Dynamic Load Allowance – A Synthesis of the State of the Art

Colin Caprani
Senior Lecturer, Monash University

Abstract: The dynamic interaction between heavy vehicles and bridges is a key consideration for bridge
loading. Over many decades, engineers have sought to understand the phenomenon and determine
suitable design rules to ensure safe designs. However, with the aging of vast networks of highway
infrastructure in many countries, there is a need to avoid overly-conservative assessment of existing
bridges. In such cases the traditional worst-case scenario approach is better replaced by a probabilistic
approach. But this approach is difficult given the subject matter and wide range of variables in the
problem. This paper summarizes the current state of the art in the modelling and assessment of the
dynamic load allowance from a first principles viewpoint. It presents recent research in numerical
modelling and experimental work. It shows that there are some consistent findings by many researchers
that could be implemented in bridge assessment codes of practice to reduce rehabilitation costs while
ensuring safety.
Keywords: Bridge assessment, Dynamic load allowance, Vehicle-bridge interaction.

1 Introduction

1.1 Basic Definitions

The load effects that a vehicle generates in a bridge arise from a combination of the static weight of the
vehicle and potentially inertial loads due to dynamic interaction of the vehicle and the bridge. The
consequent total load effect, T, is typically larger than that which would result from a wholly static
analysis, S. The dynamic amplification factor (DAF) represents the ratio of these two load effects and
is defined as: 
DAF  T
S (1)

An alternative definition of the same concept used by some authors (McLean and Marsh 1998) and
codes of practice (AASHTO 2002, AS 5100) is the Dynamic Load Allowance (DLA), Dynamic Increment
(DI), or Impact Factor (IF) given by:    S 
DLA, DI, or IF   T   100%
 S  (2)

1.2 Motivation
The motivations for studying the level of dynamic interaction that occurs between a bridge and traffic
should be clear. There are two main reasons.
For the design of new bridges, it is comparatively cheap to add additional strength, and so a conservative
approach to most design aspects, including the level of dynamic interaction, is prudent. However, for
the assessment of existing bridges this is not the case: it is very expensive to increase the strength of
an existing bridge, both in terms of the construction, but also in terms of traffic disruption costs.
Therefore, overly conservative approaches are not suited to the assessment of existing bridges. Instead
such assessments should be highly engineered to achieve a prescribed (acceptable) level of safety and
service from the structure. It is against this backdrop that the assessment rules for dynamic interaction
should be continually assessed against the current state of scientific knowledge in the area.
Free-flowing traffic which is travelling at full speed dynamically interacts with the bridge, whereas
congested traffic does not do this, due to its characteristic low speed. However, congested traffic is
denser. For shorter spans (which very few vehicles can occupy) it is clear that free-flowing traffic with
dynamic effects governs the loading, but for longer spans (which multiple vehicles can occupy)
congested traffic conditions governs. The length at which congested traffic becomes the governing
regime is not easy to determine, and depends on the level of dynamic interaction considered to be
acting. Thus for medium-length bridges, it may not be clear whether free-flow traffic, allowing for dynamic
effects, or congested traffic, which is denser, is the governing loading condition. To ascertain the correct
form of traffic, it is therefore essential to know the level of dynamic interaction expected at the return
period of interest, in order to make an accurate determination of bridge safety.
For these important reasons, much research has been conducted to examine the level of dynamic
interaction that occurs. There are many past reviews conducted of the literature on dynamic amplification
(Deng et al 2014; Pape et al 2014; AustRoads 2003). This paper instead presents a synthesis of the
state of the art, and outlines how there is a new and consistent paradigm emerging through multiple
research results.

1.3 Typical Observations

Before exploring the phenomenon of dynamic interaction and the resulting allowance to be made, it is
useful to view some field observations as a precursor to the discussion. As part of the SAMARIS project
(EU 7th Framework), a bridge weigh-in-motion (B-WIM) system (SiWIM) algorithm was improved and
used on several bridges. Roughly, by calibrating the bridge influence line using known trucks, the
dynamic signals from unknown trucks are filtered and their static weight (and other quantities) can be
back-calculated. This procedure renders the DAF intrinsically. Figure 1 provides data from the 2-week
measurements on the Hrastnik bridge (Slovenia). In total, 10 770 loading events, 370 of which were
multiple presence events with another light and 166 with another heavy vehicle, were captured. The
figure also shows some DAF code models for comparison. This experiment was conducted before and
after resurfacing of the pavement, and the average DAFs before and after this are shown in Figure 2.

1,8
One vehicle - Lane 1
Semi-trailer 400 kN

1,7
One vehicle - Lane 2
1,6 MP with a light vehicile
MP of heavy vehicles
1,5
Slovene BD code
1,4 Incorportated in Eurocode
DAF

Danish RBBA code

1,3

1,2

1,1

1,0

0,9
0 100 200 300 400 500 600 700 800 900

Weight (kN)

Figure 1. Hrastnik Bridge – Measured DAF values of loading events after resurfacing of the
pavement. Also shown code values for: Slovenia (BD), Eurocode, and Danish Road Directorate
(after SAMARIS 2006).

112%
Total Total
110%
To river To river
108% To town To town
DAF

106%

104%

102%

100%
50 150 250 350 450 550 650 50 150 250 350 450 550 650
Total w eight (kN) Total w eight (kN)

Figure 2. Hrastnik bridge – Average DAF before (left) and after resurfacing of the pavement
(right) (after SAMARIS 2006).
2 The Vehicle-Bridge Interaction Phenomenon

2.1 The Road Surface

It is consistently found that the road surface condition is a strong indicator for the level of dynamic
interaction that occurs (Deng et al 2014). Smooth road profiles produce lower DAFs than rough road
profiles. With this in mind many authors have explored the relationship between road surface quality
indices, such as the International Roughness Index (IRI), and resulting DLAs (e.g. dynamic amplification
estimator, bridge roughness index, etc.). However, due to the many other factors affecting DAF no
general relationship has been found.
One certainty that can be stated is that regular pavement maintenance reduces the level of DAF with
consequent benefits for bridge maintenance needs (see Figure 2). This is a cost-effective approach to
increasing bridge safety. Particular attention should be paid to bumps at the start of the bridge.

2.2 The Bridge

Bridge length and frequency

There is a strong inverse relationship between bridge length and first natural frequency (e.g. Chan and
O’Connor 1990). Hence one may be used as a proxy for the other. The literature (see Deng et al 2014,
Li 2006) is consistent that short bridges suffer from higher dynamic interactions than longer bridges.
One exception is that when frequency matching occurs between the vehicle and bridge, a higher DAF
results.

Bridge Type
Most research has been done on short to medium span bridges (around 4 – 8 Hz) due to their prevalence
and higher dynamic interactions. I-girder bridges seem to be the most well studied (Deng et al 2014).
The main indicator of the level of dynamic interaction is the level of static load carried by the girder:
more highly statically loaded girders tend to have lower DAFs.

Bridge Material and Damping

As with all vibration problems, increased damping usually leads to reduced response, and DAF is no
different. Timber bridges (for example) are found to have lower DAF than steel bridges (for example).
Interestingly new materials, such as fibre-reinforced polymer (FRP) do not exhibit the expected trend,
and further research is needed on these forms of construction (Deng et al 2014).

Pre-Excitation
Rattigan (2007) studies the influence of pre-existing vibration of the bridge (from a preceding loading
event) on the resulting DAF of a new loading event. In some cases, the dynamic increment can more
than double, but this requires a very particular set of circumstances (gap and damping).

2.3 The Vehicle

Vehicle Type
Gonzalez et al (2011) examine the DAFs for a range of bridges when subject to heavy articulated 5-axle
trucks or cranes, even allowing for bumps at the start of the bridge and meeting events. For shorter
spans (< about 20 m) the articulated truck yields higher average DAFs than the cranes, but this
difference dissolves for longer spans.

Vehicle Velocity and Acceleration

The effect of velocity on DAF has received much attention due to the potential for frequency matching
between the crossing frequency (velocity/length) and bridge frequency. Indeed, it has been shown that
there can exist very high levels of DAF, but usually only at very high speeds. However, these studies
are usually theoretical and simplified, and when more realistic simulations (or field tests) are conducted,
these predicted high DAFs are difficult to recreate (Zhu and Law 2002). Finally, accelerations and
decelerations are also found to affect DAF, with increased DAFs found in some cases for high
decelerations (Deng et al 2014, Li 2006).
Vehicle Weight
The mass of the vehicle is a significant influence on dynamic interaction. In particular, it is the ratio of
the vehicle to bridge mass that is relevant. High ratios change the overall dynamic properties of the
system, and also result in higher static stresses. It is widely found that dynamic amplifications decrease
with increasing weight of vehicles or static stress (see e.g. Figure 1).
Since the influence of vehicle weight or static stress on DAF is vital to later discussion, a more complete
list of third-party literature supporting this is given: Hwang and Nowak (1991); Huang et al. (1993);
Chang and Lee (1994); Abdel-Rohman and Al-Duaij (1996); Kim and Nowak (1997); DIVINE (1997);
McLean and Marsh (1998); Laman et al. (1999); Broquet et al. (2004); SAMARIS (2006); Li et al. (2008).
These works are consistent and in addition to those of the team at University College Dublin that studied
this problem (Cantero, Caprani, Dowling, Gonzalez, Harris, Li, OBrien, Rattigan).

Number of Axles
Despite the body of research, there is no clearly identified link between DAF and the number of axles
(Deng et al 2014). Indeed, it is difficult to isolate the effect of number of axles as distinct from those
more significant ones such as vehicle weight.

Vehicle Suspension
Different truck suspension systems are known to influence the dynamic amplification. For example, the
DIVINE (1997) project reports that air suspension cause less (5-10%) dynamic amplification of static
wheel loads than spring-leaf suspensions (20-40%). Indeed, Harris et al (2007) show how a bridge-
friendly vehicle suspension system can be designed that minimizes dynamic amplification.

Vehicle Load Position

The transverse location of the vehicle affects DAF only insofar as it affects the level of static load in any
particular girder or element due to the transverse stiffness of the deck. Then, the reduction of DAF with
increasing static load is again observed, and vice versa (Huang et al 1993, Deng et al 2014).

Pre-Excitation
In both field trials and numerical simulations, it is important that the vehicle has a sufficient approach
length on the road surface before the bridge so that initial excitation of the vehicle can take place. This
is merely a prudent measure to obtain realistic DAFs, but does not affect DAF directly otherwise.

2.4 The Loading Event

Number of Vehicles
The literature consistently shows that the DAFs for multiple truck presences are lower than for single
trucks (Deng et al 2014, AustRoads 2003). This is strongly related to the increased static load (and
consequent lower DAF) that has been noted previously. For example, Gonzalez et al (2011) find that
for two 5-axle trucks meeting on any bridge over 12 m, the largest DAF for shear is 5% and for bending
is 1%. Similarly, Rattigan used a calibrated finite element model of the Mura River Bridge (SAMARIS
2006) to find that in the worst case, two 5-axle truck meetings events on a 32 m bridge, give a DAF of
15%.

Inter-Vehicle Spacing
Some theoretical studies have found that for some idealized situations a higher DAF can result for
multiple moving point loads at particular spacings, than for a single point load (Li 2007). However, for
more realistic scenarios, DAF is found to consistently reduce due to the presence of more ‘load’ on the
structure.

2.5 The Structure Response

Load Effect
There are many responses of the structure that may be of interest, such as bending moment, shear
force, torsion, deflection, stress, strain, etc. The reported DAFs vary, depending on the response
selected, even for the same bridge and truck (e.g. Huang 2008, Gonzalez et al 2011). Validation of
simulation models against field trials is commonly conducted but these usually only report on strain or
deflection DAFs. Therefore, simulation models (validated as best as possible) for other effects must be
relied upon.

Location
Interestingly, the critical location for static load effect is not necessarily that for critical total (static +
dynamic) load effect (Li 2007). Cantero et al (2009) find that the maximum total bending moment occurs
away from the mid-span of a simply supported beam, and define a ‘full’ DAF (FDAF) as the ratio of the
maximum total response anywhere on the beam to the mid-span static bending moment. They find this
is always great than 1.0 (unlike DAF) and can be greater than DAF. Huang (2008) finds similar results
for a curved box girder. In short, it matters where on the bridge DAF is measured.

3 The Current Paradigm

3.1 Codes of Practice

Against the backdrop of the research that has been summarized earlier, rules for dynamic allowance
have been developed and incorporated into codes of practice for many years. These rules are
summarized in Table 1, though it should be noted that the rules for each code are not exhaustively given
– only key results for medium spans.
Table 1: Summary of some international codified rules for dynamic allowance.

Country/Region Phenomenon Key Rule

Australia Number of axles DLA  0.1  0.4

Canada (CHBDC) Number of axles DLA  0.25  0.5


DLA  0.1767 ln f  0.0157 1.5  f  14 Hz
0.45 f  14 Hz
China (MTPRC) Frequency 

DLASF  0.2 L  25 m
Europe Bridge length + Load Effect

DLA 
Japan Length + Material 50  L

DLA   0.3
New Zealand Length + Load Effect L  38

Ontario (OHBDC) Frequency DLA  0.2  0.4

Switzerland Frequency DLA  0.2  0.4

United Kingdom (BS5400) - DLA  0.25

DLA   0.3
United States Bridge length L  38.10

Consideration of Table 1 shows a wide disparity of rules. Ignoring this for a moment, more
fundamentally, there is also a disparity in the phenomena upon which the code rule is based, or aiming
to replicate. If, however, as noted earlier, bridge length and bridge frequency are considered analogous,
then the majority of codes shown (7/10) reflect this phenomenon. Interestingly, the Australian and
Canadian codes do not reflect this, but account for the number of axles instead (e.g. different DLA for
W80, A160, etc. of AS5100.2). Only a few codes account for factors such as the load effect considered
or the bridge material. More interestingly, from the discussion of the VBI phenomenon earlier, there are
many significant issues that are widely ignored, such as multiple truck presence, and reducing DAF with
increasing static load.
3.2 Static Bridge Loading
Before offering an analysis of the current paradigm in dynamic allowance, it is necessary to consider
that to which the DLA is to be applied: the static loading. Over the past few decades there has been
much research work done on the development of accurate static load models, which has informed recent
codes of practice (e.g. Eurocode, Danish Road Directorate). To summarize this broad field briefly, the
critical forms of loading are:
 Point locations (e.g. expansion joints), or lengths < ~1.0 m: single heavy axles govern;
 Lengths < ~8.0 m: axle groups (tandem, tridem, etc.) govern;
 Lengths < ~20.0 m: single-truck per lane governs;
 Lengths < ~40.0 m*: multiple trucks per lane govern
 Lengths > ~40 m*: congested traffic.
Beyond the indicated length of ~40 m congested traffic is denser, and so although no dynamic allowance
should apply (e.g. S1600), it offers more extreme load than free-flowing traffic with a dynamic allowance.
This dimension of 40 m is marked (*) because it varies depending on the load effect considered (e.g.
bending, shear, etc.), and the level of dynamic allowance (Caprani 2013).
Note that for lengths over about 8.0 m, as long as there are multiple lanes, multiple truck events will
govern the extreme loading at the design return period (see OBrien and Caprani 2005). An example is
shown in Figure 3, where the upper tail of the overall (composite) distribution is clearly governed by the
3-truck loading event distribution. This finding is consistent across many countries and bridges.

12
Standard Extremal Variate

10
1-truck
8
6
2-truck
4
3-truck
2 4-truck Composite distribution
0
-2
5-truck
0 1000 2000 3000 4000 5000 6000
Load Effect 1 (kNm)
Figure 3. Example daily maximum load effect extrapolation for mid-span bending moment in a
40 m bridge, showing different numbers of trucks involved (after Caprani 2013).

3.3 A Critique
By now from this discussion it should be clear that there is a problem with the existing paradigm.
Dynamic interaction is inherently variable (Figure 1), with many complicating factors (Section 2),
resulting in a wide variety of approaches to accommodate it (Section 3.1).
In the past, during the initial research work on dynamic allowance, when faced with uncertainty,
researchers responded by adopting a worst-case-scenario approach (e.g. see review by Chan &
O’Connor 1990). The extremely influential work by Cantieni (1983) is notable for the manner in which
the observed DAFs are enveloped by the proposed (and then adopted) design curves. Indeed, this
legacy can be seen still today throughout Table 1. Much of the following work, strongly influenced by
Cantieni’s, also saw a search for the worst-case-scenario of DAF. Invariably then, from the discussion
of Section 2, and as may be seen from Figure 1, the critical DAFs that were identified were associated
with single light weight trucks. Even further, some researchers (see Deng et al 2014) explored the
influence of velocity finding that at ever greater speeds (> 300 km/h) the DAF continues to increase (as
more and more modes become active). While it is certainly worthwhile to explore the bounds of a
phenomenon, this worst-case-scenario philosophy has permeated through codes and practice, resulting
in unreasonably high DAFs, borne of single light vehicles, being applied to the heavy multiple vehicle
presence static loading that governs the extremes.
Of the many factors discussed that influence DAF (Section 2), it is the wide omission of the knowledge
of multiple truck presence DAFs from codes of practice that is most egregious. Simply put, the research
categorically finds that as loading goes up, DAF goes down. And since the critical loading is heavy,
DAFs at the extreme end should be small. The data of Figure 1 should help to convince on this point.
As a pertinent example, in the development of the DLA for SM1600, AustRoads (2003) points out that
the DLA is developed for single vehicle events only, even though it is acknowledged that multiple vehicle
events are the critical load case for many bridges. However, since the SM1600 truck is a single truck, it
was deemed appropriate to use a DLA based on that of a single vehicle. In spite of this, there is surely
a distinction between a notional load model (intended only to replicate the effects of real traffic), such
as SM1600, and real vehicles. And indeed it is vital to maintain this separation of concepts, lest freight
operators request to operate a real M1600 truck!
Besides the inappropriateness of the worst case scenario approach, there is little acknowledgement of
the inherent variability of the DAF. Indeed, in the light of modern codes of practice, where parameters
have partial factors that are calibrated using statistical information to achieve a target level of safety, it
is clear that the rules on dynamic allowance remain mired in a “working stress”-type philosophy. There
is no distinction between serviceability limit states, or ultimate limit states, and there are certainly no
partial factors to apply to some type of ‘characteristic DAF’.
Finally, there are some other crucial considerations that have barely been touched on by researchers:
 It is well known that material responds differently at higher strain rates – in the main they are stiffer
and stronger during dynamic loading, such as vehicle-induced vibrations. Implicitly field tests allow
for this, but applying the DAFs from strain or deflection to elastic bending moment is flawed. Hussein
and Gonzalez (2012) show that when strain rate effect is allowed for, the DAF is significantly smaller.
No other works on this topic are known.
 At the ultimate limit state (ULS), ductile failures are sought, and these have tremendous energy
absorbing capabilities. Vibration interaction of the extreme multiple-truck event and a yielding bridge
will be minimal due to plastic hinge formation. Thus there can be no elastic-added dynamic effect
and so, rather simply, DAF should not apply for ductile failures at ULS. Brühwiler and Herwig (2008)
demonstrate a simple model that is consistent with this intuition. In contrast, for brittle ULS failures,
depending on the level of deformation possible, some dynamic allowance should be made
(Ludescher and Brühwiler 2009). There is no further work on this topic known.
 Earlier it was noted that the DAF (ratio of total to static) actually varies along the length of a beam
from the results of Cantero et al (2009). This raises some questions: is it the maximum DAF that is
required, for any point? Or is the ratio between the maximum total effect anywhere and the maximum
static effect anywhere? Or is it something different? Other than the cited work, this problem has not
been addressed, but is clearly fundamental to what is needed from the research for practice.
In summary, a new approach to the conceptualization of allowing for dynamics is required. One that is
probabilistic and can account for many complicating factors. This new approach has begun to emerge.

4 A New Paradigm in Dynamic Allowance

4.1 A First Principles Approach

At the outset, it is useful to remind ourselves of the intent of a dynamic allowance. Chan and O’Connor
(1990) express it concisely as:

An increase in the design traffic load resulting from the interaction of moving vehicles
and the bridge structure and is described in terms of the static equivalent of the
dynamic and vibratory effects.

With this in mind, it seems we wish to achieve something like the following:
Extreme Total Load  Extreme Static Load  Dynamic Allowance
(3)

in which we use the word “total” to mean the simultaneous response due to static effects and the
dynamic interaction. For static load effect on its own, the notion of “extreme” is well defined: it is that
loading which is due to be exceeded with an acceptably small probability. Thus, writing the statistical
distribution of static loading as GS(x), for a probability of non-exceedance, p (0.999, for example), the
extreme (or characteristic) static load is given by:
 S , k  GS1  p 
(4)

Analogously, putting aside the difficulty of obtaining it, the extreme total load could be written:
 T , k  GT1  p 
(5)

From which our first-principles definition of a characteristic dynamic allowance can be obtained from (3)
as:
  T , k GT1  p 
k  
 S , k GS1  p 
(6)

Caprani (2005) puts forward this concept, calling it an “assessment dynamic ratio” (ADR) to make it
clear that it is more suited to the assessment of existing bridges than for the design of new bridges (see
Section 1.2). Interestingly, the ADR can be seen to vary with the return period as required (through p),
and there is no presumption on any of the factors seen earlier as affecting the dynamic allowance – they
are all incorporated into the distribution of total load effect, GT(x). However, this, of course, is the
problem: the estimation of a distribution of total load effect is clearly complex.

4.2 Determining the ADR

Initial Calibrated Numerical Approach

As part of SAMARIS (2006), the Mura River bridge in Slovenia was used to provide a sample application
of the statistical analysis for ADR. Monthly maximum load effects were identified from static simulations
of European traffic. These were then simulated using a validated 3D FE VBI model to determine the
total load effect (Gonzalez et al 2008). The population of total and static load effects were then
processed statistically, an illustration of which is in Figure 4 and is described by Caprani et al (2012)
and Caprani (2005). As can be seen, the expected level of lifetime dynamic interaction, for this site and
bridge, is a dynamic allowance of about 6% at the lifetime level.

9.0
9.0

ADR = 1.058
8.5

Static = Total
Static Stress (MPa)
Static Stress (MPa)

8.5
8.0
7.5

8.0
7.0
6.5

7.5
6.0

6.0 6.5 7.0 7.5 8.0 8.5 9.0 7.5 8.0 8.5 9.0

Total Stress (MPa) Total Stress (MPa)

(a) Parent and Extreme Populations (b) Lifetime Total and Static Load Effect
Figure 4. Multivariate extreme value extrapolation for lifetime DLA (1 in 100 years).
This first attempt to re-imagine the concept of dynamic allowance, and the startlingly low figure of 6%
that emerged, meets with much scepticism. We were informed that “this conclusion will lead engineers
in the wrong direction” and that “bridge testing suggests much higher DAFs”, and these comments are
typical. However, against the backdrop of the phenomenon described in Section 2, and the first
principles concept of a dynamic allowance, this result is entirely consistent. Indeed, an ADR can never
be ‘measured’ in a live load test, just as the 1000-year return period static loading event is not observed
in a field test.

Field Measurements
Further numerical work based on the ADR concept has continued to find similar results for many bridges
and traffic scenarios (OBrien et al 2010, Rattigan 2007). Recently the results from a field trial have been
published which use a bridge weigh in motion system to measure the total and static bridge strains
simultaneously (Carey et al 2017). Through extrapolation of both, the ADR (at p = 0.999) is found to be
6.2%. This figure implicitly includes the road surface roughness, vehicle weight, and multiple presence
events, amongst other relevant parameters. While it is site-specific, it closely matches similar results
from other sites (e.g. Figure 1).

4.3 Impact of the New Approach

The main impact of the revised approach and its consequently reduced dynamic allowances, is the more
accurate assessment of existing bridge structures. For marginal cases of either strengthening or
offsetting maintenance works for another budgetary cycle, such a refinement is invaluable. In such
assessments, the site-specific nature of all the key phenomenon affecting the dynamic allowance
(Section 2) can be incorporated into the ADR concept, as demonstrated in Carey et al (2017), for
example. It then just remains for bridge owners and consulting engineers to be given the scope to make
such analyses in bridge assessment codes, such as AS5100.7.
Although stated earlier, it is emphasized that for the design of new bridges, since it is relatively cheap
to add capacity during construction, there is only a marginal direct benefit from a reduced dynamic
allowance. However, with a reducing dynamic allowance, the currently accepted governing forms of
bridge traffic loading (see Section 3.2) must also be modified. The free-flowing traffic load effect for a
range of bridge lengths is reduced and congested traffic then becomes the critical loading case for more
bridge lengths than previously assumed. Further, with the possibility of different dynamic allowances for
different load effects (at the one bridge) and at different sites, a more complete conceptualization of
“governing traffic” is needed. To address this, Caprani (2013) introduces the Critical Traffic Index,
showing that the lifetime extreme loading, allowing for dynamics, can be free-flow, congested-flow or a
mixture of either, for different bridge lengths, load effects, car proportions, and DAF statistical models.
The idea of a mixture of free-flow (with dynamics) and congested flow governing is non-intuitive but
simply means that the load effect may arise from either source, with some probability (e.g. 50-50, 70-
30, etc.).

5 Summary
In the Victorian 1936 Country Roads Board specification for bridge design (A36), it is remarked about
dynamic interaction, with prescience, that:
The momentary strength of other than brittle substances is many times the strength
indicated by testing materials with slowly applied loads.
It seems that to take dynamic allowance forward we can seek inspiration from the past, and should not
be hesitant about putting into practice where the new knowledge is taking us. After all, the goal of bridge
engineering is to ensure bridges are safe, but with maximum economy and efficiency.

6 References
AustRoads (2003), Dynamic Interaction of Vehicles and Bridges, Publication No. AP-T23-03, Sydney.
Abdel-Rohman, M. and Al-Duaij, J. (1996). Dynamic response of hinged-hinged single span bridges
with uneven deck. Computers & Structures, 59(2):291-299.
Broquet, C., Bailey, S.F., Fafard, M. and Brühwiler, E. (2004). ‘Dynamic Behavior of Deck Slabs of
Concrete Road Bridges’. Journal of Bridge Engineering, 9(2), 137-146.
Brühwiler, E., and Herwig, A. (2008). “Consideration of dynamic traffic action effects on existing bridges
at ultimate limit state.” Proc. of IABMAS’08: 4th Int. Conf. on Bridge Maintenance, Safety and
Management, H. M. Koh and D. M. Frangopol, eds., Taylor & Francis, London, 3675–3682.
Cantero, D., González, A., and OBrien, E.J. (2009), ‘Maximum dynamic stress on bridges traversed by
moving loads, ICE Bridge Engineering, Vol. 162, No. 2, pp. 75-85.
Cantieni R (1983), Dynamic Load Tests on Highway Bridges in Switzerland - 60 Years Experience of
EMPA. Technical Report EMPA-Report Nr.211, EMPA.
Caprani, C.C. (2005), Probabilistic Analysis of Highway Bridge Traffic Loading, PhD Dissertation, Dept,
of Civil Engineering, University College Dublin, Ireland.
Caprani, C.C. (2013). Lifetime highway bridge traffic load effect from a combination of traffic states
allowing for dynamic amplification. ASCE Journal of Bridge Engineering, 18(9):901–909.
Caprani, C. C., González, A., Rattigan, P. H., and OBrien, E. J. (2012), “Assessment dynamic ratio for
traffic loading on highway bridges’, Structure & Infrastructure Engineering, 8(3), 295–304.
Caprani, C.C., González, A., Rattigan, P.H. and OBrien, E.J. (2012). Assessment dynamic ratio for
traffic loading on bridges. Structure and Infrastructure Engineering 8(3):295-304.
Carey, C., OBrien, E.J., Malekjafarian, A., Lydon, M., Taylor, S. (2017), ‘Direct field measurement of
the dynamic amplification in a bridge’, Mechanical Systems and Signal Processing, 85, pp. 601-609.
Chan, T.H.T. and O’Connor, C. (1990). “Vehicle model for highway bridge impact”, Journal of Structural
Engineering, ASCE, 116 (7), 1772–1793.
Chang, D. and Lee, H. (1994). ‘Impact factors for simple-span highway girder bridges’. Journal of
Structural Engineering, 120(3), 704-715.
Danish Roads Directorate (2004): “Reliability-Based Classification of the Load Carrying Capacity of
Existing Bridges. Guideline Document“, Report 291, Copenhagen.
Deng, L., Yu, Y., Zou, Q., and Cai, C.S. (2014), ‘State-of-the-Art Review of Dynamic Impact Factors of
Highway Bridges’, ASCE Journal of Bridge Engineering, doi: 10.1061/(ASCE)BE.1943-5592.0000672.
DIVINE (1997). Dynamic interaction of heavy vehicles with roads and bridges. Technical report, OCED,
Ottawa, Canada. DIVINE Concluding Conference.
Gonzalez, A., Cantero, D, OBrien, E.J. (2011), ‘Dynamic increment for shear force due to heavy
vehicles crossing a highway bridge’, Computers and Structures, 89(23–24), December, pp 2261–2272.
González, A., Rattigan, P., OBrien, E. J., and Caprani, C. (2008). “Determination of bridge lifetime
dynamic amplification factor using finite element analysis of critical loading scenarios.” Engineering
Structures, 30(9), 2330–2337.
Harris, N.K., OBrien, E.J., Gonzalez, A. (2007), ‘Reduction of bridge dynamic amplification through
adjustment of vehicle suspension damping’, Journal of Sound and Vibration, 302, pp. 471-485.
Huang D. (2008), ‘Dynamic loading of curved steel box girder bridges due to moving vehicles’, Structural
Engineering International, 4:365–72.
Huang, D. H., Wang, T. L., and Shahawy, M. (1993). Impact studies of multigirder concrete bridges.
ASCE Journal of Structural Engineering, 119(8):2387-2402.
Hussein, A. and Gonzalez, A. (2012), ‘Response of a Simply Supported Beam with a Strain Rate
Dependent Elasticity Modulus when Subjected to a Moving Load’, Bridge and Concrete Research in
Ireland, BCRI, 6-7 September, Eds. C. Caprani and A. O’Connor.
Hwang, E. S. and Nowak, A. S. (1991). Simulation of dynamic load for bridges. ASCE Journal of
Structural Engineering, 117(5):1413-1434.
Kim, S. and Nowak, A.S. (1997). ‘Load distribution and impact factors for I-girder bridges’. Journal of
Bridge Engineering, 2(3), 97-104.
Laman, J.A. Pechar, J. S. and Boothby, T. E. (1999). ‘Dynamic load allowance for through-truss
bridges’. Journal of Bridge Engineering, 4(4), 231-241.
Li., H., Wekezer, J. and Kwasniewski, L. (2008). ‘Dynamic response of a highway bridge subjected to
moving vehicles’. Journal of Bridge Engineering, 13(5), 439-448.
Li, Y. (2006), Factors Affecting the Dynamic Interaction of Bridges and Vehicle Loads, PhD Dissertation,
Dept. of Civil Engineering, University College Dublin.
Ludescher, H. and Brühwiler, E. (2009), ‘Dynamic Amplification of Traffic Loads on Road Bridges’,
Structural Engineering International, May, pp. 190-197.
McLean D., and Marsh M.L. (1998). Dynamic impact factors for bridges. National Cooperative Highway
Research Program (NCHRP) Synthesis 266. Transportation Research Board, National Research
Council (US).
OBrien, E.J., Cantero, D., Enright, B., and Gonzalez, A. (2010), ‘Characteristic Dynamic Increment for
Extreme Traffic Loading Events on Short and Medium Span Highway Bridges’, Engineering Structures,
32(12), pp.3827-3835.
Pape, T., Kotze, R. and Ngo, H. (2014), ‘Dynamic bridge-vehicle interactions’, 9th AustRoads Bridge
Conference, Sydney, available at: https://www.onlinepublications.austroads.com.au/items/ABC-
SAS301-14
Rattigan, P.H. (2007), The Assessment of Bridge Traffic Loading Allowing for Vehicle-Bridge Dynamic
Interaction, PhD Dissertation, Dept. of Civil Engineering, University College Dublin.
SAMARIS (2006), Guideline for the Optimal Assessment of Highway Structures, Sustainable and
Advanced MAterials for Road InfraStructure (SAMARIS), Deliverable SAM-GE-D30, EU 6th Framework
Report, 2006. http://samaris.zag.si/
Zhu, X. Q. and Law, S. S. (2002). Dynamic load on continuous multi-lane bridge deck from moving
vehicles. Journal of Sound and Vibration, 251(4):697-716.