IMPROVEMENT OF MATERIAL FEEDING
SYSTEM FOR AN EXCAVATOR ASSEMBLY
LINE THROUGH LEAN PRINCIPLES
Raghavendra Ramappa1, *K. M. Sharath Kumar2, Ashoka K. Nairy3
1
Department of Mechanical and Manufacturing Engineering, M. S. Ramaiah School of Advanced Studies, Bangalore,
2
Department of Management Studies, M. S. Ramaiah School of Advanced Studies, Bangalore,
3
LEaP Caterpillar Logistics Technology Services India, Bangalore
*Contact Author e-mail: [email protected]
Abstract
Today many industries are categorised by end customer demand for a wide variety of product models and variants.
Moreover, large number of product variants has implications on the material flows with in the assembly plants. This
requires logistics system that can support production in small lot size for excavator industry. Since, continuous material
feeding system resulted in increased inventory value, operator walk and search time, an attempt to increase production
throughput is made.
In this paper, best practice for material feeding system has been tested to support the assembly line with more
variants in end product. Existing continuous supply material feeding system has been studied to find the effects on
production hindering volume growth. Part flow, inventory and volume growth plan data have been collected and analysed
for effects on production time. Subsequently, operators walk time and distance have been analysed. Consequently,
alternate material feeding systems have been explored through literature. Model station analysis has been performed on
engine line to find the suitable material feeding system using simulation techniques. Finally, lean kitting supply system has
been tested through warehouse zoning and re-profiling activity.
Hence, continuous supply system has been replaced by lean kitting supply system. Results showed reduction of line
side storage space by 69%, inventory value by 54.3%, operator walk time by 78.3% and line side replenishment by 29.6%.
This helped in achieving the takt time of four machines per day. Further, kitting gave flexibility to introduce five new
variants in the existing assembly line without any additional investments.
Keywords: Continuous Supply, Kitting, Inventory Level, Warehouse Zoning
Abbreviations
DES
Discrete Event Simulation
JIT
Just in Time
MHE
Material Handling Equipment
POU
Point of Use
SCM
Supply Chain Management
VSM
Value Stream Mapping
searching time. According to TPS and Lean theories,
these issues are defined as waste and have to be reduced
or eliminated especially to increase the line capacity [2,
3]. Hence, company wants to increase their assembly
line capacity in order to manage model variation along
with satisfying customer demand. In order to address
line side storage space and motion waste issues,
company is looking at new techniques to deliver
material. This project focuses mainly on developing
lean material feeding system to eliminate waste
operations at assembly line in order to achieve desired
takt time [2, 4].
1. INTRODUCTION
The growing number of product variants that is
reality for many manufacturing companies results in
more number of parts. These parts are to be delivered to
the assembly lines. Delivering them in the traditional
way with continuous supply becomes a problem since
increase in number of parts demands an increase in line
side storage space [1]. An increase in storage space and
part numbers create longer operator walking and search
time for parts at assembly line. Currently, Caterpillar
Excavator Division is assembling one of the product
types. The machine is assembled in an assembly line
and method of delivering parts to the assembly line is
continuous supply system. Meaning all parts that are
used for assembly are stored in Kanban system at line
side. The variation in the end product causes the line
side stores to keep high inventory of different part
numbers even though usage of some parts are very low.
Increase in assembly capacity demands more line side
space, which currently does not exists. Besides using
big line side space, increases operator walking and
TECH Journal
SAS
2. PROBLEM
Caterpillar manufacturing introduced a new
product called excavator with single variant. Due to
increased customer demand for more variants, the
company decided to introduce four more variants and
increase assembly line capacity from two machines per
day to four machines per day. However, the existing
material feeding system proved to be difficult to achieve
the given takt time. Continuous supply system or line
side storing system proved impractical to cope up with
500 extra parts, thus, leading to increase in search time
and walk movements. As the line side stock increased,
the inventory value increased more than the estimated
plan. So, the company is looking for alternate material
feeding system, which supports the production. The
27
Volume 12, Issue 2, September 2013
work reported in this paper aims at improving the
material feeding system for an excavator assembly line
to assemble four machines per day and reduce line side
inventory value by 50%.
excessive storing which according to Lean theory is
called as waste.
Decanting (Re-packaging) of Parts: Suppliers were not
delivering parts in right kanban quantities and in right
packaging that fits line side stores. This was the reason
why parts had to be replaced after receiving. This not
only involves time but also involves cost for this
additional operation. This additional handling may
cause quality defects.
The scope of the work is to analyse and
recommend reliable material feeding system, which
helps to reduce line side storage space by 70%, line side
replenishment by 20%, line side inventory value by
50% and operator walking time by 75%. The paper also
brings out methods to improve warehouse process that
supports new material feeding system. Further the
feasibility of the process has been validated before
implementation.
Big Line side Stores: Since company is following
continuous supply system, all parts were stored at line
side and even some parts were used less frequently. This
was a main problem having more parts at line side
which was consuming more space. This results in time
consuming operations for operators to walk and search
for parts required for assembly. This was against the
theory of 5S vioalting Lean rules. This was directly
affecting end product throughput time.
2.1 Methodology
Product, process, materials and inventory
related data has been collected and analysed to
understand the current situation. Alternate lean material
feeding system has been analysed and tested for benefits
using model station analysis. In continuation,
warehouse-zoning activity has been performed to
eliminate waste replenishment movement to line side.
Finally, Material Handling Equipment (MHE) analysis
has been carried out for best utilisation. Lastly, viability
of the attained solution has been validated using
Discrete Event Simulation (DES) for aisle congestion
and material flow.
Unstandardised Work: Since parts picking is a laborious
job for the operators, they started collecting parts in
batches required for one shift and this affected the
standard work which in turn affected end product
quality.
As a part of next step, space analysis has been
done for the assembly line and found that 75% of the
stations have space constraints due to addition of new
variants as depicted in Figure 2.
3. PROBLEM ANALYSIS
Data has been collected to analyse the current
continuous material feeding system in an excavator
assembly line. Parts, inventory and volume plan
information became the primary data for the analysis.
Figure 1 shows the effects of current material feeding
system.
Fig. 2 Space analysis report
In addition, storage space constraint in engine line has
been shown in Fig. 3. By virtue of this analysis, root
cause of the problem has been identified. So, finding
alternate material feeding system has been considered as
the solution procedure, which can be tested on one
model station to analyse improvements.
Fig. 1 Effects of current material feeding system
Part Shortages: The biggest issue with which assembly
had been getting affected was getting right part at right
time in Point of Use (POU). This has got many reasons
inbuilt with the feeding system and led to poor material
planning due to lack of visibility, supplier delays, poor
response time when replenishing line side and poor
inventory control. Many times part shortages were
known when the part was required for assembly.
Poor Kanban Quantity: Line side storage is poorly
balanced. Some parts had one week worth of inventory
at line side and the next part had only 3 days worth of
inventory. This poorly balanced inventory level leads to
TECH Journal
SAS
Fig. 3 Storage space constraint in engine line
28
Volume 12, Issue 2, September 2013
Now, collated data has been used for
comparing the kitting benefits with different category of
parts.
4. SOLUTION PROCEDURE
Engine line consisting of 197 parts has been
considered for model station analysis. Now the question
is to decide on the parts, which are eligible for kitting.
As discussed earlier, 100% kitting is not a good
decision, unless company has some reason to bare
additional logistics cost [5, 6]. So, decision to select the
category of parts [7-13] for kitting analysis has been
reported in table 1.
Number of times parts getting replenished at line side
(LR):
If parts are stored at line side:
LR=N Kp D / Q
If parts are kitted
Table 1. Category of parts taken for kitting analysis
Sl. No.
................. (1)
LR=Kpk PD/∑(Kpk N)
Scenario
................. (2)
Stores replenishment per day (SR):
1
All selected parts kitted
If part stored at line side
2
All standard parts kitted
SR=NKpD/Q
3
All hand weight mandatory parts kitted
4
All mandatory and selected parts kitted
5
All parts less than Rs. 50,000/- kitted
6
All hand weight parts kitted
7
All parts kitted
If part is kitted
8
No parts kitted
LS=0
.................. (3)
If parts are kitted
SR=Kpk D/Q
.................. (4)
2
Line side storage space (LS) in m :
If part stored at line side
LS=NKpS
................... (5)
.................... (6)
2
Kitting Space (KS) in m :
Where,
Hand weight parts = Parts < 8kg
Standard parts = Parts that goes on every engine
Mandatory parts = Parts that goes on every engine but
different variables are available
Selectable parts = Parts that can be selected to go on to
engine (E.g. A/C)
For model data input, let
PD:
S:
V:
Q:
U:
N:
D:
W:
KW:
TU:
BP:
FP:
OV:
AP:
WS:
Kp:
Kpk:
Npk:
If part stored at line side
KS=0
..................... (7)
If part is kitted
KS=Kpk S
..................... (8)
Operator walking distance to pick the part (WT):
If part is stored at line side
WT=2NKp D (W/OV) / 60 ............... (9)
End product demand per day in numbers
Line side storage in m2
Value in Rupees of part, i.e. purchasing cost
for supplier
Kanban quantity of parts in numbers
Part usage in %
Number of parts
Demand of parts per day,
Where, D=U N PD/100
Average walking distance in meters for the
operator to pick part from line side store
Average walking distance in m for the
operator to pick from kit (Estimated Distance:
0.5 m)
Type of usage for parts
Business process parts
Flow path parts
Average walking speed of an operator (TPS
Standard -1m/sec)
Average time in the area to move the kit
container between different component
containers (Standard-0.5m/sec)
Number of work stations for engine assembly
Kitted parts
if part is kitted then Kpk = 1 else Kpk = 0
If part is not kitted Npk = 1 else Nkp = 0
If part is kitted
WT=2Kpk D (KW/OV) / 60 ............. (10)
Value of inventory of the parts (LV) assuming JIT
supply without kit buffer):
If parts stored at line side
LV=NKp Q V .....................(11)
If part is kitted
LV=Kpk VD WS/2PD...........(12)
Number of times part has to be physically handled (PH):
If parts stored at line side
PH=NKpD
.........................(13)
If part is kitted
PH=2Kpk D ....................... (14)
Kitting time for part per day (KT):
If part stored at line side
KT=0
................... (15)
If part is kitted
TECH Journal
SAS
29
Volume 12, Issue 2, September 2013
KT=PDKpkS/2KS + Kpk D AP/60 .... (16)
Analysis Report:
Fig. 9 Physical part handling per day
Fig. 4 Line side replenishment per day
Fig. 10 Kitting space
With above analysis, three scenarios has been
proved to be beneficial namely, “All parts kitted”, “All
hand weight parts kitted” and “All parts less than
Rs.50,000/- kitted”. Based on the logistics cost [14]
involved in additional material handling of all parts
kitted, the second scenario proved to be the optimal
solution for kitting operation. Comparing with the
existing continuous supply system, the result have been
validated and is shown in Figure 11.
Fig. 5 Stores replenishment per day
Fig. 6 Line side storage space
Fig. 11 Hand weight parts vs. continuous supply
a)
b)
c)
d)
e)
f)
Fig. 7 Operator walking time and kitting time
Analysis solutions revealed:
Reduction in line side replenishment by 39.13%
Reduction in line side storage space by 85.29%
Reduced operator walking time by 79%
Reduced line side inventory value by 71.42%
Stores replenishments remained same
Increase in physical part handling by 80%
The next step was to define the replenishment
process which follows kanban system where empty kit
cart or pallet will be trigger for next replenishment [15].
The mode of kitting all hand weight parts is called as kit
cart, which has been designed and reported in this paper
as shown in Figure 12.
Fig. 8 Line side inventory value
TECH Journal
SAS
30
Volume 12, Issue 2, September 2013
Fig. 12 Kit cart
The implementation of kitting showed the benefits
in engine line. This type of material feeding system as a
combination of kitting and kanban replenishment has
been implemented in all other stations of assembly line.
But, before that how much efficient is this kitting
process is a big question. Since, kitting by itself is a
non-value added operation. There is a requirement to
plan the process in an effective way. Thus, Lean kitting
has been viewed as next approach.
Fig. 13 Parts Zoning Approach in Warehouse
An analysis was carried out to find the parts, which
fall under three categories namely: fast moving parts,
medium and slow moving parts. Based on this, fast
moving parts contributed up to 75% of the warehouse
parts. Accordingly, warehouse rack sections were
modified with three zones (Figure 14)
Approach to Lean Kitting
Kitting process can be made effective by following
Lean principles of reducing wastes in the existing
system [16- 18]. Some of the action items have been:
a)
b)
c)
Reduce kitting cycle time
Reduce kitting man power
Reduce number of partial material packages
returned to stores
Elimination of the causes of kitting errors has been
caused by the following criteria:
a) Insufficient quantity of kanban
b) Excessive quantity of parts
c) Wrong parts
d) Incomplete kits
e) Insufficient quantity of packages
Fig. 14 Three zoned warehouse rack section
Gold Zone: The opening gold zones were easily
reachable parts by using order pickers and also hand
reachable. When pick tags were generated with station
wise zoning, the tags get bundled and kitting operator
finds it easy to pick all parts at this reachable level.
Thus, reducing the kitting time.
Warehouse Zoning Solution
To reduce the picking time for the kitting
member, zone wise arrangements of parts were planned
in warehouse. Racks configurations were designed to
support FIFO for all the parts.
Silver Zone: The opening silver zones were just above
the gold zone where the medium fast moving parts were
placed. These parts were replenished either in pallets or
as-is-condition from supplier with quantity more than
one. Order pickers or reach trucks were used to pick the
parts.
In earlier condition, whenever kitting operator
gets the list to pick the parts, operator used to run
around the whole rack sections. In what follows,
searching parts used to consume most of the time.
Moreover, parts have been arranged in alphabetical
order for easy identification. But, operator has been
using order picker and reach truck to pick even fast
moving parts. Since, this process consumes more time,
it warranted immediate attention. Parts in the rack bins
were named as “Add to existing stock” whenever parts
were not picked in FIFO way. More than that, all
manual process was followed with no tracking in place.
To improve this, a new process was adopted known as
zoning of warehouse parts. This approach was divided
into two ways, one to address productivity and second
one for quality as reported in Figure 13.
TECH Journal
SAS
Bronze Zone: The opening bronze zones were at the
top level of the racks in which slow moving parts were
placed. The frequencies of picking these parts were very
low. Reach trucks were used to pick the parts in this
zone.
This activity resulted in reduction of cycle
time by 45% and reduced MHE movements. Further,
manpower will be reduced once the process gets
stabilised.
Warehouse Lean Assessment Summary:
•
•
•
31
System driven put away location
100% FIFO for all parts, including floor locations
Kanban with barcode scanning followed for D class
parts in each station
Volume 12, Issue 2, September 2013
•
•
•
Bar coded pick and put away tags
Separate storage types to identify the parts that
require painting with the system based provision to
move the parts back after painting
Visibility of inventory and Work-in-Progress
Simulation Analysis Result:
Simulation analysis was conducted to validate the
developed solution for the following parameters using
DES. Aisle Congestion analysis was been depicted in
Figure 15 and Figure 16 respectively. MHE utilisation
summary and cycle time validation for reduced
inventory at line side was documented in Figure 17 and
Figure 18.
Fig. 15 2012 Aisle traffic
Fig. 16 2015 Aisle traffic
Fig. 18 MHE utilisation summary
Also, parameters like reduced inventory at
line side, MHE breakdown effect were analysed using
simulation tool and observed good results. This
recommendation was implemented from the developed
solution. Due to confidentiality issues with the
company, simulation analysis report has not been
reported in this paper.
4. RESULTS AND CONCLUSIONS
The project was carried out with an objective
of reducing the identified MUDA in material feeding
system by implementing the best practices using Lean
concepts. To accommodate the growth in volume due to
market demand, the bottleneck was identified to be
meeting system, which hindered the assembly line
expansion. Instead of adding new process and resource
to achieve the desired takt time, company encouraged
the right approach of clearing bottlenecks in the system
to utilise the existing resources and line set up to its full
capacity. The engine line after implementation of kitting
Fig. 17 MHE utilisation summary
TECH Journal
SAS
32
Volume 12, Issue 2, September 2013
•
has been shown in Figure 19. Also, UF03 station after
kitting implementation has been depicted in Figure 20.
Line side inventory level: 54.3% against the target
of 50%
In parallel, Value Stream Mapping (VSM) [19] after
solution implementation has been reported in Figure 22.
Fig. 19 Engine line after implementation of kitting
Fig. 22 VSM after complete solution implementation
Conclusions:
The solution developed has been validated
after implementing for 30 days through pilot study for
evaluating kitting process accuracy. The result has been
“Zero” POU miss and 100% accuracy in pick and put
away process during the month of January 2013. The
output has been validated by steering committee, which
comprised of Manufacturing, Industrial Engineering,
Supply Chain Management (SCM), Plant Manager and
Industrial Guide. Standard work sheets has been
developed for standardising the process and
documented. Further, record accuracy management
team validated the material flow and storage accuracy,
which yielded positive result.
Fig. 20 UF03 station after kitting implementation
Finally, this solution was implemented in all
assembly and sub assembly stations and observed an
improvement by 85% in all workstations. Nevertheless,
in some stations, due to pending management decision
and big and pallet parts handling, kitting process did not
show much benefit. In summary, the attained result has
been compared with target results as shown in Figure
21.
Based on the study, following conclusions can be
drawn:
•
•
The study carried out in this paper recommends a
lean way of material feeding system to support
assembly line in order to achieve desired volume of
four machines per day
Following output were observed against the target
planned when kitting was implemented:
Line side replenishment: 29.6% against the
target of 20%
Line side storage space: 68.3% against the
target of 70%
Operator walk time: 78.4% against the target
of 75%
Fig. 21 Result – target vs. achieved
Line side inventory level: 54.3% against the
target of 50%
Consolidated results after kitting implementation
have been collated as follows:
•
•
•
•
Line side replenishment: 29.6% against the target of
20%
Line side storage space: 68.3% against the target of
70%
Operator walk time: 78.4% against the target of
75%
TECH Journal
SAS
33
To support Lean kitting, MUDA should be
addressed from warehouse, starting from inbound
till outbound. Zoning concept proved to be a good
solution to address such issues and improve kitting
process
Volume 12, Issue 2, September 2013
[11] Johansson B., Johansson M.I., High automated
kitting system for small parts – A case study from the
Uddevalla plant, Automotive Technology and
Automation, Vienna, pp. 75-78, 1990.
5. RECOMMENDATIONS FOR FUTURE
STUDY
Warehouse can be further optimised and wellplanned using Lean ways of material storage like twobin kanban system which supports direct kitting from
warehouse.
[12] Johansson M.I., Kitting systems for small parts in
manual assembly systems: Production Research
Approaching the 21st Century, pp. 25-30, Taylor &
Francis, 1991.
POU material storage space can still be
reduced by using some of the storage techniques like
mixed profiled storage, hybrid racks and hose trolleys to
reduce storage space.
[13] Medbo L., Assembly work execution and materials
kit functionality in parallel flow assembly systems.
International Journal of Industrial Ergonomics, Vol. 31,
pp, 263-281, 2003
By exploring some advanced features in SAP,
human intervention during pick, put away and kitting
can be reduced to achieve more accuracy.
[14] Anonymous, Is third party logistics in your future?
Modern material handling, ProQuest Science Journals,
Vol. 55, Issue 14, pp. 3-15, 2000
SCM, Global Purchase and Packaging team
should work together from the start of the project to
achieve best inbound and outbound material flow.
[15] Schwind G.F., How storage systems keep kits
moving, Material Handling Engineering, Vol. 47, Issue
12, pp. 43-45, 1992.
REFERENCES
[16] Van Landeghem and Limere (2008), A Decision
Model for Line Feeding,
http://orbe122.hallot.net/docs/BookletORBEL.22.pdf,
Retrieved on May 2012.
[1] Anonymous, (2008), Eight secrets to perfect order
picking,
http://www.elologistics101.com/Article/8Secretsof
PerfectPicking.htm, Retrieved on June 2012.
[17] Sellers C.J., Performance analysis of robotic kitting
systems, Robotic and Computer-integrated
manufacturing, Vol. 6, Issue 1, pp. 15-24, 1989.
[2] Anonymous, (2008), Lean Jobshop,
http://www.leanadvisor.com /Lean/
articles/lean_jobshop.cfm, Retrieved on June 2012.
[18] Ranko Vujosevic, (2008). Lean Kitting, A case
study, http://www.optelco.com/pdf/Lean-Kitting.pdf,
Retrieved on June 2012.
[3] Art Smalley, (1997). Toyota’s New Material
handling Systems Shows TPS’s Flexibility,
http://www.leaninstitute.nl/publications/1106/Toyota
New Material HandlingSystem.odf, Retrieved on June
2012.
[19] Rother M., Shook, J., Learning to see- Value
stream mapping to add value and eliminate muda,
Version 1.2, Brookline, MA, USA: The Lean Enterprise
Institute, Inc. 1999.
[4] Bozer Y.A., Mc. Ginnis L.F., Kitting versus line
stocking: A conceptual framework and a descriptive
model, International Journal of Production Economics,
Vol. 28, Issue 1, pp. 1-19, 1992.
[5] Brynzer H., Evaluation of kitting systems –
Implications for kitting systems design, Department of
Transportation and Logistics, Chalmers University of
Technology, 1995.
[6] Brynzer H., Johansson M.I., Design and
performance of kitting and order picking systems,
International Journal of Production Economics, Vol. 4,
Issue 1, pp. 115-125, 1995.
[7] Christmansson M., Medbo L., Hansson G.A.,
Ohlsson K., A case study of a principally new way of
material kitting-an evaluation of time consumption and
physical workload, International Journal of Industrial
Ergonomics, Vol. 30, pp. 49-65, 2002.
[8] Ding F.Y., Balakrishnan P., Kitting in Just-In-Time
production, Production and Inventory Management
Journal, Vol. 31, Issue 4, pp. 25-28, 1990.
[9] Ding F.Y., Kitting in JIT production: A kitting
project at tractor plant, Industrial Engineering, Vol. 24,
Issue 9, pp. 42-43, 1992.
[10] Ford Henry, (2007), In Encyclopedia Britannica,
Website:http://search.eb.com. proxy.lib.ltu.se/eb/article22464, Retrieved on July 2012.
TECH Journal
SAS
34
Volume 12, Issue 2, September 2013