Road Roughness Measurements

7/24/2019 Road Roughness Measurements

1/98


2/98

WORLD BANK TECHNICAL

PAPERS

No. 1.

Increasing Agricultural

Productivity

No. 2. A Model for

the Development of a Self-Help Water Supply Program

No. 3. Ventilated Improved Pit Latrines: Recent Developments in Zimbabwe

No. 4. The African Trypanosomiases: Methods and Concepts of Control and Eradication

in Relation to Development

(No. 5.) Structural Changes in World Industry: A Quantitative Analysis of Recent Development

No. 6. Laboratory Evaluation of Hand-Operated Water Pumps for Use in Developing Countries

No. 7. Notes on the Design and Operation

of Waste Stabilization Ponds in Warm Climates

of Developing Countries

No. 8. Institution Building for Traffic Management

(No. 9.) Meeting the

Needs of the Poor for Water Supply and Waste Disposal

No. 10. Appraising Poultry Enterprises

for Profitability: A Manual for Investors

No. 11. Opportunities for Biological Control of Agricultural

Pests in Developing Countries

No. 12. Water Supply and Sanitation Project

Preparation Handbook: Guidelines

No. 13. Water Supply and

Sanitation Project Preparation Handbook: Case Studies

No. 14. Water Supply and Sanitation Project Preparation Handbook: Case Study

(No. 15.)Sheep and Goats in Developing

Countries: Their Present and Potential

Role

(No. 16.)Managing

Elephant Depredation in Agricultural and

Forestry Projects

(No.

17.)Energy Efficiency

and Fuel Substitution in the Cement Industry with

Emphasis

on Developing Countries

No. 18. Urban Sanitation Planning

Manual Based

on the Jakarta Case

Study

No. 19. Laboratory Testing of Handpumps

for Developing

Countries: Final

Technical Report

No. 20. Water

Quality in Hydroelectric

Projects:

Considerations for Planning in Tropical

Forest Regions

No. 21.

Industrial Restructuring:

Issues and Experiences in

Selected Developed

Economies

No. 22. Energy Efficiency in the Steel Industry with Emphasis on Developing Countries

No. 23. The Twinning

of Institutions:

Its Use as a Technical

Assistance Delivery System

No. 24. World Sulphur Survey

( ) Indicates number assigned after publication.

(List continues

on the inside back cover)


3/98

WORLD

ANKTECHNICAL

APER

UMBER6

Guidelines

for

Conducting

and

Calibrating

RoadRoughnessMeasurements

Michael

W:Sayers, Thomas

D.

Gillespie,

and WilliamD.

0. Paterson

The

World Bank

Washington,

D.C.,

U.S.A.


4/98

Copyright

(

1986

The

International

Bank

for

Reconstruction

and

Development/THE

ORLD

BANK

1818

H

Street,

N.W,

Washington,

D.C.

20433,

U.S.A.

All

rights

reserved

Manufactured in the United States of America

First

printing

January

1986

This is

a document

published

informally

by

the

World

Bank.

In order

that

the

information

contained

in

it

can be

presented

with

the least

possible

delay,

the

typescript

has

not

been

prepared

in accordance

with

the

procedures

appropriate

to

formal

printed

texts,

and the

World

Bank

accepts

no

responsibility

for errors.

The

publication

is

supplied

at a

token

charge

to

defray

part

of

the cost

of manufacture

and

distribution.

The

World

Bank

does

not

accept

responsibility

for

the

views

expressed

herein,

which

are those

of the

author(s)

and

should

not be

attributed

to

the

World

Bank

or to its

affiliated

organizations.

The findings,

interpretations,

and condusions

are

the

results

of

research

supported

by

the Bank,

hey

do

not

necessarily

represent

official

policy

of

the Bank.The designations employed, the presentation

of

material,

and

any maps

used

in

this

document

are

solely

for the

convenience

of

the reader

and do

not

imply

the

expression

of

any

opinion

whatsoever

on

the

part

of the

World

Bank

or its

affiliates

concerning

the

legal

status

of

any country,

territory,

city,

area,

or

of its

authorities,

or

concerning

the

delimitation

of

its boundaries

or national

affiliation.

The most

recent

World

Bank

publications

are

described

in

the annual

spring

and

fall

lists;

the continuing

research

program

is

described

in the

annual

Abstracts

of

Current

Studies.

The latest

edition

of

each

is available

free of

charge

from

the

Publications

Sales

Unit,

Department

T,

The

World

Bank,

1818

H

Street,

N.W,

Washington,

D.C.

20433,

U.S.A.,

or

from

the

European

Office

of

the

Bank,

66

avenue

d'1ena,

75116

Paris,

France.

Mfichael

W Sayers

is

assistant

research

scientist

and

Thomas

D.

Gillespie

is

research

scientist

at the

Transportation

Research

Institute

of

the

University

of Michigan

in Ann

Arbor. WilliamD. 0. Paterson is senior highway engineer with the Transportation

Department

of the World

Bank.

Library

of

Congress

Cataloging-in-Publication

Data

Sayers,

M. W.

(Michael

W.)

Guidelines

or

conducting

nd

calibrating

oad

roughness

easurements.

(World

Bank

technical

aper,

ISSN

0253-7494

no.

46)

Bibliography:

.

1.

Roads--Riding

ualities--Testing.

.

Road

meters--Calibration.

.

Gillespie,

.

D.

(Thomas

.)

II. Paterson, illiam D. 0. III. Title. IV. Series.

TE251.5.S29

1986

625.8

85-17806

ISBN

0-8213-0590-5


5/98

ABSTRACr

Road roughness s gaining increasing mportance s an indicator f road

condition, oth in terms

of road pavement performance, nd as a major deter-

minant

of road user costs. This need to measure roughness as brought

a

plethora of instruments n the market, covering the range from rather simple

devices to quite complicated

ystems. The

difficulty s the

correlation nd

transferability f measures from various instruments nd the calibration o a

common scale, a situation

hat is exacerbated

hrough a large number

of

factors that cause variations etween readings of similar instruments, nd

even for the same instrument. This need to correlate nd calibrate ed to

the International oad Roughness xperiment (IRRE) in Brazil in 1982, which

is documented n a companion olume in this Series, entitled he International

Road Roughness xperiment: Establishing orrelation nd a Calibration tan-

dard for Measurements (World Bank Technical aper Number 45).

This paper defines roughness easurement ystems hierachically nto four

groups, ranging from profilometric ethods (2 groups) being accurate and

most amenable to detailed analysis through response-type oad roughness

measuring ystems (RTRRMS's) representing he most widely used, practical

and fast instruments to subjective valuation allowing assessments o be

made without use of instruments. The general planning of road roughness

measurement rograms is outlined, s well as the criteria or selection f

measurement ystem to meet the objective. The procedures or carrying out

surveys in the four groups of systems are explained, including nstrument

characteristics, he

need for adequate

checking and verification, nd the

importance

f travelling peed,

as well as the methodology

or data analysis.

The international oughness Index (IRI) is defined, nd the programs for

its calculation re provided. The IRI is

based on simulation

f the rough-

ness response f a car travelling t 80 km/h it is the Reference verage

Rectified lope,

which expresses ratio of the accumulated

uspension otion

of a vehicle, divided by the distance travelled uring the test. The report

explains ow all roughness easurements

an be related to this

scale, also

when travelling t lower speeds than 80 km/h. The IRI therefore merges as a

scale that can be used both for calibration nd for comparative

urposes.

iii


6/98

ACKNOWLEDGEMENTS

These guidelines ave their technical oundations n the published

findingsof two major research rojects:

The international oad Roughness xperiment (IRRE) [1], eld in

Brasilia in 1982, and funded by a number of agencies, ncluding he Brazilian

Transportation lanning Company (GEIPOT), he Brazilian oad Research

Institute (IPR/DNER),

he World Bank (IBRD), he French Bridge

and Pavement

Laboratory LCPC), and the British

Transport nd Road Research aboratory

(TRRL); and

The NCHRP (National ooperative ighway Research rogram) Project 1-18,

documented y NCHRP Report No. 228 [2].

Per Fossberg (IBRD) nd Cesar Queiroz (IPR/DNER) re acknowledged or

their contributions n the development f these guidelines. Also, grateful

acknowledgement s extended to Clell Harral (IBRD), ho conceived he idea of

the IRRE and arranged for the participation f the various agencies nd the

subsequent reparation

f these guidelines.

iv


7/98

TABLE

OF CONTENTS

CHAPTER 1: SCOPE ....................................

.

I..*.........**.

CHAPTER 2: PLANNING A ROUGHNESS

MEASUREMENT PROJECT......................

3

2.1 Overview of the IRI

Road Roughness Scale

........................

3

2.2 Roughness Measurement Methods....

............

.... 6

2.2.1 Class 1:

Precision profiles................................

6

2.2.2 Class 2: Other profilometric

methods.......................

7

2.2.3

Class

3: IRI

estimates from correlation

equations..........

8

2.2.4 Class 4: Subjective

ratings and

uncalibrated measures......

9

2.3 Factors

Affecting

ALccuracy o............................e..............e.

2.3.1 Repeatability error........................................1l

2.3.2 Calibration error..........................................l2

2.3.3 Reproducibilty error.......................................13

2.4 Planning

the Measurement Project

..................

........ 14

2.4.1 Long-term network monitoring

..............................

14

2.4.2 Short-term project

monitoring.

........................ .. ... 15

2.4.3 Precise

monitoring

for

research............................17

CHAPTER

3: MEASUREMENT OR IRI USING PROFILOMETRIC

METHODS (CLASSES 1 &

2).19

3.1 Description

of Method ...........................................

9

3.2 Accuracy Requirement ..

...........................................

l19

3.3

Measurement of

Profile.o......s

........o.........

......2

3.3.1

Rod and Level Survey.......................................22

3.3.2

TRRL Beam Static Profilometer...o...............

.........

.26

3.3.3

APL Inertial Profilometer.

. .... ................. 27

3.3.4 K.

J. Law Inertial Profilometers............................29

3.3.5 Other profilometers .......................................

1

3.4 Computation

of IRI

....................

3.4.1

Equations

................... .............................

1

3.4.2

Example

program for computing

IRI ....

3.4.3 Tables

of coefficients

for

the IRI equations

..............35

3.4.4 Program for computing

coefficients

for the IRI equations...35

3.4.5 Test

input for

checking computation.........................40

v


8/98

CHAPTER 4: ESTIMATION OF IRI USING A CALIBRATED RTRRMS (CLASS 3) .......... 5

4.1 Selection nd Maintenance f a RTRRMS ............................5

4.1.1 The

ramtr...................................4

4.1.2

The vehicle ............

4

4.1.3 Installation f the roadmeter n the vehicle..............o.47

4.1.4 Operating

peed.............................................47

4.1.5 Shock

absorber selection ... e...

..

. .. ............ e.. .48

4ol.6 Vehiclelodn.........................

4.1.7 Tire pressureo.....o............ e o

o.o @ *.o**.*

..... **49

4.1.8

Mechanical inkages n the roadmeter........................49

4.1.9 Tire

imbalance nd

out-of-roundness

..... o................

9

4.1.10 Temperature

ffects..........................................49

4.1.11 Water and moisture

effects

..................................

50

4.2 Calibration f a RTRRMS.............................................50

4.2.1 Calibration

method-.o.oo ...

.*o....oooo

.. .....

.

. .

.51

4.2.2

Calibration equation.. ....................... .............

53

4.2.3 Selection f calibration ites.... ..... o..o.... .o......53

4.2.4 Determining RI

of calibration ites.........................58

4.2.5

Compensation

for non-standard

speed.............o.........o.58

4.3 Operating nd Control est Procedures.............................60

4.3.1 Vehicle and roadmeter peration .......... e................0

4.3.2 Data processing .............................................2

4.3.3 Temperature ensitivity est

..............................

3

4.3.4 Control tests for RTRRMS time stability.....................63

CHAPTER 5: ESTIMIATIONF IRI BY SUBJECTIVE VALUATION CLASS 4) ...........1

5.1 Descriptive

valuation ethod

........ ............................

1

5.1.1 Method......................................................71

5.1.2 Description f the IRI Scale................................71

5.1.3 Personnel ...............o....o...o...oo.........oo.......o.o...o....75

5.1.4

Calibration o...oo .............

.o..o.............

o.o.-

......o....5

5.1. Survey eoo ooooogso*eoo.....o.....................oo....o.........o

5.1.6 Data Processingo...

........................................

76

5.2 Panel Rating of Ride Quality.

... . . . ...... . . .. . . .. . . . . . . .o.. . . . . .. o..76

GLOSSARY

.*.oo oooo.oo.e.go.o.o....oo...o......................

.o79

REFERENCES

....... ....................................

vi


9/98

CHAPTER 1

SCOPE

This documentpresents guidelines or use by personnel

n highway

organizations esponsible or setting up or operating oad roughness

monitoring rograms.

It provides guidance on:

* Choosing a method for measuring road roughness;

*

Calibrating he measurement quipment to a standard oughness

scale;

* Using procedures hat ensure reliable easurements n routine

daily

use.

The suggestions nd procedures resented ere are intended to

guide the practitioner n acquiring oad roughness ata from which to

build a roughness

ata base for a road network. Adherence to these

guidelines ill help ensure:

* That the roughness ata indicate road condition s it affects

using vehicles

in terms of ride quality,

user cost, and safety;

* That data acquiredin routine measurement

perations ill

be

related to a standard roughness cale, and that erroneous ata can

be identified rior to entry into the data base;

*

That

the roughness ata

can be compared directly

to data acquired

by other highway organizations lso following he guidelines; nd

* That the roughness easures have the same meaning on all types of

roads used by highway trucks and passenger ars, including

asphalt, oncrete, urface treatment,

ravel, and earth

surfaces.

The procedures resented n this document are primarily pplicable

to roughness easurements f two types:

*

Direct measurement f roughness n the standard cale, derived

from the longitudinal

rofile of the road

* Estimation of the standard roughness easure, using calibrated

response-type oad roughness

easurement ystems

(RTRRMSs)

1


10/98


11/98

CHAPTER 2

PLANNINGA ROUGHNESSMEASUREMENT

ROJECT

The

design

of a project for

surveying

he roughness f

a road

network should start with a clear

understanding f

the objectives o

be

achieved from

the measurement ffort.

A substantial nvestment f

manpower

and money

can be consumed

n a typical roject,

thus it

is

desirable o design the program carefully. The design itself is a

synthesis

rocess taking into

account the project

goals, the resources

available,

nd the environment

f the project. Perhaps

the most

critical element in

the design is the

selection f a roughness

measurement

ethod that is

practicable, et suitably ccurate or

the

purposes

of the project.

This section reviewsthe various measurement

methods available,

lassified ccording

o how directly

they measure

roughness

n a standard cale (Generally, he

more direct methods

are

also the most

accurate). In addition,

t explains

the types of errors

to

be anticipated, nd their importance o

various kinds of measurement

projects.

2.1 Overview f the IRI

Road Roughness

cale

In order

to address

specifics

f roughness easurement,

r issues

of accuracy,

t is first necessary o define

the roughness cale.

In

the interest

f encouraging

se of a common

roughness easurein all

significant

rojects throughout

he world, an International

oughness

Index (IRI)

has been selected.

The IRI is so-named ecause

it was a

product of

the International oad

Roughness xperiment

IRRE),

conducted

by research

eams from Brazil,England, rance,

the United

States, and

Belgium for the purpose of identifying

uch an index. The

IRRE was held

in Brasilia, razilin 1982 [11 and involved the controlled easurement

of

road roughness or

a number of roads under

a variety of conditions

and by

a variety of instruments nd

methods. The roughness

cale

selected

s the IRI was the

one that

best satisfied

he criteriaof

being time-stable, ransportable,

nd relevant, hile

also being readily

measurable y all practitioners

The IRI is a standardized

oughness easurement

elatedto those

obtained

y response-type

oad roughness easurement

ystems(RTRRMS),

with recommended

nits: meters

per kilometer

(m/km)

=

millimeters er

meter (mm/m)=

slope x 1000.

The measure

obtained

from a RTRRMS

is

called either

by its technical

ame of average

rectified

lope (ARS), or

more commonly,

y the units used

(mm/km,

in/mi, etc.).

The ARS measure

is a

ratio of

the accumulated

uspension otion

of a vehicle

(in,

mm,

etc.), dividedby the

distance travelled

y the vehicle

during the

test

(mi, km,

etc.). The reference

TRRMS

used for

the IRI is a mathematical

model,

rather

than a mechanical

ystem,

and exists

as a computation

3


12/98


13/98

IRI (m/km =

mm/mr)

NUSEAL

A

.~~~~~S

16 EROSION ULLEYS

AND

DEEP

DEPRESSIONS

14

_ :^^^

..

. 50

km/h

12

-

FREQUENTHALLOW

DEPRESSIONS,

OME _

. 60 km/h

10

DEEP.

.

ROUGH

S

:.*^UNPAVED

....

8 FREOUENTUNAE

ROADS

~80 km/h

MINOR

DEPRESSIONS

.

6

rDAMAGED

SURFACE.

PAENT

o _

OLDERAVEMENTS)

0= ABSOLUTE

I NEWPAVEMENTS)

4 ERFECTION AIRO W

USSUPERHIGHWAYSJ

Fig.

1. The

IRI roughness

cale.

5


14/98

*

overall

ride

quality

*

dynamic

wheel

loads

(damage

to

the

road rom

heavy

trucks;

braking

and

cornering

afety

limits

available

to

passenger

cars)

*

overall

surface

condition

The

IRI

is

also recommended

henever

the measurements

ill

be

obtained

using a RTRRMS at highway speeds (50 100 km/h), regardless f the use

made of

the

data.

However,

hen

profilometric

ethods

are

used

to

measure

wheeltrack

roughness,

hen

other

measures

may

serve

as

better

indicators

or

some

qualities

f

pavement

condition,

r

for

specific

components

f

vehicle

response

encompassed

y

the

IRI. These

guidelines

ddress

only

the

measurement

nd

estimation

f

the IRI.

2.2 Roughness

Measurement

Methods

The many approaches or measuringroad roughness n use throughout

the

world can

be

grouped

into

four

generic

classes

on

the basis

of how

directly

their

measures

pertain

to

the

IRI,

which

in turn

affects

the

calibration

equirements

nd

the

accuracy

associated

ith

their

use.

2.2.1

Class

1:

Precision

rofiles.

This

class

represents

he

highest

standards

f accuracy

for measurement

f IRI.

A

Class

1 method

requires

that

the longitudinal

rofile

of

a

wheeltrack

e

measured

(as

a

series

of

accurate

elevation

oints

closely-spaced

long

the

travelled

wheelpath)

as a

basis

for calculating

he

IRI

value.

For

static

profilometric

ethods,

the

distance

between

samples

should

be no

greater

than

250

mm (4

measures/meter)

nd

the

precision

n

the

elevation

measures ust be

0.5 mm

for

very

smooth

pavements.

(Less

precise

measurements

re acceptable

or

rougher

surfaces,

s

specified

n

Section

3.2.)

High-speed

rofilometers

ffer

a potential

eans

for

measuring

IRI

quickly;

owever,

the profilometer

ust

be

validated

t

some

time

against

an established

rocedure

uch

as

rod and

level

to

prove

its

accuracy.

At

the

present

time,

only

rod and

level

(Section

3.3.1)

and

the

TRRL

Beam

(Section

.3.2)

methods

have

been demonstrated

to be

valid

Class

1

methods

for

determining

RI

over

a broad

range

of

roughness

evels

and

road

types

for the

320 m site

length

used

in

the

IRRE.

Methods in this class are thosethat producemeasuresof such high

quality

that

reproducibility

f

the

IRI

numeric

could

not

be

improved.

While

this

definition

ight

at

first

appear

to

imply

an unreachable

ideal,

there

is usually

a practical

imit

to the

repeatability

hat

can

be obtained

in measuring

road

roughness,

ven

with

a "perfect"

ethod

and/or

instrument.

The

practical

imit

results

from

the inability

o

6


15/98

measure

roughness epeatedly

n

exactly

the same wheeltrack.

Therefore,

a method qualifies

s

Class 1 if measurement

rror

is negligible

n

comparison

ith

the uncertainty

ssociated

ith

trying to locate

exactly

the

same

wheeltrack

wice.

In

the IRRE

the methods found to

qualify

as Class

1 had negligible

measurement

rror

for sites 320 m

long, when

the wheeltracks

ere marked

with

painted

reference pots

spaced at

about 20

m intervals. The

repeatability nder these conditions s about 0.3 m/km IRI on paved

roads, and about

0.5 m/km

for all other road

types.

For wheelpaths

marked

even more

precisely, ethods

described

n these

guidelines s

Class

1 could

perhaps not

qualify

as Class 1 (although

t is

uncommonto

have an application

here

such a

high levelof accuracy

is

needed).

On

the

other hand,

less

stringent pecifications

ight

be suitableif

longer test

sites were

used,

or if the wheeltracks

ere not

marked

at

all.

In many

cases, a

method that

yields this level

of accuracy

ill

have an

associated

isadvantage f

requiring

great

deal of effort to

make the roughness easurement for example, y the rod and level

method).

The accuracyobtained

sing

a Class 1

method by definition

matches

or exceedsthe

requirements

f a

given application,

nd thus

the

Class

1 method

is viewed

as having primary

utility

for validating

ther

methods,

or

when special high-accuracy

ata are required.

2.2.2 Class

2: Other

profilometric

ethods.

This

class

includes

all other

methods in which profile

is

measured as the basis

for direct

computation f

the IRI,

but which

are not capable

of the

accuracy

required

for

a Class 1 measurement.

Though

the hardware

and methods

used for

profile

measurement re functionally

erified

by an independent

calibration rocess,

they

are

limited to accuracy

r bandwidth

less

than

that needed to qualify as a Class 1 method. Consequently, he IRI value

computedfrom a

Class 2 profile

measurement

ay not

be accurate

to

the

practical imit

due to

random

or bias errors

over some

range

of

conditions.

This

class presently

ncludes

IRI values

computedfrom

profiles easured

with

high-speed

rofilometers

nd with

static

methods

that

do not satisfy

the

precision nd/or

measurement

nterval

requirements

pecified

n

Section 3.2.

At the present

time,

the APL Trailer

(Section .3.3)

is

the only

dynamic

profilometer

hat

has been experimentally

alidated

ver the

range

of roughness overed

in

the IRRE. The

GMR-type

Inertial

Profilometer ith follower heels has been validated or roadswith

roughness

evels

less than an IRI

value

of about 3 m/km

[2], above

which

errors

are introduced

ue to bounce

of

the follower heels.

This

type

of

design is no longer

commercially

vailable n

the United

States,

however,

as the

follower heels

have

been replaced

ith non-contacting

sensorsto eliminate

the

bounce problem.

Two high-speed

profilometers

7


16/98

are

presently

old by K.J. Law,

Inc. (Section

.3.4),

and both are

designed

to provide

the

IRI roughness

uring

measurement.

Both are

considered s

Class

2 systems

at this time,

although

their accuracy

and

range

of operation

ave

not been

verified against

rod and level yet.

Tests with these and

other profilometers

ave been performed,

ut

the

analysesof

the data

have not

yet been completed

ufficiently

o

quantify

their ability

to measure

IRI

High-speed rofilometers

ave

the disadvantage

f being

the

most

expensive

nd complex

instrumentation

ystems

used

to measure road

roughness,

nd

generally require

operators

ith engineering

raining.

Yet, they

offer

a great

advantage n

being able

to obtain

high-quality

measurements

apidly,

ithout

requiring

hat great

effort

be spent

in

maintaining

alibration.

Detailed

procedures

or operating

profilometer

o

measure

IRI are highly

specific

to the design

of the

profilometer;

ence, the manufacturer

hould

be consulted.

Sections

3.3.3

and 3.3.4

briefly

describe several

of

the high-speed

rofilometers

that have

been used

to measureIRI.

2.2.3Class 3: IRI

estimates

rom correlation

quations. By far,

the majorityof

road roughness

ata that is collected

hroughout

he

world

today

is obtained

with RTRRMSs.

The RTRRMS

measure

depends

on the

dynamics

of

a vehicle

to scale

the measurements

o

yield

roughness

properties

omparable

o the

IRI.

The dynamicproperties

re

uniquefor

each

vehicle,

however,

and change

with

time,

Thus, the "raw"

measures

of

ARS obtained

from

the RTRRMSmust

be corrected

o the IRI

scale using

a calibration

quation

that

is obtained

experimentally

or

that specific

RTRRMS. Because

the dynamics

of

a vehicle change

easily, very

rigorous

maintenance nd operating

rocedures ust be

employed

for the

vehicles

used,

and control

testingmust

be

made

a routine

part

of normal

operations.

When changes

occur, there

is no simple

correction

hat can

be applied;

instead,

the entire

roadmeter-vehicle

ystem

must be

re-

calibrated.

This

class

also includes

other roughness

easuring

instruments

capable of

generating

roughness

umeric reasonably

orrelated

o the

IRI

(e.g., a rolling straightedge).

The measures

obtainedcan

be used

to estimate

IRI throughregression

quations

if a correlation

xperiment

is performed.

This approach

is

usually

more troublethan

it's

worth

In 1984 a Road Profilometer eetingwas held in Ann Arbor, Michigan,

to determine

the performance

haracteristics

f

a number of

profilometers,

ncluding

oth of

the currentnon-contacting

ystems

from

K.J.Law,

Inc. (USA), the

Swedish

VTI laser system,

the

APL Trailer,

and

several

non-commercial

ystems.

The

study,which was

funded

by the U.S.

Federal

Highway

Administration

FHWA)

and conducted

y UMTRI,

is still

underway

31.

8


17/98

(better easures

can be

obtained

ith less effort),

nless there

is a

need to convert

a large

amount

of past data to

the IRI

scale.

A method for measuring

roughness

ualifies

s Class 3

if it uses

the "calibration

y

correlation"

pproach described

n Section

4.2,

regardless f

what type

of instrumentation

r vehicle

is

used to

obtain

the

uncorrected oughness

easure.

While

most Class

3 methods

will

employa roadmeter

hat

accumulates

uspension

otion

to measure

ARS as

describedin Section4.1, other systemsare in use that employ

accelerometers

r other

types

of instrumentation.

However,

the

roadmeter-based

TRRMS that measures

ARS

most closelymatches

the IRI

concept,

and these

guidelines oncentrate

n

the calibrated

TRRMS as

the

principle lass

3 method.

Unless a

RTRRMS

is calibrated

y correlation,

t does not

qualify

as

a Class 3

method. Without the

calibration,

here

is no verifiable

link

between

the measuresobtained

ith any

two RTRRMSs, or

to the IRI

scale.

The reproducibility

ssociated

ith a calibrated TRRMS is about

0.5 m/km (14%)

for paved

roads for

sections 20 m long,

and about

1.0

m/km (18%)

for unpaved surfaces

f

that length.These

accuracy

figures

are only

approximate verages,

s the

errors generally

ary both with

roughness

nd surface

type.

Better accuracy

is possible

by using

longer

test sections.

2.2.4 Class

4:

Subjective

atingsand

uncalibrated

easures.

There are situations

n which

a roughness

ata base

is needed,

but high

accuracy

is not

essential,

r cannot

be afforded.

Still,it is

desirable

o relate

the

measuresto

the IRI

scale.

In those

cases,

a

subjective

valuation nvolving

ither

a ride experience

n the

road or

a visualinspection ould be used. Another possibility s to use the

measurements

rom

an uncalibrated

nstrument. Conversion

f these

observations

o the IRI scale

is

limited

to an approximate

quivalence,

which

can

best

be established

y comparison

o verbal

and/or

pictorial

descriptions

f roads

identified

ith

their associated

RI

values,as

described

n Section

5.0.

Essentially,

he estimates

f equivalence

re

the

calibration, owever

approximate,

nd they may

be considered

o be

"calibration

y description."

When these

subjective

stimates

f roughness

re

converted

o the

IRI scale

the resolution

s limitedto about

six

levelsof roughness

with accuracy

ranging

from 2

-

6 m/km

(about

35%) on the IRI scale.

(Roughness

ccuracy, xpressed

ither

in absolute nits

of m/km

or as a

percentage,

ill

generally

ary with

roughness

evel and surface

type.)

9


18/98

Note that

unless

a

valid calibration

y

correlation

s used

with

a

RTRRMS,

thereis no way

to

link the

measureto

the

standard

scale.

Thus,

an uncalibrated

TRRMS

falls within

Class

4.

2.3

Factors

Affecting ccuracy

Roughness ata

are

normally

utilized

in applications

epresenting

two extremes:(1) statistical nalysesinvolving oughness easurements

on

major segments

f

a road

network,

and (2)

individual

tudies

related

to

roughness

t specific

road

sites.

The

roughness ata

will

necessarily

nclude

some

errors

arising from

random

and systematic

effects.

The

significance

f these

errors depends

on the

nature of

the

application

or

which the

data

are intended.

An

example

of

the first type

of

application

s a

road-user

ost

study,

in which

the data

base

of operating

osts

for a fleet

of

vehicles

is regressed

gainst

the data

base

of roughness

or

the roads

on

which

those

vehicles

were

operated.

In that

case,

the

need is to

determine

levelsof roughness or comparison ith trends of costs,using

regression

ethods.

Random

errors in

individual

oughness easurements,

caused

by poor precision

r a

peculiarroad

characteristic,

ill tend

to

average

out

if the study

includes

a large

number

of

road sites.

Thus,

random

error is not

of great

concern

for this

type

of study.

On the

other

hand,

systematic

rrors

will

bias the cost

relationships

btained.

Therefore,

teps should

be taken to

keep the

systematic

rrors

to

minimal

levels.

The results

of the study

will not

be transportable

unless a standard

roughness

cale

is used,

and

steps are taken

to

ensure

that

the roughness

ata more

or less

adhere

to that scale.

Studies

that

involvemonitoring

oadway

deterioration

r

the

effects of maintenance re examples of the secondtype of application.

In

these

cases, it

is of

interest

to maintain

a continuing

ecord

of

small

changes

in the roughness

ondition

t specific

road

sites.

Random

errors

in measurement

ill

reduce

the certainty

ith which

the trends

of

interest

can

be discerned.

A

constant

ias in

the

data can

be

determined

nd corrected

n

order

to compare roads

or

apply economic

criteria,

ut it is perhaps

even more critical

to ensure

that the

bias

does

not

change

with time.

Thus,

for

measurements

o be

used

for these

applications,

he practitioner

hould

employ

procedures

hat will

minimize

randomerrors

while

also maximizing

ime-stability.

This

normally

translates

nto using the

same

equipment

nd personnel

or

regularmonitoring f a road site, utilizing epeat tests to improve

repeatability,

nd

carefully

aintaining

he calibration

f

the

equipment.

The

end use

of the roughness

ata in applications

uch as these

has a direct

impact

on

the accuracy

that

will be

necessary

n the

10


19/98


20/98

as would

be

obtained

on

a

320

m test

site

after

five

repeats

(5

x 320

m

= 1.6

km).

As a

rule

of

thumb,

a

total

length

of 1.6

km

(1.0

mile)

or

longer

is

recommended

o minimize

repeatability

rror

for

instruments

used

at

highway

speeds.

Another

means

for

increasing

he

averaging

or

a RTRRMS

instrument

is

to

use a

lower

speed

for

a

given

length

of test

site;however,

this

approach

is

not

recommended,

ecause

changing

the

speed

also changes

the

meaningof the roughness easure for the RTRRMS and increases ther

errors.

2.3.2

Calibration

rror.

Systematic

rrors

exist

in

instruments.

These

cause

the

measurements

f

one

to be

consistently

ifferent

from

those

of

another,

or

cause

one instrument

o

vary

with time.

This

can

be corrected

y calibration,

o that

the roughness

easurements

re

rescaled

to cancel

systematic

ifferences

ringing

the

measures

to a

common

scale.

However,

if

the calibration

oes

not

cover

all

of

the

variables

that

affect

the

measurement,

hen

the

rescaling

ay

not be

correct,

and

a calibration

rror

remains.

Profilonetric

ethods

(Classes

1 and

2):

Calibration

rror

is

minimal

when

direct

profile

measurements

re

used

to

obtain

the IRI.

The

instruments

hat

measure

the profile

are

calibrated

t the

factory,

and

do not

change

much

when

given reasonable

are.

Nonetheless,

systematic

rrors

can

appear

in

profile-based

easures

when

(1)

the

profile

elevation

easures

contain

errors

(usually

aking

the

profile

seem

rougher

than

it

is), (2)

when

profile

measures

are

spaced

too

far

apart

such

that

some

of

the roughness

eatures

are

missed

(making

the

profile

seem

smoother),

nd

(3)

when

profile

measures

are

subjected

o a

smoothing

r a waveband

limitation

s

occurs

with

a dynamic

profilometer

(making

the

profile

seem

smoother).

The

specifications

nd

procedures

recommended n Sections3.2 and 3.3 were designed to hold theseeffects

to

negligible

levels.

RTRRMSs

(Class

3): Calibration

y

correlation

ith

a reference

(Section

.2)

is required

or

a RTRRMS

for

many

reasons,

including

hese

important

hree:

1) The

overall

dynamic

response

f

any particular

TRRMS

vehicle

will

differ

to some

degree

from

that of

the

reference.

This

effect

can

cause

the "raw"

ARS

measure

from

the

RTRRMS

to be

higher

or lower

than

corresponding

RI values,

depending

n

whether

the

RTRRMS

is

more

or

less

responsive

han

the

reference.

2)

The roadmeter

n the

RTRRMS

generally

as freeplay

and

other

forms

of

hysteresis

that

cause

it to

miss

counts,

resulting

n

lower

roughness

easures.

12


21/98


22/98

Reproducibility s

not improved by repeating easures

on the same

site, since the effect is systematic or that site.

2.4

Planning the Measurement roject

The execution

f a high-quality

oad-:oughness easuring

rogram

is critically ependent n establishing ell-thought-out rocedures that

are adhered to in a strict and consistent ashionthroughout he

project. This section outlines the

planning needs for

the three main

kinds of roughness easurement

rojects, o aid the planner in

appreciating he logistics hat are involved.

2.4.1 Long-term

etwork

sonitoring. Long-term oughness

monitoring rograms are an integral part of network condition valuation

surveys and pavement anagement systems.

Typical objectives nclude:

1) Summary of

network condition n a regular

basis for

evaluation f

policy effectiveness

2)

Input into a network-level conomic

analysis of

pavement design

standards, aintenance olicy, and transportation osts

3) Quantifying roject condition or prioritizing aintenance nd

rehabilitation

rograms.

To meet these objectives, he measurements ill usually be continuous

over

links of the network and the total length will exceed 1000

km (or

even 10,000 m). It is essential that measures made in different reas

of the network be directly comparable,

nd that the measures be

consistent ver

time. However, the accuracy requirements or individual

roughness easurements ill generally ot be as demanding s for other

types of projects, ecause data averaging ill reduce

the effects of

random errors. Of the three sources of error

described n Section 2.3,

the calibration rror is the most critical to control.

When planning a long-term onitoring program, ne should consider:

a)

Type of roughness easuring nstruments: The rapid collection

and automatic rocessing f data are paramount onsiderations o

facilitate torage in a data bank, and streamline nalysis. Only

instruments hat can be operated at the higher speeds should be

considered. (The instrument hould operate at least at a speed of 50

km/h, and preferably t 80 km/h or faster.) Any type of RTRRMS is

suitable. A high-speed rofilometer s also suitable and can provide

useful descriptive umerics in addition

to IRI.

b)

Number of instruments:

When

the network is very large or

spread-out, ore than one instrument

ay be required. If this is the

14


23/98

case, a fleet of RTRRMSs might be more affordable han a fleet of

profilometers. The vehicles sed for RTRRMSs preferably hould

be of

the

same make for the sake of interchangeability,

lthough this is not

essential hen

sound calibration rocedures re followed.

c) Calibration ections (for RTRRMSs only ): A series of eight to

twenty calibration

ections ill be needed at a central location

nd

possibly t distant

regional ocations o permit full calibration f the

test vehicles

at regular intervals Section .2).

d) Control sections:

A small number of control sections three to

five) will be needed in every region here the instruments ill operate

to permit control checks on a daily or weekly basis (Section

.3.4).

e) Measurement peed (for RTRRMSs only3): This

may be a

compromise f conflicting onsiderations. The standard speed of 80 km/h

is likely to be applicable n the majority f situations. Severe road

geometry or

congestion ill dictate a lower speed of 50 or 32 km/h on

some links, but this should not influence he choice for the majority f

the survey. The simultaneous ollection f other data during the survey

may influence he choice.

f)

Data

processing nd reporting: Data collection ust include

location

nd other event markers

for reconciliation ith other pavement

management ata. Computerization t the earliest possible tage and use

of standard

oding forms where necessary hould be considered o

facilitate ata entry. Measurements hould be recorded t intervals f

no more than 1 km. Reporting ill usually comprise ean values either

by link or homogeneous ection of 10 km or longer, with summary

histograms f roughness

istribution y road length. These reporting

units should coincide ith at least the major changes in traffic volumes

to facilitate stimates f vehicle operating osts. For efficiency, he

data can be

managed so as to permit

separating he more

detailed

reporting equirements f

simultaneous roject evaluation nd

prioritization

tudies.

2.4.2 Short-term roject monitoring.

Evaluation f specific

rehabilitationr betterment

rojects nvolves ither

short-term

observations ver periods

up to 3 years

or one-shot oughness

measurements.

Typically, he sites

will range from 5 to 50 km in length

and will not necessarily

e contiguous.

Careful consideration hould

be

given to the detail

and accuracy required,

s accuracy

requirements an

sometimes

e more stringent han for long-term

etwork monitoring

2

Profilometers re calibrated t the factory r in a laboratory.

3

Speed requirements or profilometers re specific o the profilometer

design.

15


24/98

(Section

2.4.1).

On the other hand,

if only

approximate

oughness

measures

are

needed, considerable

conomy

can be achieved.

If a history

of surface

roughness s

desired,

then the instrument

should

be capable

of providing

repeatable

easures over

a period

of

time, and it

will be important

o

maintain calibration

rror

to small

levels. Also,

if high accuracy

is desired, repeated

easurements

an be

averaged

to reduce the repeatability

rror that might

otherwise

ask

small changesin roughness. In general,efficiency n data acquisition

is not critical for

short-term

rojects, nd therefore

mphasis

should

be placed

on obtaining

ata with quality

as high

as possible

from the

instrumentation.

In some cases,

transportability

f the

data (obtained

y using

the

standard

IRI scale)

may

not be as critical

as maintaining

high

standard

of

internal

consistency. In

practice,

owever,

the careful

controls eeded to

maintain

internal consistency

ill often

result

in

adherence o the

IRI

scale anyway

(particularly

or RTRRMSs).

a)

Profilometric ethods

(Classes

1 and 2):

Profilometric

methods

are suitable

and

can optionally rovide

useful

descriptive

numerics

in addition to

IRI, which

can be used to

diagnose

the nature

and

probable

sources

of distress. (For

example,

the APL

72 system

normally

provides

three waveband roughness

ndices.

The predominantly

long

wavelength

roughness

ndicates ubgrade

r

foundation

nstability,

whereas

shortwavelength

oughness

ndicates

ase or surfacing

distress.)

If

a profilometer

s available,

t can

probably

be appliedwith

little

modification

n procedure,

requiring

nly a

more detailed

reporting ormat

and possibly

ore

carefulmarking

of test

sites. If

no

roughness easuringinstrumentations available, profilometer ight

be imported

temporarily

ith

less overallcost

than the

purchaseof less

sophisticated

ystems

that require

extensive

alibration

ffort.

b)

Calibrated

TRRMS

(Class

3):

These methods

need to be under

rigorous

control

to be satisfactory

hen high

accuracy is desired.

If

possible,

singleinstrument

hould

be used

to performall of

the

measurements

n order to

minimizereproducibility

rror.

The

complete

calibration see Section

4.2)

may need

to be repeated ore

frequently

than

for other applications,

s even

small

changes in the response

properties

f the

RTRRMS may

mask the desired

roughness

nformation.

If a fully equipped TRRMS is available(froma long-term

project),

t

can possibly e applied

with

little

modification n

procedure,

equiring nly

more detailed

measurement

nd reporting

formats.

However,

if the accuracy

requirements

re significantly

ore

stringent han

for the other

project,

then the procedures

ill need

to

16


25/98


26/98

often been used for this application, ut generally ack adequate

precision nd give rise to uncertainty n the trend data. It should be

noted that profile

measures can also be processed to yield a variety of

surface

condition ndicators ther than the IRI, whereas RTRRMSs are

capable of only the single type of measurement.

A trade-off n the frequency f measurement

s possible: Class 1

or 2 measures need be made only annually and in conjunction ith major

maintenance ctivities, ecause of their higher accuracy. However,

Class 3 RTRRMS measures should be made at least two to three times per

year, in order to ensure confidence n the data trends. Portability f

the system is important: Class 3 methods require the

establishment f

supporting ontrol sections in distant regions, hereas Class 1 or 2

systems do not.

Data processing nd analytical ethods will usuallybe project-

specific; hus, these

topics are not addressed ere. Reporting

hould

include computation f the IRI

for the purposes of transferability, ven

if other numerics are used more directly in the research.

18


27/98


28/98

TABLE 1. Accuracy

requirements

for Class

1 and

2 prof lometric

measurement

of IRI

Maximum

convenient

sample

interval

between points

Precision

of elevation

Roughness

range

(mm)l/

measures (mm)

2

/

IRI (m/km)

Class

1 Class 2

Class

1 Class 2

1.0

-

3.0

250

500

0.5

1.0

3.0 - 5.0

250

500

1.0 1.5

5.0

- 7.0

250 500

1.5 2.5

7.0 -

10.

250

500

2.0

4.0

10 - 20 250 500 3.0 6.0

1/

For tapes marked

in

foot units,

the maximum

convenient

intervals

are

respectively

Class 1:

1 ft.

Class 2: 2 ft.

2/

Precision Class

1 yields

less

than 1.5% bias

in IRI.

Precision Class

2 yields

less

than 5% bias

in IRI.

Note:

Precision

Class

2 is adequate

for the

calibration

of response-type

systems (RTRPMS's).

20


29/98


30/98

applications

ut

not

for others.

Thus,

accuracy

requirements

etermined

for

other

applications

re not

necessarily

alid

for

measurement

f

IRI.

3.3 Measurement

f Profile

3.3.1

Rod

and Level

Survey.

The most

well-known

ay

to measure

profile

is

with conventional

urveying

quipment.

The equipment

consistsof a precision od marked in convenient nits of elevation

(typically

ajor

divisions

re cm or ft),

a

level that is

used

to

establish

a horizontal

atum

line,

and

a tape used

to

mark the

longitudinal

istance

along

the

wheelpath.

This

equipment

s widely

available,

nd can

usually

be

rentedor

purchased

t

a cost that

compares ery

favorably

ith

other

roughness

easuring

quipment.

However,

the method

requires

great deal

of

labor,

and is generally

best to

use when

only a

few profiles

re to be

measured.

Detailed

instructions

or

using a rod

and

level

are beyond

the scope

of these

guidelines;

owever,

the measurement

f

a road

profile

is not

a routine

application

f these

instruments,

nd

therefore n

overview

f the

procedure s providedbelow along with guidance

specific

for this

application.

a) Equipment.

In order

to measure

relativeelevation

ith the

required

recision

or paved

roads,

it

is necessary

o obtain

precision

instrumentation

sed

in construction,

s the

rod and

level equipment

used

for routine

land surveying

ork

cannot

provide

the

required

accuracy.

With

the precision

instrumentation,

n

which the rod

and

level

are calibrated

together,

he

level

usually includes

a built-in

micrometerto

interpolate

etween

marks on the

rod.

Note

that

the

accuracy

requirements

n

Table

1

are straightforward

with regard to rod and level:the elevation precision s generally

equivalent

o

the resolution

ith

which

the rod

can be

read through

the

level,

while the

sample

interval

is the distance

(marked

on the tape)

between

adjacent

elevation

easures.

When

a

tape is marked

in

meters,

an

interval

f 0.25

m

is convenient

or

Class

1 measures,

and

an

interval

f 0.50 m

is convenient

or Class

2 measures.

When

the

tape

is

marked in

feet,

an interval

f 2

ft (610

mm) can be

used

for Class

2

measures,

hile

the

largest

convenient

ncrement

or Class

1 measures

is

0.5 ft

(152.4 m).

b)

Field measurements.

The exact

methodology

dopted to

measure

and record

the

elevation

oints

is not

critical,

nd can

be matched

to

the

local

situation

egarding

vailable

ime,

equipment,

nd manpow4er.

Recent

improvements

n

procedure

eveloped

y

Queiroz

and

others

in

Brazil

in

obtaining

rod

and

level profiles

for the

explicit

purpose

of

measuring

roughness

ave proven

helpful,

and are

suggested

ere.

22


31/98


32/98


33/98

bl.25

M1

M

61.

Q 74

Qe

36.5

99

otom.

61

75

74.25

86.75

99.25

Z

s

~~62

74.5

9

87

99.5

:

of

JOiM

62.25

74.75

87.25

9975

(400-500)

62.5 75 87.5 100

Site Desc

2 L2.

Date

I Stort:

0=

OLD

-7*O4

_

top

o

V599

AS

of

form

.25

75

25.25

37.75

(500-600)

75

1 .2

I

__

13.5 ff

6

so 38.s

_

1.25

t[1

13.75:r

26.25

23

1.5

1

4

Ill

26.5

S

1 7 5

1425

-79

26

7

a) Pre-printed field forms for recording

rod

readings.

-.

profile

Tape

Rod

Eleu.

8o0

497.5

7047

6939

497.75

7044

6942

498

7039

6947

498.25

7045

6941

Ele.u

498.5

7044

6942

498.75

7046

6940

499

7044

6942

499.5

7041

6945

499.75

7044

6942

400

Tape

Position

600

500

7044

6942

500.25 7591 6948

r

' ontrol 26.2

500.75

7586

6953

10000

0

100

Tape

Int

|

500.7

7586

6953

12616

100.25

175

pen..5

501

75R.

6953

11378

175.25

200

Tape

=

501.25

501

.251

75

12545

200.25

300

501.5

13224

300.25

400

new

digits

= 2

501.75

13986

400.25

500

502

14539

500.25

end

Total

length

501

502.25

GL

used

502.

tI

used

502.75

lT used

b) Display

of the

microcomputer

screen

when typing

date into

the

computer

from

the field

form.

Fig.

2.

Example

of field

forms

and

special

computer

program

used to

record

and

enter

data

from

rod

and

level.

25


34/98

readings

o elevation alues should be deferred.

Instead, hese

tasks

can all be performed

y the computer fter the rod readings

re entered.

If possible,

he computer program

should present

a display to the

person entering ata

that approximately atches the field form, to allow

the quick

detection f any

typing errors. To

help detect errors,the

computer an be programmed o check

for differences

n adjacent

elevation

alues exceeding level that would indicate rroneous

ata.

An even better check is to plot the elevation rofileat a scale that

will reveal any obviously rroneous ata values. Figure

2b shows the

display

of a data entry program that was used together ith the

field

form shown

in Figure 2a [31. This is, in fact,

an exact replica

of the

screen of the Apple Macintosh icrocomputer

hen running this particular

program, hen the typist has finished ntering

the rod reading at tape

position 501 and is about to enter the reading

for 501.25. The screen

is shown to indicate

ow the data entry task has been been streamlined

in one project.

In this example,

he tape

distance s shown in the

left-most.

column on the computer screen. The numbers match those of the field

forms, allowing the typist to easily see

the correspondence

etween the

position n the computer

screen and the field form. As each rod reading

is entered,

t is shown in the second

column on the left. The elevation

is computed nd shown

in the third column. At the same time, the

elevation s added to the plot shown in the upper-right

and corner of

the screen. Any erroneous ata points

can be seen as "glitches"

n the

plot,

so errors are easily

detected nd.corrected. (The two

boxes in

the

lower right corner were

used to store the changes in the levelling

instrument eight

and to control the flow of the program.)

Using microcomputers

ith "user-friendly" rograms ritten

specifically or entering rod and level data, a typist an enter about

1000 measures per hour (including hecking or

errors).

f) Computer

selection.

The computer elected

to process the

rod

and level data should

ideally have the ability to store the

profile data

permanently

n tape or disk, the ability to plot

profile, nd the

ability to

transmit iles to other

computers. An often

overlooked

consideration

hat should receive high priority

is the availability

f

the computer for

the project. A $500 "home" computer

hat is available

100% of the time

can be much more usefulthan a $100,000 ain-frame

computer shared by a large group that

is neither readily

available or

easily programmed.

3.3.2 TRRL Beam

Static Profilometer.

An automated eam

profilometer

uch as the TRRL

Beam can reducethe survey effort required

for profile

measurement

onsiderably. A two-mancrew can measure

elevations t 100 mm intervals n

two wheeltracks 20

m long in

26


35/98


36/98

wavelengths s long as 100 m when towed at 150 km/h, or as short as

0.3 m when towed at 21.6 km/h.

The APL Trailer

is the only high-speed rofilometer

hat has been

proven to measure IRI over a full range of roughness, ncluding ough

unpaved roads.

Although the APL Trailer can be used to measure IRI, it was

developed or other purposes y LCPC, and is routinely sed in Europe

for other applications. It is normally packaged ith special

instrumentation n one of two configurations: PL 72 for routine survey

work; and APL 25 for precision ork involving uality control and

acceptance, roject evaluation, nd research.

a) APL 72. The APL 72 system employs a powerful odern station

wagon as towing vehicle (sustaining 00,000 km per year for testing and

transfers) [4]. Single-wheeltrack ystems are the norm, although ual-

track systems (two APL Trailers, ne towed in each of the travelled

wheelpaths) ave been used. In normal survey usage in Europe the wheel

travels between the wheeltracks. The profile signal from the trailer,

the speed, distance travelled, nd manually entered event comments are

all recorded n magnetic tape in the towing vehicle. Data processing s

performed ater in the laboratory. Traditional rocessing ethods

classify the roughness n a ten-point cale of signal energy in three

wavelength ranges, i.e., 1

-

3.3 m/cycle, 3.3

-

13 m/cycle, and 13

-

40

m/cycle.

Site lengths need to

be selected in multiples

of 100 m, generally

with a minimum of 200 m and normal length for APL 72 of 1000 m.

Adequate allowance f approach length is necessary or the faster test

speed.

The APL 72 system

can be easily adapted to measure IRI, by

processing the recorded ata differently n the laboratory. The

manufacturer's nstructions hould be followed for details of test

operation. In the laboratory, he analog signals stored on the tape

recorder should be digitized sing standard microcomputer ardware (also

available s part of the APL 72 system, or available n different forms

from various commercial ources). Once the profile signal is digitized

and stored on a microcomputer, t can then be processed s any other

profile data, as described in Section 3.4. When this is done, the

APL 72 can be considered o be a Class 2 method for measuring IRI.

b)

APL

25. The APL 25 system consists of a towing vehicle and

only one trailer, and is used at a slower speed of 21.6 km/h [5]. A

different nstrumentation ystem is used. It digitizes he profile

signal and stores the numerical alues on digital cassette tape along

28


37/98

with

a single summary roughness ndex called

CAPL 25, calculated or

every 25 m of wheeltrack hat is covered.

Because

of the relatively ow towing speed

used with the APL

25,

it does not sense the longer wavelengths o which to the IRI numeric is

sensitive. Thus, the APL 25 cannot be used to directly easure IRI

without introducing ias. It can, however, e used to estimate

IRI

through the use of experimentally

erived regression

quations hat

relate IRI to other numerics that can be measured by the APL 25. This

approach ould qualify as a Class 3

method. It should not be expected

to be as accurate s the direct measurement hat can be made with the

APL 72 or with other systems sing the APL Trailer at higher speeds,

however.

Therefore, he APL 25 data collection

ystem is not the system

of choice

for measuring RI with the APL Trailer.

3.3.4 K.

J. Law Inertial rofiloseters.

These profilometers,

manufactured

y K. J. Law Engineers, nc. in the United States,

are

modern versions of the original GMR-type nertial rofilometer,

developed

n the 1960's [6]. The profilometer

s an instrumented an

that measures profile in both wheeltracks s it is driven along the

road. Vertical accelerometers rovide the inertial eference. The

distance to the road surface is sensed, riginally y mechanical

follower heels, but more recently ith non-contacting ensors (optical

or acoustic, epending n the model). The accelerometer ignals re

double-integrated

o determine he position f the profilometer ody.

When this position is added to

the road-follower osition ignal, the

profile is obtained.

The original profilometers sed analog electronics o perform the

double-integration nd other processing,

nd the operator as required

to maintain constant travel speed during easurement.

In the late

1970's, the design was upgradedto replace

the analog processing

ith

digital

methods.

With the conversion o digital

methods, new

computation rocedure as

developed o make

the profile

measurement

independent f speed. This allows

the profilometer

o be operated ith

greater ease in traffic.

In addition to measuring he road

profile, these profilometers

routinely

alculate ummary

statistics ssociated ith

quarter-car

simulations. Originally,

he BPR Roughometer

imulation

as used. In

1979, the QCS

model used for the IRI was added

to these profilometers,

and has been in use since that time. Thus, both

models have the IRI

measuring apability

uilt in and automated, nd can be considered s

Class 2 methodsat this time. Neither model has been validated gainst

rod and level yet, although

alidation or a wide

variety of paved road

types is underway [3].

Two versions of a profilometer re currently vailable:

29


38/98

a)

Model 69ODNC Road Profiloneter.

This version is the more

expensive nd offers the greater capability. It includes van, the

full instrumentation eeded to measure profiles in both wheeltracks, n

on-board minicomputer, 9-track digital tape system, and various

software ptions for computing umerous profile numerics (including

IRI). The road-following eight is detected by a noncontacting ensor

using a visible light beam, replacing he mechanical ollower wheels

used in earlier versions.

The software that calculates he IRI type of roughness as

developed uring the NCHRP 1-18 project [2], and is called the Maysmeter

simulation. It differs from the IRI in that it is computed from both

wheeltracks emulating passenger car with an installed roadmeter)

rather than the single-track RI. Some of these profilometers an

measure the roughness f the wheeltracks eparately s required for the

IRI; if not, the software can be enhanced readily by the manufacturer.

When obtained from the Maysmeter simulation, he IRI is reported ith

units of inches/mile, ather than

m/km (1 m/km = 63.36 in/mi).

The performance f the Model 69ODNC has not yet been validated

against

a static method for measurement f the IRI. The earlier designs

with mechanical follower heels were validated p to roughness evels of

about 3 m/km on the IRI scale in NCHRP Report 228 [2]. With the

noncontacting ensors, peration to higher roughness evels should be

possible. Three Model 69ODNC systems participated n the 1984 Road

Profilometer eeting in Ann Arbor, and validation f the profilometer

will be provided from that study [31. The Model 69ODNC has not been

tested on unpaved roads, and is not likely to be tested soon, since

there is little interest in measuring the roughness f unpaved roads at

the present time in the United States.

b) Model 8300 Roughness urveyor.

The Model 8300 is a single-

track profilometric nstrument esigned specifically o measure IRI. In

order to minimize its cost,

the instrumentation s used to provide

an

internal profile signal only as input to the IRI calculations, hereby

eliminating he need for many of the expensive omputer nd recording

components ncluded in the Model 69ODNC. Although the system provides

IRI roughness s the default, other roughness ndices can be obtained s

options

from the manufacturer.

The Model 8300 utilizes a bumper-mounted nstrumentation ackage

containing

n ultrasonic oad-follower

ystem and a vertical

accelerometer. The system can be mounted on most passenger ars. It

has not

yet been validated or measurement

f the IRI, but did

participate n the 1984 Road Profilometer eeting; hence, information n

its validity (on paved roads) is expected from that study [3].

30


39/98


40/98

V=

(Y -Y) /

dx = slope input (8)

and

Zj'=

Z from previous osition,

=1,4

(9)

and sij and p are coefficients hat are fixed for a given sample

interval, x. Thus, Equations - 7 are solved for each position long

the wheeltrack. After they are solved for one position, qn. 9 is used

to reset the values of Z1 , Z

2

, Z

3

1

, and Z

4

' for the next position.

Also for each position, he rectified

lope

(RS)

f the filtered rofile

is computed s:

RSi = IZ3 Zli (10)

The IRI statistic s the average of the RS variable over the length of

the site. Thus after the above equations ave been solved for all

profile points, the IRI is calculated s:

IRI 1 t RSi

(11)

The above procedure s valid

for any sample interval

etween

dx-.25 m and dx=.61 m (2.0 ft). For shorter sample intervals, he

additional tep of smoothing he profile ith an average value is

recommended o better represent he way in which the tire of a vehicle

envelops the ground. The baselength or averaging s 0.25 m long. The

IRI can then be calculated n either of two ways:

1) The elevation oints falling ithin each .25 m of length ay be

averaged o obtain an equivalent rofile point for the .25 m

interval. Then the IRI is calculated rom the above equations

based on a .25 m interval sing the coefficients or the .25 m

interval.

2) A "moving average" s obtained s the average of all points

falling ithin a .25 m interval entered n the profile elevation

point. Then the IRI is calculated y solving the equations or

each averaged oint using coefficients n the equations

appropriate or the smaller interval.

The algorithm sed in the example computer rogram listed in

Figure

3 in Section 3.4.2 is validfor any baselength ver the

range 10

- 610 mm. When dx is less then 0.25 m, it applies the proper moving

average to the input.

32


41/98


42/98


43/98

Lines

1260

1360 computethe

slope

input from

the entered

elevation oints.

The

Y array is a buffer

used

for temporary

torage

of

up

to 26 profile

points. Only

the first

K elements

are ever used,

however.

Thus, when DX is

0.25 m or

greater, hich will

be the

case for

most

applications

here profile

is

measured manually,

=2

and only the

first

two elementsin

the Y array

are needed. For

very short

sample

intervals,

owever,the

Y buffer

is needed

for the moving

average.

When

DX

=

.01 m, then

all 26

elementsin

the Y

buffer

are used.

Lines

1380

-

1490

are straightforward

translations

of Eqs. 4

-

10.

A major

change

that is

recommended o

make

the programmore

practical

s to provide

for reading the

measured

profile

from disk or

tape. Since

file

structures re specific

to different

achines,

he

example program

does

not do

this, but instead requires

hat

the user

enter

each

profile elevation

n sequence.

Lines

1160, 1170, 1280,

and

1290

can be replaced

ith equivalents

hat read

data

from stored

files.

Details concerning

the characteristics

f the

reference

nd this

particular

omputation ethod are readilyavailable [1, 2].

3.4.3 Tables of

coefficients

or the

IBI equations.

The

coefficients

o

be used in Eqns.

4 -

7 and in the example

IRI

computation

rogram depend

on

the interval t

which

the elevation

measurements

re

obtained.

Table 2

provides

the coefficient

alues

for

the commonly-used

ntervals hat are

convenient

or manual

measurement

of profile. When

an interval

is

used that is not

covered

in the table,

then

the coefficients

an be computed

sing

the algorithm

isted in

Figure

4 in Section

3.4.4.

3.4.4 Program

for computing

oefficients

or the IRI equations.

Coefficients or use in Eqns. 4

-

7 can be determined or any profile

interval

by using the computer

program

listed in

Figure 4.

The language

is BASIC,which

was discussed

in Section3.4.2.

The details

of the

vehicle

simulation

re coveredelsewhere

[1],

so only

the actual

equations

sed

in the programare included

ere.

The coefficients

sed in

Eqs. 4 -

7 are derived from

the dynamic

properties

f the reference

ehicle

model.

These dynamic

properties

re

described

y four

differential

quations,

hich

have the matrix

form:

dz(t)/dt

=

A

* z(t)

+

B

*

y(t)

(12)

where z is a vector containingthe four Z variables f Eqs.

I - 7;

A is

a 4 x 4 matrix

that

describes

the dynamics

of the model;

B

is a 4 x 1

vector

that describes ow

the profile

interacts

ith the

vehicle;

and

y(t) is

the profile

input,

as perceived

y a

moving vehicle.

These

matrices

are defined as:

35


44/98

Table 2.

Coefficients for the IRI Equations.

dx=

50 n, dt= 00225

sec

.9998452

2.235208E-03

1.062545E-04 1.476399E-05

4.858894E-05

ST

= -.1352583

.9870245

7.098568E-02

1.292695E-02 PR

6.427258E-02

1.030173E-03 9.842664E-05

.9882941

2.143501E-03 1.067582E-02

.8983268 8.617964E-02

-10.2297

.9031446

9.331372

dx = 100 mm, dt= .0045 sec

.9994014 4.442351E-03 2.188854E-04 5.72179E-05 3.793992E-04

ST = -. 2570548

.975036

7.966216E-03

2.458427E-02 PR

.2490886

3.960378E-03

3.814527E-04

.9548048

4.055587E-03

4.123478E-02

1.687312

.1638951

-19.34264

.7948701

17.65532

dx = 152.4 mm 0.50 ft), dt = .006858 sec

.9986576

6.727609E-03 3.30789E-05

1.281116E-04

1 309621E-03

ST = -. 3717946

.9634164

-. 1859178

3.527427E-02

PR

=

5577123

8.791381E-03 8.540772E-04

.8992078

5.787373E-03

9.200091E-02

2.388208

.2351618 -27.58257

.6728373

25.19436

dx =

166.7 mm,

dt = .0075015 sec

.9984089 7.346592E-03

-1.096989E-04

1.516632E-04

1.70055E-03

ST = -.4010374

.9603959

-.2592032 3.790333E-02 PR

= .6602406

1.038282E-02

1.011088E-03

.8808076 6.209313E-03

.1088096

2.556328 .2526888 -29.58754 .6385015 27.03121

dx = 200 mm, dt

= .009 sec

.9977588 8.780606E-03

-6.436089E-04

2.127641E-04

2.885245E-03

ST =

-.4660258

Road Roughness Measurements

Documents

Transcript of Road Roughness Measurements