Order Description
Read the first half (before the ‘future-state map’ section) of the “Development of lean model for house construction using value stream mapping”
article and write a reflection (minimum two pages, double-space).
Please make sure you follow the instruction. Please let me know if you have any question about it.
Development of Lean Model for House Construction Using
Value Stream Mapping
Haitao Yu
1
; Tarry Tweed
2
; Mohamed Al-Hussein
3
; and Reza Nasseri
4
Abstract: Lean construction has recently attracted considerable attention in the home building industry. Lengthy delivery time and
significant waste in the construction process have caused many home builders to seek a more effective production model that will increase
process reliability, reduce total lead time, and improve quality. However, although housing construction provides the closest analogy to
manufacturing, a high level of variability prevents the direct transplantation of lean paradigm and techniques. In collaboration with a local
home builder, a systematic approach based on value stream mapping technique is developed in this research to analyze the current process
and to formulate a lean production model. The model has four main features: synchronized first-in, first-out lane-based flow, production
leveling at pacemaker, work restructuring, and improved operation reliability. A simulation template is built to verify the model and to
assist in the development of interim implementation models. This paper presents data collection and value stream selection, current
practice analysis, and specific changes proposed for the lean production model.
DOI:10.1061/ !ASCE”0733-9364!2009″135:8!782″
CE Database subject headings: Buildings, residential; Housing; Lean construction; Production management; Construction
management; Process control .
Introduction
In North America, the home building industry has changed little
since the 1920s, when the wood platform-frame structure
emerged as the standard building technology. Although advances
in tools and materials have led to some incremental improve-ments, the fundamentals of the construction process remain al-most identical and no
significant improvements in productivity
have been observed!Zhang et al. 2005″. In contrast, other indus-tries have experienced remarkable productivity improvements
through the application of innovative technologies and operation
management tools. A recent example is the automobile industry,
where manufacturers have dramatically improved productivity
through the adoption of a new production philosophy which has
led to the “lean production system.” Popularized by Womack’s
book, The Machine That Changed the World !1990″, “lean
theory” has been widely employed by a range of industrial sectors
in the past decade.
Research on the implementation of lean theory in construction
began in the early 1990s, when Koskela !1992″ wrote a ground-breaking paper, “Application of the New Production Philosophy
to Construction.” This notion quickly attracted the attention of
researchers in residential construction. Gann !1996″ and Barlow
et al. !2003″ compared industrialized housing with automobile
manufacturing in Japan, highlighting the similarities existing in
their production strategies, but the focus of the papers was on
product development, supply-chain coordination, marketing, and
sales, rather than on fundamental construction practice. Zhang
et al.!2005″ proposed a waste-based management approach that
considered all process inputs, including labor, equipment, materi-als, data and information, work space, and time, as potential
sources of waste. Two case studies in housing construction were
presented to demonstrate that minimizing resources waste would
significantly improve productivity and quality. Ballard !2001″
suggested that variability was the major source of waste in con-struction and that even flow production could increase the reli-ability of work flow
and thus reduce cycle time in home building.
Bashford et al. !2003″ further discussed implications of even flow
production and concluded that the strategy had minor impact on
construction duration but could significantly reduce work flow
variability.
The research presented in this paper continues Ballad and
Bashford’s efforts on house production flow management, but
proposes a new lean production model by utilizing value stream
mapping !VSM”. VSM, referred to at Toyota as material and in-formation flow mapping, is the most commonly used tool in lean
planning. It helps lean system practitioners to think about flow
instead of isolated wastes and to implement a lean system instead
of individual lean techniques. Some research has been done in the
application of VSM to construction, but these efforts have either
focused on macroprocess level, such as supply chains !Arbulu
and Tommelein 2002; Fontanini and Pichi 2004 ” and project de-livery !Mastroianni and Abdelhamid 2003″, or on single opera-tion, such as masonry !
Pasqualini and Zawislak 2005″ and
components manufacturing !Alves et al. 2005 “. No report has yet
been found on the use of VSM for fundamental construction pro-cess improvement.
1
Ph.D. Candidate, Dept. of Civil and Environmental Engineering,
Univ. of Alberta, Edmonton AB, Canada T6G 2W2. E-mail: hyu@
ualberta.ca
2
General Manager, Landmark Homes, 9765-54 Ave., Edmonton AB,
Canada T6E 5J4.
3
Associate Professor, Dept. of Civil and Environmental Engineering,
Univ. of Alberta, Edmonton AB, Canada T6G 2W2.
4
President, Landmark Group, 9765-54 Ave., Edmonton AB, Canada
T6E 5J4.
Note. This manuscript was submitted on November 1, 2007; approved
on March 11, 2009; published online on July 15, 2009. Discussion period
open until January 1, 2010; separate discussions must be submitted for
individual papers. This paper is part of the Journal of Construction
Engineering and Management, Vol. 135, No. 8, August 1, 2009.
©ASCE, ISSN 0733-9364/2009/8-782–790/$25.00.
782 /JOURNALOFCONSTRUCTIONENGINEERINGANDMANAGEMENT©ASCE/AUGUST2009
A number of factors impede the application of VSM to the
main construction stream at the operational level. First, an under-lying prerequisite for VSM is the repetition of the production
process. In manufacturing, hundreds of thousands of products in a
product family pass through similar processing steps, so it is fa-vorable to develop and implement a lean system to continually
improve that process. A construction project, in contrast, presents
a unique design, specifications, and context, and must be con-structed accordingly—following a unique construction process
!value steam “. As VSM requires diligent management commit-ment coupled with massive efforts in systematic data capture and
analysis, lean training, core implementation team assembly, and
working process transformation, practitioners have been hesitant
to invest such efforts to improve a process that may not recur.
Second, VSM is a quantitative tool that uses a list of process data
to depict the current state of the process and to determine what
the future state will be. Construction companies, however, gener-ally do not fully track construction processes. Moreover, most of
the construction steps are lengthy and subject to numerous vari-ables. Site investigation is useful to assist researchers in under-standing the
process, but it has proven nearly impossible to
collect statistically meaningful data in a short time period. Third,
key concepts/elements used in VSM, such as cycle time, change-over time, up-time, and inventory, are defined in the context of
manufacturing and seem nonapplicable to construction.
Although construction industry as a whole is defined very
differently from manufacturing, home building, as a unique sector
in construction, provides the closest analogy to automobile pro-duction !Winch 2003”. Its distinctive features, including high
production volume !repetitiveness “, controllable production flow,
and large inventory of work in process, make the application of
VSM here favorable !Yu et al. 2007″. In the research, a seven-step methodology is adopted: !1″ data collection and value stream
selection; !2″ current-state mapping;!3″ existing practice analysis
and lean metrics development; !4″ formulation of lean production
model; !5″ laboratory testing of the model using simulation; !6″
kaizen !continuous improvement” plan development; and !7″
implementation of kaizen plans and results evaluation. This paper
focuses exclusively on the first four steps, especially current-state mapping and formulation of the lean production model;
simulation and the lean implementation are briefed to provide
integrality.
Data Collection and Value Stream Selection
An important rule in VSM is to bring the stopwatch when walk-ing along the actual pathways of material and information flow,
and to rely only on information obtained firsthand !Rother and
Shook 2003 “. This rule, however, is not applicable in construc-tion. High variability in task durations as well as queuing times
and complexity in the construction process make it impossible for
a small team of researchers to collect sufficient data merely
through site observations. Instead, an intranet-based production
tracking system was developed, in which site managers recorded
the booking date, confirmed start date, actual start date, and actual
finish date of every task in the construction process. This research
used production data collected in 2006—approximately 400
houses for each task. Based on the data, four basic measurements
were calculated: !1″ cycle time !CT”; !2″ booking time !BT”; !3″
lead time between tasks !LT”; and !4″ percent started on schedule
!PSS “. Table 1 shows the definitions and calculation formula of
these measurements and other key VSM elements used in the
research. The writers have redefined most of the concepts used in
traditional VSM to suit the context of construction, and have de-signed two new measurements !BT and PSS” specifically for the
application of VSM in construction.
In addition to data collection, two management decisions must
be made prior to the commencement of VSM:!1″ select a value
stream; and !2″ decide on the level of mapping. A value stream is
a series of activities required to bring a product or service from
raw state through to the customer. As a manufacturing plant usu-ally fabricates multiple products—each by a unique transforming
process, the product family with the highest production volume is
generally selected as the target on which to focus improvements
!Tapping et al. 2002 “. The products of a home building company,
on the contrary, can be seen as a single product family, because
they are generally constructed following similar processing steps
and utilizing the same subtrade pool. However, the housing con-struction process is a complex system that involves tens of trade
contractors and consists of hundreds of construction activities. A
single map encompassing the entire process would be too large
and cumbersome for a VSM team to handle. In fact, a rule of
thumb for VSM is that the target value stream should include no
more than 12 tasks !or process stations “. A feasible solution is to
divide the entire production flow into stages, with each stage con-sidered as an independent value stream in a supply chain. Another
reason to compartmentalize the construction process is that the
process is too long to be synchronized with one takt time. For
example, it is obvious that the capacity of excavation trade work-ing at the beginning of the process needs to match the current
pace of sales, but for finishing trades at the tail end of the process,
their production pace should be synchronized with the pace of
sales that had occurred 6 months ago. Based on the analysis of
historical data, the writers recommend that the total production
duration of a stage be shorter than 2 months.
Level of mapping is another important issue to be considered
in defining the value stream. In manufacturing, mapping generally
begins at the level of the production process in a single plant, with
the activity box indicating a continuous product flow. In other
words, the tasks in the map are divided at the places where the
product flow stops and in-process inventory accumulates. In con-struction, the houses !products” do not move along a production
line but, rather, workers move from one house !product” to an-other. Thus, the operations of a trade crew can be regarded as a
continuous flow and would be shown with one activity box on the
map.
As shown in Table 2, the entire house production process was
divided into five stages after considering the size requirement of
the value stream map, total production duration, and the logical
relationship between the construction activities. The summary
data shown in Table 2 clearly demonstrate that Stage 1 is the most
problematic segment. The scheduled duration of Stage 1 is
20 days, but houses actually spend an average of 73 days !365%
of the scheduled duration” in this stage. In addition, a large stan-dard deviation ! 35 days” indicates that the construction process in
Stage 1 was not effectively controlled and that a high potential
exists to reduce construction time by redesigning the process.
Based on the above-presented analysis, Stage 1 was selected by
the VSM team as the target value stream.
Current-State Mapping
The goal of current-state mapping is to create a clear picture of
the existing production process and to expose wastes. The con-JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2009 / 783
ventional approach of residential construction management is
based upon a management model which views the construction
process as a series of tasks to be completed in sequence !Bashford
et al. 2005″, and each house is scheduled and managed individu-ally as a small project using a Gantt chart or critical path method.
Fig. 1 illustrates the conventional approach using predefined
icons of VSM. It was drawn up in a 2 day VSM session attended
by the lean team and an external lean expert. The information
recorded in the data box was estimated by site managers in the
session and then verified by researchers using historical data
which had been collected through the cooperating company’s pro-duction tracking system. A statistical software, Stat::Fit, was used
to process the data. The map shows the main flow of Stage 1,
where 11 trade crews !represented by activity boxes ” are in-volved. The process can be thought of as one house passing
through a line of process stations. Unlike with manufacturing,
attributes of each task are not a constant, but are expressed in
the form of distribution in order to reflect the variability of the
construction process. Site managers are the center of production
control. Due to the unpredictability of both the market and the
construction process, home building is essentially a “make-to-order” business. No overall production schedule exists in the
home building company, and construction is triggered when the
file of a new house is released by sales. Subsequently, the respon-sible site manager starts booking material and subtrades and tries
to push the process as quickly as possible. Meanwhile, no look-Table 1. VSM Key Elements
Key concepts Definitions Formula
Cycle time !CT” The duration that a subtrade needs to complete its work
package.
CT!actual finish date “actual start date
Booking time !BT” The time that a subtrade needs to deploy its crew to a
given job.
BT!confirmed start date”booking date
Lead time!LT” The time that elapses between one task being completed
to the next task being started. In lean system, LT serves
as a time buffer to shield downstream crews from
upstream variability.
LT!actual start date of task i +1″actual finish date
of task i
Percent started on
schedule !PSS ”
A measure of the proportion of start-date promises made
by subtrades that are delivered on time, in percentage.
PSS !number of tasks started on schedule/
total number of tasks
Changeover time The time that a crew needs to switch from working at
one house to another, including demobilization and
mobilization.
Uptime A measure of the proportion of available production
time !APT ” that is actually used on construction, in
percentage.
Uptime = ! APT ! bad weather days ! changeover time ” /APT
Work in process !WIP” Number of uncompleted houses in the value stream,
including the houses in construction and those standing
idle waiting for the start of next activity.
In-process inventory/
supermarket yield
The backlog of ready houses that stands idle waiting for
the start of a given task. In lean system, it serves as a
buffer to protect the continuous work flow of
downstream crews.
Number of houses in inventory = LT / operational
takt time
Yield The percentage of houses that go through an operation
correctly, without any rework.
Takt time The rate at which a home builder must build the house
to satisfy customer demand.
Table 2. Construction Stages and Descriptive Statistics
Attributes
Stage 1
foundation
Stage 2
lock-up
Stage 3
interior and siding
Stage 4
prefinals
Stage 5
finals
Start task Stake out Framing Verandahs, smart board,
and rear deck
Plumbing prefinal
and water test
Tile flooring
End task Drill and place piles Roofing Drywall taping House clean and vacuum House possession
Number of tasks 11 11 11 12 14
Actual duration average
!calender day”
73 31 54 42 26
Standard deviation 35 14 20 14 7
Coefficient of variation !%” 48 45 37 33 27
Scheduled duration
!calendar day”
20 25 22 18 16
Difference!actual”schedule ” 53 6 32 24 10
784 /JOURNALOFCONSTRUCTIONENGINEERINGANDMANAGEMENT©ASCE/AUGUST2009
ahead schedule is available for trade contractors. The booking
information is generally issued by site managers via phone or fax
on a task-by-task basis.
Existing Practice Analysis and Lean Metrics
Development
Upon drawing up the current-state map, several wastes can be
identified immediately. In this case, the first observation was that
lead times were very lengthy. The total duration of Stage 1 was
found to be 65.5 workdays, but the total lead time accounted
for 50 workdays. This means that houses in construction stand
idle about 76% of the time, with no construction activity on site.
One apparent cause of long lead time is the high level of variabil-ity of the process. The booking times and cycle times in the map
are very variable. Six of 11 booking times have a standard devia-tion in excess of 5 days and cycle times of five tasks must be
described using statistical distributions. In the current practice,
site managers book the downstream subtrade immediately fol-lowing confirmation of the start date of the upstream tasks. The
intention of this practice is to crash construction duration by
overlapping booking time and task cycle time, but the actual re-sult is that nearly half of the tasks cannot begin on the scheduled
start date !the average PSS on the current-state map is 54%”.
Temporary contract relationship between home builder and trade
contractors make any delay in schedule be magnified by ripple-through effect. For example, bad weather !e.g., heavy rain” pre-vents the excavation
from commencing on the confirmed date for
a given house. As the excavation subtrade has already scheduled
other jobs on consecutive days for other home builders, the de-layed job has to be rescheduled to the end of its working sched-ule. Moreover, as the
downstream task !in this example, pouring
of footings” cannot begin until the excavation is complete, the site
manager has to cancel original bookings and attempt to get new
commitments based on the newly scheduled excavation date.
However, from the perspective of the footing contractor, a sudden
schedule change means that it must find a new job fitting in that
time slot in a very short period of time. Then, overbooking !i.e.,
subtrades accept jobs exceeding their capacity ” has become com-mon practice. Consequently, a greater number of tasks fail to
begin on the scheduled start date and lead time becomes even
more unpredictable.
Second, variations in cycle time are relatively high, especially
for tasks whose cycle times are described in distributions. Site
managers have reported that the major cause of high variation is
not work-load differences between house models but the manner
in which subtrades carry out their respective jobs. They tend to
deploy their crews continuously on new jobs where large quanti-ties of work are available, leaving uncompleted, minor details to
rework crews. These rework crews follow separate working
schedules and usually arrive several days later to finish the job.
Quality problems are another cause of high variation in cycle
time. It has not been rare, for example, that the crews who install
the main floor spend 1 day cleaning the beam pockets and level-ing the top of the foundation walls.
Based on the analysis of current practice, the aim of lean
initiatives for the cooperating company was set as increasing
Fig. 1. Current-state map of home building process !Stage 1″
JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / AUGUST 2009 / 785
productivity by stabilizing the process, reducing lead time,
and eliminating defects. Accordingly, the lean metrics shown in
Table 3 were developed in order to clarify these goals and track
progress.
Future-State Map
The focus of future-state mapping is to eliminate the root causes
of wastes and to link the value stream in a smooth flow. Unlike
manufacturing, where the fundamental problem is overproduction
caused by “batch and push” !Womack and Jones 1996″, the home
building industry suffers most from variability. Unpredictability
of the process causes all kinds of waste, not just of long lead
times and excess inventory. Uncompleted houses are vulnerable
to weather, requiring temporary protection; to pilferage, requiring
security and extra materials; and to vandalism, causing rework.
Variability also results in fluctuation of the production flow. This
means that home builders need to sustain a large workforce pool
and cannot provide stable work flows to trade contractors. In
order to reduce the variability of the process, the following four
measures were taken in the future-state mapping!Fig. 2″: estab-lishing a production flow and synchronizing it to takt time; lev-eling production at
pacemaker task; restructuring work; and
improving operation reliability with work standardization and
total quality management.
First-in, First-out Lane-Based Flow and Its
Synchronization
In manufacturing, continuous flow forms the centerpiece of the
lean production system and is regarded as the most effective way
of production. Nevertheless, the production system built on con-tinuous flow can only be used for a reliable process. As the sys-tem is fully
synchronized, any small delay or breakdown in one
operation will result in halting the entire system. Housing con-struction is a site-based production !as opposed to factory-based
manufacturing “. Weather and site conditions have a significant
impact on the execution of construction activities, so variation in
task duration is unavoidable. In addition, the housing construction
process is a long process involving more than 60 work “pack-ages” !tasks”. Connecting all the tasks into a continuous flow
would make the system very fragile. Finally, construction work is
performed by various trade contractors who have individual inter-ests and are almost exclusively concerned with the efficient ex-ecution of their
individual tasks !Bashford 2004″. Therefore,
keeping an excess capacity buffer to overcome minor flow fluc-tuation is not practical for home builders.
Another important lean tool, supermarket-based pull flow, is
also nonapplicable in house construction. A pull-flow system is
Table 3. Lean Metrics
PLACE THIS ORDER OR A SIMILAR ORDER WITH US TODAY AND GET AN AMAZING DISCOUNT 🙂