You are on page 1of 13

th

9 Austroads Bridge Conference, Sydney, New South Wales 2014

NEW FATIGUE DESIGN LOADINGS FOR NEW


ZEALAND ROAD BRIDGES
Dr Michael Beamish, Beca Ltd, New Zealand

ABSTRACT
Fatigue design of new bridges in New Zealand must consider the future demands from
increasing heavy freight volumes and higher productivity vehicles. The new vehicles are
designed to carry more freight with fewer trips to offset the reduced productivity arising from
traffic congestion and to reduce overall fuel usage and carbon emissions. This has prompted a
review of bridge design standards, as was required in Australia in the late 1990’s and early
2000’s.

This research project for the NZ Transport Agency has investigated the current bridge fatigue
loadings on main highways, and potential increases arising from adoption of higher mass
vehicle configurations over the next decade or so.

For assessment of existing bridges, a vehicle spectrum representing the seven most common
vehicle types has been developed. The vehicle spectrum format, with 2 or 3 weight bands per
vehicle type, is suitable for use with any type of material fatigue endurance equation.

The project has found that a higher mass truck-and-trailer fatigue vehicle is more suitable for
design use in New Zealand than the M1600 vehicle specified in Australian standard AS5100.2,
and modifications to the AS5100.2 application rules are proposed to cover combined effects
from more than one lane.

INTRODUCTION
This paper presents key findings from a research project report [1] prepared for the NZ
Transport Agency. The intent of this research project was to provide the basis for an
amendment to fatigue loadings for road bridges in the New Zealand Bridge Manual [2] based on
New Zealand heavy vehicle characteristics, with appropriate allowances for long-term growth
forecasts and recommendations for methods of application.

Prior editions of the Bridge Manual deferred to BS 5400: Part 10 [3] for guidance on fatigue
design loads, and the 3rd edition [2] references interim guidance for steel bridges [4] based on
AS 5100.2 [5] and AS 5100.6 [6]. The interim guidance [4] drew on assessments of 2005 heavy
traffic data [7] prior to introduction of higher productivity motor vehicle (HPMV) standards in
2010, and adopted the AS5100.2 fatigue vehicle and axle models with approximately 10%
reduction in the ‘route factors’. Given that the M1600 fatigue vehicle model from AS5100.2 now
represents the ‘status quo’ for new bridge design [2], the agreed project scope included
provision for continuing with the AS5100.2 model if required by the Transport Agency, along with
development of an alternative new model. Also, because the Bridge Manual now references AS
5100.6 for steel bridge design, we defer to that standard for fatigue design methods.

In this paper we summarise the outcomes of the main steps of the research, as listed below:

• A survey of bridge fatigue design load models in selected international codes and methods
for determining equivalence to measured heavy vehicle data.
• Analysis of Weigh-in-Motion (WIM) data from New Zealand sites, to establish the fatigue
loading characteristics of representative vehicle datasets and equivalent vehicle spectra.
• Estimates of the increased fatigue loading per vehicle with take-up of HPMV higher mass
limits.
• Selection and calibration of a new standard fatigue vehicle to suit the heavy vehicle mix on
New Zealand roads.

© ARRB Group Ltd and Authors 2014 1


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

• Selection of allowances for long term heavy traffic volume and vehicle mass growth.
• Fatigue loading implementation, including application rules.

The status of the research project is that the Transport Agency is considering the outcomes and
recommendations, and further work to prepare case studies and evaluate the cost implications
is likely to be requested prior to preparation of a draft amendment to the Bridge Manual.

BRIDGE FATIGUE DESIGN LOAD MODELS


Vehicle fatigue load models generally comprise either a single vehicle, a selection of common
vehicle types, or site specific vehicle records; with the repetition counts over the design fatigue
life. International practices (Eurocodes [8], UK annex [9]) are to provide both a design vehicle
spectrum made up of common heavy vehicle configurations, in addition to a single standard
vehicle method suitable for steel bridges. The Australian [5], US [10] and Canadian [11] codes
all adopt the vehicle specified for the design live loads, with varying scale factors. The Australian
code is unique in its use of a very high vehicle loading weighing considerably more than
common vehicles (the M1600 vehicle) but with lower variable cycle counts.

Single fatigue vehicle versus vehicle spectrum models


This research project investigated both vehicle spectrum methods, and single vehicle options.

Single vehicle methods in the current international codes all use the maximum stress range for
one fatigue vehicle passage (ie a single cycle), and apply an adjustment factor varying with
span length, component type, and other parameters. Rainflow counting is not required because
additional cycles per vehicle in shorter spans are included for in the adjustment factor(s). The
single vehicle methods are normally applicable to the fatigue life formulae for steel and
composite steel-concrete bridges only, and the development of a single design fatigue vehicle
model for New Zealand in the research project assumed that it would only be applied to such
bridges.

A comprehensive vehicle spectrum fatigue load model (such as in BS5400 Part 10 [3] or the UK
Eurocode annex [9]) substitutes a set of representative vehicles with multiple weight levels and
corresponding repetition counts for the entire heavy vehicle population, allowing general
application to any form of material fatigue life formulae and to assessment of complex
structures.

Single vehicle models are convenient for design of typical steel bridges but lack the generality
that may be required for detailed assessments, complex structures (which may fall outside the
scope of the application rules and adjustment factors provided in the codes), and materials
other than structural steel.

Calculation of equivalent fatigue vehicle loadings


The research used raw heavy vehicle data from Weigh-in-Motion stations (axle weights and
spacings) to estimate the equivalent repetitions of single fatigue vehicles. The basic vehicle data
processing methodology was similar to work by Grundy and Boully [12] for development of the
AS 5100.2 fatigue loads, thereby providing the means to compare Australian results for the
Hume Freeway traffic to the New Zealand results. The Grundy and Boully paper showed the
effect of span length on the equivalent fatigue damage per truck in terms of average cycles of
the unfactored M1600 vehicle effect (moments in simply supported spans).

The evaluation of the vehicle fleet to estimate equivalent repetitions of any design fatigue
vehicle can be expressed as:

fatigue damage average fatigue single cycle equivalent


heavy vehicle
� � for each heavy � = � damage per � × � � = � fatigue damage � × � number �
traffic volume of cycles
vehicle in fleet heavy vehicle for design vehicle

© ARRB Group Ltd and Authors 2014 2


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

For structural steel components in road bridges the appropriate fatigue damage measures are
th
cumulative summations of the 5 powers of the “stress” cycle ranges (represented by bending
moment, shear force and reaction range values). This is consistent with the Eurocodes [13],
th
where 5 power weighted averages of vehicle gross weights are used in vehicle mix
th
adjustments and nationally determined parameters. The reason for reliance on 5 powers is
seen in the form of S-N curves for steel [6,14] illustrated in figure 1. It is expected that numbers
6
of axle sets will exceed 5x10 over the design life of most bridges, as will the numbers of trucks
on most highways. Thus the stresses for all normal legal vehicles should fall below the constant
amplitude fatigue limit and be in the region of the S-N curve where the exponent m=5 is
applicable.

500
400
300

200

1
Stress Range S (MPa)

m=3
100
80
80
Constant Amplitude Fatigue Limit at N = 5x106
60
50
40 1
m=5
30 Cut-off at N = 1x108

20

10
1.E+05 1.E+06 2 5 1.E+07 1.E+08 1.E+09
Endurance, number of cycles N

Figure 1: Design S-N curve showing fatigue strength for detail category 80

NEW ZEALAND HEAVY VEHICLE CHARACTERISTICS


Weigh-In-Motion and vehicle classification facilities
The Transport Agency provided data for the study from 6 WIM sites on the state highway
network, including the Auckland Harbour Bridge (AHB), plus heavy vehicle counts from a
network of vehicle counter telemetry stations. The AHB is the only site with data for an urban
area, while the others represent high and medium volume inter-city routes or a bridge site
dominated by logging trucks. Short duration samples of classified vehicle data were available
from telemetry sites on state highways in rural areas, from which the variations in truck mixes on
other highway categories were assessed. Table 1 lists locations of the WIM sites considered
and 2011 heavy vehicle volumes. The three rural freight route sites were found to be sufficiently
similar in terms of fatigue effects per vehicle to adopt one (Te Puke) to represent all similar sites
leaving four with quite different heavy traffic mixes to evaluate.

Table 1: WIM stations in New Zealand and daily truck counts for data samples

Location Highway, area Lanes Heavies/Day Comments


Name /Direction
Drury SH1N, South of Auckland 4 2150 Motorway, freight route
Eskdale SH5, Napier 2 290 Freight, logging route
Te Puke SH2, East of Tauranga 2 970 Freight route, rural
Tokoroa SH1N, South Waikato 2 800 Freight route, rural
Waipara SH1S, North Canterbury 2 670 Freight route, rural
AHB SH1N, Auckland 5x2 3200 Urban motorway

© ARRB Group Ltd and Authors 2014 3


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Truck Weight Distributions


The truck weight distributions for selected dated sets from the WIM sites were examined by
vehicle type and direction to determine which should be used to represent typical New Zealand
freight routes. All sites show directional bias in the proportions of fully laden vehicles but sites
with high logging content are extreme, with empty trucks piggy-backing the trailers in one
direction (figure 2 shows the highest peak for fully laden vehicles above the 44 tonne legal limit
eastbound, and a peak in 4-axle trucks at 17 tonne westbound). Examples of the weight
distributions for both directions combined are presented in the NZ Transport Agency’s annual
reports [15]. The Eskdale site data was used to derive estimates of the highest fatigue loading
per vehicle per lane for New Zealand highways. The Drury site (figure 3) had the highest heavy
vehicle count per lane, but carries a higher proportion of volume constrained freight vehicles
and rigid trucks.

Heavy Vehicle Mass Distribution - SH5 Eskdale


12%

10%
% Directional Count

8%

6%
Eastbound
4% Westbound

2%

0%
4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54
GVM bin (tonnes)

Figure 2: Vehicle weight distribution with logging content

Heavy Vehicle Mass Distribution - SH1 Drury outside lanes


8%
7%
6%
% Lane Count

5%
4%
Southbound
3%
Northbound
2%
1%
0%
4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54
GVM bin (tonnes)

Figure 3: Vehicle weight distribution, mixed freight

Vehicle spectra before HPMV introduction


Vehicle spectra representing the most common vehicle types were developed to provide the
starting point for estimating the additional effects of higher mass vehicles and for estimating
historic loading on existing structures prior to the mass limit increases. The approach adopted
was to:

• Provide a reasonably comprehensive spectrum in the style of the UK Annex to Eurocode 1


[9], covering the most common vehicle types with fixed, average dimensions.
• Link the heavy vehicle type counts to the Transport Agency’s classification scheme.

© ARRB Group Ltd and Authors 2014 4


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

• Allow for significantly different route types through variations in the vehicle type counts, with
fixed axle weights representing each weight band.
• Include directional bias by adopting data for the more heavily loaded directions.
• Tune the spectra to fit the fatigue damage characteristics at all span lengths while
maintaining an approximate fit to the total weights, with minimal conservatism.
• Exclude infrequent heavy overload vehicles with non-standard axle configurations.
• Exclude dynamic impact effects other than the inherent scatter and road roughness effect.
• Rationalise the results, through rounding of proportions per weight band, and presenting in a
style that allows for changing the truck class proportions to estimate spectra for other sites.

The truck weight spectra in table 2 below represent heavy vehicle traffic prior to significant take-
up of higher mass permits, so the spectra are not appropriate for design of new structures, but
would be of assistance for fatigue assessment of existing structures (to estimate fatigue life
consumed to date), and their applicability is not limited to steel structures. It is anticipated that a
future revision will include a few additional spectrum vehicles to represent the common higher
mass configurations.

Table 2: Truck weight spectra at WIM sites (for 2010-11 before approval of HPMV)

AHB Drury Te Puke Eskdale


(high volume (motorway, (general (high bias,
Standard Vehicle Details motorway) expressway) freight) logging)
Vehicle
NZTA Load Vehicle Group Vehicle Group Vehicle Group Vehicle Group
Set Configuration Weight
Classes Group Mix Split Mix Split Mix Split Mix Split
(kN)
L 50 70% 60% 60% 60%
Rigid Truck
1 4 M 90 50.0% 20% 27.2% 35% 28.9% 35% 21.1% 35%
o--o
H 150 10% 5% 5% 5%
L 110 65% 40% 40% 40%
Rigid Truck
2 5 M 170 25.0% 15% 10.3% 40% 8.9% 40% 6.0% 40%
o--oo
H 210 20% 20% 20% 20%
Rigid Truck L 160 75% 75% 75% 75%
3 6+7 5.5% 6.1% 6.2% 4.3%
oo--oo H 260 25% 25% 25% 25%
L 200 40% 22% 22% 22%
Artics
4 8+9 M 310 9.0% 34% 20.0% 48% 13.1% 48% 11.4% 48%
o-oo--ooo
H 420 26% 30% 30% 30%
Truck+Trailer L 180 25% 25% 25% 25%
5 10+11 4.0% 8.1% 9.5% 7.4%
o-oo--oo--oo H 450 75% 75% 75% 75%
L 210 20% 25% 15% 5%
Truck+Trailer
6 12 M 350 5.5% 30% 19.9% 35% 27.2% 15% 42.5% 15%
oo--oo-oo--oo
H 450 50% 40% 70% 80%
L 220 15% 15% 15% 15%
B-Train
7 13 M 360 1.0% 30% 8.4% 30% 6.2% 30% 7.3% 30%
o-oo--ooo--oo
H 450 55% 55% 55% 55%
Average weight (5th power weighted) 280kN 348kN 366kN 391kN

The presentation in the format above rather than counts for each weight band is intended to
show the distribution of loading within each vehicle class grouping. For example, 70% of the
class 12 truck-and-trailers at the Te Puke site are fully laden in one direction, while the Drury
site has 40% fully laden and 35% at medium weight (and possibly volume constrained).

Short term vehicle mass growth – HPMV effect


At the time the WIM data considered in this study was collected, higher mass HPMV introduced
in 2010 regulations had not been approved for access to any of the highways with WIM sites,

© ARRB Group Ltd and Authors 2014 5


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

and take-up of the new limits by new and existing vehicles is currently in progress. Because it
was essential to allow for HPMV in the new design fatigue loading, a methodology was
developed to estimate the increase in fatigue damage per heavy vehicle, using information
provided in the business case reports [16, 17] and their supporting documents.

The studies on bridge live loading effects following HPMV introduction included details of likely
vehicle configurations [18] at the new gross mass limits in the 2010 regulations, including
increased gross and axle masses for existing vehicle lengths, and new vehicle types with
additional axles. For the new longer vehicle configurations we considered the likely maximum
gross masses based on axle set limits, with axle set spacings increased to the shortest lengths
complying with the new mass limits. This provided two potential upgrade options for each
vehicle class – existing length, or longer length at maximum masses.

Having distilled the WIM site vehicle data into standardised truck weight spectra, the
subsequent steps to estimate the effect of higher mass vehicles were simplified in two respects:

• The numbers of vehicles deemed to be candidates for eventual take-up of the higher mass
limits can be taken from the high mass band proportions in table 2 (4H, 5H, 6H, 7H). The
calculations assumed that all vehicles in these bands would be upgraded (eg 40% of class
12 at the Drury site or 70% at the Te Puke site would convert to HPMV limits eventually).
• The vehicle spectra weights and counts allow for effects of mass variance and overweight
incidence (and dynamic weighing scatter) relative to the existing limits, so it was assumed
that scaling the vehicle spectra fatigue effects would preserve the underlying spread in the
mass distribution, avoiding the need to estimate new truck weight distributions.
The key assumptions for the estimates of increased fatigue damage were:

• Scaling of average fatigue effects per vehicle uses the ratios of effects of maximum legal
mass vehicles, ie new HPMV limit versus Class 1 limit (44 tonne maximum gross mass). This
may over-estimate the increase if mass limit compliance is improved.
• Fewer vehicles are required to move the same freight volume. Thus a lower increase in
damage per year arises from reducing fully laden vehicle counts in proportion to the payload
increase.
• The higher estimates of average damage increase per vehicle obtained by not reducing
counts as noted above are most relevant to the future fatigue design loading, because future
annual vehicle counts will eventually be based on full take up of the new vehicle standards.

Class 1 vehicles (44 tonne) versus Longer Higher Mass Vehicles


1.4
Equivalent Moment Ratio

1.3
R12T22->52t
R22T22->55t
1.2 B1232->53t
B1232->B1233 52t
B1232->B1233 57t

1.1 B1233->57t

1.0
0 10 20 30 40 50 60
Span Length (m)

Figure 4: Equivalent moment increase per vehicle for transition to full size HPMV

A range of vehicle upgrade scenarios was applied to quantify the increases in damage using
existing vehicle configurations, or longer and heavier rigs with additional axles. Figure 4 shows
th
the ratio of damage equivalent moments (from summation of 5 powers of moment ranges) for
upgrading 44 tonne vehicles to the new maximum mass truck-and-trailers or B-Trains. For long

© ARRB Group Ltd and Authors 2014 6


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

spans the effects are proportional to gross mass ratio, while for shorter spans there are higher
increases because axle set weights for the 44 tonne vehicles are constrained by the gross mass
limit and longer axle set spacings can increase moment ranges (through smaller trough values).

For the most common freight vehicle (R22T22) the moment increase is up to 21% for 10m-30m
th
spans, giving up to 150% increase in fatigue damage using a 5 power rule. These potentially
large increases for the top weight band vehicles are moderated by aggregation over all vehicles
in the spectra, giving 136% increase for the Eskdale spectrum and or 98% for the Drury
spectrum. These increases would occur over several years, and are conservative because not
all vehicle upgrades would be to the maximum size configuration, and reductions in vehicle
movements for a fixed freight task are anticipated.

SINGLE FATIGUE VEHICLE MODEL OPTIONS


It was apparent from the dominance of truck-and-trailer vehicles in the New Zealand fleet
compared to semi-trailers (in the North American and European fleets) or semi-trailers plus B-
doubles (in the Australian) fleet, that it was appropriate to consider a new fatigue design vehicle
for New Zealand. The effects of mass limit increases noted above indicated that this vehicle
should be capable of representing the future vehicle fleet including routes with high proportions
of HPMV, and therefore the proposed new vehicle represents a higher mass truck-and-trailer.

Table 3 below compares a new vehicle option to the M1600 vehicle in AS5100.2 or a modified
version. Vehicles from other codes [10, 11] were also considered but those were found to be
unsuitable for medium to long spans. The Eurocode fatigue load models with national
parameters to suit New Zealand were briefly considered due to advantages in adopting the
Eurocode simplified method but the different vehicle axle configurations, higher weights, and
lack of provision for long term vehicle mass growth ruled it out of contention.

Table 3: Comparison of M1600 options with a truck-and-trailer fatigue vehicle

Vehicle Advantages Disadvantages


M1600 Same as AS5100.2. Unrealistically high for future NZ
as per Opportunity to apply with minimal loading (Taplin et al [19] propose
AS5100.2 changes to AS5100.2 . 0.80SM1600 for future LL standard).
Simple to apply if NZ adopts AS5100.2 Design cycle counts at medium – long
with minor mods. spans do not fit 5th power evaluations
(overly conservative).
Simple relationship between design
cycles, daily truck counts, and span A single axle load is necessary to cover
length. short spans (<5m), but load and cycle
counts appear to be excessive.
Fatigue vehicle effects may be derived
from live load analysis in simple cases Axle set effects are much higher than
only (separate fatigue analyses are current (and future) triple or quad axle
usually necessary). sets – short span effects are not
representative of typical trucks.
Reduced Opportunity to modify existing Alternative axle load (or tandem set) is
M1600 AS5100.2 format of cycle count necessary to cover short spans.
(eg 60%) formulae to fit NZ spectra and likely All M1600 variants (including the 1
st

growth parameters. Could specify with option above) may over-estimate


less onerous but more realistic method hogging moment effects over supports
of application. for continuous span lengths in the
Stress ranges are closer to envelope of range where the variable spacing
expected higher mass vehicles. places 2 tri-axle sets near the middle of
Similar axle weights to current vehicles. adjacent spans.

HPMV Representative of future NZ higher Not directly compatible with SM1600


vehicles: mass vehicles (medium term). based LL standard.
R22T22 Rigid trucks on short spans are well Requires new rules for application (but
54 tonne represented by part of the vehicle. this is also an opportunity to address
or concerns with the AS5100.2 rules
Stress ranges similar to higher mass
R23T23 relating to multiple lanes, multiple
vehicles. Axle and axle set weights

© ARRB Group Ltd and Authors 2014 7


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Vehicle Advantages Disadvantages


57 tonne similar to current vehicles. presence).
Separate analysis with a single axle Separate fatigue vehicle analysis
load is not necessary. always required.
Well suited to short span deck New vehicle not in existing codes or
components (eg ladder decks). existing analysis software.

Fatigue vehicle selection


There was a trade-off required between a perceived desire for close alignment with AS5100.2
and a standard vehicle model that aligns with the most common large vehicles in New Zealand
(truck-and-trailers) and appropriately represents loading on short, medium and long span
lengths; and transverse girders. It should be noted that analysis cases for fatigue design are
normally separate from live load cases for strength design and exclude the uniformly distributed
portion, so there is no significant benefit from choosing a vehicle included in the design live load
model and no disadvantage from selecting a new vehicle, if that is technically appropriate.

Proposed truck & trailer fatigue vehicle


A R22T22 truck-and-trailer with 530kN total weight was found to be the preferred standard
fatigue vehicle model from a technical standpoint (considering points noted in table 3), as it is
more representative of New Zealand vehicles and provides a more consistent fit to fatigue
effects over a wide range of span lengths compared to the other options considered. Minor
variations can also be considered but the proposed vehicle shown in figure 5 below is a
maximum mass variant of vehicle 6H in table 2 with axle spacings based on averages for
current R22T22 vehicles, which was a convenient starting choice.

50kN 50kN 75kN 75kN 70kN 70kN 70kN 70kN

1.8m 3.3m 1.3m 4.2m 1.25m 4.3m 1.25m


17.4m

0.22m
0.20m 0.5m
2.0m

2.3m

1.8m

(Typ.) (Typ.)
0.22m

Figure 5: Proposed 530kN fatigue vehicle (twin steer truck & trailer, higher mass HPMV)

Proposed fatigue vehicle calibration


The method used for fitting equivalent cycle counts per heavy vehicle to the New Zealand
vehicle loading data follows a similar methodology to Grundy and Boully (2004) for calculation of
equivalent cycles of the new fatigue vehicle. This relies on damage equivalent moment, shear
th
force and reaction calculations (5 power) for simply supported spans and assumes a similar
implementation to the AS5100.2 fatigue loading, but with new cycle count formulae.

© ARRB Group Ltd and Authors 2014 8


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Equivalent cycles per heavy vehicle for future HPMV vehicle mixes (5th power)
10

Damage x 530kN R22T22 Standard Truck SH5 Eskdale


Proposed fit to
SH1 Drury
HPMV full take-up SH2 Te Puke
SH1 AHB
Proposed Fit
Eskdale 2011
1

x 1.0
x 0.8
x 0.6
x 0.4
x 0.3
x 0.2
0.1
1 10 100
Simply Supported Span Length (m)

Figure 6: Fatigue vehicle calibration versus New Zealand loading with HPMV scaling

The outcome is shown in figure 6 above, with results for the four spectrum sets scaled up to
represent the maximum estimates for HPMV take up, and the result for the current Eskdale
spectrum in table 2 (no HPMV). The dashed lines indicate possible route factor values, following
the AS5100.2 approach of scaling the cycle counts to suit other route categories.

The base (x 1.0) equivalent cycle counts for this option (530kN truck-and-trailer) are based on 2
cycles per truck for span L ≤ 5m, reducing to 0.6 cycles per truck for span L > 16.67m. This
means that for 10m span, the average heavy vehicle effect is equivalent to one cycle of the
maximum stress range for a passage of the fatigue vehicle.

At short spans less than 5m, the equivalent count per truck increases to 2 cycles (of the 2x75kN
axle set), and the flat line fit suggests a suitable alternative to the A160 axle fatigue load model
in AS5100.2 using a tandem axle set to provide more consistent coverage at the shortest spans.
It also indicates that a separate axle set model may be unnecessary where the proposed
vehicle is used, as the envelope of the vehicle effects covers the tandem axle set.

The results shown in figure 6 may be re-calculated for any chosen fatigue vehicle. Figure 7
compares the Eskdale results for M1600 damage equivalence (with full take-up of HPMV limits)
and a suggested alternative fit to the AS 5100.2 design cycle count assumption with the results
for the Hume Freeway (Grundy and Boully [8]) and the fit adopted for AS5100.2. This shows
that the AS 5100.2 equation is conservative at span lengths over 6m and unsuitable for short
spans (where the A160 axle loading governs). It is also seen that the 2010/11 results for the
Eskdale site are comparable to the 2001 Hume Freeway results at long spans.

© ARRB Group Ltd and Authors 2014 9


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Equivalent cycles per heavy vehicle (5th power)


1
Alternative fit to HPMV full take-up,
Damage x M1600 Vehicle
used with reduced M1600 options

AS5100.2 design
0.1 0.125L-0.5

0.01 SH5 Eskdale (HPMV)


Alternative M1600 fit
AS5100.2 design
SH5 Eskdale 2010/11
Hume freeway 2001 [12]
0.001
1 10 100
Simply Supported Span Length (m)

Figure 7: Comparison with M1600 vehicle calibration

LONG TERM FATIGUE DAMAGE GROWTH ALLOWANCE


The long term growth allowance requirements are driven by government policy to allow higher
mass vehicles with heavier axles, and the anticipated future mass increases discussed in the
future live load study report [19], plus general growth in freight task.

The cycle counts specified in AS5100.2 clause 6.9 allow for:

• 75 year fatigue design period (as for the North American codes [10,11], considering that
fatigue life estimates normally represent the onset of cracking and mean lives are
considerably longer). This may be compared with 100 years minimum design life.
• Approximately 4% compound growth rate in equivalent cycle counts.
• Total cycle counts of approximately 440 times the year 1 count (average per year of 5.8
times the presumed year 1 fatigue damage), which provides for a combination of volume
growth and increasing mass per vehicle.

For the NZ vehicle growth, a range of combinations of volume growth and long term vehicle
mass scenarios was explored and it was observed that:

• The base fatigue vehicle loading assumes full take-up of the HPMV higher mass limits.
• An on-going compound growth rate for damage is necessary to adequately represent a
combination of linear mass growth and linear volume growth.
• A scenario with 0.7% pa vehicle mass linear growth (capped at 30% increase beyond a 10
year initial HPMV take-up period) with 3% pa linear volume growth fits the AS5100.2
multiplier for fatigue damage cycles (440 x base year). A cap at 4000 trucks/lane/day was
also adopted.
• Uncapped mass limit growth (at 0.7% pa) would result in much higher damage at 75+ years
(700 x base year at 75 years) but it is expected that caps on axle mass growth would be
necessary to manage pavement impacts.
• If the most optimistic HPMV take-up scenarios are to be covered, a higher growth rate may
be necessary.
• A 100 year fatigue design life with ongoing growth may require 60%-75% more design cycles
(roughly 10%-12% reduction in stress range for the fatigue vehicle) compared to the 75 year
period used in AS5100.2.

© ARRB Group Ltd and Authors 2014 10


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

• The fatigue strength reduction factor φ f specified in the steel design code can be used to
address reliability aspects such as ease of inspection and repair, and a reduction from say
1.0 to 0.85 would have greater effect than an increase to 100 year fatigue design life.

The rate and extent of future growth in fatigue damage rate is uncertain. The growth multipliers
incorporated in the AS5100.2 cycle count formulae (4% per annum compound growth for 75
years – a total of 440 times the first year loading estimate) are considered to be the minimum
that should be considered in a New Zealand fatigue loading model.

FATIGUE LOADING IMPLEMENTATION


The proposed application methods for the truck & trailer fatigue vehicle were adapted from
AS5100.2 but reviews of UK and Eurocode practice indicated that improvements were
necessary to include allowances for multiple lane effects and additional impact near expansion
joints. The proposed changes include:

• Cycle counts for 530kN R22T22 vehicle tuned to New Zealand loadings (figure 6).
• Fatigue vehicles centred in design lanes.
• No fixed stress reduction factor (0.7 in AS5100.2) or additional dynamic allowance except for
a recommendation to apply the Eurocode allowance of 30% at sections within 6m of
expansion joints.
• Combined effects from adjacent lanes (or opposing direction) for significantly affected
components (eg transverse girders, box structures) using a Miner’s summation.
• Side-by-side running (multiple presence) or combined additive stress ranges for opposing
lanes (eg transverse bracing with stresses of opposite sign from 2 lanes) addressed using a
modification factor from the UK national annex to Eurocode 1.
• Additions to the 65% share of trucks per direction in any one lane for urban routes – 80% in
one lane where only 2 lanes are available, 45% in an adjacent lane in the same direction
where effects from 2 lanes are combined.

Multiple presence effects for vehicles in the same lane on long spans are considered to be
adequately covered by the form of the design cycle count equations with conservative fits to the
tails beyond 60m allowing for the maximum size HPMV vehicles.

The complete form of the proposed cycle count formula for the maximum stress range due to a
passage of the 530kN R22T22 vehicle is:

2.0 for L ≤ 5m, number of heavy vehicles


10 4
� for 5m < L <16.67m, or� × 16×10 × � per lane per day for � × (route factor)
L first year of service
0.6 for L ≥ 16.67m

where definitions of effective span L and heavies/day are as given in AS5100.2.

Route factors for New Zealand roads are covered the research report, with 1.0 being relevant to
exceptional routes and 0.8 being relevant to general high volume freight routes.

Discussion
AS 5100.2 was the only code found to include a fixed reduction factor to allow for differences
between calculated and real stress ranges, and although there is ample evidence to confirm
small reductions, a 0.7 factor (30% stress reduction) is considered to be too low. However
AS 5100.2 also the only code to use the dynamic load allowance from design live loads in the
fatigue loading, whereas other codes apply a reduced allowance compared to design live loads
or allow for dynamic effects in the base fatigue load models.

On balance, it was considered that there is insufficient evidence to support fixed allowances in
fatigue loadings tailored to New Zealand bridges and derived from recorded vehicle weights
including dynamic scatter. The recommended modifications to AS5100.2 application rules

© ARRB Group Ltd and Authors 2014 11


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

exclude both those factors and the net effect is a small increase in loadings. An additional 30%
dynamic amplification allowance was retained within 6m of expansion joints. A small decrease in
loadings may arise from the vehicle centred in lanes versus placement at the worst case offset
position specified in AS5100.2. The combined effect of these changes is minor, and there are
opportunities for designers to determine appropriate stress reductions applicable to service load
levels.

The additional application rules to allow for combined lane effects address a potential deficiency
in the unmodified AS5100.2 procedures. It is suggested that additional investigation be carried
out by the AS5100.2 committee to confirm suitable additions to allow for combined effects from
multiple lanes on susceptible components such as cross beams, transoms and transverse
bracing.

SUMMARY
This research project found that a new fatigue vehicle based on the most common freight
vehicle on New Zealand main highways (8 axle truck-and-trailer) is preferable to the M1600
vehicle from AS5100.2 (as specified in the current Bridge Manual edition [2]). The preferred
vehicle weighing 530kN (figure 5) was more representative of New Zealand vehicles and
provided a more consistent fit to fatigue effects of vehicle fleets at the WIM sites for a wide
range of girder span lengths. Through inclusion of tandem axle sets, consistent fits to effects on
spans less than 5m were achieved without requiring an additional since axle fatigue load model.

Given that a separate study [19] has recommended adoption of 0.8SM1600 live loading from
AS5100.2 for bridge design, it was considered appropriate to consider an option to continue
using the M1600 model. The existing AS5100.2 fatigue loading can be adjusted for use in New
Zealand with minor amendments such as appropriate route factors, but a reduced weight
M1600 vehicle with modified application rules would be preferable to adequately cover
combined effects from more than one lane and reduce the vehicle effects to more realistic
service load levels. All M1600 options were considered to have significant disadvantages as
they are a poor fit to effects on short spans and may overestimate effects on continuous spans.

The recommended application rules for the proposed fatigue vehicle were similar to AS5100.2
section 6.9, with additional application rules intended to address effects on components
significantly affected by loadings in two adjacent lanes or opposing directions.

The proposed cycle counts are tuned to New Zealand vehicle loadings and full take-up of the
HPMV higher mass limits on main highways. The growth multipliers incorporated in the
AS5100.2 cycle count formulae (a total of 440 times the first year loading estimate) were
considered to be the minimum that should be considered in a New Zealand fatigue loading
model.

Long term growth allowance requirements are driven by government policies allowing higher
mass vehicles with heavier axles and anticipated future vehicle mass increases identified in the
parallel Transport Agency research project on future bridge live load design standards [19]. This
research confirmed the applicability of the new fatigue design criteria for future bridges in New
Zealand.

ACKNOWLEDGEMENTS
The financial support of the NZ Transport Agency for this research project is gratefully
acknowledged.

REFERENCES
1. Beamish, M, Fatigue design criteria for road bridges in New Zealand, NZ Transport Agency
research report 547, June 2014. www.nzta.govt.nz/resources/research/reports/547
2. NZ Transport Agency, Bridge manual SP/M/022. 3rd edition 2013.
3. British Standards Institution, BS 5400 Steel, concrete and composite bridges - Part 10:
Code of practice for fatigue, 1980.

© ARRB Group Ltd and Authors 2014 12


th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

4. Clifton, GC, Recommended draft fatigue design criteria for bridges, version 3. New Zealand
Heavy Engineering Research Association, 2007.
5. Standards Australia, AS 5100.2-2004, Bridge Design, Part 2: Design Loads
6. Standards Australia, AS 5100.6-2004, Bridge Design, Part 6: Steel and Composite
Construction
7. Beamish, M, P Tindall, and I Billings, Auckland Harbour Bridge fatigue assessments,
Austroads 6th Bridge Conference, Perth 2006. Presentation to conference.
8. BS EN 1991-2:2003 Eurocode 1: Actions on structures - Part 2: Traffic loads on bridges
9. UK National Annex to Eurocode 1: Actions on structures, Part 2: Traffic loads on bridges,
2008
10. American Association of State Highway and Transportation Officials, AASHTO LRFD Bridge
Design Specifications, 5th edition, 2010.
11. Canadian Standards Association, CAN/CSA-S6-06, Canadian Highway Bridge Design
Code.
12. Grundy, P and G Boully, Fatigue Design in the New Australian Bridge Design Code.
Austroads 5th Bridge Conference, Hobart, Australia, 2004.
13. BS EN 1993-2:2006 Eurocode 3 - Design of steel structures - Part 2: Steel bridges
14. BS EN 1993-1-9:2005 Eurocode 3: Design of steel structures - Part 1-9: Fatigue
15. NZ Transport Agency, Annual Weigh-In-Motion (WiM) report 2012.
16. NZ Transport Agency, Vehicle dimensions and mass rule amendment 2010, funding and
investment guidelines, 2010.
17. Stimpson, D, Business case for lower bound high productivity motor vehicles, 2012.
18. Waldin, J, Assessment of HPMV load limits for bridges, an alternative approach – discussion
paper, Opus International Consultants Ltd report, 2011.
19. Taplin, G, M Deery, T van Geldermalsen, J Gilbert and R Grace, A new vehicle loading
standard for road bridges in New Zealand, NZ Transport Agency research report 539, 2013.

AUTHOR BIOGRAPHY
Mike Beamish is a senior associate in the bridge engineering team at Beca, with over 25 years
experience in bridge and structural engineering. Notable projects have included superstructure
design for the Otira Viaduct in Arthur’s Pass, seismic assessment and retrofit of the Auckland
Harbour Bridge, and design of the Sky City casino and hotel. For the past 15 years Mike has
been a structural engineer for the Auckland Harbour Bridge maintenance contracts with
particular focus on live load and fatigue assessments, repairs, and strengthening of the truss
and box girder structures.

Copyright Licence Agreement

The Author allows ARRB Group Ltd to publish the work/s submitted for the 9th Austroads Bridge
Conference, granting ARRB the non-exclusive right to:

• publish the work in printed format


• publish the work in electronic format
• publish the work online.

The Author retains the right to use their work, illustrations (line art, photographs, figures, plates) and
research data in their own future works

The Author warrants that they are entitled to deal with the Intellectual Property Rights in the works
submitted, including clearing all third party intellectual property rights and obtaining formal permission from
their respective institutions or employers before submission, where necessary.

© ARRB Group Ltd and Authors 2014 13

You might also like