Académique Documents
Professionnel Documents
Culture Documents
Applied Flow
Technology
"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.
Graphical Interface
Input Output
Yes
Optim izer
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).
Model Data
Visual Report
Workspace Output
Graph Results
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.
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).
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).
Summary
Topics covered
Engineering parameter (pipe weight) optimization
Pipe linking options
Control Valve constraints
Required knowledge
No prior knowledge is required for this example.
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
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.
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.)
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
Repeat this process for junction J3 (Discharge Tank), but use a Pressure
of 200 psia.
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
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.
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.
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.
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.
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.
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
Figure 3.7 Pipe size range sets are created and modified in the
Pipe Optimization Parameters window.
Figure 3.8 Pipe Constraint Sets are created and modified in the
Pipe Optimization Parameters window.
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.
Figure 3.11 Constraint sets for control valves are created in the
Compressor/Fan Optimization Parameters window.
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
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
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.
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
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
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
Figure 3.16 Pipe sizes selected by AFT Titan for first cost, and life
cycle cost over 5 years.
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.
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.
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.
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.
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.
Figure 4.3 Optimization Summary Window with all the pipe mains
linked to each other and all distribution pipes linked
Figure 4.6 Workspace with Primary Design Case (at top) and one
Dependent Design Case (at bottom)
Figure 4.7 Global Junction Edit is used to change the FCV flows
to 6.25 lbm/sec to the burners in the DDC.
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).
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.
Figure 4.12 Optimum pipe sizes for the case with eight link basis
pipes
Figure 4.13 Cost Report for the case with eight link basis pipes
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.
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
Constraint 7, 34
E
Active and Inactive 7
Elevation 4
General junction 40
Pipes 34
F
Constraint Set
Fan Curve 39
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
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