Not Just-In-Time: Structural Steel Supply and Construction Processes

MORE JUST-IN-TIME:
LOCATION OF BUFFERS IN STRUCTURAL STEEL
SUPPLY AND CONSTRUCTION PROCESSES
Iris D. Tommelein1 and Markus Weissenberger2

ABSTRACT
The erection of a buildings structural steel frame is a major construction phase on many a
project. The main resource in this process, the steel erectors crane, defines not only the pace
of erection of steel, but also the pace for handling and installing many other structural and
non-structural materials. This production system cannot afford any delays. Some claim that
structural steel therefore is managed as a just-in-time (JIT) process with materials being
delivered to site as needed and installed promptly. This is the case only in appearance as is
clear when one considers the JIT principles that were developed as part of Toyotas lean
production philosophy. To illustrate the point, this paper draws on examples of typical
structural steel supply chains from the industrial- and building construction sector. The use of
symbols from manufacturing is investigated to map key production steps as well as buffers
in-between them in order to elucidate where resources do and do not flow. Industry practices
in these two construction sectors vary significantly. Neither one is lean. This paper reports on
a preliminary investigation into the location of buffers in the structural steel supply and
construction process. The reasons for having buffers at various locations are explored. A
more in-depth investigation is recommended to gain a deeper understanding of the buffer
sizing criteria and steel component sequencing rules that govern current practices. Insight
into these will then help determine which buffers can be trimmed in order to reduce work in
progress and cycle time. This will support the effort of achieving more JIT by making
processes within individual companies as well as across the entire steel supply chain leaner.

KEY WORDS
Supply-chain mapping, integration, structural steel, fabrication, erection, pull-driven
scheduling, just in time, JIT, kanban, buffer, lean construction.

1
Associate Professor, Constr. Engrg. and Mgmt. Program, Civil and Envir. Engrg.
Department, 215-A McLaughlin Hall, Univ. of California, Berkeley, CA 94720-1712 and
Research Associate, Lean Construction Institute, phone: 510/643-8678, fax: 510/643-
8919, [email protected]
2
Exchange student in the Constr. Engrg. and Mgmt. Program at U.C. Berkeley in
1997-98, now Senior (5-year Civil Engineering Degree), Technical Univ. Hamburg-
Harburg, Germany, [email protected]

Proceedings IGLC-7 109

Tommelein and Weissenberger

INTRODUCTION
The erection of a buildings structural steel frame is a major construction phase on many a
project. The main resource in this process, the steel erectors crane, defines not only the pace
of erection of steel, but also the pace for handling and installing other structural and non-
structural materials such as decking or grating, reinforcing bars, spray-on fireproofing, that
is, work done by others in the parade of trades (Tommelein et al. 1998, 1999). This
production system cannot afford any delays. Some claim that structural steel therefore is
managed as a just-in-time (JIT) process, with materials being delivered to site as needed and
installed promptly.
We set out to validate this claim by reviewing the literature on steel erection and
interviewing steel fabricators, erectors, and contractors. In part, we were eager to find
examples of JIT in current construction practice. A companion paper describes our findings
regarding existing JIT systems for batching and delivering concrete (Tommelein and Li
1999). To make the present paper stand on its own, a few sections from that paper had to be
repeated here.

JUST-IN-TIME SYSTEMS
The term Just-In-Time (JIT), used for instance to describe the delivery of materials to a
construction site, suggests that materials will be brought to their location for final installation
and be installed immediately upon arrival without incurring any delay due to storage in a
laydown or staging area. JIT is a concept developed by the Japanese who created the Toyota
Production System, later translated into English as the lean production system. The ultimate
objective of JIT production is to supply the right materials at the right time and in the right
amount at every step in the process.
Toyota achieves JIT production by implementing a pull system using kanban, loosely
translated from Japanese as cards. Kanban are designed to prevent overproduction and
ensure that parts are drawn from process to process, in reverse order. They thus implement a
replenishment system designed to control production quantities. Parts are taken and
replenished only when needed and in the right amount.
The pull of product from upstream is illustrated by means of a withdrawal kanban as
shown in Figure 1. The customer process withdraws parts from a supermarket, which is a
place of limited capacity for storing product that is output by the supplying process. The
supermarket gets replenished by issuing a production kanban when inventory is too low.
The production kanban thereby instructs the supplying process to produce more product.
The supplying process produces only enough product to replace what was withdrawn. This
method prevents overproduction but allows a tightly-bound inventory to build up between the
supplying and the customer process (for more detail, see Hopp and Spearman 1996 and
Rother and Shook 1998).
In a more general sense, a pull system may comprise a feedback mechanism for a station
downstream to notify a station upstream that more product is needed. Product then is
fabricated and shipped to meet this specific need. This mechanism purposefully constrains
the amount of work in process in the system, which tends to reduce the product cycle time.
The alternative is to produce product in anticipation of a need, i.e., based on a forecast need,

110 26-28 July 1999, University of California, Berkeley, CA, USA

 Not Just-In-Time: Structural Steel Supply and Construction Processes

as is the case in push systems. Product pushed into the production process stands a greater
chance of sitting idle, thereby not only tying up resources but also increasing its cycle time.
Tommelein (1998) compared and contrasted the delivery and installation of pipe spool at
a construction site subject to various uncertainties using pull- and push-driven scheduling
methods. Push systems tend to increase the amount of waste (e.g., long wait times,
excessively large inventories) because they are estimate driven and include fudge factors to
compensate for anticipated uncertainty. Uncertainty may or may not manifest itself during
project execution. When it does, push systems have no means to quickly adapt to changing
system requirements whereas pull systems do.

production withdrawal
KANBAN KANBAN

SUPPLYING CUSTOMER
product product
PROCESS PROCESS

supermarket

Figure 1: Kanban System (after Rother and Shook 1998 p. 42)

This notion of pull is an essential part of a JIT system. Associated questions of system design
are Where and when to pull? and What quantities of product to pull and allow into buffers
at various stages in the process? Our preliminary investigation of possible JIT practices in
the steel supply chain thus led us to map key production steps as well as buffers in-between
them in order to elucidate where resources are held up vs. where they are able to flow.
Qualitative findings of this investigation are reported here. Further investigation will lead to a
more quantitative analysis of industry practices.

VALUE STREAM MAPPING

Koskela (1992) pointed out that architects, engineers, and construction practitioners have for
the longest time focused on conversion activities and overlooked issues of flow. Flow is
important because work or materials that do not flow sit idle in inventory, tying up money
(including the procurement cost of ingredients plus labor and machine time to bring them to
the stage of completion they are in) as well as space. They stand the risk of being damaged or
becoming obsolete due to design changes or market competition. Inventory means product
waits: its cycle time increases, that is, it takes longer for the product to traverse all production
steps it needs to go through before reaching its customer. As a result, project durations are
larger than they would have been had flow not been inhibited.
Most tools used today by practitioners who manage construction, such as those for
design, planning, scheduling, and costing, do not acknowledge flow: they do not explicitly
capture changes of resource characteristics over time. Process modeling tools for discrete-

Proceedings IGLC-7 111

Tommelein and Weissenberger

event simulation are an important exception and warrant more attention by the lean
construction community. Such models can incorporate input regarding individually-
characterized components, uncertainties of numerous kinds, and sequencing rules (e.g.,
Tommelein 1997) and then produce output data regarding buffer sizes, cycle times, idle
times, production rates, etc. The symbols commonly used to depict process models for
construction, however, have yet to distinguish how processes are being managed, for
instance, whether or not a JIT system has been implemented. Practitioners in manufacturing,
working for Toyota and then later for other companies going lean developed their own
pictorial language to help focus attention on what matters in their transition.
We borrowed such symbols from Rother and Shook (1998) and used them to map
CUSTOMER

structural steel supply chains. Boxes denote value-adding processes or tasks, such as
PROCESS

ordering raw materials, fabricating steel, and transporting shipments to a site. A triangle
denotes work in progress or inventory. It represents an accumulation of product (materials or
information) possibly of unlimited amount and for an indeterminate duration. An inverted
triangle is an order to batch. Kanban (introduced in Figure 1) denote orders to withdraw
or produce product, in order to deplete or replenish a supermarket. A supermarket,
represented by , refers to controlled inventory in terms of how much material is kept on
FIFO
hand and how replenishment takes place. The FIFO symbol denotes the first-in-first-
out release of resources output by a task. The circular arrow denotes a physical pull of
materials from a supermarket. It differs from the withdrawal kanban in that it pertains to the
amount of product needed at the time of the withdrawal and not necessarily a predetermined
fixed quantity. A dashed line with an arrow designates the flow of product. A solid
white line is transportation of product to the customer site. A black-and-white dotted
line shows that product is pushed into inventory.
Rother and Shook (1998) use these symbols for so-called value stream mapping where
the term value pertains mainly to reducing work in process inventories and product cycle
times. Our use of this notation stems as much from our desire to engage in mapping the
structural steel supply chain as it does from our desire to test the adequacy of those symbols
in representing architecture/engineering/construction processes and in capturing value.

STRUCTURAL STEEL SUPPLY CHAINS

In order to compare and contrast structural steel supply chains, different examples must be
presented. In the examples that follow, we illustrate what we understand to be common
industry practices though they are not practices applied on any one specific project. We take a
birds-eye view of supply chains, showing fabrication as a process but no specifics thereof,
and including companies that may be involved for months if not years on a single project.
Industry practices on industrial projects, including construction of power plants and
petrochemical facilities, are described first.

112 26-28 July 1999, University of California, Berkeley, CA, USA

 Not Just-In-Time: Structural Steel Supply and Construction Processes

EXAMPLE STRUCTURAL STEEL SUPPLY CHAIN FOR INDUSTRIAL CONSTRUCTION

Many industrial projects are delivered using engineer-procure-construct (EPC) contracts.
This contract may be held by a single company or be divided up among several with design
and procurement being done by one and construction by another company for instance.
Figure 2 depicts a situation where the structural designer (line II) is different from the steel
erector (line VII). Other parties involved with the project are the owner (I), the fabricator
(III), the raw materials supplier (IV)(such as one or several steel mills and distributors), the
shipping agent(s) (V), and the galvanizer (VI). The latter is to provide corrosion protection
for steel that will remain exposed. Steel that needs fireproofing (not shown) is usually
shipped to yet another facility prior to reaching the site.
SPECIFY
OWNER (I)

DESIGN APPROVE
STRUCTURAL DESIGNER (II)

TAKE - OFF PROCURE DETAIL TAKE - OFF PROC . RUN - FABRICATE

MAJOR PCS . OF - MILL PCS .
FABRICATOR (III)

MILL
RAW MATLS. PROVIDER (IV)

SHIP SHIP SHIP

SHIPPING AGENT (V)

GALVANIZE
GALVANIZER (VI)

ERECT
ERECTOR (VII)

Figure 2: Example Structural Steel Supply Chain for Industrial Construction

The owner and designer go through a lengthy iterative process of articulating the project
requirements and then translating them into design documents, including drawings and
specifications. The final design needs approval not only from the owner but also from
regulatory agencies (not shown). A fabricator is then selected and a batch of design
documents released (line III). Especially on large projects this release of drawings will be
phased. Doing so makes it possible to fast-track the process, provided permits can be secured
in the same fashion.
Consideration must be given to the time it will take the fabricator to procure materials
(procurement lead time). Large structural steel sections are not run-of-the-mill items. Their
TAKE-OFF3 and PROCURE[ment] OF MAJOR PIECES is done early, possibly as soon as
design concepts have been decided on, as they must be mill ordered. Steel mills typically

3
Items in CAPS denote process steps shown in the figure.

Proceedings IGLC-7 113

Tommelein and Weissenberger

work on 6-week rolling schedules so there are windows of opportunity to hit, or steel
delivery and thus fabrication will be delayed. When it comes to jumbo sections shipped from
overseas mills, procurement lead time is an even bigger concern. In any case, major pieces
are produced according to the mills schedule and shipped soon thereafter to the fabricator.
Meanwhile, the fabricator proceeds with DETAIL[ing] connections and creating shop
drawings for engineering APPROV[al]. Many steel components are custom designed and
hundreds of different ones may be needed on any one project. The drawings await approval
(unless the fabricator is willing to take a risk) prior to a second round of TAKE-OFF and
PROCURE[ment of] RUN-OF-THE-MILL PIECES. Depending on how typical these pieces
are, the fabricator may carry them in inventory or have to order a new load. These materials
are called run-of-the-mill because mills produce them on a regular basis. They either keep
them in their own supermarket-like inventory (as shown in figure 2) awaiting customers or
sell them to distributors to store and resell at a higher price. Fabricators often buy them from
the mill by the train load in order to keep procurement and shipping costs low even when not
all materials in a load will be used on a given project.
Materials received by the fabricator are stored in a laydown yard, then retrieved one piece
at a time for FABRICAT[ion] including cutting to size, drilling, welding, etc. Some materials
may be partially fabricated then moved out of the way for some time to make room on the
shop floor for other pieces to take priority. A significant amount of handling back and forth is
the result of fabricators objective of maximizing machine and labor utilization.
Finished pieces are loaded on flatbed trucks and shipped to the GALVANIZE[r] or
fireproofer (not shown) if needed, and then to the site. The fabricator, who may already have
been paid in part for materials procurement, gets paid in full for site delivery. The contractor,
in turn, may also earn value for receiving materials in the laydown yard, which helps to offset
their cost for materials management. Steel erection is done by the contractors direct-hire
labor. Cranes may be rented or provided by the owner if the owner has nearby facilities with
equipment not needed on a full-time basis for plant maintenance.
As mentioned, design and construction may be handled by the same or different
companies but fabrication is nearly always done by a third party. Because industrial projects
tend to be complex and large, and numerous uncertainties remain unresolved at the start of
the projectthough more could be resolved if designers were made aware of contractors
sequencing decisionsa significant amount of real-time communication is needed between
all parties to coordinate their production systems. Cooperative fabricators may provide
transparency into their part of the supply chain by for instance revealing status data on a
weekly basis including: (1) Release Number, (2) Drawing/Piece mark(s), (3) Date Received,
(4) Tons of the Release, (5) Date Details Sent for Approval, (6) Date Details Approved, (7)
Fabrication Start Date, (8) Fabrication Finish Date, (9) Date Shipped to Galvanizer, (10) Date
Shipped to Fireproofer, and (11) Date Shipped to Site. This way, the contractor can see
work coming towards them and plan for steel erection accordingly. Note, however, that work
is pushed onto the contractor, so it may not be sequenced optimally for construction.
Nevertheless, because of uncertainties in the release and sequencing of drawings early in
the supply chain, followed by uncertainties in the release and sequencing of fabricated and
shipped materials, contractors consider it good practice to make sure they have enough
materials on hand to keep their crews busy for a few weeks if not months. Howell and

114 26-28 July 1999, University of California, Berkeley, CA, USA

 Not Just-In-Time: Structural Steel Supply and Construction Processes

Ballard (1996) and Tommelein (1998) describe similar practices in the piping function. The
contractors strategy is to try to de-couple their field operations to the largest extent possible
from the flow of materials upstream in the supply chain (Howell et al. 1993) in order to
provide their work force with the greatest flexibility in choosing which work to do next and
thereby enable them presumably to be optimally productive. Large laydown areas are thus
commonplace especially on greenfield sites where space appears to be free and abundant.
Off-site laydown space is rented for brownfield projects where space is tight. Extra handling
of course costs extra money and can be avoided when the flow of materials off- and on-site is
more thoughtfully routed and pieces orderly sequenced.

EXAMPLE STRUCTURAL STEEL SUPPLY CHAIN FOR BUILDING CONSTRUCTION

The structural steel supply chain for building construction differs in several regards from the
one for industrial construction. The buildings frame may in fact be more complex, especially
when it supports a very tall structure, so the major steel sections require extra procurement
effort.
The industry is also organized differently. On design-bid-build projects, a common
delivery method for buildings, the owner typically hires an architectural engineering (AE)
firm, which in turn hires a structural designer. When the AE has prepared all bid documents,
the project is put out for bid. A general contractor (GC) is then selected. The GC subcontracts
the steel work to the fabricator, who in turn subcontracts field installation work to a structural
steel erector. The latter essentially provides the crane and skilled labor, whereas the former is
responsible for acquiring, fabricating, and shipping the materials to site in the sequence
needed for erection. The fabricator may also subcontract the structural steel detailing work.
Fabricator and erector work as a tightly knit team. The GC will meet with them during
bid preparation. They must assess the project site constraints to position the erectors crane,
as it determines not only the steel erection sequence but also the layout of other temporary
facilities and thus the flow of many construction resources. This sequencing in turn drives the
fabrication schedule. It must of course meet the GCs master schedule but must also be
efficient4. The fabricator who subcontracts the erection work has an incentive to minimize
that work and does so by thoroughly planning the sequencing and site delivery of steel pieces
in the order they will be needed.
As for logistics, a big difference between the industrial and the building sector is that
more often than not building space is very tight, especially on projects located in an urban
environment. Industrial projects tend to be more remotely sited. Materials deliveries to
building projects accordingly are constrained by traffic patterns and transportation permit
requirements. Trucks parked in the street along the edge of a site ready to off-load steel may
not remain there for any extended time. When deliveries take place, the crane gets dedicated
to off-loading and moving pieces to a staging area, namely the highest floor with decking,
4
The measure of efficiency would include not just a single resources utilization (e.g.,
crane or fabrication facility), but possibly a combination of resource utilization rates and
their variability, augmented with factors accounting for the rate at which the project
progresses and cost. The measure will be different, depending on what one considers of
value.

Proceedings IGLC-7 115

Tommelein and Weissenberger

from where steel will subsequently be picked up and moved into final position. Only in
exceptional cases will steel be erected directly off the flatbed truck. This saves extra handling
steps but can be done only when it is acceptable to tie up the truck longer and provided the
steel has been loaded in inverse order needed.
Differences in value stream maps between industrial and building construction are
therefore expected at least near the end of the chain, especially in the way delivery to the
project site is organized. If JIT is practiced in industry today one possible way is depicted in
figure 3.
Figure 3 includes two supermarkets, which illustrate the presence of pull mechanisms.
The steel mill (IV) still takes special orders. The resulting output is stored in a generic buffer
(triangle). The buffer is not specifically controlled in size but it is filled only based on firm
customer orders. That product is sold so it is unlikely to become obsolete (waste). The mill
also produces run-of-the-mill product in anticipation of customer orders. As was the case in
figure 2, this is denoted by a supermarket where quantity-on-hand will not exceed a threshold
value and gets replenished at appropriate times. A second supermarket is shown in figure 3 to
handle output from fabrication (III). For instance, the fabricator of a 20-story building could
complete the steel for stories 1 through 3, then await orders from the construction site
(withdrawal kanban) to ship them steel for story 1 before starting work on story 4
(production kanban). A smaller inventory buffer may be well suited provided fabrication can
keep pace with erection. Since there is virtually no storage space on site, no buffer of
materials is shown preceding ERECT[ion] (VI).
SPECIFY
OWNER (I)

DESIGN APPROVE
STRUCTURAL DESIGNER (II)

TAKE - OFF PROCURE DETAIL TAKE - OFF PROC . RUN - FABRICATE

MAJOR PCS . OF - MILL PCS .
FABRICATOR (III)

MILL
RAW MATLS.
PROVIDER (IV)

SHIP SHIP
SHIPPING
AGENT (V)

ERECT
ERECTOR (VI)

Figure 3: Idealized Example of Structural Steel Supply Chain for Building Construction

Note that this figure idealizes current practice in terms of its implementation of a JIT
production system. The timely erection and therefore fabrication of steel is so critical on just
about any project that the owner may not want to risk any delay. Since no significant buffer
of steel can be established on site, it is created off-site, most commonly at the steel
fabricators output end. A triangle therefore more appropriately depicts the actual practice of

116 26-28 July 1999, University of California, Berkeley, CA, USA

 Not Just-In-Time: Structural Steel Supply and Construction Processes

buffering rather than a supermarket. The owner pays the fabricator for finishing work even if
the site is not yet ready to receive it. If project delays are incurred, the owner may even
award an extra fee for longer-term storage, so the fabricator has an incentive to complete as
much work as early as possible. Large buffers of ready-to-ship materials create a feel-good
situation, but at a cost!

DISCUSSION
The creation of large buffers either at the contractors or fabricators site is contrary to JIT
production system design. As the word JIT suggests, materials must be fabricated or
delivered on time, which means not too late but not too early either. This implies that
variability regarding timing, actual pieces released, as well as quality must be limited and
controlled.
In a true JIT system, this timeliness pertains not only to a single hand-off between two
production steps, but rather, one aims at achieving JIT flow between all production steps. In
the idealistic extreme, this means having no buffers at all! In practice, this means buffers
must be determined strategically. Admittedly, doing so is not an easy task, especially in a
production system of complex products that involve several organizations as is the case for
structural steel. But this is what lean production systems design is all about: achieving flow
where possible and thoughtfully locating buffers and sizing them to achieve cost-effective
decoupling with minimal impact on cycle time.
A significantly amount of additional data is needed to describe the structural steel supply
chain at a level of detail that makes it possible to strategically locate and size buffers. The
difficulty is doing so stems in part from the fact that structural steel is custom designed and
every project requires a variety of components that change as construction progresses.
Because of constructions one-of-a-kind project nature, the structural steel supply chain
differs from manufacturing systems for more standardized products, which can be likened to
the Toyota Production System. The manufacturing symbols used here for mapping the steel
supply chain provide no room for defining individually-characterized resources or detailed
sequencing rules. While their ability to show processing durations and delay times has not
been used in this paper (we expect to do so in future work), we doubt that using only
deterministic values will suffice. Some representation of uncertainty will have to be
incorporated in the maps.
The current practice of buffering stems from the desire to optimize labor and machine
utilization and from admitting that uncertainties exist in the supply chain. Many uncertainties
are the result of variability, which could be understood better if at least it were measured and
explicitly accounted for. This is not the current practice in construction. Uncertainties and
variability should not be taken for granted. They should be acknowledged, managed, and
minimized to a reasonable degree. In fact, one technique to identify them is to reduce buffer
sizes in-between various production steps in order to see and learn the extent to which they
are needed. Process improvement efforts can then focus on those steps where the impact on
throughput of the system as a whole will be most significant. Example improvements in
construction are those that aim at reducing uncertainty, as is done for instance through
reliable planning by the Last Planner (Ballard and Howell 1998), work methods design, and
work structuring (Ballard et al. 1999).

Proceedings IGLC-7 117

Tommelein and Weissenberger

Existing construction industry practices and contractual incentives counter the objective
of achieving flow. A few examples were given in this paper. Nevertheless, a number of
fabricators as well as parts and components suppliers to the construction industry have
successfully implemented lean practices in their own shops (e.g., WireMold, Trane, Square-
D). But this success is constrained by their upstream suppliers and downstream customers.
JIT and other lean principles are aimed at streamlining the entire process across all
organizational boundaries and allowing product to flow from conception to installation.
Opportunities for change in the lean direction exist, especially with the construction
industry actively seeking alternative delivery methods. Frank exchange of production plans
and performance data are a step in the right direction. Although a fabricator providing
production transparency is useful, it is only a small step towards trimming the supply chain.
As it is often done now in the situation described, work gets pushed onto the contractor.
However, to minimize work in process, product must be made to order, upon customer
demand. This means that the contractor should be able to pull materials through all upstream
processes, including not only fabrication but also permitting and design. Pulling goes beyond
mere coordination of phasing of permits and releasing of design documents with
construction. Whereas coordination may be limited to revealing sequences and timing to
others, pull implies that there is agreement to mutually adjust to each others needs and to
synchronize production steps in order to improve system performance. Pull will be easier to
implement when needs are well defined and can be revealed ahead of time, that is, when all
supply chain partners work according to reliable plans. This is where the real opportunity lies
for fast-tracking a project and cutting cycle time as well as cost.
EPC projects hold the greatest potential for implementing incentives that promote
improvements to the overall production process spanning across organizational and, within
organizations, across departmental boundaries. Construction schedules of reliable quality and
created early can drive design and pull the release of drawings through the fabricator.
Fabricator capacity can then be committed to (including payment for the reservation)
accordingly.
Design-build (DB) projects, like EPC projects, also provide opportunities in this regard
that remain unrecognized and thus unexploited. Team members on these projects have a
vested interest in making their joint effort succeed. They therefore may be more willing and
able to cross their traditional organizational boundaries and try to meet production objectives
that stress supply chain performance. For example, a design-build developer (responsible for
design, construction, as well as complete financing during construction) engaged in building
design, then realized they could not meet the owners schedule because they could not get the
needed steel rolled in a timely fashion. To cut cycle time, steel mills and steel resellers were
visited around the country and procured sections that were readily available and more-or-less
matched the design. The design then was adapted and further developed to suit the
dimensions of the procured steel.
Note that while EPC and DB projects provide great opportunities for the implementation
of lean construction practices, lean construction practices can apply to any project.

118 26-28 July 1999, University of California, Berkeley, CA, USA

 Not Just-In-Time: Structural Steel Supply and Construction Processes

SUMMARY
This paper has given examples of structural steel supply chains in two sectors of the
construction industry, namely industrial construction and building construction. Value stream
symbols from manufacturing were tried as a means to map them. The resulting maps show a
very simplified view of the structural steel supply chain. They highlight that buffers are used
at nearly all stages in the process and that steel therefore cannot possibly flow. Flow is
advocated in JIT production. Only a minimal amount of it is achieved in building
construction regarding site deliveries of structural steel, though one can barely call this the
implementation of JIT practices.
Buffering is done to allow processes to achieve high equipment and labor utilization and
to shield downstream processes from upstream uncertainties. Individual process performance
may therefore look good but it results in increased product cycle time and thus extended
project duration as well as increased costs for maintaining work in process. When not
judiciously located and sized, buffers may grow to be disproportionately large or anxiety-
provoking small.
We suggest that further study be conducted into the intricacies of the structural steel as
well as other supply chains, so that the rationales for buffering and the associated risks and
rewards can be elucidated. Whether they are contract- or production driven will guide further
analysis. Insight into these will then help determine which construction principles best apply
so the production system can be shaped in order to obtain more benefits of JIT practices.

ACKNOWLEDGMENTS
The authors would like to acknowledge the contributions made by several industry
practitioners in the course of this research. They gave generous amounts of their time and
knowledge about the structural steel supply chain. Markus Weissenberger also wishes to
thank all faculty and staff in the Construction Engineering and Management Program at U.C.
Berkeley for their warmth, support, and dedication to education he experienced during his
stay. This work was funded in part by a seed grant from U.C. Berkeleys Committee on
Research and by grant CMS-9622308 from the National Science Foundation (NSF) whose
support is gratefully acknowledged. Any opinions, findings, conclusions, or
recommendations expressed in this paper are those of the authors and do not necessarily
reflect the views of NSF.

REFERENCES
Alarcon, L. (ed.)(1997). Lean Construction. A.A. Balkema, Rotterdam, The Netherlands, 497
pp.
Ballard, G. and Howell, G. (1998). Shielding Production: Essential Step in Production
Control. J. Constr. Engrg. and Mgmt., ASCE, 124 (1) 11-17.
Ballard, G., Howell, G.A., Tommelein, I.D., and Zabelle, T.R. (1999). Lean Construction: A
Revolution in the Design and Building of Capital Facilities. Lean Construction Institute,
Ketchum, ID, www.leanconstruction.org.
Bernold, L.E. and Salim, Md. (1993). Placement-oriented design and delivery of concrete
reinforcement. J. Constr. Engrg. and Mgmt., ASCE, 119 (2) 323-335.

Proceedings IGLC-7 119

Tommelein and Weissenberger

Hopp, W. and Spearman, M. (1996). Factory Physics. McGraw Hill, Boston.

Howell, G.A. and Ballard, H.G. (1996). Managing Uncertainty in the Piping Process. RR 47-
13, Constr. Industry Institute, Univ. of Texas, Austin, TX, Sept., 103 pp.
Howell, G., Laufer, A., and Ballard, G. (1993). Interaction between Sub-cycles: One Key to
Improved Methods. J. Constr. Engrg. and Mgmt., ASCE, 119 (4) 714-728.
Koskela, L. (1992). Application of the New Production Philosophy to Construction. Tech.
Report 72, CIFE, Stanford Univ., CA, September, 75 pp.
OBrien, W.J. (1995). Construction Supply Chains: Case Study and Integrated Cost and
Performance Analysis. Proc. IGLC-3, Albuquerque, NM, 187-222 in Alarcon (1997).
Rother, M. and Shook, J. (1998). Learning to See: Value Stream Mapping to Create Value
and Eliminate Muda. v.1.1, Oct., The Lean Enterprise Institute, Brookline, Mass.
Thomas, H.R., Sanvido, V.E., and Sanders, S.R. (1989). Impact of Material Management on
Productivitya Case Study. J. Constr. Engrg. and Mgmt., ASCE, 115 (3) 370-384.
Tommelein, I.D. (1997). Discrete-event Simulation of Lean Construction Processes. Proc.
Fifth Ann. Conf. Intl. Group for Lean Constr., IGLC-5, 16-17 July at Griffith Univ.-Gold
Coast Campus, Queensland, Australia, 121-135.
Tommelein, I.D. (1998). Pull-driven Scheduling for Pipe-Spool Installation: Simulation of
Lean Construction Technique. ASCE, J. Constr. Engrg. and Mgmt., 124 (4) 279-288.
Tommelein, I.D. and Li, A.E.Y. (1999). Just-in-Time Concrete Delivery: Mapping
Alternatives for Vertical Supply Chain Integration. Proc. 7th Ann. Conf. Intl. Group for
Lean Constr. IGLC-7, 26-28 July, held at U.C. Berkeley, CA.
Tommelein, I.D., Riley, D., and Howell, G.A. (1998). Parade Game: Impact of Work Flow
Variability on Succeeding Trade Performance. Proc. 6th Ann. Conf. Intl. Group for Lean
Constr., IGLC-6, 13-15 August held in Guaruja, Brazil, 14 pp.
Tommelein, I.D., Riley, D., and Howell, G.A. (1999). Parade Game: Impact of Work Flow
Variability on Trade Performance. ASCE, J. Constr. Engrg. and Mgmt., Sept./Oct. Issue.

120 26-28 July 1999, University of California, Berkeley, CA, USA