Professional Documents
Culture Documents
Design-Integrated Process Planner For Rebar Placement
Design-Integrated Process Planner For Rebar Placement
Design-Integrated Process Planner For Rebar Placement
REBAR PLACEMENT
ABSTRACT: To remain competitive in the global market, the u.s. construction industry
is looking at drastic productivity increases. In recent years, developments in reinforced-
concrete construction have been rapid and sweeping. Therefore, reinforced-concrete
construction is an area of construction in which improvements in work methods and labor
efficiency could result in major cost savings. The placement of reinforcing bars (rebar)
is an integral part of this type of construction. This is an assembly-type operation with
fabricated rebar being placed and tied one by one. Process-oriented intelligent planning
for the placement of rebar aims to automate development of detailed plans that depict
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.
the assembly sequences of individual bar or set(s) of bars. Traditionally, rebar is fabri-
cated, bundled, and delivered without considering their sequential placement. As a result,
a considerable amount of time is wasted for searching, identifying, rehandling, etc., of
rebar. To eliminate this waste, the rebar has to be delivered according to the need of
the assembly process. The two main objectives of this paper are to present: (I) A feature-
based and process-oriented planning framework; and (2) an intelligent and computer-
aided-design-integrated (CAD-integrated) process-planning system for the placement of
rebar. This paper demonstrates that data generated within a computer-aided design
(CAD) package can be extracted and utilized by other applications needed for automated
process planning. A prototype that has been implemented on a 486 personal computer
(PC-486) using LEVELS OBJECT, AutoCAD, and dBASE III PLUS will be used to
discuss some specific elements of such a system. The paper will conclude with a short
description of a successful field assessments of such a system.
INTRODUCTION
Reinforced-concrete structures constitute a major portion of U.S. construction. The instal-
lation of rebar is an integral part of this type of construction. Although reinforcing steel occupies
a small portion of the volume of reinforced concrete, its contribution to the cost is approximately
equal to that of concrete. The placement of rebar that includes tying of the bars is basically an
assembly operation with an individual bar or set(s) of bars. Traditionally, final placement of
rebar in the concrete form is done by hand at job site. A detailer usually develops the placing
drawings and bar lists. Placing drawings that show the concrete elements, including size, length,
and location of rebar, are used by the field personnel (e.g., foreman) and ironworkers for placing
fabricated rebar. However, the actual placement or installation is not planned and controlled
by the detailer. In the eastern part of the United States, a separate placing contractor usually
accomplishes the tasks of actual placement, while in the western part, rebar fabricators are often
responsible for placement. Bar lists include both bent and straight bars with their dimensions,
bends, and grade of steel. In today's shop practice, fabrication, bundling, and delivery of rebar
are based on the bar lists that do not consider sequences of rebar placement (Salim and Bernold
1992). Consequently, a considerable amount of time and other resources are wasted for search-
ing, identifying, and rehandling of the needed bars for sequential assembly operation.
Rebar should be fabricated and delivered according to the rebar-placement sequences. There-
fore, a basic requirement to achieve a placement-oriented (driven) fabrication and delivery is
the availability of a placement plan before the start of fabrication. Due to the complexities of
reinforced-concrete construction, and the many different rules and objectives of rebar placement,
the planning of such an operation can be very complex. According to Bernold and Reinhart
(1990), an intelligent process-planning system could be instrumental in addressing such problems.
Both design and construction have embraced the advantages provided by computers for
improving productivity and the quality of work. With the increased use of computer-aided-
design (CAD) systems and motivation for computer-integrated construction (CIC), automated
process planning has become more important in realizing the full potential of other tools (Zozaya-
Gorostiza et at. 1989). The fact that the detailers increasingly take advantage of computer
capabilities and software (e.g., CAD), serves as a key element in integrating detailing with other
functions (e.g., placement planning). Automatic evaluation of CAD-represented rebar designs
with a linked process-planning system using artificial intelligence (AI) could assist the detailer
'Asst. Prof., Dept. of Industrial Technol., Univ. of Northern Iowa, Cedar Falls, IA 50614-0178.
°Assoc. Prof., Dept. of Civ. Engrg., North Carolina State Univ., Raleigh, NC 27695-7908.
Note. Discussion open until September I, 1995. To extend the closing date one month, a written request must
be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible
publication on July 28, 1993. This paper is part of the Journal a/Computing in Civil Engineering, Vol. 9, No.2,
April, 1995. ©ASCE, ISSN 0887-3801/95/0002-0157-0167/$2.00 + $.25 per page. Paper No. 6655.
processing. In today's practice, a human process planner (scheduler) who is intimately familiar
with all aspects of the respective manufacturing or assembly operation performs the process-
planning functions. Continuous attempts are being made to enhance the process-planning func-
tions by augmenting the abilities of human expert with those of computers. According to Ansaldi
et al. (1989), developments in manufacturing industry aim at an intelligent process-planning
system that would integrate design and production data for generating useable process plans.
Despite significant efforts to develop automatic process-planning systems using different meth-
odologies, the manufacturing industry has yet to develop a system that will fulfill the ultimate
objective of intelligent process planning. Existing computer-aided process-planning systems can
be classified into: (1) Variant process planning; and (2) generative process planning (Chang and
Wysk 1985).
milling operation in manufacturing, the sequential placement of rebar depends on the feature
of rebar combined with the features of concrete elements. A general definition of feature found
CONCRElB
COMPONENTFEA'IURB
CONCRETE RElNFORCEMENT
FOR
OFFlCE BUlLDlNG
IlIl
Ill:
i=
E~
e !E
Z
82
IlIl
~-< ~
IlIl
=
IlIl
:!
~ ...
8Ill:
@:
Ill:
l!l
!1,11
have only one path between them. In the search of a tree, movement occurs because some
condition or set of conditions triggers an operator. The operators can be viewed as if-then
statements. The work breakdown structure (WBS) for a construction project can be considered
a simple application of a search tree in that physical elements are logically connected. Fig. 1
presents a hierarchical tree structure based on the critical features of concrete building elements
(Bernold and Salim 1993). For example, concrete floor systems are organized into classes of
beam(s), slab(s), girder(s), and connection(s). The class beam(s) has many subcomponents such
as T-beams and rectangular beams, that characterize a beam in more detail and with it provide
a framework for linking a particular building element with an appropriate construction method.
To facilitate tree searches necessary for creating placement plans artificially, all the concrete
elements are represented as structured objects. The hierarchical structure for organizing the
concrete-building elements alone is not sufficient for the intelligent planning of the assembly
sequences of rebar placement. Thus, it is also necessary to physically organize the rebar system
of a building based on the critical features of both concrete and rebar elements. An integrated
tree structure to relate building elements with appropriate construction methods for rebar as-
sembly is presented in Fig. 2. As shown, this tree structure enables the physical organization
of the rebar based on the critical concrete element and rebar features. A common descriptor
(e.g., inverted T-beam) is able to provide a natural connector between the process oriented.
tree structure of rebar placement (Fig. 1) and the hierarchical decomposition of the rebar system.
The reinforcement for an office building is decomposed into distinct classifications (super classes)
of the building structure (e.g., FOUNDATION) and then into classes (e.g.. INVERTED
T-BEAM). Moving further down the tree (Fig. 2) vertically, three other classifications: (1)
Subassembly (e.g., LONGITUDINAL BARS), (2) set of bars (e.g., SET 1); and (3) individual
rebar element (e.g., BAR BISIT1) can be easily recognized. The individual rebar element is
identified with a code (e.g., B1S3TS1). Different segments of this code are used to represent
different aspects of the rebar element such as type of stirrup, set number, etc. The individual
rebar elements are grouped logically together into set(s) of bars based on the commonality of
certain features (e.g., type/shape). The bars or set(s) collectively represent a subassembly. For
example, the subassembly, STIRRUPS of a beam may comprise several sets of stirrups (e.g.,
five No. S, six No.3). Each individual stirrup receives its individual code that helps in identifying
it easily. For example, BIS3TSI (Fig. 2) is based on the following coding scheme: (1) B(eam)l
relates to inverted T-beam 1; (2) S(et)3 identifies the bars belonging to set 3; and (3) T(ype)
S(tirrup)l, shown as last three digits in the code, specifies the type of individual stirrup.
The integrated tree structure (Fig. 2) enables the building reinforcement to be methodically
arranged on the basis of critical concrete and rebar element features so that it can be searched
by the intelligent process planner. The following sections describe a CAD-integrated process-
planning system for rebar that utilizes the developed tree structures.
§ DB's
EXTRACTION
MANUAL
I'ROCESS
I'LAN
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.
l'ROCESS
ORIENTED
REBAR
LIST
DESCRIPTIVE
RP-PP
III
GRAPHICAL
Kl'-l'1'
I~~~~::~~il
KBS
ti
FIG. 3. Architecture of CAD-Integrated Rebar Placement Planner (CRPP)
In the CRPP, the designer is able to pick any standard bar from the library and insert it into
the design drawing. A customized CAD menu facilitates the designer to easily create detailed
design drawings by utilizing standard rebar from the feature library. Frame-based attribute lists
enable the designer to specify size, lengths of different sections of rebar, and so forth. Upon
completion of the feature-oriented rebar design, CAD data is made available to a database
management system (DBMS) as shown in Fig. 3. The extracted CAD data is stored in the main
database file (Rebar.dbf). CAD data is reorganized within the DBMS and two other separate
database files: (1) Design.dbf; and (2) Process.dbf are established. Design.dbf contains relevant
CAD data used by DBMS to generate a quantity takeoff (QT) rebar lists as shown in Table 1.
Instead of sending all the detail design data to the intelligent process-planner (PP) module, the
data is massaged and reorganized within the DBMS and required input data for process-planning
functions is stored in Process.dbf file. The outputs of the PP are placement plans for rebar (RP-
PP). As indicated in the Fig. 3, process plans can be developed manual1y (without the use of
PP) sequencing individual or set of bars using the tools provided within the CAD system. In
this case, a human planner labels the placement sequence of each individual bar or set of bars
with ascending numbers. "Placement sequence" is an attribute attached to each individual rebar
element in the feature library. Thus, when data is extracted from CAD, the list of rebar shows
not only sizes, lengths, etc., but also their position in the placement sequence.
entation, and location. These transferred data are utilized by the PP to develop the desired
plans based on expert knowledge stored in the knowledge bases. The next section will discuss
some of the rules used in the CRPP.
BUNDLE F.Cl.1.1.0
(ON TOP OF F.CI.1.4.01
BUNDLE F.Cl.1.4.0
STAGING AREA CI.2 (GREEN) (UNDERNEATH F.CI.1.I.O)
BUNDLE C.CI.1.3.0
BUNDLE F.CI.2.2.0 (DOWELS SEPARATE)
(ON TOP OF F.Cl.2.S.0)
SECTION A· A BUNDLE F.CI.2.S.0
(UNDERNEATH F.Cl.2.2.0)
top mat: all # 6 bars
bottom mat: all # 7 bars
FIG. 4. Design and Staging of Rebar for Simple Building Structure
~KJE assembly I
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.
WITH p_sequence := I
FIND END
based approach to automatically generate rebar placement plans for a building structure. A
simple building structure with four reinforced-concrete columns supporting a roof slab is pre-
sented in Fig. 4. As mentioned, LEVEL5 OBJECT has been utilized to structure the knowledge
bases for rebar placement planning of the building structure. The architecture of LEVELS
OBJECT is designed to isolate major areas of functionality into libraries of modules. A control
layer contains the root of the program as a shell, which is the main-event processor. It supports
the Dynamic Data Exchange (DOE) format for intertask communications and data exchange
(Kerry 1990). For example, the rule modeled in LEVEL5 OBJECT and presented in Fig. 5 was
able to sequence one of the main sets of rebar for an individual column footing having two (top
and bottom) mats of reinforcement as shown in Fig. 4. In this rule (Fig. 5), PP reads data from
the data-base Process.dbf. Once the record has values BLl (bottom layer 1), 1, and Cl for
bar-spec bar specification, set, and location fields, respectively, an instance (assembly 1) is
created. The attributes for assembly 1 are bar-spec, set, location and p_sequence as shown in
the premise clauses of the rule. When the record in the database file (Process.dbf) is located
using the fields and corresponding values, and an instance named "assembly 1" is created with
the attributes and values as dictated by the premise clauses of the rule, all of the conclusion
clauses of the rule become true. The conclusion clause (then) of the rule represents a common
practice in construction where bottom layer (BL1) of an individual column footing is placed in
the first step before the placement of other layers of rebar. The database file (Process.dbf)
record that contains values of bar-spec, set, and location as BL1, 1, and Cl, respectively appears
on the screen during KBS application and the field rebar placement sequence# (p-sequence)
is given a value 1 by the PP. The value of p-sequence as obtained from KBS application can
be updated. Records can be read, updated and deleted from within KBS application in LEVEL5
OBJECT. As mentioned earlier, data-driven forward-chaining rule bases have been utilized to
create rebar placement sequences or placement plan. LEVELS OBJECT facilitates the database
search for the antecedents of rules. When antecedents for a rule become true, it will conclude
and the actions are dictated by the consequents of the rule, such as the sequence for placement
of bottom layer (BL1) of the column footing. The result of a successful sequencing of the rebar
for an individual column footing is shown in Fig. 6.
As depicted, the PP assigned the number 1 to the placement sequence# (PS#) of BLl, the
number 2 to PS# of BL2, and the number 4 to PS# of TLI. PP assigned the number 3 to PS#
rI.l=-
:;
:;~
O=-
~I~
u
~
=-
rI.l
1 trI.l
E-
-=
~
<
~z
~9
-
0
E-
~
c::l
:; =~~
~UQ
U
y r...'
<~z ~
... ;.;l
QQ
~8
c.?
'"'
z~~
-~E-
~Cl'.~
E:S~U
~U
Cl'. ~ ;E- U W;1
~ ...
Cl'.<O
9
°... =
Q ~ ~
0 ~ rI.l
~
U rI.l = 0 ~ =-rI.l
TION FOOTING
SHALLOW IND. RECTAN BL2 2 0.30 90,0 Cl IN
FOUNDA- COLUMN PLACE
TION FOOTING
.. ........ .. ...... .. .. ....................
SHALLOW IND. RECTAN TLI 3 0.40 0.0 CI
FOUNDA- COLUMN
TION FOOTING
.....••..•. .•.•......
SHALLOW IND. RECTAN BLI 6 0.40 0,0 C2 IN :.
FOUNDA- COLUMN PLACE :
TION FOOTING <
of column dowels, which is not shown in Fig. 6. While the sequencing of bars within a feature
class would provide valuable information to the rebar crew. the placement plan offers an excellent
opportunity to organize the delivery of the fabricated rebar according to the assembly sequence.
As mentioned, rebar is bundled and tagged depending upon size and shape, and stored in a
haphazard manner in the site laydown or staging area. This bundling method and on-site storage
does not take into account the order of assembly. One alternative approach to this traditional
method is to bundle and deliver set(s) of rebar according to the placement plan. For example,
rebar for an individual column footing, having only two layers of reinforcement for the bottom
mat, could be bundled separately, each bundle consisting of rebar for one layer. While unloading,
the set(s) of bundled rebar could be directed to a predesignated staging area closest to the
column footing. Thus, bundling and on-site staging could take advantage of sequencing infor-
mation for the automatic generation of placement-oriented bundling and staging schemes.
To take advantage of this opportunity, rule bases for bundling and staging have been estab-
lished within CRPP. While the available data transferred from the CAD system provides relevant
data (see Fig. 6), additional conditions and constraints have to be considered. For example, the
available rigging machinery will limit the capacity to be lifted at a time. However, sets of rebar
belonging to the same building element (e.g., column footing) should be combined as much as
possible into one lift to secure the lowest possible crane cycles for unloading. Also, available
storage areas constrain the bundling options. Overall, the objective of a bundling rule base is
to create bundles that will minimize the rigging and assembly time by combining the bars
according to their placement sequences. In other words, rebar needed next in the sequence
should always be immediately accessible to eliminate search time on site. The realization of
these on-site savings should not cause additional cost in the fabrication shop. In the shop,
shearing and bending of straight bars require setup time for every new set of bars. To minimize
these additional costs, careful planning and organization of the operations are crucial. Personal
observations at the fabrication shop during the experimental work revealed only minimal impact
of a modified bundling scenario on the fabrication of rebar.
The bundle code for the first set of bars shown in Fig. 6 consists of five different identifiers,
separated by periods. The first segment of the code indicates the feature class. For example,
the bundling code for BLl (bottom layer 1 of individual column footing) is F.C1.1.l.0 where
F., C1.1., and 1.0 stand for foundation, staging area I for column CI, and placement sequence
number 1, respectively. The 0 at the end of the code indicates that the set does not need a more
detailed sequencing (i.e., sequencing for each individual bar in the set). Thus, the staging area
for BLl is designated as CI.I, which identifies staging area I assigned to the rebar for column
according to its orientation in the formwork. For example, a set is stored parallel to its final
location, thus minimizing time and space consumed in handling.
The sequencing of the individual or set(s) of rebar, as obtained from process-planning action,
can be fed back to AutoCAD to generate three-dimensional (3-D) drawings of the critical steps
of assembly operation. The bottom line of the CRPP is to provide 3-D CAD models of the
rebar placement sequences together with a bundling scheme and staging assignments to minimize
wastage of resources for searching, identifying, and rehandling of the needed bars for assembly
operations.
ratings. A good sample design is particularly important for productivity ratings of construction
crews because it is physically impossible to observe and record details of a construction operation
every minute. For sampling of construction operations, there is a general agreement that a
confidence limit of 95%, and a limit of error of plus or minus 5% give a good indication of the
overall effectiveness of an operation. A category proportion 50% for construction labor sampling
is usually considered. With these limits, a minimum of 384 observations are required to measure
and to draw statistically sound conclusions (Oglesby et al. 1989). It was decided to make 432
observations per floor. The main reason for the decision of making 432 observations was to
have an allowance to ensure the attainment of minimum-required observations under any adverse
situations (e.g., loss of observation data). For the purpose of final assessment of rebar-placing
operation, it is necessary to analyze the three main work categories: (1) Direct or effective
work; (2) essential contributory work; and (3) ineffective work. Fig. 7 presents a report from
the study (Salim 1993). The direct or effective work is represented by a "hill-like" dashed curve
ranging approximately from 25% of crew time for the second floor to 55% of crew time for the
fourth floor. Contributory and ineffective works are represented by two other curves. Contrib-
utory works for the second through fifth floors were 23%, 27%, 21 %, and 22% respectively.
Ineffective works for the second through fifth floors accounted for 52%, 45%, 24%, and 34%,
respectively. Results from the study revealed that crew time spent in effective work for the
traditional method was approximately 26% (based on an average of the second- and the third-
floor values), whereas for placement-oriented bundling, delivery, and staging it was about 55%
(second-floor value). For the fifth floor, crew time spent in effective work was found to be 44%.
Productivity ratings for ironworkers conducted by a large construction firm provides the following
information: (1) Direct or effective work (31 %); (2) contributory work (36%); and (3) ineffective
work (33%) (Oglesby et al. 1989). It is also stated that the ratings given represent good per-
formance (Oglesby et al. 1989). This information provides a good reference point although the
nature of construction for the studies are not known. Compared to the studies conducted by
.~/. -_ _ .
,
...................................... _ _....~
. .
~---
__----=e
.................... ._ .. ,. _._ ._ .. _._ ....•..••.•..•1 _~.•..•. .....
CONCLUSIONS
This paper presented a process-oriented planning model for the placement of rebar. A com-
puter-integrated framework for implementing the model has provided a promising technique
for automating the planning of rebar placement. The primary motivation for this part of the
work was the need for a CAD-integrated process-planning system that was able to create
placement sequences of rebar for assembly operations. This paper also presented the results of
field experiments conducted to measure the effects of process-oriented bundling, delivery, and
staging on crew-level productivity. Productivity ratings revealed that crew time spent in effective
work for the traditional method was approximately 26% (an average of the second and third
floor), whereas for placement-oriented bundling, delivery, and staging it was about 55% (fourth
floor). For the fifth floor where traditional bundling with modified delivery and staging schemes
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.
was used, crew time in effective work was found to be 44%. The data suggests that the concept
of placement-oriented bundling, delivery, and staging is sound and very effective in improving
crew-level productivity for the placement of rebar in reinforced-concrete construction with beam-
joist floor system.
ACKNOWLEDGMENT
The material in this paper is based upon the work supported by the National Science Foundation under grant
No. MSM-9196052. The writers gratefully acknowledge the support of the National Science Foundation. The
writers also extend their grateful thanks to National Erectors and Owen Steel of North Carolina for their co-
operation in the field studies.
APPENDIX. REFERENCES
Ansaldi. S., Boato, L., Canto, M., Fusconi, F., and Giannini, F. (1989). "Integration of AI techniques and CAD
solid modeling for process planning applications." Computer application in production and engineering, Elsevier
Science publishers, North Holland, Amsterdam, The Netherlands.
Bernold, L. E., and Livingston, E. E. (1990). "A prototype for intelligent computer integrated wood truss
fabrication." Robotics and autonomous systems 6, Elsevier Science Publishers, North-Holland, Amsterdam,
The Netherlands.
Bernold. L. E .. and Reinhart, D. B. (1990). "Process planning for automated stone cutting." 1. Comput. in Civ.
Engrg.. ASCE, 4(2), 225-268.
Bernold. L. E., and Salim. M. (1993). "Placement-oriented design and delivery of concrete reinforcement." J.
Constr. Engrg. and Mgmt., ASCE, 119(2),323-335.
Chang. T. C, and Wysk, R. A. (1985). An introduction to automated process planning systems. Prentice-Hall,
Englewood Cliffs, N.J.
Cunningham, J. J., and Dixon. J. R. (1988). "Designing with features." Compo in Engrg., 237-243.
Dym, C L., and Levitt, R. E. (1991). Knowledge-based systems in engineering. McGraw Hill, New York, N.Y.
Kerry, R. (1990). Integrating knowledge-based and database management systems. Ellis Horwood, U.K.
Oglesby. C H., Parker, H. W., and Howell, G. A. (1989). Productivity improvement in construction. McGraw
Hill. New York, N.Y.
Salim, M. (1993). "Process-oriented planning for the placement of reinforcing bars," PhD thesis, Dept. of Civ.
Engrg., North Carolina State Univ., Raleigh, N.C
Salim. M., and Bernold, L. E. (1992). "AI supported process planning for automated rebar fabrication." Proc.
8th Con!- CompUi. Civ. Engrg., ASCE, New York. N.Y.
Zozaya-Gorostiza, C, Hendrickson, C, and Rehak. D. R. (1989). Knowledge-based process planning for con-
struction and manufacturing. Academic Press, San Diego, Calif.