1. BY ARC ADVISORY GROUP DECEMBER 2001
Trends in Process Field Devices
for Automation
Executive Overview ................................................................................ 3
New Field Measurements Needed .......................................................... 4
Soft Sensors Stage a Comeback ............................................................. 6
Innovative Uses of Ultrasonic Waves ....................................................... 8
Old Measurements Go Mainstream ...................................................... 10
Vertical Applications Key to Success of Model Based Systems.................. 11
Benefits and Drawbacks of Inferential Measurement .............................. 13
Enterprise & Automation Strategies for Industry Executives

2. ARC Strategies • December 2001
NEW MEASUREMENTS
Real-time Corrosion Soft Sensors
Interface
Leak Detection
Pressure Flow
Model-based
Multiphase Flow TRADITIONAL MEASUREMENTS Vertical Solutions
Level Temperature
Analytical
Product Quality
Viscosity
Real-time Analytical Entrained Air
2 • ARCweb.com • Copyright © ARC Advisory Group

3. ARC Strategies • December 2001
Executive Overview
Demanding measurement applications are growing rapidly across all
industries. Suppliers recognizing the enormous opportunity are developing
new and enhanced measurements. New measurement solutions hitting the
market include viscosity, corrosion, and vertical solutions.
Soft sensors are a rapidly growing solution for measurement applications.
Soft sensors are the only answer for many applications where hard
measurements are unavailable. Today’s soft sensors are widely used in the
power industry to continuously monitor environmental dis- Despite the high level of
charges such as the emissions from combustion units to meet commonality of measurements in
federal, state, and local regulations. Soft sensors, however, are process systems between industries,
gaining acceptance across all the process industries in a wide va- there are many unique process
riety of applications. Obstacles still remain, however, including measurements needed in each.
issues such as measurement validation.
Ultrasonic technology is findings its way into virtually every process
measurement. Most commonly used to measure flow rate, level, and density,
ultrasonic techniques are now making inroads into temperature, and inter-
face detection applications. Not only is ultrasonic technology capable of
measuring these parameters, it often exceeds the performance of the current
mainstream technologies.
Corrosion measurement has been employed for decades to monitor corrosion
rates of critical pipelines. Old corrosion instruments relied on off-line sample
techniques by measuring the reduction of weight of disposable coupons. The
impracticality of making off-line measurements limited its use until Intercorr
International met the challenge by releasing the first on-line real-time
corrosion monitor to measure localized corrosion of pipelines and equip-
ment.
Despite the high level of commonality of measurements in process systems
between industries, there are many unique process measurements needed in
each. Direct instrumentation, process analyzers or soft sensor sensors cannot
measure many production specific parameters. Intangible parameters can be
very complex properties of a substance such as taste, creaminess, color or
softness, which can only be subjectively defined. Comprehensive model
based process applications are meeting the demand for industry specific
Copyright © ARC Advisory Group • ARCweb.com • 3

4. ARC Strategies • December 2001
solutions. When adopting these solutions, however, drawbacks of the solu-
tion must be considered from the onset to have a successful implementation.
New Field Measurements Needed
Traditional process control field devices commonly known as the “four
horsemen” provide four main categories of measurements: pressure,
temperature, flow, and level. New faster response online analytical devices
are providing near real-time measurements and increasingly have been used
as a mainstream process measurement however, as broad as these measure-
ments are, collectively they still cannot provide the total insight needed for
today’s complex processes.
Demanding applications are growing rapidly in all industries particu-
larly in the emerging biotech industry, which now has over 4000
companies world-wide up from just a few hundred less than 3 years
ago. This unprecedented growth offers an enormous opportunity for
suppliers to develop new and enhanced measurements beyond the
traditional four and analytical. Increasing cost reduction targets oc-
curring in all industries is compounding the situation and rendering
once satisfactory measurements and processes obsolete in the quest to
increase margins. Additionally, recent biological terrorist events have
renewed the push for faster development of vaccines and new drugs
to combat the suddenly heightened worldwide threat.
Pressure, Temperature, Flow,
and Level Asset management solutions continue to be a hot topic due to its po-
tential savings in time and labor in the day-to-day operation of the plant.
Coined Plant Asset Management (PAM), asset management solutions have
become a critical strategy in many plants that are trying to squeeze that extra
bit of production from their plants. PAM applications have progressed sig-
nificantly in recent years. Once used strictly to analyze trends using the
automation system’s historian and identify abnormal trends, PAM ap-
plications have expanded to embedding artificial intelligent model based
solutions to obtain more accurate health status and, in some cases, extend the
useful life of the measurement by quantifying the problems, correcting the
output, and validating the signal while sending a low level alert highlighting
potential issues with the unit.
4 • ARCweb.com • Copyright © ARC Advisory Group

5. ARC Strategies • December 2001
Field device suppliers are taking up the slack by offering new innovative
real-time hard sensors coupled with intelligent software to provide the
measurement of desired property. More sophisticated field device solutions
are expected with the increasing adoption of fieldbus, which allows easy
access to every online measurement on the bus. The rich data set offered by
Fieldbus, coupled with increased intelligence, will enable the local field de-
vice to use model based techniques to infer the measurement of new parame-
ters previously unavailable.
Field devices are following a natural product life cycle evolution as they ex-
pand functionally. For example, ultrasonic level principle is being used in an
increasing number of hard measurements ranging from flow, and tempera-
ture, to leak detection and nondestructive vessel testing. Field device
suppliers are breaking away from traditional roles and using
technology to broaden their appeal in the ever-tightening market.
By developing products to go beyond their traditional roles, vi-
sionary field device suppliers are creating a strong launch pad
into future markets and other lines of business. Expect to see in-
creasing growth in combined hardware and the software
solutions in this area.
Needed Measurements
Despite the improvements in analytical measurements, the lack
Viscosity
of appropriate on-line instrumentation and reliability of on-line
Entrained Air
instruments have forced users to depend on laboratory analyses
resulting in infrequent and irregular and long analysis delays. Consistency
Laboratory analyses may also be subject to reliability problems. Measurement Validation
On-line sensors may be available but suffer from long measure- Leak Detection
ment delays (e.g. gas chromatographs) or subject to factors that Multi-Phase Flow
affect the reliability of the sensor (e.g. drifts and fouling). In Real-time Analytical
either case, automatic control and optimization schemes cannot Product Quality
be implemented and, as a result, process performance is de-
Bio Cell Production
graded.
Corrosion
Left unattended, the problems of delays and unreliable meas- Pipe Coating
urement manifest themselves into critical problems that affect
product quality and availability of production. Critical problems
such as these can quickly devour the last bit of shrinking margin to drive en-
terprises into the red. Fortunately, field device suppliers have recognized
this need and are working hard to provide new online real-time measure-
ments. Progress in many areas, however, continues to be slow compared to
demand.
Copyright © ARC Advisory Group • ARCweb.com • 5

6. ARC Strategies • December 2001
Soft Sensors Stage a Comeback
Automatic control enables and facilitates consistent production. To achieve
these goals, however, controllers require regular and reliable measurements
at the appropriate frequency. Some slow reacting processes can do with in-
frequent sample measurements but most demand continuous attention to
maintain a narrow band of uncertainty for consistent product quality.
Processes are tough on instru-
Technology Function Inferential Measurement
mentation. Process pressure
Ultrasonic Flow & Level Transit Time & Frequency
shift
and temperature can exceed
limits of field devices, high am-
Magnetic Flow Milli -Volts
bient temperature can cook
Orifice/Venturi/Pitot/Nozzle Flow Differential Pressure electronics, and the inaccessibil-
Radar Level Frequency ity of the measurement location
RTD Temperature Resistance may make physical measure-
ments impractical. When hard
Turbine & PD Flow Rotor Speed
measurements are unavailable,
Common Inferential Field Devices
soft sensors may be the only
answer. Soft sensors, also known as inferential model-based analyzers, are
programs that primarily use secondary measurements combined with mod-
eling methods to develop process input-output variable relationships. Soft
sensors use mathematics to estimate either the value of product properties
measured in the laboratory, properties measured infrequently on-line or that
take an inordinate amount of time to measure, or properties not measured at
all. Soft sensors utilize easy to measure process variables such as pressure,
flow, density, or temperature to estimate the value of an important property
and make the data immediately available to operators and control systems.
Because many of the desired primary measurements are directly related to
one or more secondary variables, measurement of previously unavailable
parameters is readily possible by using secondary variables to infer the state
of a quality or primary variable. For example, liquid composition can be de-
fined by pressure and temperature, concentration by density, and biomass
growth can be linked to CO2 evolution and feed rate.
The process of inferential measurement is quite common and has been in use
by suppliers for decades to measure flow and level. For example, magmeters
measure electrical potential, orifices measure differential pressure, and ultra-
sonic flowmeters measure transit time for both flow and level measurements.
6 • ARCweb.com • Copyright © ARC Advisory Group

7. ARC Strategies • December 2001
What is different in today’s solution is the complex modeling techniques ap-
plied to standard measurements.
By modeling the relationship between a primary output and secondary out-
put the response time can approach that of a typical online process meas-
urement. Suppliers such as Pavilion, Aspen Technology, Optima PSE, and
Shell Global Systems are investing in this approach to provide soft sensors
to replace analytical measurements. Instead of waiting 15-30 minutes for a
gas chromatograph to complete its analysis, the inferential measurement
system could be returning estimates of compositions every few seconds.
Field device suppliers are also investigating the opportunities for these tech-
Emissions
niques to provide enhanced performance such as increasing the low range Measurement
measurement of vortex flow meters and modeling of fluid mixtures to
enhance concentration measurements. Increasing sophistication of modeling
techniques now make them sufficiently accurate to use inferential measure-
ment as direct feedback for automatic control. Despite their performance,
many users continue to use inferential soft sensor measurement in conjunc-
tion with traditional off line techniques to optimize measurements.
Soft sensors are widely used in the power industry to continuously monitor
environmental discharges such as emissions from combustion units to meet
federal, state, and local regulations. Typically software solutions extract data
from the process historian to build the model. Once built, the soft sensor will
monitor and predict air emissions in real-time. Pavilion's Soft Sensor tech-
nology for air emissions, CEM Software, is typical of these solutions.
Working with the EPA, Pavilion has installed 250 software solutions
throughout the US. The documentation generated is a necessary step in the
approval process that, in many cases, must be replicated for each local utility.
Using similar modeling techniques suppliers are developing measurement
validation by employing rules at the lowest measurement level. To accom-
plish this, automated program models are applied to key process variables
by taking advantage of redundant information contained in process instru-
mentation. Predicted value is then compared against the actual measured
key variable to determine if it is operating correctly or not. If the input is
suspect, an alarm is triggered; signaling for maintenance attention and con-
trol action can be programmed for application specific action while
preventing the use of suspect data.
Copyright © ARC Advisory Group • ARCweb.com • 7

8. ARC Strategies • December 2001
Despite their benefits, operators typically still do not totally trust the reliabil-
ity of soft sensors. Many users point to the inability to quickly test, check,
and calibrate a soft sensor solution that is standard operating procedure rou-
tinely performed on direct instrumentation. Previous initial solutions have
also produced inconsistent readings and because instrumentation technicians
grew up with physical measurements, they are biased toward hard meas-
urements and suspect the readings of soft. Bad memories are hard to erase
and for soft sensors to continue their move into mainstream applications, ad-
ditional enhancements are needed to allow measurement checks to ease the
fears of operators.
Innovative Uses of Ultrasonic Waves
Ultrasonic technology is findings its way into virtually every process meas-
urement. Traditionally used to measure flow rate, level, and density,
ultrasonic techniques are now being applied to measure temperature, viscos-
ity, and interface detection. Not only is ultrasonic technology capable of
measuring these parameters, it often exceeds the performance of the current
mainstream technologies such as RTD temperature probes and vibration type
viscometers.
Temperature can be measured with a multitude of sensors such as thermo-
couples, infrared, thermistors, and RTDs. Caldon, a niche ultrasonic
supplier, utilizes ultrasonic technology in its LEFM meter to accurately
measure feedwater temperature in nuclear power plants. Sound velocity of
a fluid is determined by the ratio of its compressibility and its density. For
specific liquids such as pure water, compressibility and density are linked
to various characteristics like temperature and pressure. As a result of this
interdependence, it is quite easy, with the right model, to infer temperature
Ultrasonic from the sound velocity through the fluid and pressure.
Technology
Using mean pulsed time from its meter combined with an external pres-
sure input Caldon’s LEFM meter precisely calculates the temperature of high
temperature feed water exceeding 450 deg Fahrenheit. Sound velocity can be
determined within 1 part in 1000 yielding an accuracy of +/- .3 percent,
which is a significant improvement over the typical RTD performance of +/-
2.6 percent in this application. Other suppliers are investigating the use of
high speed 22 bit software in a multi-input transmitter to approach fast re-
8 • ARCweb.com • Copyright © ARC Advisory Group

9. ARC Strategies • December 2001
sponse highly accurate temperature measurement to within .0005 degrees
Celsius for the semiconductor industry in the production of the next genera-
tion of processors.
Viscosity can be determined in a similar way by measuring the attenuation
of ultrasonic signals as they pass through the viscous liquid. Despite the
simplicity of the concept for both temperature and viscosity measurement a
number of parameters such as distance of travel, sonic velocity, pressure
measurement and the allowance for transmission delays in the liquid must
be accurately determined. Ultrasonic technology is one of the fastest re-
sponding measurement methods, which makes this technology ideal for
determining the interface between two different fluids transported through
Nuclear Power
pipelines carrying finished petroleum products coast to coast. Applications
Caldon’s Sountrack measures viscosity, density and interface using one or
multiple ultrasonic transducers mounted in a simple ring that is simply in-
stalled between a pair of process pipe flanges. Interface can be detected
within one second and kinematic viscosity can be determined within +/-.04
centistokes (cSt).
Caldon, best known for pioneering approval from the Nuclear Regulatory
Commission (NRC) for power up rates in nuclear power plants based on im-
proved feedwater flow measurement, introduced a new eight-path transit
time flowmeter for custody transfer applications that may prove to be one of
the most accurate meters in the world. Multipath ultrasonic technology
gained user mass acceptance in the natural gas industry after approval by the
American Gas Association in 1998. Krohne, Daniels, FMC and Instromet all
have released products utilizing multipath technology. The combination of
no moving parts with high reliability and high performance is hard to ignore.
Ultrasonic technology is increasingly the technology of choice for leak detec-
tion of long pipelines. Krohne, Panametrics, and Controlotron offer leak
detection systems. Caldon’s new LineWatch dynamic leak detection systems
is a prime example of these solutions, which can be used on existing pipe-
lines to cost-effectively comply with environmental guidelines and regula-
tions. LineWatch can determine the location of a sudden break or leak in the
pipeline by measuring and timing the pressure waves created by such an
event. Caldon also manufactures a static leak detection system used princi-
pally for jet fuel delivery lines in commercial and military airports, which
requires significantly less down time than other technologies.
Copyright © ARC Advisory Group • ARCweb.com • 9

10. ARC Strategies • December 2001
Old Measurements Go Mainstream
Corrosion is not new and has been employed for decades to measure corro-
sion rates of critical pipelines, the main artery of the oil & gas industry.
Traditional instruments, however, relied on the reduction of weight of dis-
posable coupons. The impracticality of making this off-line measurement
limited its use to the detriment of equipment and the safety of the personnel
running the plant. The increasing popularity of equipment condition moni-
toring and plant asset management indirectly pushed corrosion monitoring
into the spotlight. Intercorr International met the challenge by releasing the
first on-line real-time corrosion monitor to measure general corrosion and,
more importantly, the localized corrosion that represents the bulk of corro-
sion failures.
Intercorr’s SmartCET is a cost-effective electrochemical multi-technique
corrosion monitoring instrument that measures corrosion using tech-
niques such as electrochemical noise, Linear Polarization Resistance (LPR),
Harmonic Distortion (HD), Galvanic Current (GC), and Hydrogen Per-
meation Current (HP). The system consists of a wetted 2- or 3- electrode
probe with a companion transmitter/computer to compute corrosion rate.
By releasing the first online monitor, Intercorr has provided users a means
to perform low cost corrosion measurements with information immedi-
Real-Time Corrosion
Monitoring ately available whether onshore or offshore, local or remote. SmartCET
also provides users the opportunity for direct interfacing between corro-
sion, chemical injection, process control, and maintenance management sys-
tems to allow personnel to take appropriate remedial action
Yamatake’s Smart Plus One gas chromatograph is the world’s smallest natu-
ral gas analyzer that fits in the palm of your hand. Previous products needed
large amounts of real estate and were typically the size of a small refrigera-
tor. SmartPlus is one of the first introductions of the new trend of
miniaturized analyzer on a chip. The small size of future field devices will
expand measurement possibilities and include variables that were previously
inconceivable.
10 • ARCweb.com • Copyright © ARC Advisory Group

11. ARC Strategies • December 2001
Vertical Applications Key to Success of
Model Based Systems
Despite the high level of commonality of measurements in process systems
between industries, there are many unique process measurements needed in
each. Some of these go unattended and left for the operator to control the
best they can with limited knowledge by juggling seemingly contradictory
needs. The resulting inefficiency wastes raw materials, finished products
and time that can never be recovered.
Direct instrumentation, process analyzers, and soft sensor sensors cannot
measure many production specific parameters. These intangible parameters
can be found in virtually every industrial segment with many concentrated
in the oil & gas industry. Intangible parameters can be very complex proper-
ties of a substance such as taste, creaminess, color or softness, which can only
be subjectively defined. Although an intangible parameter can be tied to
physical properties of the substance, there is typically no known definition.
Typically, intangible parameters must be measured without having any de-
tailed knowledge of the parameter.
Comprehensive model based process applications are meeting the demand to
measure intangible parameters with industry specific solutions. Using infer-
ential modeling techniques, industry specific systems takes soft sensor
technology to a much higher level that many times is a semi-customized so-
lution for an industrial process or particular plant.
Final product quality isn’t the only quality assurance necessary to meet
customer standards and regulatory requirements, interim product qual-
ity in the production is increasingly needed to allow time to take action
on off-spec products. With the high production rates mandated by effi-
cient production, time is of the essence, making online real-time Model Based Software
continuous measurements necessary.
Complex non-linear models can be built using neural networks, genetic pro-
gramming, partial least squares, algebraic, hybrid neural, and regression
technologies. PSE Optima offers Product Quality Estimator solutions an em-
pirical modeling tool to develop on-line models for estimating product
quality from process measurements in real-time or with engineering correla-
tions or user engineering correlations. This approach has proven to be a very
effective means for on-line estimation of polymer properties. Typically, his-
Copyright © ARC Advisory Group • ARCweb.com • 11

12. ARC Strategies • December 2001
torical process data is used to develop models of key process and quality
variables to be placed on-line. These models are updated online with lab re-
sults (or analyzers if available). Virtually all performance optimization
companies are offering non-linear model solutions today for use in oil & gas,
refining, food processing, petrochemical, power, chemicals, mining, and pulp
& paper industries.
Specific process parameters and quality measure-
Suppliers Product
ments inferred on-line are very dependent on the
AspenTech AspenIQ
specific process and product requirements. Today’s
Gensym NeurOn-line excellent examples of solutions include: Melt index,
Pavilion Technologies Property Predictor Tray Temperature, Mooney, Naphtha 95% point,
Shell Global Solutions RQE Catalyst Mileage, Polymer residence time and Sheet
color and opacity.
Honeywell Profit Sensor
Emerson Process Intelligent Sensor Toolkit Because these parameters can drastically change
Management
with process operational changes it may be neces-
Invensys/Pacific FACTNET sary to use an adaptive inferential measurement
Simulations
scheme by augmenting the estimated parameter
Typical Soft Sensor Solutions
with slow off-line measurements. This will correlate
values generated by the parameter estimator with
measured values and compensate the measurement model to more represen-
tative values. In this adaptive framework, the slower measured primary
output is fed to the faster parameter estimator, which updates the inferential
measurement model when available.
An excellent example of a model based vertical application solution is ABB’s
OptimizeIT Well Monitoring System for estimating the flow rates of oil, gas,
and water from individual wells in an oil field. OptimizeIT Well may be
used as a software multiphase flowmeter, as a Reliability Tool, and as a Pro-
duction Allocation System. OptimizeIT Well can also detect water or gas
breakthrough, predict the production rate with high accuracy, provide re-
dundancy, and validate specific measurements from the well’s
instrumentation.
OptimizeIT Well is based on data from standard well instrumentation that is
designed to handle any configuration of well instrumentation, making the
solution easily transferable from region to region and, therefore, suitable for
any user around the globe. Calculations are based on a mechanistic model
for multiphase flow and a full, compositional model of the fluid. ABB’s ap-
proach to calculating all fluid properties using the full model is superior to
12 • ARCweb.com • Copyright © ARC Advisory Group

13. ARC Strategies • December 2001
methods based on interpolation of tabulated data most commonly used in
commercially available multiphase fluid simulation programs.
Model-based measurement products typically come from software optimiza-
tion companies that take reams of data to predict the measurement without
direct knowledge of the sensor making the measurement. Direct involve-
ment of sensor manufacturers such as ABB helps to expand this technique
further. With collaboration of sensor and software suppliers, development of
new innovative measurements may result.
Benefits and Drawbacks of Inferential
Measurement
Despite the relatively simple concept of inferential measurement and the
straightforward inferential model building techniques given the abundance
of computer-aided tools, there are a number of important things to bear in
mind when adopting inferential model based solutions.
Benefits
Here are some of the benefits that you can expect from successful inferential
measurement and control scheme:
• Faster return of information. Process upsets can be detected quicker and
remedial action can be taken before it is too late.
Validation Key For
• Increasing standard “shrink wrapped” vertical solutions reduce model- Continued Adoption
ing time and lowers implementation cost.
• The inferential estimates carry a fair degree of 'feedforward' in-
formation. For example, disturbances affecting tray temperatures in a
distillation column in a refinery may show up much later in the product
compositions but will immediately be reflected in the composition esti-
mates.
• Reducing human involvement in the control loop increases consistent
production.
• Better process control can be achieved.
Copyright © ARC Advisory Group • ARCweb.com • 13

14. ARC Strategies • December 2001
• Increased scope for process optimization leading to higher profit-ability.
Drawbacks
Drawbacks are few. Here are the major issues that need consideration for
successful implementation:
• High cost model based solutions will not replace on-line field de-vices
any time soon. They can, however, provide the only real alternative.
• Data conditioning can be time consuming - Delete data with abnormali-
ties since the old adage rubbish in, rubbish out applies.
• Selection of appropriate secondary inputs can be difficult – Select the
least number of variables to develop a model of sufficient accuracy.
• Estimator testing must be rigorous- test model over a broad range and
account for changing process conditions.
• Measurement mistrust from operators and the inability to calibrate or to
check the measurement. This, however, is becoming less of an issue now
with measurement validation enhanced solutions.
14 • ARCweb.com • Copyright © ARC Advisory Group

15. ARC Strategies • December 2001
Analyst: Wil Chin
Editor: David Clayton
Distribution: All MAS-P & MAS-H Clients
Acronym Reference: For a complete list of industry acronyms, refer to our web
page at www.arcweb.com/arcweb/Community/terms/indterms.htm
AI Artificial Intelligence EPA Environmental Protection Agency
ANSIAmerican National Standards Institute ERP Enterprise Resource Planning
API Application Program Interface GC Galvanic Current
APS Advanced Planning & Scheduling HD Harmonic Distortion
B2B Business-to-Business HP Hydrogen Permeation Current
B2C Business-to-Consumer MRP Materials Resource Planning
BPR Business Process Reengineering OLE Object Linking & Embedding
CAGRCompound Annual Growth Rate OPC OLE for Process Control
CAN Controller Area Network NRC Nuclear Regulatory Commission
CMM Collaborative Manufacturing PAM Plant Asset Management
Management PD Positive Displacement
CNC Computer Numeric Control RTD Resistance Temperature Detector
CPG Consumer Packaged Goods SCE Supply Chain Execution
CRM Customer Relationship Management SPC Statistical Process Control
EAI Enterprise Application Integration TMS Transportation Management System
EAM Enterprise Asset Management WAH Web Application Hosting
EC Electronic Commerce WMS Warehouse Management System
Founded in 1986, ARC Advisory Group is the leader in providing strategic plan-
ning and technology assessment services to leading manufacturing companies,
utilities, and global logistics providers, as well as to software and solution suppli-
ers worldwide. From Global 1000 companies to small start-up firms, ARC
provides the strategic knowledge needed to succeed in today’s technology driven
economy.
ARC Strategies is published monthly by ARC. All information in this report is pro-
prietary to and copyrighted by ARC. No part of it may be reproduced without
prior permission from ARC.
You can take advantage of ARC's extensive ongoing research plus experience of
our staff members through our Advisory Services. ARC’s Advisory Services are
specifically designed for executives responsible for developing strategies and
directions for their organizations. For subscription information, please call, fax, or
write to:
ARC Advisory Group, Three Allied Drive, Dedham, MA 02026 USA
Tel: 781-471-1000, Fax: 781-471-1100, Email: info@ARCweb.com
Visit our web page at ARCweb.com
Copyright © ARC Advisory Group • ARCweb.com • 15

16. Cambridge, U.K.
Düsseldorf, Germany
Munich, Germany
Hamburg, Germany
Tokyo, Japan
Bangalore, India
Boston, MA
Pittsburgh, PA
San Francisco, CA
Visit ARCweb.com for
complete contact information
Three Allied Drive • Dedham, MA 02026 USA • 781-471-1000 • Fax 781-471-1100