Titan 3.0 Quick Start

AFT Titan
Quick Start Guide
AFT Titan version 3.0

Compressible Flow Pipe System Modeling and
Optimization
Applied Flow
Technology
5/7/2003 2:31 PM Page 1 c:\aft manuals\t3\quickstart\titan3

quickstart 1.doc
Information in this document is subject to change without notice. No part of this Quick
Start Guide may be reproduced or transmitted in any form or by any means, electronic or
mechanical, for any purpose, without the express written permission of Applied Flow
Technology.
2003 Applied Flow Technology Corporation. All rights reserved.

Printed in the United States of America.
"AFT Titan", "AFT Arrow", "Applied Flow Technology", and the AFT logo are
trademarks of Applied Flow Technology Corporation.
Windows is a registered trademarks of Microsoft Corporation.

quickstart 1.doc
Contents
1. Introducing AFT Titan .................................................... 1
How AFT Titan works.............................................................................. 1
Analysis vs. design ................................................................................... 3
Analysis.............................................................................................. 3
Design ................................................................................................ 3
Cost-based optimization vs. cost estimating............................................. 4
AFT Titan design capabilities .................................................................. 4
Types of systems that can be optimized............................................ 4
Optimization parameters available..................................................... 5
Engineering assumptions in AFT Titan.................................................... 5
AFT Titan Primary Windows ................................................................... 5
Input windows.................................................................................... 6
Output windows ................................................................................. 6
Optimization terminology......................................................................... 7
Design variables ................................................................................. 7
Design constraints .............................................................................. 7
Objective function.............................................................................. 7
Continuous vs. discrete optimization ................................................. 8
Design variable linking ...................................................................... 8
Feasible and infeasible designs.......................................................... 8
2. Weight Optimization Example ....................................... 9

Topics covered.......................................................................................... 9
Required knowledge ................................................................................. 9
Model files.............................................................................................. 10
Step 1. Create the model......................................................................... 10
Summary .......................................................................................... 10
A. Layout model............................................................................... 11
B. Specify solution method .............................................................. 11
C. Select fluid................................................................................... 12
5/7/2003 2:31 PM Page iii c:\aft manuals\t3\quickstart\titan3

quickstart 1.doc
iv AFT Titan 3.0 Quick Start Guide
D. Enter data for tanks ..................................................................... 13

E. Enter control valve data............................................................... 14
F. Enter pipe data ............................................................................. 15
Step 2. Setup the optimization data ........................................................ 16
A. Create pipe size range set ............................................................ 16
B. Create a control valve constraint set ........................................... 19
Step 3. Apply optimization data ............................................................. 20
A. Apply optimization data to P1..................................................... 20
B. Apply optimization data to P2..................................................... 20
C. Apply optimization data to J2 (FCV) .......................................... 22
Step 4. Specify Optimization Control..................................................... 23
Step 5. Run the Optimization ................................................................. 24
Step 6. Review optimization results ....................................................... 25
Conclusions ............................................................................................ 26
3. Initial and Life Cycle Cost Optimization Example ..... 27

Topics covered........................................................................................ 27
Required knowledge ............................................................................... 27
Model files.............................................................................................. 28
Optimization goals.................................................................................. 28
Getting started......................................................................................... 28
Review Optimization Control................................................................. 29
Review databases.................................................................................... 31
Review pipe optimization setup ............................................................. 32
Pipe size range sets .......................................................................... 33
Pipe constraint sets........................................................................... 34
Pipe linking ...................................................................................... 35
Creating pipe size range sets and constraint sets ............................. 35
Review junction optimization setup ....................................................... 37
Optimizing systems with compressor/fans....................................... 39
Junction costs ................................................................................... 40
Junction constraint sets .................................................................... 40
Creating junction constraint sets...................................................... 40
5/7/2003 2:31 PM Page iv c:\aft manuals\t3\quickstart\titan3

quickstart 1.doc
Table of Contents v
Understanding the model........................................................................ 41

Running the scenarios and interpreting results....................................... 43
Scenario to minimize first cost......................................................... 43
Scenario to minimize life cycle cost for 5 years .............................. 45
Optimizing with compressor/fan curve data........................................... 48
Conclusions ............................................................................................ 49
4. Multiple Design Case Example ................................... 51

Topics covered........................................................................................ 51
Required knowledge ............................................................................... 51
Model files.............................................................................................. 52
Optimization goals.................................................................................. 52
Getting started......................................................................................... 52
Review model ......................................................................................... 53
How the dependent design case was created ................................... 55
Run the optimization .............................................................................. 59
Consider the results ................................................................................ 61
Optimize for eight pipe sizes .................................................................. 62
Run the optimization .............................................................................. 63
5. Other AFT Titan Capabilities ....................................... 65

Optimize with operating costs spread over multiple cases..................... 65
Vary recurring costs over time ............................................................... 65
Time value of money .............................................................................. 66
Working with different currencies.......................................................... 66
Costs vs. size........................................................................................... 66
Optimizing rectangular duct systems...................................................... 66
Compare VFD vs. FCV optimized systems............................................ 66
Maximum cost groups ............................................................................ 67
Network databases.................................................................................. 67
5/7/2003 2:31 PM Page v

c:\aft manuals\t3\quickstart\titan3 quickstart 1.doc
vi AFT Titan 3.0 Quick Start Guide
5/7/2003 2:31 PM Page vi c:\aft manuals\t3\quickstart\titan3

quickstart 1.doc
CHAPTER 1
Introducing AFT Titan
Welcome to AFT Titan 3.0, Applied Flow Technology's powerful

compressible flow pipe system optimization tool. With AFT Titan you
can automatically size all pipes or ducts in your system to minimize
monetary cost, weight, volume, or surface area. In addition, you can
concurrently size the compressors or fans and pipes to obtain the
absolute lowest cost system that satisfies your design requirements.
Finally, by accounting for non-recurring and recurring costs, you can
optimize pipe and duct systems to minimize life cycle costs over some
specified duration.
How AFT Titan works

AFT Titan consists of three basic elements: the Graphical Interface, the
Compressible Flow Solver, and the Optimization Engine (i.e., the
Optimizer). Figure 1.1 shows the relationship between the three.

quickstart 1.doc
2 AFT Titan 3.0 Quick Start Guide
Graphical Interface
Input Output
Yes
Compressible Flow No Converged

Solver on optimum?
Optim izer
Figure 1.1 AFT Titan main component flow chart
The Compressible Flow Solver obtains a balanced flow solution for a

specific pipe or duct system. The Optimizer then modifies the design,
and the Compressible Flow Solver evaluates the modified design. The
Optimizer continues this process until it is satisfied that no further
design improvements are possible. At this point, the Optimizer has
converged on a design, and the resulting optimized design is then sent
back to the Graphical Interface where it is displayed to the user.
The Compressible Flow Solver, which functions as the prime mover in
performing an engineering analysis (e.g., AFT Arrow), becomes a
subroutine called by the Optimizer. The Optimizer is the prime mover in
AFT Titan.
The Compressible Flow Solver in AFT Titan was derived from AFT
Arrow, a leading commercial compressible pipe flow analysis product
with many years of industrial use to its credit.
The optimization engine employed by AFT Titan uses state-of-the-art
optimization technology licensed from Vanderplaats Research and
Development, the leading company in optimization technology. VR&Ds
technology has been used for many years in engineering design, with
extensive use in structural finite element analysis.

quickstart 1.doc
Chapter 1 Introducing AFT Titan 3
Analysis vs. design
Analysis
Traditional piping system engineering has employed pipe flow analysis.
Engineering analysis is the process of using accepted calculation
methods to predict the behavior of a given system. These calculation
methods may be manual or automated in a computer program.
The weakness of analytical methods is that they require the specification
of the system before the methods are applied. Specifically, the pipe or
ducts sizes, compressor, fan, valve and other equipment must be
specified in order to perform the calculation.
However, when a new pipe system is being designed, these parameters
are not known. To use the analytical methods, the engineer must guess at
the pipe sizes and required equipment, perform the analysis, then modify
his or her original selections as necessary.
The analytical methods are used iteratively to arrive at a final design.
Design
A design oriented approach to piping system engineering would allow
the selection parameters to be variables. Rather than specifying pipe
diameters, the engineer solves for pipe diameters by specifying the flows
and pressures (and other design requirements) and selecting the
appropriate pipes which minimize the overall system cost.
Within certain limits engineers do this with traditional analytical
methods, but the number of design tradeoffs that can be considered is
limited and the tradeoffs considered are indirectly tied to cost.
AFT Titan offers a true design-oriented approach to piping system
engineering by using advanced optimization methods to evaluate
competing designs vs. cost and selecting the optimum design. The
analysis method (i.e., the Compressible Flow Solver) is called repeatedly
by the Optimizer in an effort to identify design improvements (i.e.,
improvements that reduce cost).
5/7/2003 2:31 PM Page 3

Cost-based optimization vs. cost estimating

AFT Titan is not a cost estimating tool. Rather, it offers a rational and
automated approach for comparing potential pipe or duct system designs
using the common denominator of cost. It is the pipe or duct system
design that AFT Titan obtains that is of immense value, not the process
of cost estimating.
AFT Titan design capabilities

AFT Titan can be used to optimize a wide variety of engineering
systems.
Types of systems that can be optimized

Open and closed (recirculating) systems
Network systems that branch or loop, with no limit on the number of
loops
Pressure-driven systems
Compressor or fan-driven systems, including multiple
compressor/fans in parallel or in series
Compressor/fans with variable speed, controlled discharge pressure
and controlled flow
Systems with pressure and/or flow control valves
Systems with valves closed and compressor/fans turned off
Heat transfer analysis and system energy balance
Systems with non-ideal gases
Systems that experience sonic choking, including multiple sonic
choking points
Systems with non-reacting flow stream mixing and user-defined
mixtures (with optional Chempak)
Systems with elevation changes or rotation such as in
turbomachinery

quickstart 1.doc
Optimization parameters available

Pipe size
Pipe velocity, pressure gradient, pressure, and flow
Compressor/fan pressure rise, proximity to BEP (Best Efficiency
Point), power, and others
Control valve pressure drop and open percentage
Engineering assumptions in AFT Titan

AFT Titan is based on the following fundamental fluid mechanics
assumptions:
Compressible flow
All gases are superheated
Steady-state conditions
One-dimensional flow
No chemical reactions
Supersonic flow does not exist in the system
AFT Titan Primary Windows

AFT Titan has five subordinate windows that work in an integrated
fashion. You work exclusively from one of these windows at all times.
For this reason they are referred to as primary windows.
Of the five primary windows, two are input windows, two are output
windows, and one displays both input and output information. Figure 1.2
shows the relationship between the primary windows.
5/7/2003 2:31 PM Page 5

Model Data
Visual Report
Workspace Output
Graph Results
Figure 1.2 Primary window workflow in AFT Titan
Input windows
The two windows that function exclusively as input windows are the
Workspace window and the Model Data window. These two windows,
one graphical and the other text-based, work together to process model
input data with immense flexibility. The tools provided in these two
windows allow you to model a large variety of pipe networks.
The Visual Report window can function in support of both input and
output data. As an input window, it allows you to see the input data
superimposed on the pipe system schematic created on the Workspace.
Output windows
The two windows that function exclusively as output windows are the
Output window and the Graph Results window. The Output window is
text-based, while the Graph Results window is graphical. These two
windows offer a powerful and diverse range of features for reviewing
analysis results for modeling errors, gaining a deeper understanding of
the pipe system's flow behavior, and preparing the results for
documentation.
As an output window, Visual Report allows you to see the output results
superimposed on the pipe system schematic created on the Workspace.
The five primary windows form a tightly integrated, highly efficient
system for entering, processing, analyzing, and documenting
incompressible flow analyses of pipe networks.

quickstart 1.doc
Optimization terminology
General optimization terminology applied to pipe systems is as follows:
Design variables
The design variables in AFT Titan are the pipe sizes.
Design constraints
There are numerous design constraints in AFT Titan. Common
constraints are pipe velocity, control valve pressure drop and proximity
to compressor/fan BEP (Best Efficiency Point).
Active and inactive constraints

One constraint type is pipe velocity. One may set the maximum to 100
feet/sec, and the minimum to 10 feet/sec. If the final pipe velocity were
99 feet/sec, the maximum velocity constraint would be active because
the final value of the velocity is at or near the maximum velocity. On the
other hand, the 99 feet/sec pipe velocity is far away from the 10 feet/sec
minimum. If we remove the minimum velocity constraint, the result will
still be an actual pipe velocity of 99 feet/sec. Thus the minimum velocity
constraint is inactive and does not influence the pipe size selected by the
optimizer.
On the other hand, if one removed the 100 feet/sec maximum velocity
constraint, the actual velocity would probably increase above 100
feet/sec, thus resulting in a different pipe size. The maximum velocity is
active in that if we change or remove the constraint the selected
optimum would change. Changing or removing inactive constraints have
no effect on the selected optimum.
Objective function
This is the cost of the system. The cost can be monetary or can be based
on weight, volume or some other parameter. As AFT Titan varies the
pipe sizes, the cost of the pipes and associated equipment varies. The
optimization engine searches for combinations of pipe sizes that
minimizes the objective function (i.e., cost).
5/7/2003 2:31 PM Page 7

Continuous vs. discrete optimization

Most commercial pipe comes in discrete sizes (e.g., 1 inch, 2 inch, 3
inch, etc.). When dealing with discrete data, AFT Titan evaluates the
best combination of discrete pipe sizes. If, on the other hand, it is
possible to obtain the pipe in any size, the pipe sizes are continuous.
AFT Titan can find continuous optimums in addition to discrete
optimums. The continuous optimum will typically provide a better
design than the discrete, and it is much easier (and faster) for the
optimization engine to identify.
Design variable linking

The optimization process takes longer as the number of pipes of
potentially different size increases. Frequently there are groups of pipes
in your system which either you want to be of the same size for design
purposes, or must be the same size by virtue of their location in the
system. To minimize the calculation time, it is best to link pipes. When
one links pipes, one is saying that the linked group of pipes all must have
the same pipe size, and be the same material and schedule, class, or type.
A linked group thus collapses the individual pipes that are part of that
group to a single design variable for that group. There can be multiple
linked groups in a model.
Feasible and infeasible designs

A feasible design is one which satisfies all constraints, while an
infeasible design does not satisfy one or more constraints. There are
many ways you can create a model that has no feasible solution. For
example: Connect a pipe to an assigned pressure junction set to 200 psig,
then place a constraint on the pipe that it must have a pressure less than
100 psig. Since the pipe is connected to an assigned pressure junction at
200 psig, there is no way for it to satisfy the 100 psig constraint. Thus,
no feasible solution exists.

quickstart 1.doc
CHAPTER 2
Weight Optimization Example
Summary
This example shows how to use AFT Titan to perform an engineering

parameter optimization on a system. Engineering parameter
optimization involves minimizing some quantity such as the total pipe
weight or volume. Pipe weight or volume frequently relates closely to
the actual pipe cost, and is easier to setup than cost based optimization.
This example demonstrates minimizing the pipe weight in a system.
The example is for a steam transfer system that feeds steam from a
pressurized tank to a lower pressure tank at a specified flow rate. A
control valve is used to control the flow rate.
Topics covered
Engineering parameter (pipe weight) optimization
Pipe linking options
Control Valve constraints
Required knowledge
No prior knowledge is required for this example.

quickstart 1.doc
Model files
This example uses the following file, which is installed in the Examples
folder as part of the AFT Titan installation:
Control Valve.ttn - AFT Titan model file
Step 1. Create the model
Summary
Create a model as shown in Figure 2.1 with the following specifications:
Steam
Inlet tank at 250 psia and 600 deg. F and discharge tank at 200 psia
and 600 deg. F
Desired flow is 10,000 lbm/hr
Control valve must have a pressure drop of at least 20 psid
Total pipe length is 250 feet
Pipes are Schedule 40 from 1 to 10 inch, excluding 3 1-2 inch and 5
inch pipe.

quickstart 1.doc
Chapter 2 Weight Optimization Example 11
Figure 2.1 Control Valve system layout
A. Layout model
Create a model as shown in Figure 2.1.
1. The three junctions, J1, J2 and J3, can be dragged from the Toolbox
at the left and dropped on the Workspace.
2. The two pipes, P1 and P2, can be drawn on the Workspace by
clicking the Pipe Drawing Tool at the upper right of the Toolbox and
then drawing lines on the Workspace. Make sure the directional
arrows point from J1 to J2 and then J2 to J3. (The flow direction can
be reversed by use of the Reverse Direction tool on the Arrange
menu.)
B. Specify solution method

Models run more accurately with the default Solution Control methods,
but more quickly with Lumped methods.
1. Open Solution Control from the Analysis menu
5/7/2003 2:31 PM Page 11

2. Select the Lumped Adiabatic method (Figure 2.2)
3. Click OK
Figure 2.2 Solution Control for steam transfer system
C. Select fluid
1. Open the System Properties window from the Analysis menu (see
Figure 2.3)
2. In the upper left select the "AFT Standard" option
3. In the "Fluids Available in Database" list, select "Steam"
4. Click the "Add to Model" button
5. Click the OK button to close the window and accept the fluid data
for the model

quickstart 1.doc
Figure 2.3 System Properties window is where you specify the

fluid.
D. Enter data for tanks

As shown in Figure 2.1, the J1 (Inlet Tank) junction is a Tank junction.
1. Double-click the J1 junction icon to open the Tank Specifications
window (see Figure 2.4)
2. Enter a Pressure of 250 psia
3. Enter a Temperature of 600 psia
4. Enter an Elevation zero feet
5. Click the OK button
5/7/2003 2:31 PM Page 13

Figure 2.4 Input for Tank junction J1
Repeat this process for junction J3 (Discharge Tank), but use a Pressure
of 200 psia.
E. Enter control valve data

The J2 junction is a Control Valve junction. We will use this as a flow
control valve (FCV).
1. Double-click the J2 junction icon to open the Control Valve
Specifications window (see Figure 2.5)
2. Enter an elevation of 0 feet.
3. In the "Valve Type" area select "Flow Control (FCV)"
4. In the "Control Setting" area enter a "Flow Setting" of 10,000
lbm/hr.

quickstart 1.doc
Figure 2.5 Input for Control Valve junction J2
F. Enter pipe data

1. Double-click on the pipe P1 to open the Pipe Specifications window
(see Figure 2.6)
2. In the "Size" area choose the "Pipe Material" as "Steel"
3. Choose the Size and Type as "6 inch" and "Schedule 40"
4. Specify the length as 125 feet
6. Enter the same data for pipe P2 as pipe P1.
5/7/2003 2:31 PM Page 15

Figure 2.6 Pipe Specifications window for pipe P1
Step 2. Setup the optimization data
A. Create pipe size range set

In order to select an optimal pipe size, AFT Titan needs to know what
pipe materials and/or sizes should be considered. This is specified
through pipe size range sets.
1. Open the Pipe Optimization Parameters window by selecting "Pipe
Parameters" from the Optimization menu (see Figure 2.7)
2. Click the "Create Set" button and name the set "Steel Sch 40"
3. In the "Range Set Definition" area select the material as "Steel"
4. Click the "Select Pipe Sizes" button to open to open the Select Pipe
Sizes window
5. In the Sort area at the lower left select "Type, Schedule, Class" (see
Figure 2.8)

quickstart 1.doc
6. In the "Available Material Sizes and Types" list at the left expand
the "Schedule 40" listing
7. Add all sizes from 1 inch to 10 inch, except for 3 1-2 and 5 inch, by
selecting from the list on the left and then clicking the "Add >>"
button to add each one to the "Use These Sizes" list on the right
(again, see Figure 2.8)
9. The Pipe Optimization Parameters window should now look like that
in Figure 2.9
At this point, we have merely created a size range set. The size range set
will have no effect on the optimization process until we actually apply it
to specific pipes in the model. We will do this in Step 3.
Figure 2.7 Pipe Optimization Parameters window allows you to

create size range sets and constraint sets
5/7/2003 2:31 PM Page 17

Figure 2.8 The Select Pipe Sizes window allows specify the
specific pipe sizes to be included in the size range set
Figure 2.9 The "Steel Sch 40" size range set when completed

quickstart 1.doc
B. Create a control valve constraint set

The flow control valve has two design requirements. First is that it
control the flow to 4000 scfm. Second, that it operate with at least 20
psid pressure drop. The flow requirement is met by specifying the valve
as an FCV with flow control of 4000 scfm. We did this already (see
Figure 2.5).
To obtain the 20 psid pressure drop we must apply a constraint to the
control valve.
1. Open the Control Valve Optimization Parameters window by
selecting "Control Valve Parameters" from the Optimization menu
(see Figure 2.10)
2. Click the "Create Set" button and name the set "Pressure Drop Min"
3. In the grid at the bottom, select the Apply checkbox for "Pressure
Drop Static Minimum"
4. Enter a "Value" of 20 psid
Figure 2.10 Control valve constraints are created in the Control

Valve Optimization Parameters window
5/7/2003 2:31 PM Page 19

Similar to the pipe size range set we created, the control valve constraint
has been created but not applied. It thus will have no effect on the
optimization until we actually apply it to the J2 control valve. We will
do this in Step 3.
Step 3. Apply optimization data

To make use of pipe size range sets and constraint sets, they must be
applied to the relevant pipes and junctions.
Figure 2.11 Optimization data for pipe P1
A. Apply optimization data to P1

1. Double-click pipe P1 to open the Pipe Specifications window once
again
2. Select the Optimization tab (see Figure 2.11)

quickstart 1.doc
3. In the Optimization area select the "Optimize" option (this
automatically forces selection of the "Include Cost in Cost Report
and Objective" option)
4. In the Linking area (at the lower left) select the "Use as a Link Basis
Pipe" option
5. In the "Size Range" list (at the lower right) select the "Steel Sch 40"
size range set we created previously (Figure 2.9)
B. Apply optimization data to P2

1. Double-click pipe P2 to open the Pipe Specifications window
3. In the Optimization area select the "Optimize" option (this
automatically forces selection of the "Include Cost in Cost Report
and Objective" option)
4. In the Linking area (at the lower left) select the "Link to Pipe"
option, with pipe P1 as the link (this dropdown list is a list of all link
basis pipes one in our case)
5. The "Size Range" list (at the lower right) will automatically select
the size range set of the link basis pipe, which for pipe P1 was "Steel
Sch 40". This list is grayed out to indicate it is not an option for pipe
P2 because it is a linked pipe.
By linking pipe P2 to P1, we are saying that we want AFT Titan to select
the same size pipe for each. If we did not link them, the two sizes would
be selected independently, and may or may not end up being the same
size.
5/7/2003 2:31 PM Page 21

Figure 2.12 Pipe P2 is optimized and linked to pipe P1
C. Apply optimization data to J2 (FCV)

1. Double-click junction J2 to open the Control Valve Specifications
window
3. In the Constraint Sets area select the "Pressure Drop Min" constraint
set

quickstart 1.doc
Figure 2.13 Control Valve window optimization data
Step 4. Specify Optimization Control

1. Open the Optimization Control window from the Analysis menu (see
Figure 2.14)
2. In the "Optimization Type" area (at the upper left) select the
"Perform Discrete Optimization" option
3. In the Objective area (at the upper right) select "Engineering
Parameter" in the dropdown list at the top
4. Select the "Minimize" option
5. Select the "Pipe Weight" option in the list of four engineering
parameters
5/7/2003 2:31 PM Page 23

Figure 2.14 Optimization Control window accepts criteria

specification for how to optimize the pipe system
Step 5. Run the Optimization

We are now in a position to run the optimization. Before doing so, take a
moment to consider what this model is trying to accomplish. The pipe P1
and P2 sizes will be selected from the "Steel Sch 40" size range set as
the same size, such that their weight is minimized while obtaining a
minimum 20 psid pressure drop across the control valve.
1. To run the optimization, select Run from the Analysis menu. This
will start the Solution Progress window (see Figure 2.15). The
optimum for this small system is quickly found. It is 1,894 lbm. It
required 30 calls to the Compressible Flow Solver to find the
optimum.
2. Click the "View Results" button
3. The Output window will display, where the optimization results can
be viewed

quickstart 1.doc
Figure 2.15 The Solution Progress window shows the progress of

the Compressible Flow Solver and Optimizer
Step 6. Review optimization results

1. The Cost Report tab (see Figure 2.16) shows a total weight of 1,894
lbm. This should be and is consistent with the weight shown in the
Solution Progress window (in Figure 2.15). Each pipe weighs 947
lbm, which is consistent with the fact they are the same length and
linked to optimize for the same diameter.
2. The Optimization tab shows the optimal pipe size chosen was 3 inch.
The 3 inch pipe over 250 feet of length corresponds to a weight of
1894 lbm.
3. The CV Constraints tab shows that the 20 psid minimum pressure
drop is satisfied in that the pressure drop for 3 inch pipe was 35 psid.
5/7/2003 2:31 PM Page 25

Figure 2.16 The Output window shows the compressible flow

results of the optimized system, plus the optimal
system size results
Conclusions
This example demonstrates AFT Titan's optimization capabilities for a
simple system with a single constraint. Engineering parameter
optimization for weight is fast and easy to implement.

quickstart 1.doc
CHAPTER 3
Initial and Life Cycle Cost Optimization

Example
This chapter demonstrates some key features in using AFT Titan to

optimize an air distribution system for cost.
Topics covered
This example will cover the following topics:
Engineering and Cost Databases - How to connect and use
Linking Pipes - How to limit the number of independent variables
Size range sets and constraint sets - How to set requirements on the
system
Optimization Control - How to optimize for initial or life cycle cost
Optimization output - How to understand the optimization results
Required knowledge
This example assumes that the user has some familiarity with AFT Titan
such as placing junctions, connecting pipes, entering pipe and junction
specifications, and creating and using pipe size range sets and
constraints. Refer to the Weight Optimization Example in Chapter 2 for
more information on these topics.

quickstart 1.doc
Model files
This example uses the following files, which are installed in the
Examples folder as part of the AFT Titan installation:
Air Distribution.ttn AFT Titan model file
Air Distribution.dat junction engineering database
Air Distribution.cst cost database associated with Air
Distribution.dat
pipe-steel-sch40-galv-threaded.cst - cost database for galvanized
steel
Optimization goals
This example uses an existing model to investigate two optimization
cases:
1. Optimize system for initial cost with a 5-year operating period
2. Optimize system for life cycle cost with a 5-year operating period.
Getting started
To begin, start AFT Titan and load the model file Air Distribution.ttn.
This model has a number of different scenarios. If you are not familiar
with scenarios, you can review the Scenario Manager discussion in the
Help system or Chapter 5 of the AFT Titan User Guide.
Open the Scenario Manager from the View menu to see the existing
scenarios. Select and load the scenario "Five Year Life/Design for Initial
Cost". The model should be ready to run, but first lets understand what
the model is doing. See Figure 3.1.

quickstart 1.doc
Chapter 3 Initial and Life Cycle Cost Optimization Example 29
Figure 3.1 Air Distribution example model
Review Optimization Control

Open the Optimization Control window from the Analysis menu (Figure
3.2). The Optimization Control offers a number of features to control the
optimization process. In the Optimization Type area, the first selection is
"Do Not Optimize". This is equivalent to running AFT Titan as one runs
AFT Fathom.
The second selection is "Calculate Costs, Do Not Optimize". This is
identical to the first selection, with the exception that costs are
calculated for the system and displayed in the Output window.
The third selection is "Perform Continuous Optimization", which selects
pipes sizes assuming the diameters are continuous. In other words, it
ignores the fact that commercial pipe is available only in discrete sizes,
and instead assumes that any diameter is acceptable.
5/7/2003 2:31 PM Page 29

The fourth selection is more realistic than the third because it recognizes
the fact that commercial pipes are available only in discrete sizes, and
chooses the optimal combination of discrete sizes.
Figure 3.2 The Optimization Control window provides features

that control the optimization process
The Optimization Control window also provides features to characterize

the optimization objective. Frequently this will be to minimize certain
monetary cost categories. Also available is Engineering Parameter
optimization, which allows one to minimize the pipe or duct volumes,
weight, surface area, etc. (Weight optimization was used in Chapter 2.)
For calculating life cycle costs, users must provide a System Life and,
optionally, interest and inflation rates to calculate the present value of
the recurring costs.
This scenario assigns a system life of 5 years, and includes only material
and installation costs in the objective. Both of these cost categories are
non-recurring. With only these specified, the model is attempting to find
an optimum based on non-recurring (i.e., first or initial) cost.

quickstart 1.doc
Review databases
The actual material and installation costs that the Optimization Control
window specifies are contained in cost databases. The cost databases
needed for this air distribution example already exist, and just need to be
accessed.
The Database Manager (opened from the Database menu) shows all of
the available and connected databases. Databases can either be
engineering databases or cost databases. Cost databases are always
associated with an engineering database, and are thus displayed
subordinate to an engineering database in the database lists.
Here we will summarize some key aspects of databases:
Cost information for a pipe system component is accessed from a
cost database. Cost database items are based on corresponding items
in an engineering database. (The engineering databases also include
engineering information such as pipe diameters, hydraulic loss
factors, etc.)
To access a cost for a particular pipe or junction in a model, that
pipe or junction must be based on items in an engineering database.
Moreover, that database must be connected.
There can be multiple cost databases associated with and connected
to an engineering database. This makes it easier to manage costs of
items.
The Database Manager should appear as shown in Figure 3.3. With the
"Air Distribution System", both database sections are checked. The two
sections should be Junction/Component Costs and Energy Costs.
The engineering databases associated with these two cost databases are
the AFT DEFAULT INTERNAL database, and an external database
called "Air Distribution System". For the cost data in the two cost
databases to be accessed by pipes and junctions in the model, the pipe
and junctions must use these two engineering databases.
5/7/2003 2:31 PM Page 31

Figure 3.3 Database Manager shows available and connected

databases
Review pipe optimization setup

Open pipe P1 on the Workspace by double-clicking it. Figure 3.4 shows
the Pipe Specifications window with the Pipe Model tab selected, and
Figure 3.5 shows it with the Optimization tab selected.
The Optimization data offers the ability to optimize or not optimize a
pipe. If Optimize is selected, AFT Titan will treat the pipe diameter as a
variable and vary it according to certain criteria that will be discussed
shortly. Why would one choose to not optimize a pipe? There could be a
number of reasons, but one good reason is that the pipe represents a pipe
in an existing system and the design does not allow the replacement of
that pipe with a new one. Therefore its diameter is fixed, and optimizing
the pipe would serve no purpose. All pipes in this model are to be
optimized, because it is a new system and it makes sense to select the
optimal diameters for all pipes.

quickstart 1.doc
Figure 3.4 Pipe P1 showing Pipe Model tab.
Pipe size range sets

When the pipe is optimized, the optimal diameter is selected from a
group of potential diameters called a pipe size range set. A specific pipe
can only access one pipe size range set, but a specific pipe size range set
can be accessed by any number of pipes.
As shown in Figure 3.5, the Size Range list shows all the pipe size range
sets that exist for this model. In this case there are two, and one and only
one must be selected.
The pipe size range sets are created in the Pipe Optimization Parameters
window, which is opened from the Optimization menu. It can also be
opened by clicking the Optimization Parameters button at the lower right
(see Figure 3.5).
Once a size range set is selected, the list of available pipe sizes on the
pipe window is restricted to pipes in that size range set. Look in Figure
3.4 and you can see the area normally described as "Size" is now
described as "Size (From Optimization Size Range)". The actual pipe
size you select here is analogous to an initial guess at the size.
5/7/2003 2:31 PM Page 33

Figure 3.5 Pipe P1 showing Optimization tab.
Pipe constraint sets

Constraints can be applied to both optimized and unoptimized pipes.
Constraints are design limitations on certain parameters. For example,
there may be a design requirement that the velocity cannot exceed 100
feet/second. As the optimizer evaluates different pipe sizes, if a
particular size results in a velocity greater than 100 feet/second, that pipe
size is rejected because it causes a constraint violation.
When a pipe that is not optimized has a constraint, it means that the
optimized pipe sizes cannot be such that it violates the constraint for the
unoptimized pipe. There can be multiple constraints in a constraint set,
and multiple constraint sets applied to a pipe. For instance, in addition to
a maximum velocity limit, a maximum pressure limit may exist.
Constraints are contained in constraint sets, which are created in the Pipe
Optimization Parameters window. This window is opened from the
Optimization menu or by clicking the Optimization Parameters button at
the lower right (see Figure 3.5).

quickstart 1.doc
As shown in Figure 3.5, pipe P1 does not have any constraint sets
applied. If you open the Pipe Optimization Parameters window, you will
see the "Max Velocity" constraint is 150 feet/sec. Other pipes in this
model can have constraints applied according to the design
requirements. You can view the constraint sets for each pipe in the
Model Data window. In addition, you can view the pipes using each
constraint in the Optimization Summary window.
Pipe linking
Pipe linking is the process whereby certain groups of pipes are specified
to have the same pipe size. As pipes are linked it reduces the number of
design variables and allows the optimizer to run faster. It also has the
effect of simplifying the design.
Figure 3.6 Linking Information shows the linking relationships

for a model.
An unlinked pipe is one that has no links to any other pipe, and no other
pipes linked to it. A "link basis" pipe is one that is not linked to other
5/7/2003 2:31 PM Page 35

pipes, but allows other pipes to link to it. A linked pipe is one linked to
another pipe that is a link basis pipe.
There is a fourth type of linking, which is related to dependent design
cases. This is discussed in the next chapter.
In this model, pipe P1 is a link basis pipe. To find out how all the pipe
linking is setup, open the Optimization Summary window from the
Optimization menu and select the Linking tab. If you select the Expand
Tree button, the window should look like Figure 3.6. This tree
graphically shows the number of design variables that AFT Titan will
solve. This particular model has eighteen pipes that are being optimized,
but because of linking it has twelve link basis pipes, and thus twelve
design variables. You can also view the pipe linking by selecting the
Color Linked Pipes feature on the Optimization menu.
Creating pipe size range sets and constraint sets

The pipe size range sets are created in the Pipe Optimization Parameters
window, which is opened from the Optimization menu. It can also be
opened by clicking the Optimization Parameters button (see Figure 3.5).
Figure 3.7 Pipe size range sets are created and modified in the
Pipe Optimization Parameters window.

quickstart 1.doc
The source of cost data for each of these size range sets is specified in
the Cost Database Source list (Figure 3.7). In our case, all cost sources
are specified as "All Cost Databases", which means to use the data
sources as specified in the Database Manager.
The Pipe Optimization Parameters window, Constraint Sets tab, is
shown in Figure 3.8. Here constraint sets for pipes are created. These
constraint sets can be applied to the pipes of your choosing. For
example, if all pipes have a maximum velocity of 100 feet/sec, then that
constraint set should be applied to all pipes in the model. This is done on
each pipe window, in the Constraints Sets area as shown in Figure 3.5.
Figure 3.8 Pipe Constraint Sets are created and modified in the
Pipe Optimization Parameters window.
Review junction optimization setup

Open compressor/fan junction J2 on the Workspace by double-clicking
it. Figure 3.9 shows the Compressor/Fan Specifications window with the
Compressor/Fan Model tab selected, and Figure 3.10 shows it with the
Optimization tab selected.
5/7/2003 2:31 PM Page 37

Figure 3.9 Compressor/fan junction J2 Compressor/Fan Model

tab
Figure 3.10 Compressor/fan junction J3 Compressor/Fan

Optimization tab

quickstart 1.doc
Optimizing systems with compressor/fans

A compressor/fan can be modeled either as a compressor/fan curve or as
an assigned flow. The fundamental reason one would choose one
approach over the other is if one is trying to identify the pressure
generation requirements for the purposes of choosing a compressor/fan
(in which case one would model the compressor/fan as an assigned
flow), or if one already has a specific compressor/fan identified (in
which case one would model the compressor/fan as a compressor/fan
curve). It could be the case that one does not have a specific
compressor/fan selected, but does have several candidate
compressor/fans. In this case, it would be best to model each candidate
compressor/fan as a compressor/fan curve, and do this within a scenario
created for each compressor/fan.
Modeling a compressor/fan as an assigned flow

The compressor in our air distribution case is modeled as an assigned
flow of 1250 scfm. This was chosen for the following reason. Each of
the five rooms is required to receive 250 scfm of air, for a total of 1250
scfm.
As already stated, when a compressor is modeled as an assigned flow, it
is not a specific compressor from a specific manufacturer. Thus, the
costs for the compressor can only be approximated. To a first
approximation, it should be possible to estimate the non-recurring cost
(i.e., material and installation cost) for the compressor/fan as a function
of power requirements at 1250 scfm. For instance, a five horsepower
compressor/fan (of specific configuration and materials of construction)
may cost $1000, and a ten horsepower may cost $1800. Other typical
costs for different power requirements can be approximated. The actual
cost for the compressor will, of course, highly depend on the application.
These costs are entered into a cost database, and are accessed by the
optimizer. As AFT Titan evaluates different combinations of pipe sizes,
each combination will require a certain power from the compressor With
a cost assigned to this power, AFT Titan can obtain a cost to enter into
the objective function which it optimizes. Later in this chapter, we will
look at the costs for the compressor in our example air distribution
network.
5/7/2003 2:31 PM Page 39

In addition to non-recurring costs, recurring costs can also be estimated.
Specifically, the cost of the power used by the compressor over a period
of time (which AFT Titan calls the "system life") can be determined. All
AFT Titan needs is an overall compressor/fan efficiency to determine
the actual power from the ideal power. Again, since we do not have a
specific compressor selected yet, the efficiency can only be
approximated. AFT Titan calls this the "nominal efficiency" (see Figure
3.9).
A step-by-step method of compressor and fan selection proceeds in two
phases. The steps are outlines in the Help System and in Chapter 12 of
the AFT Titan User Guide.
Junction costs
There are three choices in specifying the cost of a junction. These
choices are provided on the Optimization tab in the junction's
Specifications window (see Figure 3.10).
1. Do Not Include in Cost As it says, the cost of the junction is
entirely neglected.
2. Include in Cost Report Only This reports all costs for the junction,
but does not include the cost in the objective function. The costs
thus do not impact the overall optimization process.
3. Include in Cost Report and Objective This reports all costs for the
junction, and includes the cost in the objective function. The
junction costs are thus allowed to influence the optimization process.
Junction constraint sets

Junctions do not have size range sets as do pipes, but many have
constraints. Junction constraints function similarly to pipe constraints.
Some examples of junction constraints are compressor/fan power and
control valve pressure drop. Junction constraint sets are specified
similarly to pipes, by selecting the constraint sets from the provided list.
Creating junction constraint sets

Junction constraint sets are created in a similar way to pipe constraint
sets. Constraint sets for compressor/fans, control valves, and other types
of junctions are created on three different windows opened from the

quickstart 1.doc
Optimization menu. Figure 3.11 shows a control valve constraint set
used in the air distribution example we are reviewing. This constraint is
applied to control valves J111, J121, J211 and J221.
Figure 3.11 Constraint sets for control valves are created in the
Compressor/Fan Optimization Parameters window.
Understanding the model

Before running the model, lets discuss some key aspects of the model.
You may want to review the actual model itself as we discuss these
points.
1. The system is an air distribution system, with air supplied at 70F.
2. The overall goal of the model is to size the J2 compressor for a flow
of 1250 scfm, and in the process find an optimal compressor and
pipe system combination.
3. There is a design requirement that each of the five rooms receive
250 scfm, with flow maintained by flow control valves. In addition,
5/7/2003 2:31 PM Page 41

it is required that the final design yield at least a 2 psid drop across
each control valve.
4. After the design is complete, the flow through the compressor will
be controlled by the five flow control valves. Since we are sizing the
compressor, we select the J131 control valve as the most remote and
set it to the minimum allowed pressure drop of 2 psid. The choice of
2 psid comes from the design requirement for this system that the
flow control valve shall, at a minimum, have a 2 psid pressure drop.
5. A Control Valve Constraint Set called Min Pressure Drop has
been specified as a minimum 2 psid. This has been applied to J111,
J121, J211, and J221. But not J131, which has been modeled as a
fixed pressure drop at this stage. (You can see which junctions use
this constraint in the Optimization Summary window).
6. Except for pipe P1, all other pipes are inside the building and cannot
have a velocity greater than 150 feet/sec. This is a pipe constraint
called Max Velocity, which is applied to all pipes except P1. (You
can see which pipes use this constraint in the Optimization Summary
window).
7. As shown in Figure 3.6, the pipes are linked such that there are
twelve independent pipe sizes (i.e., design variables) that will be
included. The linking is such that pipes that should have the same
size are linked together. For instance, it makes sense that the suction
and discharge pipes for the compressor have the same size.
Therefore, pipe P1 and P2 are linked together, with pipe P1 being
the link basis (either of the pipes could be chosen as the link basis,
and yield identical results).
8. The compressor has a nominal efficiency of 80%.
9. The compressor is specified to "Include in Cost Report and
Objective" (see Figure 3.10).
10. The two cost databases, shown in Figure 3.3, have cost data for the
pipes and the compressor. You can use the Cost Database window
opened from the Database menu to review the cost data for these
databases. In summary, there is cost data for the compressor as a
function of power. There is also operational cost for power of a
fixed 0.06 U.S. dollars per kW-hr. The cost is fixed in that is
assumed to be constant over the system life. It is possible to increase
or decrease the cost over time using scaling tables.

quickstart 1.doc
Running the scenarios and interpreting results

For the two scenarios we will evaluate, the only change will be in the
Optimization Control window. The first case is the one we have been
reviewing.
Scenario to minimize first cost

As shown in Figure 3.2, the system life is set to 5 years for this scenario,
and the objective only includes costs for material and installation.
The optimization results are already saved in this file. You can view the
results in the Output window by selecting Output from the Windows
menu. Alternatively, you can rerun the optimization. Depending on your
computers processor, this may take 3-10 minutes.
When the model is run, AFT Titan evaluates a range of pipe size
combinations, and finds the one that minimizes the sum of material and
installation costs (i.e., initial cost).
Figure 3.12 The Cost Report in the Output window shows the total
and individual costs for the optimized system.
The Cost Report is shown in the General Section of the Output window
(see Figure 3.12). AFT Titan shows all costs in the Cost Report, even
5/7/2003 2:31 PM Page 43

those that were not used in the optimization. The total cost for this
system is $149,313. This includes all costs over 5 years. The first cost,
which was the basis for the optimization, was the total of the material
and installation ($68,255). This total is also the cost shown as the "Items
in Objective". Again, AFT Titan minimizes the objective, and thus the
cost that was minimized was the $68,255 cost shown in the Cost Report
as "Items in Objective". Individual items whose cost contributed to the
objective are shown with a green background in the cell.
Other costs that are displayed in the Cost Report are "Items Not in
Objective". These are items that have costs associated with them, but
were not included in the objective that was minimized.
Figure 3.13 Optimized pipe sizes are shown in Output window

pipe section.
Note that the Items Not in Objective total to $81,058. Looking across at
the subtotal one can see that all of this cost is Operational (i.e.,
compressor/fan power costs). The total cost is the sum of the two, or
$149,313. Why was the Operational Cost not added to the Items in
Objective? The reason relates to the Optimization Control settings, as

quickstart 1.doc
shown in Figure 3.2. Here one sees that the objective is set to minimize
only material and installation cost. Costs related to maintenance and
operation are thus not minimized. If this is still not clear, read through to
the next section which optimizes with operation costs included.
Optimized pipe sizes

The actual pipe sizes chosen as optimum are shown in Figure 3.13. This
is the Optimization tab in the Output window pipe section area. Along
with the optimized pipe sizes, the pipe linking is shown and the cost of
each pipe.
Checking source of cost data

In both the pipe and junction sections of the Output window, there are
tables called Database Source. These show the source of all cost data for
each pipe and junction. This is useful for model verification.
Scenario to minimize life cycle cost for 5 years

Using Scenario Manager, load the scenario called "5 Year Life/Design
for LCC". Open the Optimization Control window, and you can see the
objective now has operation cost selected (Figure 3.14). This is the only
difference from the previous scenario. Let's call the previous scenario
the "first cost" scenario and this one the "life cycle cost" scenario.
By including operational costs in the objective, we are now trying to
minimize the sum of the material, installation and operation cost. This
may cause the material and installation costs to increase. Let's see what
happens.
5/7/2003 2:31 PM Page 45

Figure 3.14 Optimization Control setting for "5 Year Life/Design

for LCC" scenario
Run the model (this model should run in less than a minute) and look at
the results in the Cost Report (Figure 3.15).
Figure 3.15 Cost Report for life cycle cost optimization scenario.

quickstart 1.doc
There are a number of items of note:
1. The overall cost is now $135,790. Taken on its own, this represents
a savings of about $13,500 over the first cost scenario which was
design based on first cost. This represents a 9% cost reduction.
2. The "Items Not in Objective" category is now zero, whereas it was
substantial in the first cost scenario. The reason is that we have
included all costs in the objective based on the settings in
Optimization Control. To be complete, the Optimization Control did
not include maintenance costs, but there were no maintenance costs
in any of the cost databases. You can see this by looking at the Cost
Report column for Maintenance. If there were maintenance costs in
these cost databases, there would be values in this column and the
"Items Not in Objective" would not be zero unless Optimization
Control was modified to include maintenance cost.
3. The "Items in Objective" category now includes all cost items, and
the Cost Report breaks this down further into a Non-Recurring Sub
Total ($89,700) and a Recurring Sub Total ($46,100).
4. Because Operation costs are now included in the objective, the cells
for Operation are colored green whereas before they were not.
Looking at the source of operational costs, one sees it comes from
the compressor/fans (which will always be the case for operational
costs).
5. In the first cost scenario, the non-recurring cost was $68,300, while
the overall cost was $149,300. Now the non-recurring cost is
$89,700 while the overall cost is $135,800. The first cost thus
increased by about $21,400 in order to reduce the operating cost
from $81,100 to $46,100 (a reduction of about $35,000).
Table 3.1 Cost Summary of Optimization Runs for Air Distribution

System
Optimized for: Material Installation Total Operating Total (system + Reduction
System operation)
Initial Cost 5 yr 41,100 27,200 60,300 81,100 149,300

Life cycle cost 5 yr 55,300 34,400 89,700 46,100 135,800 13,500
6. The source of the operating cost is the cost of power for the
compressor. To reduce compressor power usage, it makes sense to
increase the pipe size and thus reduce frictional losses. For the AFT
5/7/2003 2:31 PM Page 47

Titan "Design for LCC" scenario, AFT Titan optimizes the sum of
these two costs. The larger pipe sizes can be reviewed by looking at
the Output window pipe Optimization tab, and also are summarized
in Figure 3.16.
J2
41.4 hp
J1 23.5 hp J3 J201
P1 P2 P201
6 inch 6 inch 4 inch
8 inch 8 inch P101 6 inch P202
4 inch 4 inch
6 inch 4 inch
J112 J111 J101 J202 J211 J212
P211 P212
P112 P111 2 1/2 inch 2 1/2 inch
2-1/2 inch 2-1/2 inch P102 3 inch 3 inch
2-1/2 inch 2-1/2 inch 4 inch
P203
4 inch
3 inch
4 inch
J122 J121 J102 J203 J221 J222
P122 P121 P221 P222

3 inch 3 inch 2 1/2 inch 2 1/2 inch
3 inch 3 inch P103 4 inch 4 inch
3 inch
4 inch
Legend:
J132 J131 J103
Optimized for initial cost
Optimized for 5 year life cycle
P132 P131
2 1/2 inch 2 1/2 inch
4 inch 4 inch
Figure 3.16 Pipe sizes selected by AFT Titan for first cost, and life
cycle cost over 5 years.
7. Now it should be clearer why the operational costs were included in

the Cost Report of the first cost scenario. If they were not in the Cost
Report of the first cost scenario, the total cost would be $68,300.
Since the life cycle cost scenario total cost is $135,800, it would
appear (at first glance) to be a cost increase. But that is misleading.
Just because the operating costs were not included in the first cost
scenario's objective does not mean they do not exist. To get a clear

quickstart 1.doc
comparison between designing for first cost and life cycle cost one
needs to make an "apples to apples" comparison, including all
relevant costs in each case. The approach taken in this chapter does
this.
8. If, after having performed the optimization, one still wants to design
the system for first cost, the first cost scenario results provides the
optimal system design to minimize first costs. Importantly, the
designer has quantitative data on the impact of first cost design on
operational costs of the system.
Optimizing with compressor/fan curve data

Once the compressor/fan is sized then actual compressor/fans can be
modeled. The actual compressor/fan should closely match the sizing
results in the following areas: generated pressure at the design flow,
efficiency at the design flow, and cost.
Reviewing the results for this case one can see that the optimum system
calls for a compressor/fan of about 41 hp which generates about 6.7 psid
of pressure at 1250 scfm. The nominal efficiency used in the sizing part
of the analysis was 80%. The material cost for such a compressor/fan
was about $6500, and the installation cost was about $3250. Note that if
no actual compressor/fans can be found that reflect these requirements,
then the phase 1 of the Summarizing the compressor/fan selection
process should be repeated with better performance and/or cost data for
the compressor/fan.
Conclusions
Using cost databases in the optimization process involves increased
complexity from simple engineering parameter optimization as discussed
in Chapter 2. However, it allows more powerful optimization options
including the ability to optimize costs over a system life cycle.
5/7/2003 2:31 PM Page 49


quickstart 1.doc
CHAPTER 4
Multiple Design Case Example
Summary
This example will optimize the pipe sizes for a natural gas supply system
to five burners where there are two design cases. This example uses
monetary cost optimization.
Topics covered
This example will cover the following topics:
How dependent design cases are used to satisfy two different
operating modes for a system
Pipe linking and its effect on how well AFT Titan can optimize a
system
Required knowledge
This example assumes that the user has some familiarity with AFT Titan
such as placing junctions, connecting pipes, entering pipe and junction
specifications, and creating and using pipe size range sets and
constraints. Refer to the Weight Optimization Example in Chapter 2 for
more information on these topics.

quickstart 1.doc
Model files
Natural Gas Burner.ttn - AFT Titan model file
pipe-steel-sch40-galv-threaded.cst - Pipe cost database for steel pipe
Optimization goals
This example uses an existing model to investigate a single system with
two operating cases. The cases are the following:
1. Normal flow to burners requires 5 lbm/sec controlled by flow
control valves. Pressure drop across flow control valve must be at
least 25 psid and velocity at burner must be less than 200 feet/sec.
2. One burner off case has center burner turned off. Total flow must
be maintained so each flow control valve must flow 6.25 lbm/sec.
Same pressure drop requirement at control valves, but there is no
velocity requirement.
For this example, we will evaluate both operating cases. First, we will
use AFT Titan to optimize assuming a requirement to use up to three
pipe sizes throughout the system. After this we will optimize the pipes
assuming the pipes can be six different sizes.
Getting started
Start AFT Titan and open the " Natural Gas Burner.ttn" file. Select the
scenario "Base Scenario/Optimized Cases/ One Dependent Case / 2
Design Variables (Case 1)" (see Figure 4.1).
This example consists of a natural gas supply (Tank 1) at 200 psia and
50 deg. F that supplies natural gas to five burners. All the pipes in the
system are Steel schedule 40. The supply line to the heat exchanger is
fixed at 6 inch and is not be changed. The optimization of this system
will work with all pipes in the distribution part of the system.

quickstart 1.doc
Chapter 4 Multiple Design Case Example 53
Figure 4.1 Scenario Manager for Housing Project example
Review model
The supply to each burner under normal conditions is 5 lbm/sec of
natural gas (modeled as methane see System Properties window). Each
burner has a flow control valve to maintain this flow, and it is required
that each flow control have at least 25 psid pressure drop. Finally, it is
required that the velocity at each burner be at most 200 feet/sec. The
primary design case, "Supply to 5 Burners", is shown in Figure 4.2. This
case was built in the Base Scenario, which is not the scenario we are
currently reviewing.
For the supply line to the distribution piping (pipes 1 and 2), the pipe
size is fixed to six inches and will not be optimized. All other pipes can
be changed. In this scenario, we are assuming that all main pipes (pipes
11, 12, 21 and 22) will be sized to the same size. The distribution pipes
(31-72) can be sized to a different size than the mains, but but must all
be the size. Thus there will be two fundamental pipe sizes selected. The
size of the mains and the size of the distribution pipes.
5/7/2003 2:31 PM Page 53

To do this, we have to establish two pipes as "link basis pipes" and then
link all the other pipes (except the supply pipes) one of these pipes. For
this example, Pipe 11 and Pipe 31 were chosen as the link basis pipe. If
we look at the Optimization Summary window, we can see that all the
pipes are linked to one of these two pipes. Figure 4.3 shows the
Optimization Summary Window for this scenario.
Figure 4.2. Primary Design Case layout of Natural Gas Burner,

which is in the Base Scenario.
You can also investigate the size range sets and pipe constraint sets for
this model in the Pipe Optimization Parameters window opened from the
Optimization menu. The size range set includes all Schedule 40 steel
pipe from 1 inch to 12 inches. One pipe constraint set has been set up for
a maximum velocity of 200 feet/sec to the burners. A control valve
constraint set has been set up for the minimum 25 psid pressure drop
across the control valves.

quickstart 1.doc
Figure 4.3 Optimization Summary Window with all the pipe mains
linked to each other and all distribution pipes linked
How the dependent design case was created

After we created the primary design case model shown in Figure 4.2, we
created one dependent design case. A dependent design case (or DDC)
in AFT Titan is a system that models the same physical pipes and
junctions as the primary design but with different operating
requirements. The dependent design case for this model is when one
burner is turned off. When one burner is off, the flow to each of the
other four burners is increased from 5 to 6.25 lbm/sec.
Here are the steps to create this DDC for the first time:
1. Open the Optimization Control window (from the Analysis menu)
and select the "Enabled Dependent Design Cases" option at the
lower left (Figure 4.4). Click OK.
2. Choose Select All from the Edit menu.
5/7/2003 2:31 PM Page 55

Figure 4.4 Dependent Design Case modeling is enabled in the

Optimization Control window
Figure 4.5 Duplicate Special is the easiet way to create a

dependent design case

quickstart 1.doc
3. Open Duplicate Special (from the Edit menu), enter an increment of
100 and select "Make Dependent Design Case" (Figure 4.5). Click
OK.
Figure 4.6 Workspace with Primary Design Case (at top) and one
Dependent Design Case (at bottom)
5/7/2003 2:31 PM Page 57

4. Move the duplicated pipes and junctions to distinguish them from
the original ones from the Primary Design Case (Figure 4.6).
5. Close the J151 flow control valve in the DDC by opening its
Specifications window, setting the Action if Control is Lost to Fail
Closed, and then (on the Optional tab) setting the Special Condition
to Closed. Since this valve is closed, the 25 psid pressure drop
requirement no longer applies. On the Optimization tab the
Minimum DP constraint was unselected.
6. Open the Specifications window for DDC pipes 132, 142, 152, 162
and 172 and unselect the velocity constraint, which does not apply
when one of the burners is off.
7. Use Global Junction Edit (from the Edit menu) to change all of the
flow control valve junctions in the DDC (junctions 131, 141, 151,
161, and 171) from 5 to 6.25 lbm/sec (see Figure 4.7).
Figure 4.7 Global Junction Edit is used to change the FCV flows
to 6.25 lbm/sec to the burners in the DDC.
When Duplicate Special was performed with Dependent Design Case

selected, each of the duplicated pipes was created with a special type of
linking relationship. For example, pipe 111 is linked to pipe 11 as "Link

quickstart 1.doc
to Pipe as Dependent Design Case" (see Figure 4.8). This type of link
functions similarly to a regular link, except the cost is not counted for
the pipe.
Figure 4.8 Pipes in Dependent Design Cases have a special

linking relationship to pipes in the primary case
Run the optimization

Figure 4.4 showed the Optimization Control settings for this analysis.
AFT Titan will perform a discrete optimization on Material and
Installation costs of this system. Select Run from the Analysis menu and
AFT Titan will find the optimum pipe size to meet the system
requirements.
After the run finishes, examine the optimum pipe sizes calculated by
AFT Titan. The results for pipe size are shown in Figure 4.9. One can
see that the optimum size is 4-inch pipe for the mains and 5-inch for the
distribution pipes.
5/7/2003 2:31 PM Page 59

Figure 4.9 Optimum pipe size in Output window
The cost for all optimized pipes is $38,696 (see Cost Report in Figure
4.10). This is the cost for "Items in Objective". Since we did not
optimize the supply pipes (because these were fixed at 6- inch), there is
an additional cost of $8,250 for "items not in objective". The pipes that
were optimized are distinguished in the Cost Report by having a green
background color. The total cost is merely the sum of the two ($46,946).

quickstart 1.doc
Figure 4.10 Cost Report is shown in Output window
Consider the results

The optimization we just ran had two operating cases. One was for
normal natural gas supply, and the second was for supply with one
burner turned off. The normal case is modeled by the network shown in
the upper part of Figure 4.6. The second model shown at the bottom of
Figure 4.6 models the one burner off case.
The pipes and junctions in the one burner off case are separate entities
for modeling purposes, but represent the same pipes and junctions as the
normal flow case. Thus when performing optimization, we only count
the cost of the pipes and junctions once. This occurs in the primary
design case, which in our model is the normal flow case. This can be
seen in Figure 4.9, where pipes 1-72 all have an associated cost, while
pipes 101-172 do not.
By applying different pressure and flow requirements between the
normal and one burner off cases, we performed an optimization that
selected the optimal pipe size which simultaneously met the
requirements of both cases.
5/7/2003 2:31 PM Page 61

Finally, since the previous case only selected for two pipe sizes, there
were only two design variables. By unlinking some of the pipes, a better
optimum is possible. We will investigate this in the next section.
Optimize for eight pipe sizes

Return to the Scenario Manager and select the "8 Design Variables
(Case 2)" scenario. All input data is the same in this case as the previous
case, except for:
Figure 4.11 Linking relationships for the case with eight optimized
pipes.
1. Pipes 12, 41, 42, 51 and 52 have been made link basis pipes. These
are in addition to the original link basis pipes, 11 and 31.
2. Pipe 22 is now linked 12, pipe 61 is now linked to 41, pipe 62 is now
linked 42 and pipe 72 is now linked 32.

quickstart 1.doc
3. The default search method would work fine here, but it took about
five minutes to optimize. A different search method which uses
genetic algorithm methods was used to obtain results faster for the
sake of this example (it took about one and half minutes). This was
setup in the Optimization Control window, with some additional
adjustments input after selecting the Advanced button.
There are now eight independent pipe sizes for which the optimizer will
search. Figure 4.11 shows the new linking relationships.
Run the optimization

Run this scenario and examine the results. The optimum pipe sizes for
this scenario are shown in Figure 4.12. The pipe sizes now are a
combination of sizes including 3-1/2, 4, 5 and 6 inch pipe. This design
still meets all the requirements of both scenarios but the pipe sizes are
optimized for cost.
Figure 4.12 Optimum pipe sizes for the case with eight link basis
pipes
5/7/2003 2:31 PM Page 63

The cost of the design is shown in Figure 4.13. The cost of the optimized
pipes (i.e., Items In Objective) has changed from $38,696 to $32,950, a
15% cost reduction.
Figure 4.13 Cost Report for the case with eight link basis pipes

quickstart 1.doc
CHAPTER 5
Other AFT Titan Capabilities
This Quick Start Guide necessarily omitted coverage of a number of

AFT Titan capabilities. This chapter briefly describes some of the
important capabilities not covered.
Optimize with operating costs spread over multiple cases

For systems with multiple design cases, the system may operate a
significant portion of the time in each design case mode. AFT Titan
allows one to assign a percentage of the operating cost to each relevant
design.
For example, assume you are designing a two-compressor air supply
system which uses both compressors in the summer but only one in the
winter (because of reduced demand for air). The winter operating cost is
thus much less than in the summer. You can account for this by
assigning the percentage of the time the compressor operates in each
design case. This allows a proper accounting of annual operating costs,
and a proper life cycle optimization to be performed.
Vary recurring costs over time

Through "scale tables" AFT Titan allows you to input recurring data for
pipes or junctions that vary over time. For instance, perhaps the
maintenance cost of a piece of equipment is low at first but requires
increased maintenance over some time period. The cost for maintenance
can be varied over time to match the anticipated maintenance schedule.

quickstart 1.doc
Time value of money

An important consideration in evaluating recurring cost items is the cost
of money for future expenses. AFT Titan allows you to enter an interest
rate and inflation rate to account for actual value of future expenses
based on today's currency value. This data is entered in the Optimization
Control window.
Working with different currencies

Cost data can be entered in any currency you wish. Different currencies
can be defined in the Parameter and Unit Preferences window, and then
used as a basis for cost databases.
Costs vs. size

Scale tables can be used to model how equipment costs vary with pipe
diameter. For instance, a 4-inch valve will cost more than a comparable
2-inch valve. This can be accounted for in the cost database by use of
scale tables.
Generic compressor and fan costs can be modeled as a function of
power, and generic control valve costs can be modeled as a function of
maximum Cv.
Optimizing rectangular duct systems

AFT Titan can optimize systems with only rectangular ducts or a
combination of rectangular and cylindrical ducts. Rectangular ducts have
two space dimensions, and thus two design variables are required for
each independent duct size.
Compare VFD vs. FCV optimized systems

Variable Frequency Drive (VFD) control methods are frequently
compared to Flow Control Valve (FCV) methods. However, this is
typically only evaluated in the context of a given pipe system. AFT Titan

quickstart 1.doc
Chapter 5 Other AFT Titan Capabilities 67
allows one to optimize the system for each of these methods. This allows
a more meaningful comparison.
Maximum cost groups

When evaluating multiple operating cases, one of the compressor or fan
operating modes will drive the compressor or fan selection for all cases.
This is by virtue of the fact that the cost will be based on the mode
requiring the most power and/or largest pressure rise. This can be
modeled in AFT Titan with Maximum Cost Groups. A Maximum Cost
Group allows one to couple costs together for all compressors or fans
that will eventually be of the same design (and thus cost the same). This
allows the proper cost information to be used for each design case.
Maximum cost groups can also be used for control valves.
Network databases
Engineering and cost databases can be located on local PC's or deployed
across local or wide area networks. The Database Manager allows users
to connect to relevant databases for their specific pipe system design.
5/7/2003 2:31 PM Page 67


quickstart 1.doc
Creating for pipes 36
Index Junctions 40
Pipes 34
A
Control valve 9
AFT DEFAULT INTERNAL
DATABASE 31 Control Valve constraint set 19
AFT Titan Control Valve junction 14
engineering assumptions 5 Control Valve Optimization Parameters

window 19
Overview 5
Cost Report 25, 43, 60
summary of capabilities 4
Currency units 66
C
D
Chempak 4
Database Manager 31, 67
Compresors
Databases 67
Cost of power 40
Cost 31
Compressor Curve 39
DDC See Dependent Design Case
Compressor/fan junction
Dependent Design Case 55
Compressor/fan Curve 39
Enabling in Optimization Control 55
Compressor/Fan Specifications window
Linking See
Nominal efficiency 40
Design variable 7, 35, 36, 62, 66
Compressors
linking 8
Optimizing with compressor curves
49 Duct optimization 66
sizing method 39 Duplicate Special 57
Constraint 7, 34
E
Active and Inactive 7
Elevation 4
General junction 40
Pipes 34
F
Constraint Set
Fan Curve 39
5/7/2003 2:31 PM Page 69 c:\aft manuals\t3\quickstart\titan3 quickstart 1.doc

Fans Operational costs 45
sizing method 39 Optimization
Flow Control Valve 19, 66, 67. See Continuous vs. discrete 8
Control valve
Optimization Control window 23, 29,
Maximum Cost Groups 67 55, 66
Optimization Summary window 36, 54
G Output window 6
Global Junction Edit 58 Pipe Optimization 45, 60
Graph Results window 6 Overview of AFT Titan 5
I P
Infeasible design 8 Parameter and Unit Preferences window
Installation costs 43 66
Pipe Drawing Tool 11
J Pipe linking See Linking
Junctions Pipe Optimization Parameters window
Costs 40 16, 33, 34, 36, 54
Pipe Size Range Set 16. See Size
Range Set
L
Pipe Specifications window 15, 20
Link basis pipe 35, 54
Optimization data 32
Linking 8, 35
Pipes
M Optimizing 32
Material costs 43 Primary Design Case 53
Maximum Cost Groups 67 Primary windows 5
Model Data window 6

R
O Rectangular duct optimization 66
Objective 7 Recurring costs 65

Rotational acceleration 4
5/7/2003 2:31 PM Page 70 c:\aft manuals\t3\quickstart\titan3 quickstart 1.doc

Index 71
S
Scenario Manager 28, 62
Size Range Set
Creating for pipes 36
Pipes 33
Solution Progress window 24
Supersonic flow 5
System Life 30
System Properties window 12
T
Tank junction 13
Time value of money 66
V
Variable Frequency Drive 66
Visual Report window 6
W
Workspace window 6, 11
5/7/2003 2:31 PM Page 71


Titan 3.0 Quick Start

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Titan 3.0 Quick Start

Transféré par

Droits d'auteur :

Formats disponibles

AFT Titan

Quick Start Guide

AFT Titan version 3.0

5/7/2003 2:31 PM Page 1 c:\aft manuals\t3\quickstart\titan3

2003 Applied Flow Technology Corporation. All rights reserved.

5/7/2003 2:31 PM Page 2 c:\aft manuals\t3\quickstart\titan3

2. Weight Optimization Example ....................................... 9

5/7/2003 2:31 PM Page iii c:\aft manuals\t3\quickstart\titan3

D. Enter data for tanks ..................................................................... 13

3. Initial and Life Cycle Cost Optimization Example ..... 27

5/7/2003 2:31 PM Page iv c:\aft manuals\t3\quickstart\titan3

Understanding the model........................................................................ 41

4. Multiple Design Case Example ................................... 51

5. Other AFT Titan Capabilities ....................................... 65

5/7/2003 2:31 PM Page v

5/7/2003 2:31 PM Page vi c:\aft manuals\t3\quickstart\titan3

Introducing AFT Titan

Welcome to AFT Titan 3.0, Applied Flow Technology's powerful

How AFT Titan works

5/7/2003 2:31 PM Page 1 c:\aft manuals\t3\quickstart\titan3

Compressible Flow No Converged

Figure 1.1 AFT Titan main component flow chart

The Compressible Flow Solver obtains a balanced flow solution for a

5/7/2003 2:31 PM Page 2 c:\aft manuals\t3\quickstart\titan3

Analysis vs. design

5/7/2003 2:31 PM Page 3

Cost-based optimization vs. cost estimating

AFT Titan design capabilities

Types of systems that can be optimized

5/7/2003 2:31 PM Page 4 c:\aft manuals\t3\quickstart\titan3

Optimization parameters available

Engineering assumptions in AFT Titan

AFT Titan Primary Windows

5/7/2003 2:31 PM Page 5

Figure 1.2 Primary window workflow in AFT Titan

5/7/2003 2:31 PM Page 6 c:\aft manuals\t3\quickstart\titan3

Active and inactive constraints

5/7/2003 2:31 PM Page 7

Continuous vs. discrete optimization

Design variable linking

Feasible and infeasible designs

5/7/2003 2:31 PM Page 8 c:\aft manuals\t3\quickstart\titan3

Weight Optimization Example

This example shows how to use AFT Titan to perform an engineering

5/7/2003 2:31 PM Page 9 c:\aft manuals\t3\quickstart\titan3

Step 1. Create the model

5/7/2003 2:31 PM Page 10 c:\aft manuals\t3\quickstart\titan3

Figure 2.1 Control Valve system layout

B. Specify solution method

5/7/2003 2:31 PM Page 11

Figure 2.2 Solution Control for steam transfer system

5/7/2003 2:31 PM Page 12 c:\aft manuals\t3\quickstart\titan3

Figure 2.3 System Properties window is where you specify the

D. Enter data for tanks

5/7/2003 2:31 PM Page 13

Figure 2.4 Input for Tank junction J1

E. Enter control valve data

5/7/2003 2:31 PM Page 14 c:\aft manuals\t3\quickstart\titan3

Figure 2.5 Input for Control Valve junction J2

F. Enter pipe data

5/7/2003 2:31 PM Page 15

Figure 2.6 Pipe Specifications window for pipe P1

Step 2. Setup the optimization data

A. Create pipe size range set