versatile, more complex, expand the required skills of assemblers, as they must learn to test,
and is provided with more sophisticated controls. inspect, adjust, and repair products, as well as to assemble them.
Alternatively, job exibility can be achieved at the system level Various manual and motor-propelled jigs and xtures rotate
by distributing the required capability among a variety of machines parts and assemblies through several axes and adjust to the needs
or work stations, each of which would then be specialized to cer- of a range of assemblers. Pallet xture exibility determines the
tain processing tasks. As a result, this introduces the requirement degree of freedom available to part loading schedules [15]. This
to move jobs between machines in order for all important opera- exibility also has to be taken into account when dealing with
tions to be performed and so a materials handling system such as buffer size, which is the ability to accommodate different parts of
material handling systems and conveyors must be provided [3]. different shapes and sizes, and the readjustment of paths in case of
Plant layout exibility. Flexibility in facilities design literature is expansion.
often dened as the capability of a layout to react to disturbances Production exibility. Production exibility is the universe of part
caused by future change. Four areas that can affect a shop layout types that a manufacturing system can produce without adding
are changes in product mix, changes in product volume, changes in major capital equipment. Minor resources such as new tools may
the process, and changes in raw materials used in producing these be allowed in order to dene the relevant universe. In contrast
products. Thus, the layout exibility contributes to mix and part with product exibility, note that production exibility may allow
mix, volume, process, and material and raw material exibilities. considerable setups, but not major capital equipment. In a way,
Design of production plant is related to spatial arrangements production exibility denes the set of existing and potential parts
of machines and robots, mobile robot handling devices, material- from which the candidate parts can be drawn as the new parts
product storage areas, and automated guided vehicles. Total considered in the denition of product exibility [1,2,19,27].
estimated annual material handling costs for a particular design Slack [22] emphasized the importance of labor exibility in
are used provide a quantitative measure of a design exibility. obtaining production exibility. He also surmised that product ex-
Machine exibility. Machine exibility refers to the ease of mak- ibility is a response exibility, whereas production exibility is
ing the changes required to produce a given set of part types a range exibility. This denition is similar to those proposed by
[1,2,27]. Browne et al. [2].
R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454 449

Production exibility allows the rm to compete in a market should be able to support the user in the production modeling
where new products are frequently demanded. Production exibil- activity, feeding the decision support tools with the required
ity minimizes the implementation time for new products or major data, considering exibility as a requirement in the knowledge
modications of existing products [5]. On the operational level, it management framework, i.e., the model must be adaptable in
permits an increase of part families and allows the rm to diversify order to describe many different production system scenarios,
its risk. As Browne et al. [2], and Chatterjee et al. [6] imply, pro- process and product exibilities. Integrating exibility measures
duction exibility depends on the variety and the versatility of the which serve as an input back to the production system could help
machines that are available, the exibility of the material handling the SME to operate in a Just-In-Time (JIT) environment [32].
system, and the factorys information and control system.
If a exible material handling system is in place, in terms of 6. Living lab setup and use case
part types, production exibility is more than the sum of the parts
that each individual machine can make. Production exibility is 6.1. Case study one: exible robot part handling in an iron
related to the properties of the transportation system, warehousing foundry
system, interfacing system, distributed data bases, systems control,
Preferring to remain anonymous, the iron casting is a manufac-
and software modularity.
turer of home heating ovens. The company faces seasonal customer
An open communication system as well as the use of a com-
demand patterns, and it produces eight series of ovens, with four
mon communication protocol will help to increase production
variants each. Two series have special coating processes after the
exibility.
individual assembly parts are casted, and have special handling
Operation exibility. Operation exibility is a property of the
requirements.
part, and means that the part can be produced with alternate pro-
The company wanted to look into automating the manu-
cess plans, where a process plan means a sequence of operations
facturing operations targeting at reducing the manual handling
required to produce the part. A process will be considered to have
operations by automating the CNC machining for the parts with
operation exibility if parts that are being produced in the system
a number of machining operations (>30/part) and heavy weight
possess operation exibility and if the material handling system is
(Fig. 4). A description of the automation cell is provided below.
able to deliver parts to machines in different possible orders. The
The implemented automation system consists of seven mod-
denition of the operation exibility of a process is consistent with
ules: (i) the vision module (Tordivel Scorpion), (ii) the robot module
Browne et al. [2] and Chatterjee et al. [6].
(ABB 6400), (iii) the robot end effecter part handling module, (iv)
Operation exibility of parts contributes to various system ex-
the automated storage lift, and (v) the CNC machine (CinCh-30S).
ibilities, especially routing exibility. Additionally, it allows for
(vi) The RFID tags with (vii) part family xtures (Figs. 5 and 6).
easier scheduling of parts in real time [2] and increases machine
When there is an order from ERP system to meet a request down-
availability and utilization, especially when machines are unreli-
stream, the HMI requests the bin selection from the storage, and the
able. Parts that are standardized or modular [11] are likely to exhibit
robot waits for the signal notifying that the bin is in place under the
operation exibility. The number of different processing plans for
vision system, and noties the camera via TCP/IP (through PLC), to
its fabrication can measure operation exibility of a part [16].
take the picture. The position and orientation of the part/xture is
Response exibility. Response exibility is dened as the ease of
transferred to the robot via the PLC, which then proceeds to orient
moving from one manufacturing system conguration to another,
the gripper accordingly to transfer the part on the CNC xture.
within the existing capability and capacity envelope [21].
Part routing exibility. Routing exibility refers to the dynamic 6.2. The robot grippers: three industrial versions
assignment of parts to machinesthat is, the rerouting of a given
part if a machine used in its manufacture is incapacitated [8]. Rout- Handling the parts once they arrive at the machine is also
ing exibility also delivers the ability to handle breakdowns and to important. Implementing machine exibility with the use of robots
continue producing a given set of part types [1,2,4]. with recongurable grippers and intelligent interfaces (exible
Another denition of routing exibility of a manufacturing workspace, vision system, etc.), automatic tool changers and mul-
system is its ability to produce a part by alternate routes (or paths) tiaxis robots can help enhance material handling exibility at the
through the system. Alternate routes may use different machines, most affordable price. Fig. 6 shows a exible laboratory prototype
different operations, or different sequences of operations. Typi- tested to handle the parts [28] and Figs. 79 show the three imple-
cally, these different machines (e.g., lathe and milling machines or mented designs at the foundry.
two brands of grinders) are those capable of essentially the same The rst version of gripper developed is shown in Fig. 7. This
processes. gripper works well in lifting small at parts, the weight ranging
The above analyzed exibility parameters could be incorpo- from 2 to 5 kg. The gripper consist of an off-the-shelf at elec-
rated in an integrated knowledge framework to support exible tromagnet 60 mm, attached to the robot end via a pneumatic
manufacturing decision support. The need for integration of the cylinder (Festo, DFM-20-40-B-PPV-A-GF, pmax = 10 bar). Gripper I is
exibility aspects derive from the viewpoint of manufacturing limited in its capabilities when handling complex parts with special
SMEs, who would want a view on manufacturing and management contours. Hence there was a need to develop a specialized electro-
of product lifecycle to be as complete as possible. Moreover, it magnet to handle the parts at the foundry; and the development
could serve as a guidance to handle the most critical exibility and results of the use of the special magnet is currently proprietary
issues occurring within a system problem. For example, starting information.
from the product evolution standpoint, it is possible to take The motivation for the version ii design was to have a exi-
decisions on, e.g., which degree of exibility to acquire to best ble gripper that could handle parts of size range 200320 mm. The
meet the production needs or opt for more dedicated manufac- length of the arms is 260 mm (including 40 mm height of 80 mm
turing solutions. Regarding the integration of information, Kimura magnet) with a 3 compliance at the joint. Triangular workspace
proposed a framework for product and process from a virtual ranges from 350 to 500 mm. Version II weighs 25 kg and has a pneu-
manufacturing viewpoint. Thibault et al. presented Ontoforge to matic cylinder with stroke length 30 mm (Rexroth 63 H 30, 0822
support the integrated design in forging processes. Bernard et al. 010 875) on the top center, and range for the three arms is 100 mm
proposed a meta model structure to link the function to external using linear step motor (Haydon Switch and instrument, E57 H4A,
conditions. The framework to manage manufacturing information 3.25 A).
450 R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454

Fig. 4. Metalcasting process.

6.2.1. Limitations
Stepper motor does not allow for position control for mag-
net. The range of parts that the gripper can handle is limited. The
pneumatic cylinder holding the electromagnets causes problems in
picking curved parts (electromagnet gripper ends do not touch the
part completely), for example, if the part movement due to moving
conveyor belt vibration or the part orientation.
The version iii gripper weighs 20 kg has a stroke length of
150 mm on the three sides of the gripper. The electromagnets move
over a distance of 40 mm and are attached via a pneumatic cylinder
(Festo) to the upper rails. Three servo motors 24 V (Faulhaber) each,
are used for the movement on the rails. The voltage used for the
electromagnet is 24 V DC, with a total power consumption of 12 W
at still position, and 84 W at operating condition. The two main dif-
ference between the versions II and III is the workspace covered by Fig. 6. Testing the Universal Passive Gripper with the foundry parts.
the electromagnets and the use of servo motors to overcome the
limitations of stepper motor use.
This version has the advantage that the speed of movement was vision techniques like edge detection are not efcient for casted
4 times faster than the version II because of the use of servo motors. parts which have ash on the edges and hence varying outer dimen-
Accuracy is also better and faster for the servo motor. sions. The markers could only be placed on one side of the part, as
The gripper was able to pick more parts than version II, as the the other side needs to be used for aesthetic reasons in product
magnets are able to move closer in the center handle smaller parts. assembly. Additionally, the texture of the part varies because of
The three versions of grippers were designed to assist in quick varying molten metal cooling rates and during change of material
changeovers with an automatic tool changer, to handle the product
mix at the iron foundry. The installed automation cell works with-
out human intervention. There are some issues that could interrupt
the continuous operation. For example, the company suggested the
use of markers (Fig. 10) to aid in accurate estimate of the parts 3D
orientation in the loading bin. Markers were necessary because 3D

Fig. 7. Version I gripper principle and prototype.

Fig. 5. The automated cell with the robot, CNC machine, the automatic storage lift
and the vision system. Fig. 8. Version II gripper principle (a) and prototype (b).
R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454 451

for handling automation. The company wanted to look into the


design for transport pallets and overhead conveyors for the inter-
nal transport of castings which required signicant investment and
maintenance costs as they were specially tailored for each new
product. During production, parts are transported around the fac-
tory on pallets that are attached to different assembly lines. In order
for robotic grippers to intervene parts it is important that they are
positioned securely. Moreover, handling sand cores requires addi-
tional constraints to ensure it should not be worn or cracked during
Fig. 9. Version III gripper workspace principle (a) and prototype (b). transport.
Fig. 10 shows the xture pins holding the sand core (yellow) in
place. These blocks are adapted to sand core shapes to ensure that
the sand core, both have great enough rest area and that it will be
xed properly with the pallet (Fig. 11).
Various manual and motor-propelled jigs and xtures rotate
parts and assemblies through several axes and adjust to the needs
of a range of assemblers. Pallet xture exibility determines the
degree of freedom available to part loading schedules. This exi-
bility also has to be taken into account when dealing with buffer
size, which is the ability to accommodate different parts of dif-
ferent shapes and sizes, and the readjustment of paths in case of
expansion. The solutions suggested below were based on modular
principles enabling ease of production, and the identied need for
short changeover time, reuse of components and reducing the need
to adjust the pallet for exible handling of new component designs.
Considering the above requirements one of the four imple-
mented modular solutions for handling aluminum parts and sand
cores are shown in Figs. 12 and 13. The same standard base plate
is used to transport sand cores and aluminum parts. The alu-
minum parts are supported by three exible points. Three arms
Fig. 10. Markers casted on the part. (For interpretation of the references to color in are adjustable in the xy- and z-direction ensures that all products
the text, the reader is referred to the web version of this article.)
within a product family (swing arms in this case) can make this pal-
let. The implementation presents an integrated solution in which
for different production batches (variation in molten metal shades no part must be replaced when the pallet is to be converted. Alu-
within a batch is less than between two batches). minum part location pins are adjustable in height direction (z) by a
The part orientation is recognized by the vision system when tube to move freely inside another, the same applies to the longi-
the markers are on top, but if the parts are randomly mixed in tudinal direction (x). The arms can also be turned around the z-axis
the incoming bin, the vision system slows the production by not of the tubes can be rotated independently. When arms are prop-
being able to identify the part. Other issues that could be a problem erly adjusted in all directions, they must be xed in this position
include a molding failure to the markers, or if the part has a sur- (Fig. 14).
face defect due to differential cooling of the casting. If any of these The modular part handling solutions offer routing exibility that
events occur, the container is marked as error and stored back in the aluminum foundry desired.
the lift, and a new one is pulled from the rack; operator can take
care of the issue when he returns back to the automated cell. 6.3.1. Limitation of the case studies
There are very few foundries that have raw material exibil-
6.3. Case study two: modular concept for part handling and ity and that can run iron, steel and aluminum. Once a foundry is
xture exibility in aluminum foundry operating on one of these metals, the casting nishing system is
quite different for each. Casting expertise required to produce X-
Case study two is an aluminum foundry which wanted to ana- ray quality aluminum castings for automotive is much higher than
lyze how and where the company could use modular components casting iron.

Fig. 11. Sand cores in xtures.


452 R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454

Fig. 12. Locating pin principle for handling aluminum parts.

Fig. 15. (a) Markers as seen on the part under normal lighting. (b) Screen shot from
the vision system showing the reection issues and disappearing lines at high angle
tilting, but the circles are still visible.

Fig. 15b shows that the circular markers have always a certain
part recognizable by the vision system under light, whereas the
straight line markers are almost disappearing.

Fig. 13. Aluminum parts xture on standard base plate.


7.2. System interface issues

Every component in the implemented automation cell has its


7. Challenges in implementing manufacturing exibility in own preferred communication protocols and standards. For exam-
foundries ple, the Faulhaber servos inside the griper can only use CAN-bus, so
one needed a PROFIBUS to CAN-bus gateway. As a result of this one
Fundamental research and follow up experimental studies is a needed to set up CAN bus to I/O modules inside the gripper. Or one
benecial way to approach conceptual challenges our society is would have two separate BUS lines into the gripper. The key issue
facing in order to develop and incorporate automation and techno-
logical solutions to sustain high level of standards in the future. It
is essential that the requirements of the system govern the choices
and the decisions made during the system design. The importance
and adoption of industrial standards such as STEP and PLCS as a
reference to develop the manufacturing process framework. The
following array of automation challenges were identied during
the analysis and implementation of automated solutions working
with the consortia.

7.1. The vision system challenges: picking from conveyor v/s bin
picking

The functionality of the vision module is to perform object


recognition of a 3D part in space and locate the coordinates. Mark-
ers (Fig. 15a) were casted to aid in accurate estimate of the parts
orientation for the purposes of pick and place operation from a con-
Fig. 16. Clamping attached to the base of the CNC milling machine. The RFID tags
veyor belt and the stationary bin. The iron foundry production parts ensure that the correct xture is loaded and thus requesting the appropriate G code.
were at in nature with a dimensions varying between 200 mm2
and 600 mm2 and thickness between 5 and 8 mm. Picking from a
conveyor belt was conveniently implemented for automation. But,
when parts are together in a random order in a bin, there was a
reection issue with the straight line markers. The straight line
markers were discovered to have a limitation of reection with
a tilt angle greater than 30 for iron castings.

Fig. 17. Fixture markers for vision recognition and locators corresponding to the
Fig. 14. Sand cores located on the standard base plate. xture rotation and sliding locators.
R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454 453

Fig. 18. Ensuring poka-yoke for vision system identication.

in communication between systems is that the protocol sometimes require process modication and increase implementation costs
differs a little from the documentation. (Figs. 1618).

8. Technical guidelines for exible automation 10. Conclusions

The basic philosophy for a foundry automation engineer should In recent years, exibility has attracted signicant attention
be that the automation of manual tasks to enable exibility should from both SME industry and academia due to the external factors
be as easy as possible. Based on the living lab experience, following resulting from varying customer demands and increasing competi-
are some guidelines to enable a exible foundry automation system tion. Changing operating conditions are forcing rms to be exible
helping in reducing non value adding activities: in terms of handling variations in demand, product variations in
the internal processes, and uncertainty and changes in the envi-
The automation cell for major foundry processes, i.e., machin- ronment. Such factors have affected manufacturing companies for
ing ash, transferring parts, etc. should enable minimum human a long time, but the inuence of them has escalated during the past
intervention. 20 years as a result of advances in manufacturing technology, and
Scrap handling operation caused due to defects like rat tails, demand for mass customization. Collaborative engineering is the
hot tear in aluminum casting, and non homogenous mix in iron preferred mode of operation especially for SMEs that provide spe-
casting should be integrated as an automated activity whenever cialized one of a kind product and services to demanding clients. In
possible. the present competitive manufacturing environment, exible con-
Eliminate in-process automation equipment adjustments by guration of a manufacturing system could help fullling the ever
using markers for vision systems in iron foundry operations. changing demands of the customers. Metalcasting industries work
Provide sliding locators to account for part size variations due in lot sizes for economical production and automation is neces-
to varied cooling rates and thus assisting in ash removal dur- sary for efcient changeovers between the lot sizes and for the part
ing automated CNC milling. Enable ofine checking of parts by handling areas; to enhance the ability to respond quickly to the
improving the quality of automation settings in the cell. Use servo changing customer needs.
motors for faster motion and better control. This paper presented a living lab case study and some ex-
Enable checks for ensuring fool-proong the automation enablers ible automation challenges of foundry SMEs. The development
such as automated selection and loading of the correct xture and evaluation of robust analytical models dening technological-
for foundry parts ensuring the use of industrial communication machine and plant level exibility suitable to the foundry SME
standards for CNC and RFID enabled automation. environment constitutes the central objective for the future phases
Design machine feeders based on modular design principles. of the project.
The use of pneumatic or electric actuators is recommended in a
foundry environment, as compared to hydraulic driven ones. Acknowledgments
Complex die designs require specialized tters for setup and
changeover. Consider making it as simple as possible. Financial support from the AutoCast Foundry Consortium and
Enforce law of gravity while designing part feeders for assembly. the Norwegian Research Council is gratefully acknowledged.
Apply principle of standardization for work procedure while
designing automation cells. References

[1] Abdel-Malek L, Wolf C. Evaluating exibility of alternative FMS designs.


9. Poka yoke principles
A comparative measure. International Journal of Production Economics
1991;23(October):310.
Poka-yoke, a term coined by Shingo in Japan in 1960s, implies [2] Browne J, Dubois D, Rathmill K, Sethi SP, Stecke KE. Classication of exible
manufacturing systems. The F.M.S. Magazine; April 1984. p. 1147.
mistake or error-proong. In a foundry automation context this
[3] Buzacott JA. The fundamental principles of exibility in manufacturing systems.
could mean ensuring proper conditions before an automation pro- In: Proceedings of the 1st international conference on F.M.S. 1982. p. 1322.
cess proceeds further. Ideally, poka-yoke should be considered [4] Chandra P, Tombak MM. Models for the evaluation of routing and machine
during the development of automation process methodologies exibility. European Journal of Operational Research 1992;60:15665.
[5] Chatterjee A, Cohen MA, Maxwell WL. A planning framework for exible man-
to maximize opportunities for error proong. Once the process ufacturing systems (WP #87-07-04). Philadelphia: University of Pennsylvania;
is selected, mistake proong opportunities are more limited or July 1987.
454 R.S. Wadhwa / Journal of Manufacturing Systems 31 (2012) 444454

[6] Chatterjee A, Cohen MA, Maxwell WL, Miller LW. Manufacturing exibility: [19] Ribiero L, Cabral J. A Benchmarking methodology for the metalcasting industry.
models and measurements. In: Stecke KE, Suri R, editors. Proceedings of the Benchmarking, ProQuest; 2006. p. 23.
rst ORSA/TIMS special interest conference on FMS. 1984. p. 4964. [21] Shewchuk JP, Moodie CL. Denition and classication of manufacturing exi-
[7] Fernandes R, Gouveia JB, Pinho C. Product mix strategy and man- bility types and measures. The International Journal of Flexible Manufacturing
ufacturing exibility. Journal of Manufacturing Systems March 2012, Systems 1998;10:32549.
http://dx.doi.org/10.1016/j.jmsy.2012.02.001. [22] Slack N. The exibility of manufacturing systems. International Journal of Oper-
[8] Gerwin D. An agenda for research on the exibility of manufacturing pro- ations and Production Management 1987;7(4):3445.
cesses. International Journal of Operations and Production Management [24] Storey D, Westhead P. Management training and small rm performance: why
1987;7(1):3849. is the link so weak? International Small Business Journal 1996;14(July):13
[9] Gerwin D, Tarondeau JC. International comparisons of manufacturing exibil- 24.
ity. In: Ferdows K, editor. Managing international manufacturing. Amsterdam, [25] Tarondeau JC. Technologies exibles et performances. Working paper. Cergy,
The Netherlands: Elsevier; 1989. p. 16985. France: ESSEC; 1986.
[10] Gupta D, Buzacott JA. A framework for understanding exibility of manufac- [27] Voss CA. Flexibility in exible manufacturing systemsa comparative study
turing systems. Journal of Manufacturing Systems 1989;8(2):8997. of three systems: managing advance manufacturing technology. In: Voss CA,
[11] Gustavsson S. Flexibility and productivity in complex production processes. editor. Proceedings of the UK operations management association conference.
International Journal of Production Research 1984;22(5):8018. UK: University of Warwick; 1986. p. 1119.
[12] Hutchinson GK. Flexibility is key to economic feasibility of automated small [28] Wadhwa RS, Monkman GJ, Lien TK. Robust prehension for ferrous metalcasted
batch manufacturing. Industrial Engineering 1984;16(6):7786. product families: a conceptual overview. In: MITIP 2011. NTNU; 2011.
[13] Kolodny HF. Work organisation design in Sweden: some impressions from [29] Wadhwa RS. Towards measuring investment in exible foundry manufactur-
198283. Human Systems Management 1985;5:20717. ing. International Journal of Computer Science Issues August 2012.
[14] Koren Y. Recongurable manufacturing systems: key to future manufacturing. [30] Wadhwa RS. Flexible Manufacturability and handling automation in foundry
Journal of Intelligent Manufacturing 2010;11. SMEs. International Journal of Metalcasting June 2012.
[15] Newman Jr WE. Models to evaluate the benets of FMS pallet exibility. In: [31] Wang W, Koren Y. Scalability planning for recongurable manufacturing sys-
Proceedings of the second ORSA/TIMS conference on exible manufacturing tems. Journal of Manufacturing Systems 2012;31(April (2)):8391.
systems. 1986. p. 20920. [32] Abdul-Nour G, Drolet J, Lambert S. Mixed production, exibility and SME. Com-
[16] Pyoun YS, Choi BK. Quantifying the exibility value in automated manufactur- puters and Industrial Engineering 1999;37:42932.
ing systems. Journal of Manufacturing Systems 1994;13(2):10818. [33] Lien TK. Industri Robot Teknikk. Tapir Forlag; 1993.

You might also like