Research suggests conservative design of concrete box culverts
By Paul Smeltzer, P.Eng.
and Evan Bentz, Ph.D.
Specimen placed in test frame and
being prepared for experiment.
Precast concrete box culverts
are primarily used to carry
streams beneath roadways, as
conduits for buried stormwater
or sanitary sewers, and as access tunnels
under road and rail infrastructure.
It is critical to public safety that these
structures not undergo brittle shear
failures. If the shear reinforcement in
these structures is excessive, however,
it is a waste of resources and contrary
to principles of sustainable development.
To resist bending moments caused
by earth pressures and vehicle loads,
precast concrete box units contain longitudinal
and circumferential steel
reinforcement near the inside face and
the outside face of the box. Where
excessive loads are experienced, due to
depth of cover or other factors, shear
steel reinforcement is added to the
design. While such shear reinforcement
requires more steel, it is also difficult
to place and can significantly
increase production time and energy
resources during the dry cast production
process.
Preliminary research at Centennial
Concrete Pipe (now Hanson Pipe &
Products Canada, Inc.) in late 2000,
witnessed by Gamsby and Mannerow
Limited, suggested that box specimens
produced with, and without a new type
of shear reinforcement failed at a load
between 820 and 1000 KN. The engineer
reported that the testing did not
establish the usefulness of the new
type of shear reinforcement. Testing
also indicated that there was a need to
investigate traditional formulae and
codes calling for the level of shear
reinforcement currently required in
box units.
Canada is fortunate to have
Professor Michael Collins, a world
leader in the consideration of shear in
concrete and as a faculty member at
the University of Toronto. Prompted
by the research at Centennial, Dr.
Collins was approached by the Ontario
Concrete Pipe Association (OCPA) to
propose a research project on shear
steel design.
In addition, after Harvey Pelligrini
of Materials and Manufacturing
Ontario (MMO) was contacted on this
landmark research to determine the
interest of the province, MMO quickly
agreed to partner on the project.
Rounding out the research partners
were the Natural Sciences and
Engineering Research Council
(NSERC), the OCPA, Con Cast Pipe
Limited, Hanson Pipe & Products
Canada, Inc., M-Con Products Inc.,
and Munro Concrete Products. A
research proposal presented by Dr.
Collins was accepted in 2003 and the
results released in January 2004.
University of Toronto research
involved 12 major experiments on
actual box culverts to develop an
understanding of the shear resisting
mechanisms for such structures. The
crack development, reinforcement
strains, and specimen deformation
were compared to the results of extensive
nonlinear finite element analysis
using the computer modeling techniques
developed at the University of
Toronto. Discussion presented in the
report called, “Shear Behaviour of
Concrete Box Culverts: A Preliminary
Study” by R.A. Yee, E.C. Bentz, and
M.P. Collins, identifies areas of weakness
and lack of clarity in the current
codes governing box culvert design.
The study developed an experimental
procedure to determine the adequacy
of the current shear design procedures
for a range of commercial box
culverts. By comparing the experimental
results to the analytical predictions
from the shear strength equations in
North American codes, the ability of
these provisions to predict the shear
behavior of the test specimens was
examined.
Ultimately, the objective is to use
the information to provide recommendations
to industry outlining the adequacy
of commercial box culvert
designs in shear, and to more accurate-
ly identify where shear reinforcement
is required and where it is not. The
analytical and experimental results are
useful to future box culvert studies and
the precast industry, as well as to
industry associations and academics
involved in researching the shear
behaviour of concrete structures.
In addition to the observations at
Centennial, there were other reasons
for investigating the adequacy of the
current shear design procedures. In
2000, the Canadian Highway Bridge
Design Code (CHBDC) was updated
and became the principal culvert
design guideline for Ontario, formerly
under the jurisdiction of the Ontario
Highway Bridge Design Code
(OHBDC). This change in code
prompted concern by the Ontario concrete
pipe industry that products of
some OCPA producers would no
longer appear to be adequate in shear
under the revised code provisions,
even though no deficiency in shear
behaviour had been observed in the
field.
Some uncertainty remains as to the
precise load causing shear failure
under ideal worst-case conditions due
to the presence of axial load as a result
of the test set-up conditions. Axial
loads in the slab were several times
those specified by design specifications
but were necessary to ensure a
statically determinate loading scheme.
Most code predictions anticipate that
the presence of these additional axial
loads would enhance the shear capacity
of the sections by 10 percent or less,
with the exception of the method presented
in CHBDC 7.8.8.2.1. The adoption
of the expression
in CHBDC Clause 7.8.8.2.1 to account
for axial load has a significant influence
on shear strength predictions,
increasing these predictions on average
1.5 times in this experimental program.
This may lead to un-conservative
predictions of shear strength if
axial loads are significant.
In addition, stirrups, which are
commonly used in the reinforcing
design of box units for shear design,
were not used in any of the specimens
cast for this testing.
The code shear provisions should
be able to predict the behaviour of the
loaded culverts in the experiments
regardless of the presence of axial
load. All the code shear provisions
were found to underestimate the shear
capacity of the culvert members. Three
shear code equations from CHBDC
and the AASHTO box culvert shear
equation were studied. Calculations
were performed assuming an elastic
member response. The test data confirmed
that the point of inflection
between the regions of the positive and
negative moment in the slab shifts
inwards toward the mid-span at higher
loads. As an elastic analysis predicts a
stationary inflection point, it was
found that actual post-cracking
moments in the sections of higher
shear stress were lower. This would
suggest that the use of elastic methods
in design is somewhat conservative,
and hence the code predictions should
be slightly conservative.
Figure 1:1: Size Effect
CHBDC code provisions predicted
shear failures on average 56 percent of
actual observed shear failures. The
AASHTO Box equation was less conservative
predicting shear failure on
average 71 percent of actual observed
shear failures. Care must be taken,
however, to recognize the limitations
of the AASHTO equation. Studies
have shown that the shear capacity of a
reinforced concrete member is subject
to a size effect depending on the depth
of the section. Shear critical sections
tested in this study are plotted with a
much larger box culvert section tested
at the University of Toronto Structural
Laboratories for the Toronto Transit
Commission in Figure 1.1. The
CHBDC General Method was found to
give an accurate prediction of the failure
load of the TTC box (Kuzmanovic
1998). Notice, however, that the empirical
AASHTO minimum limit gave a
highly un-conservative prediction of
strength for the TTC box.
Tests on larger depth members
show that the AASHTO minimum
limit can be dangerous for application
to specimens with larger section
depths due to the size effect in shear.
Curiously, the lower limit on the
AASHTO equations of
is
the same as the upper limit on the
CHBDC Clause 7.8.8.2.1. This could
imply that CHBDC Clause 7.8.8.2.1
Method is limited by an overly conservative
expression.
The CHBDC General Method gave
the most consistent predictions. Since
the CHBDC General Method is known
to give accurate predictions in other
applications, clearly there are factors
not being considered when applying it
to culverts.
Figure 1:2: Specimens considering load paths of forces.
One possible source of error could
arise from an inappropriate consideration
of the loaded length that influences
the critical shear section. In
Figure 1.2, the tested culvert sizes are
shown to scale with the critical section
at dv from the haunch and crack
extending to 2d from the haunch.
When considering the load path of the
applied load, it would be reasonable to
assume that all of the load on the
haunch side of the break would flow
directly into the culvert wall. Thus, the
actual load acting on the critical section
at dv could be less than anticipated,
indicating that the section would
fail at higher loads.
Issues such as concrete cracking
strength (which has a profound influence
on the shear strength of sections
without transverse shear reinforcement),
the superior bond characteristics
of the reinforcement mesh, and the
previously mentioned conservatism
associated with internal loads determined
from elastic analysis results,
may all contribute to the overall conservatism
of the predictions.
Although the study represents a
preliminary investigation, there is substantial
evidence included in the study
report to suggest that shear reinforcement
may have little influence in the
design of most culvert sections conforming
to Ontario Provincial
Standards Specification (OPSS) 1821
– Material Specification for Precast
Concrete Box Culverts and Box
Sewers, if concrete is allowed to reach
sufficient compressive strength.
Results also indicated that shear reinforcement
for many deeper earth cover
applications may not be required to
satisfy the CHBDC design load
requirement. The study recommends,
however, that further testing confirm
any such inference, particularly for
box members with lower concrete
strengths.
Preliminary investigations into
shear behaviour of concrete box culverts
gave an indication of the design’s
susceptibility to shear failure in the
box slab. Experimentation also clearly
established the load range at which
shear failure is to be expected in the
slab, the reliability of analytical techniques
at predicting member internal
forces, and the conservatism of both
the box member design and code shear
prediction methods outlined in
CHBDC and AASHTO codes. These
findings provide a basis from which to
expand the reliability and scope of the
investigation so that specific recommendations
can be made to codes
reflecting the conservatism of the culvert
designs. With a greater understanding
of the behaviour of the culverts
in shear it would also be possible
to expand the range of culvert sizes
and design depths standardized in
OPSS 1821.
Investigations into shear behaviour
of concrete box culverts continue. The
Ontario Concrete Pipe Association is
working with its industry partners
including the American Concrete Pipe
Association and the Canadian
Concrete Pipe Association to acquire
funds required for the next level of
research. When completed, this
research is expected to have a significant
impact on the cost of buried infrastructure
and the use of resources for
producing precast concrete boxes. But
most significantly, research will have a
profound impact on design methodology
and principles used in industry and
academia for concrete structures that
will ensure public safety. Contact Paul Smeltzer, Ontario
Concrete Pipe Association, Tel: 905-
631-9696. Evan Bentz is an Assistant
Professor, Department of Civil
Engineering at the University of
Toronto.
See our home page on how to order your subscription. We regret we can
only accept orders from Canada and the United States.
