Chemical Engineering November 2010 - 52

Feature Report
PC
Proposed addition
End users
fidelity model, the result should see
downstream temperatures go up (or
down) as they would in the real process.
With a low-fidelity model, the
downstream temperature may be explicitly
specified as the output of a heat
exchanger. Changing the valve position
in such a model will not affect the
downstream temperature - although
it may result in a calculated heat duty
that is not feasible for the process.
It's not necessary to model an entire
facility. Look for units where a benefit
can be derived by connecting operations
data with a model. Consider
starting with units that are easier to
model, such as utilities.
Regardless of the reason you plan
to connect your simulator with DCS
data, be sure to validate the model
with historical data. Don't simply assume
that your model is accurate.
Connect DCS data with the simulator
model. You'll need to determine
the best method to obtain data from
the DCS; try to avoid manual entry,
which is tedious and error-prone. Arguably
the easiest data interface to set
up depends on the ability of both software
packages to support data transfer
with Excel. In this setup, have your
DCS transfer data to Excel, and have
your simulator receive data from Excel.
Consider using Excel Visual Basic for
Applications (VBA) or scripting languages
from your DCS or simulator to
automate the process. If this is an automated
process, you will need to define
the conditions used to send data to the
simulator and trigger the calculations.
Other options for data transfer
include open process control (OPC),
a direct Microsoft component-object
model (COM) interface between programs,
or a custom program to handle
moving data from one source to
the other as needed.
Implement additional features to
your project. Depending on your intended
application, you might want to
reconcile the model with sensor data,
send calculated values to the control
system, build a custom user interface,
or even develop a custom program to
coordinate the interaction. Even simple
implementations with Excel can
use a custom interface, with a drawing
similar to the supervisory control and
data acquisition (SCADA) displays and
PCV 3
PC
Vaporizer
6.7 MPa
PCV 2
PC
5.6 MPa
PCV 1
push buttons to send data to and from
the simulator. More complicated implementations
might involve modification
of the operator panel displays to include
simulation results, or custom programs
with streamlined user interfaces.
Determining capabilities
Leading process simulators have the
ability to interface to data via Excel, Microsoft
COM interface (allowing other
programs to automate the simulator),
and even OPC (allowing direct connection
to DCS/data historian). It is also
possible to use programming to send
keystrokes to your simulator if it does
not support links to other programs.
Most of the modern offerings for
control system and data-historian
software support connections via OPC
and Excel. Even older versions often
support connections via Excel.
Who does the implementation?
Although it's possible for one person to
implement such a data connection, in
most cases a team of two to four people
will be involved. The process simulation
models will need to be developed
by someone well skilled in simulation.
Typically, a few years' experience in
modeling will be required to be able
to develop a high-fidelity simulation.
A deep understanding of the control
system is not required. Good skill
with computers is a must for the team
member who will be connecting information
between the simulator and
DCS. Depending on the implementation,
some degree of programming
knowledge may be necessary.
Expected economic benefits
The economic benefits vary by process,
and are sometimes difficult to
36 ChemiCal engineering www.Che.Com november 2010
quantify. How much money could you
save if you had better control of your
process? What is the value of tuning
the control system in a virtual plant
rather than in a real plant? How would
your bottom line improve if your DCS
registered alarm conditions further in
advance, based on predictive models?
The real economic potential lies in
your answers to these questions.
Conclusions
Improvements to computer hardware,
software, and data-interface standards
now make it possible to combine
process simulators with control system
data for small- and medium-sized
projects. This allows engineers to have
better models for their processes, and
better control of their processes. As
technology continues to improve, we
should expect to see more interfaces
between these systems.
■
Edited by Rebekkah Marshall
Authors
David Hill is the manager of
technical support for Chemstations
(11490 Westheimer
RD Suite 900, Houston, TX
77042; Phone: 713-978-7700;
Email: davidh@chemstations.
com), developer of the Chemcad
simulator. He has 13 years
experience working technical
support for a process simulator
company. He holds a
B.S.Ch.E. from the University
of Houston and is a member of AIChE.
Kenichi Matsuoka is senior
system engineer at INPEX
Corp. (Tokyo; Email: kenichi.
matsuoka@inpex.co.jp) He is
currently working on a liquified
natural gas (LNG) project
in Indonesia as a process control
engineer. Prior to that, he
has had ten years of experience
as a process engineer for
LNG, liquified petroleum gas
(LPG) and ethylene process at
JGC Corp. and more than 15 years experience as
advanced process control engineer for Yamatake
Corp. He holds a B.S.Ch.E and M.S.Ch.E. from
Tokyo University of Agriculture and Technology.
PC
Figure 2.
Dynamic simulation,
using
historical data,
proved useful
for expansion
of this flow network
and the
development of
its control logic
http://www.Che.Com
Chemical Engineering November 2010

Table of Contents for the Digital Edition of Chemical Engineering November 2010
