Transient - Chapter 5 - Cooling Water System Part 2

PIPENET VISION
TRAINING MANUAL
TRANSIENT MODULE
CHAPTER 5
A COOLING WATER SYSTEM INCLUDING A

CONTROLLER – PART 2
Revision 2.5, September 2017

PIPENET VISION TRAINING MANUAL TRANSIENT: CHAPTER 5
PAGE 1 OF 76 REVISION 2.5, SEP 2017
1. INTRODUCTION....................................................................................................................................... 2
1.1 INTRODUCTION TO CONTROL SYSTEMS .................................................................................................... 2
1.2 CASCADE CONTROL ................................................................................................................................ 3
1.3 CONCEPTS OF TURBO PUMPS.................................................................................................................. 4
1.4 TURBO PUMPS IN PIPENET VISION ....................................................................................................... 5
2. NETWORK SCHEMATICS ....................................................................................................................... 6
3. INPUT DATA ............................................................................................................................................ 7
3.1 INITIALISATION DATA ............................................................................................................................... 7
3.2 UNITS ..................................................................................................................................................... 8
3.3 FLUID PROPERTIES ............................................................................................................................... 10
3.4 LIBRARY DATA ...................................................................................................................................... 10
3.5 INITIALISATION DATA REVISITED............................................................................................................. 13
4. NETWORK DATA................................................................................................................................... 13
4.1 SHORT PIPES ....................................................................................................................................... 13
4.2 ELASTIC PIPES...................................................................................................................................... 14
4.3 SPECIFICATIONS ................................................................................................................................... 16
5. THE SIMULATION CASES .................................................................................................................... 27
5.1 CASE 1: BASE CASE (PUMP TRIP, NO VACUUM BREAKERS, AND NO FLY W HEEL)...................................... 27
5.2 CASE 2 (PUMP TRIP, NO VACUUM BREAKER, AND W ITH A FLY W HEEL) .................................................... 32
5.3 CASE 3 (PUMP TRIP, SMALL VACUUM BREAKER, AND NO FLY W HEEL)..................................................... 34
5.4 CASE 4 (PUMP TRIP, LARGE VACUUM BREAKER, AND NO FLY W HEEL) .................................................... 36
6. CONCLUSIONS...................................................................................................................................... 37
APPENDIX A. PRINCIPLES OF CONTROL SYSTEMS .............................................................................. 38
1 CASCADE PID CONTROLLER ................................................................................................................. 38
2. DIGITAL SENSORS ................................................................................................................................. 39
3. IMPROVED TRANSFER FUNCTION ........................................................................................................... 40
APPENDIX B. HOW TO CREATE SUTER CURVES FROM PUMP PERFORMANCE CURVES .............. 52
1. INTRODUCTION...................................................................................................................................... 52
2. BACKGROUND ....................................................................................................................................... 52
3. METHODOLOGY .................................................................................................................................... 53
4. EXAMPLE 1: DEDUCE THE SUTER CURVE FROM A KNOWN P-Q CURVE ................................................... 54
5. EXAMPLE 2: DEDUCE AN UNKNOWN SUTER CURVE FROM THE KNOWN PART .......................................... 56
6. PROGRAM ............................................................................................................................................ 59
7. SUMMARY ............................................................................................................................................. 59
APPENDIX C. THE CALCULATION OF HYDRAULIC TRANSIENT FORCES IN A PIPE NETWORK ..... 62
1. INTRODUCTION...................................................................................................................................... 62
2. BASIC CONCEPTS AND PRINCIPLES ........................................................................................................ 62
3. EXAMPLE: A SIMPLE STEAM NETWORK .................................................................................................. 64
4. RESULTS AND DISCUSSION .................................................................................................................... 69
5. CONCLUSIONS ...................................................................................................................................... 76
1. Introduction
The PIPENET VISION Transient Module can be used to illustrate how to reduce and
control severe pressure oscillations, which can sometimes be caused by a pump trip.
When a pump trips, the effects are similar to those caused by a valve closing; namely, a
vacuum pressure can be created downstream of the pump, which can subsequently result
in pressure oscillations.
In this chapter of the training manual, PIPENET VISION is used to model a cooling water
system. Cooling water systems are increasingly constructed using GRP pipes, the main
reasons being that GRP is less susceptible to corrosion and is cheaper. However, in order
to harness these highly desirable qualities, when a cooling water system is built using
GRP material, more attention must be given to the design aspects. The systems generally
have to operate within a relatively narrow pressure range, and pressure oscillations should
be minimised.
In a cooling water network, GRP pipes may need to work under vacuum conditions after a
pump trip. The GRP pipes may leak, or even collapse, as a GRP pipe is generally able to
withstand only a limited negative pressure. Further, it is important to minimise pressure
oscillations. Pipe fittings are attached to pipes using adhesives, rather than by welding.
Severe pressure oscillations can result in relative movement between the pipe material
and the fitting, and ultimately cause the pipe and fitting to become detached.
The PIPENET VISION Transient Module is used to compare two alternative ways of
resolving the difficulties caused by a pump trip. In the first approach, vacuum breakers are
used, and, in the second, a flywheel is attached to the pump rotor. PIPENET VISION is
used to investigate the moment of inertia of the flywheel, which is the main factor of
interest. Also, when vacuum breakers are added to the network, PIPENET VISION is
used to determine the optimal size of the vacuum breakers.
One key aspect of the model is the following. In the usual pump performance curve
supplied by manufacturers, the downstream pressure is higher than the upstream
pressure. However, when a pump trips, the downstream pressure becomes lower than the
upstream pressure. PIPENET VISION contains a turbo-pump component for modelling
such situations, and this model must be used in this situation.
In the network, the pump speed is controlled by a pressure-control system and the valve
position is controlled by a flow-control system.
The system generally has to operate within a relatively narrow pressure range, and
pressure oscillations should be avoided if possible.
1.1 Introduction to Control Systems
1.1.1 Proportional Term
For a given change in the error, a large proportional gain results in a large change in the
output. If the proportional gain is too high, the system can become unstable. In contrast, a
small gain results in a small output response to a large input error, and a less responsive
(or less sensitive) controller. If the proportional gain is too low, the control action may be
too small when responding to system disturbances.
In the absence of disturbances, pure proportional control will not settle at its target value,
but will retain a steady-state error, which is a function of the proportional gain and the
process gain. Despite the steady-state offset, both tuning theory and industrial
practice indicate that it is the proportional term that should contribute to the bulk of
the output change.
1.1.2 Integral Term
The integral term (when added to the proportional term) causes the progress towards the
set point to be accelerated, and eliminates the residual steady-state error that occurs
with a proportional-only controller. However, since the integral term is responding to
accumulated errors from the past, it can cause the present value to overshoot the set-point
value. If the integral gain is too high, the control system can become unstable.
1.1.3 Derivative Term
The derivative term slows the rate of change of the controller output, and this effect is most
noticeable close to the controller set point. Hence, derivative control is used to reduce
the magnitude of the overshoot produced by the integral component. In essence, the
speed at which the system is approaching the set point is detected, and the controller
output is changed accordingly. However, the differentiation of a signal causes the noise to
be amplified, and may result in a process becoming unstable, especially if the noise and
the derivative gain are sufficiently large.
The characteristics of the P, I, and D Terms are summarised in the following table.
Control term Response time Overshoot Settling time Steady State Error
Proportional Decrease Increase Small Change Decrease
Integral Decrease Increase Increase Eliminate
Derivative Small Change Decrease Decrease Small Change
1.2 Cascade Control
One distinctive advantage of PID controllers is that two PID controllers can be used
together to yield a better dynamic performance. This control is called cascaded PID
control, in which the PIDs are arranged so that one PID controls the set point of another.
1.3 Concepts of Turbo Pumps
Consider the relationships
where
is the flow rate,

is the head,
is the rotational speed,
is the torque of the pump and motor combined.
The suffix represents values at the Rated Point of Operation, which is the point of
maximum efficiency.
The angle, , which is used to define the operating zones for the pump, is given by
1.3.1 Operating Zones for the Pump
The eight operating zones are illustrated in the following diagram.

1.3.2 Characterisation of Turbo Pumps by Suter Curves
For the purpose of standardisation, the operating zones are characterised using Suter
Curves. The following two dimensionless parameters are used for defining these curves.
The above parameters are normally called the Head and Torque Suter parameters. These
parameters can be plotted against the angle θ, and the resulting curves are called Suter
Curves.
The Suter Curves depend on the specific speed of the pump. The specific speed is one of
the main parameters that characterises turbo machines, and is defined as follows.
in which
is the power (kW),

is the head (m),
is the pump speed (rpm).
The specific speed depends predominantly on the impeller type (axial, radial or mixed), the
shape of the impeller, the casing, and other parameters.
1.4 Turbo Pumps in PIPENET VISION
In the PIPENET VISION Transient Module, turbo pumps are used to model pump trip,
particularly if the vacuum pressure downstream of the pump is of concern. In the
operating-zones diagram above, the turbo pump usually functions in the modes of normal
pumping, energy dissipation, and reverse turbine.
PIPENET VISION has built-in Suter Curves for the following specific speeds.
= 261 - axial,
= 147 - mixed,
= 25 - radial.
Appendix 2 of this chapter of the Training Manual contains a description of how to

generate approximations for the four-quadrant Suter Curves from a known pump
performance curve that covers only the positive quadrant.
Sunrise Systems has also developed an Excel spreadsheet that can be used for
generating Suter Curves from known pump performance curves. This spreadsheet can be
supplied free of charge to customers of the PIPENET VISION Transient Module.
2. Network Schematics
The network that we shall be using for this part of the training course is generally as
follows. However, it will vary slightly, depending on the exact scenario, but the basic
system will remain unchanged. The input stage of this network will begin in Section 3.
The pipe bores and elevations are depicted in the following diagram.
3. Input Data
3.1 Initialisation Data
From the Options Menu, choose “Module options” (which is represented in PIPENET
VISION as Options | Module options).
The preliminary options for the Transient Module can be entered using the following dialog
box.
3.2 Units
The units that are to be used are tabulated below.
Variable Unit
Length metres
Diameter mm
Velocity m/sec
Temperature Celsius
Density kg/m3
Viscosity cP
Time seconds
Mass kg
Mass rate kg/s
Torque Nm
Inertia kg m2
Force N
Volume m3
Surface Tension N/m
Thermal Conductivity W/(m K)
Heat Capacity J/kg K
Young’s Modulus G Pa
Pressure Bar Gauge
Flow type Volumetric
Flow rate m3/hr
Input these units into the Transient Module using the following dialog box (which is
obtained from Options | Units, with “User defined” selected). Note that, to see the flow-
rate units, double-click on “Flow type”.
3.3 Fluid Properties

The following fluid properties are to be used.
At this point we take a diversion, and enter the user-defined data into the library.
3.4 Library Data

The reason that we need to take this diversion is because we cannot select the pipe type
(which is required during the initialisation stage) until after the user-defined pipe schedules
have been input into the library.
3.4.1 Pipe Schedule Data

The pipe schedules dialog can be obtained via Libraries | Schedules, as indicated below.
Only two pipe schedules, called “CS flake line 1” and “GRP 1” are to be used in the
network. So, the dialog boxes for only those pipe schedules are shown.
CS flake line 1
Enter the following data for the pipe type CS flake line 1.
GRP 1
Enter the following data for GRP 1.

3.4.2 Turbo Pump Data
It is assumed that the built-in Suter Curve for the Axial Type impeller (with a specific speed
of 261 rpm) can be used to represent the behaviour of the pumps. Please refer to
Appendix 2, which contains a description of how to create user-defined Suter Curves.
3.4.3 Fittings data
The following user defined fittings are needed for the network. Go to Libraries | Fittings
and enter the following fittings in the library:
45BE Type K-factor, K-factor=0.11

90BE Type K-factor, K-factor=0.26
DE06 Type Device, Head loss=1.37 bar Flow rate 24500.02 m3/h
DE1 Type Device, Head loss=1.55 bar Flow rate 18666.68 m3/h
MI3 Type Mitre, Angle= 3
3.5 Initialisation Data Revisited
Having input the user-defined pipe schedules into the library, we now need to select them
for use in the network (via Options | Pipe types).
4. Network Data
As the network in this chapter is quite complex and requires much precision and attention
to input, users may want to use the ready network from data files available in Help |
Learning | Training Materials | Data Files. However, in order to practise using PIPENET,
it is recommended to input the network, following the guidance below.
4.1 Short Pipes
The short pipes in the model are highlighted in the following diagram.
The data for these short pipes is tabulated below.
Label Input Output Diameter Length Net Height Roughness

Node Node (mm) (m) Change (m) (mm)
1 2 28 1500 15 0 0.06
2 5 29 1500 15 0 0.06
3 517 55 1500 3 0 0.06
4 181 8 1500 3 0 0.06
5 9 135 1500 3 0 0.06
6 188 12 1500 10 0 0.06
10 37 55 1500 20 0 0.06
11 39 56 1500 20 0 0.06
79 247 190 1500 10 0 0.06
272 56 471 1500 3 0 0.06
As an example, a completed dialog box for Pipe 1 is shown below.
4.2 Elastic Pipes
The elastic pipes of the model are highlighted below.

The properties of these pipes are presented in the table below. Note that the additional
K-factor for all of these pipes is zero.
Label Input Output Diameter Length Net Height Roughness Type Fittings
Node Node (mm) (m) Change (m) (mm)
7 16 117 950 800 -3.12 0.2 CS flake DE1
line 1
8 17 111 950 800 -3.12 0.2 CS flake DE1
line 1
9 7 135 950 164.308 0 0.2 CS flake 90BE x2
line 1
12 66 250 1500 120.344 0 0.06 GRP 1 90BE x2
13 250 199 1500 318.666 -4.083 0.06 GRP 1 90BE x2
MI3 x2
14 199 216 1200 600 3.786 0.06 GRP 1 45BE x4
90BE x2
24 64 87 1500 120.344 0 0.06 GRP 1 90BE x2
25 87 207 1500 318.666 -4.083 0.06 GRP 1 90BE x2
MI3 x2
36 207 17 950 600 2.67 0.2 CS flake DE1
line 1
38 199 16 950 600 2.67 0.2 CS flake DE1
line 1
42 123 138 950 224.171 4.197 0.2 CS flake 45BE x2
line 1 90BE x2
Tee (run) x2
44 127 7 950 224.171 4.197 0.2 CS flake 45BE x2
line 1 90BE x2
Tee (run) x2
45 138 135 950 164.308 0 0.2 CS flake 90BE x2
line 1
76 172 181 1200 164.308 0 0.06 GRP 1 90BE x2
91 207 214 1200 600 3.786 0.06 GRP 1 45BE x4
90BE x2
96 214 218 1200 1000 -5.285 0.06 GRP 1 DE06
97 216 221 1200 1000 -5.285 0.06 GRP 1 DE06
98 224 245 1200 331.31 5.246 0.06 GRP 1 90BE x3
Tee (run) x2
99 226 172 1200 331.31 5.246 0.06 GRP 1 90BE x3
Tee (run) x2
118 245 181 1200 164.308 0 0.06 GRP 1 90BE x2
A completed dialog box for Pipe 36 is shown below.
4.3 Specifications
The nodes that require specifications are highlighted in red in the figure below.
The numbers of these nodes can be seen on the following diagram.

The specifications for the nodes are as follows.
Node Specification Type Time Function Constant Value

1 Pressure Constant 0.3 bar G
4 Pressure Constant 0.3 bar G
10 Information Constant 1
13 Information Power ramp See below
189 Information Power ramp See below
Power ramps have been used for the specifications at Nodes 13 and 189. The position of
the corresponding valves has been set to a constant value of 0.3. This setting is merely for
convenience, and, if we subsequently wished to simulate a valve-closure scenario, this
scenario could be modelled simply by setting the stop value to 0.
The transient phenomenon results from either the pumps tripping or the valves closing.
Whether it is the pumps that trip or the valves that close depends on the scenario under
consideration, and this point is highlighted below in the appropriate scenario. For the
pump-trip case, the trip time is specified as part of the input data for the pump; for the
valve-closure case, the specification is applied to the information node of the valve.
4.3.1 Turbo Pumps
There are two turbo pumps in the network, and these are highlighted in the diagram below.
The data for these pumps is as follows. Note that the turbo pumps trip 20 seconds into the
simulation.
Turbo Input Output Information Pump Rated Rated Speed Torque MoI Trip
-1 2
Pump Node Node Node Type Pressure Flowrate (sec ) (Nm) (Kgm ) time
Label (Bar) (m3/hr) (sec)
RLC
1 1 2 3 Pump 4.2 16840 6.167 57700 765 20
RLC
2 4 5 6 Pump 4.2 16840 6.167 57700 765 20
A sample Properties Dialog Box is as follows.
4.3.2 Receiving Vessels
The two receiving vessels are highlighted in the diagram below.

The data for the components is shown in the Properties Dialog Box below. The data is
identical for both receiving vessels.
Note that the diameter for the receiving vessel is the hydraulic mean diameter (which
depends on the shape of cross section). The hydraulic diameter, , is defined by
where A is the cross-sectional area and P is the wetted perimeter. Also, in the above
dialog box, the cross-sectional area is the physical cross-sectional area. It is important to
note that the cross-sectional area is not necessarily the area calculated from the hydraulic
mean diameter.
4.3.3 Flow Sensors
There are four flow sensors in the network, and they are highlighted in the figure below.
The node numbers for the sensors are as follows.

Label Input Node Output Node Information Node

1 112 123 124
3 118 127 128
4 219 224 225
5 222 226 227
Note that a flow sensor has to be in-line in a network. It has an input node, an output
node, and an information node (which corresponds to the output signal).
The Properties Dialog Box for Flow Sensor 1 is as follows.
4.3.4 Pressure Sensors
The two pressure sensors are highlighted in the following schematic diagram of the
network.
The properties of the first sensor are indicated below.

4.3.5 Transfer Functions
There are twelve transfer functions in the network, and these are highlighted below.
The data for these transfer functions is tabulated below.
Label Input Output Input Type Output Type Order Gain Time Constant
Node Node (seconds)
1 124 129 Flowrate Flowrate First 1 0.5
2 130 113 Information Information First 1 5
31 62 488 Pressure Pressure First 1 0.5
33 482 490 Pressure Pressure First 1 0.5
It is important to note that a transfer function represents the dynamics of measuring

instruments and items such as valves.
A sample Properties Dialog Box is shown below.
4.3.6 PID Controller Data
The locations of the PID controllers in the network are illustrated in the following diagram.
The data for the controllers is as follows.
Label Input Output Variable Controller Set Output Gain Reset Track- Rate
Node Node Type Type Point (bar-1 Time ing time
or (sec) Time (sec)
m-3 hr) (sec)
1 129 130 Flowrate PID 3000 0.5 -2E-04 10 4.47 2
2 490 491 Pressure PID 4 0.7 -0.05 10 4.47 2
3 133 134 Flowrate PID 3000 0.5 -2E-04 10 4.47 2
4 230 231 Flowrate PID 5000 0.3 -1E-04 10 4.47 2
5 232 233 Flowrate PID 5000 0.3 -1E-04 10 4.47 2
16 488 489 Pressure PID 4 0.7 -0.05 10 4.47 2
Below is a sample Properties Dialog Box.

4.3.7 Operating Valves and Control Valves

Create the following valves in your local library.
Note that (in the above figure), when the s-value is 1.0, the flow coefficient is 10000 (m 3/h,
Bar1/2).
The values for the DN1500 Valve are tabulated below.
s-value Flow Coefficient

0 0
0.111 233.581007
0.1667 1263.065559
0.2222 3508.0411
0.2778 6500.469032
0.3333 10523.24672
0.3889 16228.63186
0.4444 23384.05667
0.5 31708.98029
0.5556 42091.25646
0.6111 54952.89746
0.6667 70152.05616
0.7222 88719.11646
0.7778 116920.2833
0.8333 161326.6306
0.8889 210456.2823
0.9444 232507.4769
1.0 233840.5667
Once the values for s and the flow coefficient have been entered, the dCv/ds values can
be obtained by clicking on the Linear Button.
Now, associate the appropriate library type with each operating valve.
Label Input Node Output Node Information Node Valve Characteristic

1 28 37 38 DN1500
2 29 39 40 DN1500
3 8 9 10 DN1500
4 135 247 13 DN1500
10 55 64 65 DN1500
11 56 66 67 DN1500
25 111 112 113 FCV950
27 117 118 119 FCV950
36 181 188 189 DN1500
43 218 219 220 FCV1200
44 221 222 223 FCV1200
79 471 517 473 DN1500
5. The Simulation Cases

5.1 Case 1: Base case (Pump Trip, no Vacuum Breakers, and no Fly
Wheel)
In this scenario, both pumps trip when the simulation time is 20 seconds.
As can be seen from the figures below, the calculated minimum and maximum pressures
of the GRP pipes are -0.95 bar G and 4.37 bar G respectively. In the base case, there is
no fly wheel, and the pump and motor assembly have a moment of inertia of 765 kg m 2.
Moment of Inertia
In order to obtain a good initial state for the system, it is necessary to increase the run-in
time (via Calculation | Options… | Initial state).
Run the simulation (either by (a) clicking on the red Calculate Button and selecting “OK” or
(b) choosing Calculation | Go… and pressing the OK Button).
Open the graph viewer (either (a) by clicking on the Graph Viewer Button or (b) via Output
| Graphs…). Next, plot a graph of the speed of Turbo Pump 1 with time (by clicking on the
prompt next to the directory tree, opening “All variables of pump 1”, and ticking the box for
“Speed of turbo pump 1”) , to obtain the following graph.
Open the output file for the simulation (using either the Browse Output Button or Output |
Report…). This file contains the following results for the global pressure.
PRESSURE EXTREMA
----------------
Maximum pressure is 4.37468 bar G

on pipe 38 at the inlet
at time 4.300000 seconds
Minimum pressure is -0.950000 bar G

on pipe 98 110.4 metres from the inlet of the pipe
The following values (for the maximum and minimum pressure in the pipes of the system)
are also present in the output file, an example of which is as follows.
MAXIMUM/MINIMUM PRESSURE
------------------------
Maximum Pressure Minimum Pressure

Component Pressure Time Position Pressure Time Position Pipe Type
bar G seconds metres bar G seconds metres
------------------------------------------------------------------------------------------------------------------------
pipe
36 4.375 4.1 Inlet -0.743 32.7 Outlet CS flake line 1
38 4.375 4.3 Inlet -0.743 32.7 Outlet CS flake line 1
42 2.243 30.8 Inlet -0.151 22.8 Inlet CS flake line 1
44 2.243 30.8 Inlet -0.151 22.8 Inlet CS flake line 1
45 1.565 4.6 Inlet 0.168 32.7 Inlet CS flake line 1
7 4.190 4.9 Outlet -0.761 32.7 114.286 CS flake line 1
8 4.190 4.7 Outlet -0.761 32.7 114.286 CS flake line 1
9 1.565 4.7 Inlet 0.168 32.7 Inlet CS flake line 1
118 1.559 3.3 Inlet -0.609 25.1 Inlet GRP 1
76 1.559 16.0 Inlet -0.609 25.1 Inlet GRP 1
24 3.999 6.5 Inlet -0.282 28.0 Outlet GRP 1
25 4.375 4.1 Outlet -0.306 28.0 53.111 GRP 1
91 4.375 4.1 Inlet -0.688 27.7 436.364 GRP 1
96 4.342 2.6 1000.000 -0.674 26.2 526.316 GRP 1
97 4.342 3.0 1000.000 -0.674 26.2 526.316 GRP 1
98 2.121 2.6 Inlet -0.950 25.3 110.437 GRP 1
99 2.121 3.6 Inlet -0.950 25.3 110.437 GRP 1
12 3.999 6.7 Inlet -0.282 28.0 Outlet GRP 1
13 4.375 4.3 Outlet -0.306 28.0 53.111 GRP 1
14 4.375 4.3 Inlet -0.688 27.7 436.364 GRP 1
--------------------
short pipe
1 4.003 6.0 Inlet 0.123 28.2 Outlet
2 4.003 6.4 Inlet 0.123 28.2 Outlet
10 4.001 6.5 Inlet 0.060 28.1 Outlet
11 4.001 6.5 Inlet 0.060 28.1 Outlet
272 4.000 6.7 Inlet 0.060 28.1 Inlet
79 0.515 38.5 Inlet 0.346 200.0 Inlet
3 4.000 6.7 Inlet 0.060 28.1 Inlet
4 1.543 0.1 Inlet 0.322 200.0 Outlet
5 1.543 0.1 Inlet 0.322 200.0 Outlet
6 0.518 38.3 Inlet 0.346 200.0 Inlet
5.2 Case 2 (Pump Trip, no Vacuum Breaker, and With a Fly Wheel)
This case is similar to the base case, except that a fly wheel is used, causing the moment
of inertia of the pumps to be increased 5-fold (from 765 kg m2 to 3825 kg m2).
Moment of Inertia
The calculated minimum and maximum pressures of the GRP pipes are -0.028 bar G and
4.37 bar G respectively. In this scenario, the oscillation is reduced significantly in
comparison to the base case. The moment of inertia of fly wheel can be optimised, based
on the results of PIPENET VISION simulations.
In this case, the graph of the speed of Turbo Pump 1 with time is as follows.
The following text appears in the output file.
PRESSURE EXTREMA
----------------

Minimum pressure is -0.275884E-01 bar G

5.3 Case 3 (Pump Trip, Small Vacuum Breaker, and no Fly Wheel)
This case is the same as the base case, except that small vacuum breakers (with
diameters of 50 mm at the air inlet valve and 25 mm at the air outlet valve) are installed in
the system. Please see the figure below for the locations of the vacuum breakers.
The data for the vacuum breakers is tabulated below.
Label Input Output Air Inlet Air Outlet

Node Node Diameter Trigger Pressure Diameter Trigger Pressure
(mm) (bar G) (mm) (bar G)
1 17 18 50 1 0 25 1 0
2 39 14 50 1 0 25 1 0
3 16 19 50 1 0 25 1 0
4 7 20 50 1 0 25 1 0
5 245 21 50 1 0 25 1 0
6 172 22 50 1 0 25 1 0
7 37 23 50 1 0 25 1 0
8 87 24 50 1 0 25 1 0
9 250 25 50 1 0 25 1 0
10 123 26 50 1 0 25 1 0
11 224 27 50 1 0 25 1 0
12 138 513 50 1 0 25 1 0
13 127 30 50 1 0 25 1 0
14 226 31 50 1 0 25 1 0
15 214 516 50 1 0 25 1 0
17 216 518 50 1 0 25 1 0
Note that vacuum breakers are not used in all of the scenarios. One of our objectives is to
select the optimal size of the vacuum breakers. Therefore, the above data is not used in
all of the cases considered.
The dialog box for Vacuum Breaker 1 is illustrated below.
We reduce the time step up to 0.02 seconds and run the simulation. From the output file, it
can be seen that the calculated minimum and maximum pressures of the GRP pipes are -
0.48 bar G and 4.38 bar G respectively.
PRESSURE EXTREMA
----------------


The graph of the speed of Turbo Pump 1 with time is as follows.
The effect of vacuum breakers is to reduce the oscillations in pressure, which can be
observed if pressure graphs for this case are compared with corresponding graphs for the
base case.
5.4 Case 4 (Pump Trip, Large Vacuum Breaker, and no Fly Wheel)
In this case, larger vacuum breakers (with diameters of 200 mm for the air inlet valve and
100 mm for the air outlet valve) are used in the system. The system schematic is same as
for Case 3, and the data for the vacuum breakers is tabulated below.
Label Input Output Air Inlet Air Outlet

Node Node Diameter Trigger Pressure Diameter Trigger Pressure
(mm) (bar G) (mm) (bar G)
1 17 18 200 1 0 100 1 0
2 39 14 200 1 0 100 1 0
3 16 19 200 1 0 100 1 0
4 7 20 200 1 0 100 1 0
5 245 21 200 1 0 100 1 0
6 172 22 200 1 0 100 1 0
7 37 23 200 1 0 100 1 0
8 87 24 200 1 0 100 1 0
9 250 25 200 1 0 100 1 0
10 123 26 200 1 0 100 1 0
11 224 27 200 1 0 100 1 0
12 138 513 200 1 0 100 1 0
13 127 30 200 1 0 100 1 0
14 226 31 200 1 0 100 1 0
15 214 516 200 1 0 100 1 0
17 216 518 200 1 0 100 1 0
The calculated minimum and maximum pressures in the GRP pipes are -0.95 bar G and
7.81 bar G respectively.
PRESSURE EXTREMA
----------------

on pipe 97 at the outlet

Although the minimum pressure is the same as in the base case, the magnitude of the
oscillations is larger than that in Case 3 (in which smaller vacuum breakers were used).
Therefore, there is an optimal size for the vacuum-breaker valves, and the PIPENET
VISION Transient Module can be used to find this optimum. For this network, the results
from the 50/25 mm vacuum-breaker valves are better than those from the 200/100 mm
vacuum-breaker valves.
6. Conclusions
Pump trip and its associated problems are discussed in this document. Including a fly
wheel and including vacuum breakers are both effective methods for reducing and
controlling pressure oscillations. With reference to the simulation results, it can be seen
that the size of the vacuum-breaker valves and fly wheel can also be optimised.
APPENDIX A. PRINCIPLES OF CONTROL SYSTEMS
1 Cascade PID Controller
1.1 Introduction
Unlike the set point of the traditional PID controller, the set point of a cascade PID
controller can be changed with the time, or based on the signal from another cascade PID
controller. If a time type cascade PID controller is given a constant set point, it can work
as a traditional PID controller.
1.2 Typical Connection
Based on the control type of the set point, a cascade PID controller can be classified as
being of type time, pressure, differential pressure, or flow rate.
Type 1 Cascade Control: Time type Cascade PID Controller
Reset the input set point
X 1/0 Y
Sensor Cascade PID Controller Transfer function Equipment
?
Type 2 Cascade Control: Pressure, differential pressure and flowrate type Cascade PID
Controller
Primary Cascade
PID Controller Sensor2
Reset the input

set point
X Secondary Cascade 1/0 Y

Sensor1 PID Controller Transfer function Equipment
?
1.3 Typical Applications

1.3.1 Time Type Controller
1.3.2 Pressure, Differential Pressure and Flow Type Controller
2. Digital Sensors
2.1 Introduction
Digital sensors are used to simulate the behaviour of a discrete data system. The
designed components include the digital pressure sensor, digital flow sensor and digital
differential pressure sensor.
2.2 Symbol and Interface
The digital sensors have the same symbols and interfaces as the conventional (analogue)
sensors. In addition, the digital sensors have a new parameter, the scan frequency, which
has the unit of 1/time. Note that the reciprocal of the scan frequency must not be less than
the calculation time step.
The input and output of a sensor is illustrated in the following diagram.

3. Improved Transfer Function
3.1 Introduction
The order of the original transfer function in PIPENET VISION can be either zeroth, first or
second.
The improved transfer function includes bias, and users can set the minimum or maximum
values of the output. The new transfer function can, therefore, be used for split range
functionality control, and for setting minimum or maximum values in a control valve.
3.2 Interface
3.3 Example Scenarios

3.3.1 Scenario 1: Split Range Functionality Controller
In the above network, the settings for the PID controllers A and B are the same as the
settings for the transfer functions that appear before the PID controllers. Therefore, they
have the same output; i.e.,
X1 = X2 = X.
Please note that the option “Calculate initial steady state” must not be selected; otherwise,
the initial output of the PID controllers A and B may be different.
The purpose of the control system is to maintain a pressure of 2 bar G at the location of
the sensors. The philosophy is that Valve A opens first, and Valve B begins to open after
Valve A is fully open (as is indicated in the table below).
PID Controller output (X) Setting of Valve A (Y1) Setting of Valve B (Y2)
0% 0% 0%
25 % 50 % 0%
50 % 100 % 0%
75 % 100 % 50 %
100 % 100 % 100 %
Y1 = 2*X + 0 Y2 = 2*X - 1
X
( 0% ≤ Y1 ≤ 100% ) ( 0% ≤ Y2 ≤ 100% )
The settings of the Transfer Functions A and B are as follows.
Transfer Function A Transfer Function B
The settings for the PID controllers and operating valves are as follows.
Results
The following graphical results are obtained. Please note that the graphs in this appendix
were not obtained using the most recent version of the PIPENET VISION Graph Viewer.
Figure 1: The inlet pressure of the system and the pressure at the control point
Figure 2: The input and output of the transfer function A (Y=2*X & 0% ≤ Y ≤ 100%)
Figure 3: The inlet and output of the transfer function B (Y=2*X-1 & 0% ≤ Y ≤ 100%)
3.3.2 Scenario 2: Split Range Functionality Controller With Limitation- Case 1
In this scenario, which is based on Scenario 1, the operating valves have minimum and
maximum output values; i.e., 20% ≤ Y ≤ 80%. The other settings are the same as in the
first scenario.
PID Controller Output (X) Setting of Valve A (Y1) Setting of Valve B (Y2)
0% 0% 0%
25 % 50 % 0%
50 % 100 % 0%
75 % 100 % 50 %
100 % 100 % 100 %
Y1 = 2*X + 0 Y2 = 2*X - 1
X
(20% ≤ Y ≤ 80%) (20% ≤ Y ≤ 80%)
The settings for the Transfer Functions A and B are shown below.

Results
The following graphical results are obtained.
Figure 4: The inlet pressure of the system and the pressure at the control point
Figure 5: The input and output of the transfer function A (Y=2*X & 20% ≤ Y ≤ 80%)
Figure 6: The input and output of the transfer function B (Y=2*X-1 & 20% ≤ Y ≤ 80%)
3.3.3 Scenario 3: Split Range Functionality Controller With Limitation- Case 2
This network has two sources, a main source and a spare source. The spare source is
used only when the pressure of the main source is not high enough to maintain the desired
pressure at the sensors. In this scenario, the spare source has constant pressure of 5 bar
G. Further, the initial pressure of the main source is 0 bar G, and this value increases to
10 bar G in 10 seconds. The philosophy is that Valve A is open initially, but is closed when
the pressure of the main supply recovers (as is shown in the table below). In this scenario,
only the operation of Valve B is restricted; i.e., 20% ≤ Y ≤ 80%.
PID Controller Output (X) Setting of Valve A (Y1) Setting of Valve B (Y2)
0% 100 % 0%
25 % 50 % 0%
50 % 0% 0%
75 % 0% 50 %
100 % 0% 100 %
Y1 = -2*X + 1 Y2 = 2*X - 1
X
(0% ≤ Y ≤ 100%) (20% ≤ Y ≤ 80%)
The settings of the Transfer Functions A and B are illustrated below.

Results
The following graphical results are obtained.
Figure 7: The input pressure of the main supply and the pressure at the control point
Figure 8: The inlet and output of the transfer function A (Y=-2*X+1 & 0% ≤ Y ≤ 100%)
Figure 9: The inlet and output of the transfer function B (Y=2*X-1 & 20% ≤ Y ≤ 80%)
APPENDIX B. HOW TO CREATE SUTER CURVES FROM PUMP

PERFORMANCE CURVES
1. Introduction
It is usually possible to obtain the performance curve in the positive quadrant for a pump
without difficulty. In this quadrant, both the flowrate and the head are positive. However,
in order to model the behaviour of a pump in other situations, it may be necessary to
consider all four quadrants. In other words, both positive and negative values for the head
and the flowrate need to be considered. However, this kind of data is often extremely
difficult to obtain, and even pump manufacturers do not generally have such data.
The classic textbook called “Fluid Transients”, by Wylie and Streeter, contains an outline of
a technique based on the use of Suter Curves for modelling the behaviour of a pump in all
four quadrants. It also contains the Suter Curves for a limited range of pumps.
Nevertheless, unless pump data is available for all four quadrants, it is not possible to
determine the Suter Curves for other pumps accurately.
In this document, we show how to use the built-in Suter Curves of PIPENET VISION and
the known pump performance curve (in the positive quadrant) to achieve the following:
1. In the positive quadrant, where the pump curve is known, the Suter Curve follows
the pump curve.
2. In the other three quadrants, where the pump behaviour is not known, we
extrapolate the built-in curves.
2. Background
During transient flow, a pump may experience a reversal in flow through the pump, or a
change in its rotational speed, or both. Furthermore, it may also experience negative
torque values and/or negative pressures. Therefore, in order to accurately model a turbo
pump, more performance data is needed, and this data should encompass regions of
abnormal operation. The Suter Curves graphically represent all operating conditions of a
turbo pump. For a detailed description of the Suter Curves, and the corresponding
concepts and equations, please refer to either the web site of Sunrise Systems Limited or
Chapter 1 of the Transient Module Technical Manual (entitled “Modelling Aspects”).
The figure below contains typical Suter Curves for a Radial Pump. A whole Suter Curve
should cover all four quadrants; i.e. a phase angle lying between 0 and 2π radians. There
are eight possible zones of pump operation: four occur during normal operation and four
are abnormal zones. During a transient event, a pump may enter most, if not all, of the
regions in the figure, depending on the circumstances. However, not all manufacturers can
provide the whole range for the Suter curves. In most cases, only the information at the
normal operating condition can be obtained; i.e., Zone D in the figure below. The key
question is “Is it possible to deduce a full-range Suter Curve reasonably in this case?” A
simple way to solve this problem is described in this document.
3. Methodology
It is well known that there are traditionally three types of pump: radial pump, mixed pump
and axial pump. The classification is closely related to the specific speed, NS, in the
following equation and in the figure below.
where
is the specific speed,

is the rated pump speed (in rpm),
is the rated pump flow rate (in m3/s),
is the rated pump piezometric head (in metres).
Three built-in Suter Curves represent three typical pumps: (1) Radial Pump with =25, (2)
Mixed Pump with =147 and (3) Axial Pump with =261. Please note the units when
using the above equation to calculate the specific speed. In order to supply the missing
part of Suter Curve for a certain pump, we must base it on these three built-in Suter
Curves, and make the following assumptions.
1. A specific speed is only mapped to a pump type, and the value shift from a
radial pump to an axial pump is continual and uninterrupted.
2. A Suter Curve is only matched with a pump type, and the shift from a radial
pump to an axial pump is continual and uninterrupted.
3. Any Suter Curve between the two built-in Suter curves has a linear relationship
with the specific speeds.
In the next section, we shall consider a specific example to illustrate the method.
4. Example 1: Deduce the Suter Curve From a Known P-Q Curve
In this section, we show how to derive the Suter Curve from the known performance curve
in the positive quadrant. We also show that the Suter Curve derived in this manner follows
the performance curve of the pump. In the last table of this section, the performance curve
that we input is compared with the results obtained from the corresponding Suter Curve.
4.1 Known Parameters
The following values are known.
= 47.5 m,
= 42840 m3/hr,
= 6.16667 1/s = 370 rpm,
= 164 KN.m,
= 11530 kg.m2.
The Turbo Pump Characteristic and Suter Curve are tabulated below.
    

3
m /h m rpm % KW KN.m
42840 0.00 0.00 47.50 79.50 1.67 370 1.00 0.0% 5007 164.0 129.2 0.79 1.00 1.67 0.79 3.14
42840 5000 0.12 47.50 75.20 1.58 370 1.00 20.5% 5126 164.0 132.3 0.81 1.01 1.56 0.80 3.26
42840 10000 0.23 47.50 71.1 1.50 370 1.00 38.0% 5229 164.0 135.0 0.82 1.05 1.42 0.78 3.37
42840 15000 0.35 47.50 66.6 1.40 370 1.00 51.0% 5475 164.0 141.3 0.86 1.12 1.25 0.77 3.48
42840 20000 0.47 47.50 61.7 1.30 370 1.00 62.0% 5563 164.0 143.6 0.88 1.22 1.07 0.72 3.58
42840 25000 0.58 47.50 58.2 1.23 370 1.00 71.0% 5728 164.0 147.8 0.90 1.34 0.91 0.67 3.67
42840 30000 0.70 47.50 56 1.18 370 1.00 79.0% 5944 164.0 153.4 0.94 1.49 0.79 0.63 3.75
42840 35000 0.82 47.50 54.3 1.14 370 1.00 85.4% 6220 164.0 160.5 0.98 1.67 0.69 0.59 3.83
42840 40000 0.93 47.50 50.8 1.07 370 1.00 89.1% 6374 164.0 164.5 1.00 1.87 0.57 0.54 3.89
42840 42840 1.00 47.50 47.5 1.00 370 1.00 89.5% 6355 164.0 164.0 1.00 2.00 0.50 0.50 3.93
42840 45000 1.05 47.50 44.6 0.94 370 1.00 88.6% 6331 164.0 163.4 1.00 2.10 0.45 0.47 3.95
42840 50400 1.18 47.50 37 0.78 370 1.00 82.8% 6295 164.0 162.5 0.99 2.38 0.33 0.42 4.01
4.2 Scenario
The test-bed for checking that the Suter Curve follows the pump performance curve in the
positive quadrant is described here. Note that this pump performance curve is normally
available from the pump manufacturer.
4.3 Results
In this section we show that the results from the test-bed follow the actual pump curve that
was used to generate the Suter Curve.
We change the flowrate specification at Node 5 to check the relationship between the flow
rate and pressure lift of the turbo pump. From the table below, we can see that the input
data is roughly equal to the calculated results.
Flow Rate m3/h 0.00 5000 10000 15000 20000 25000 30000 35000 40000 42840 45000 50400
Input Lift m 79.5 75.2 71.1 66.6 61.7 58.2 56.0 54.3 50.8 47.5 44.6 37.0
Calculated Lift m 79.26 75.21 71.06 66.79 62.07 57.97 55.71 54.99 50.26 48 44.65 37.85
Error % -0.3% 0.0% -0.1% 0.3% 0.6% -0.4% -0.5% 1.3% -1.1% 1.1% 0.1% 2.3%
5. Example 2: Deduce an Unknown Suter Curve From the Known

Part
In this section, we show how to extrapolate the built-in Suter Curves to the cases where
the Suter Curve is not known. It is important to emphasise that this is a technique of
extrapolating known data to unknown data, and so should be regarded as an

approximation.
5.1 Known Parameters
The following parameters are known.
= 47.5 m,
= 42840 m3/hr,
= 6.16667 1/s = 370 rpm,
= 164 KN.m,
= 11530 kg.m2.
The known part of Suter Curve is tabulated below.
3.14 3.26 3.37 3.48 3.58 3.67 3.75 3.83 3.89 3.93 3.95 4.01
1.67 1.56 1.42 1.25 1.07 0.91 0.79 0.69 0.57 0.50 0.45 0.33
0.79 0.80 0.78 0.77 0.74 0.67 0.63 0.59 0.54 0.50 0.47 0.42
5.2 The Pseudo-specific Speeds at the Boundaries
The real specific speed can be calculated using Equation 1 (from Section 3), and its value
is shown in the table below. The calculated pump (for which is 71.5) is between a radial
pump ( =25) and a mixed pump ( =147). Therefore, the deduced Suter Curve will be
based on the data for these two built-in Suter curves.
Name Symbol Value Unit

Rated pump speed 370 rpm
Rated pump flowrate 11.90 m3/s
Rated pump piezometric head 47.5 m
Specific speed 70.54
From the known Suter curve, we can interpolate to obtain the pseudo-specific speeds at
the boundaries (i.e., at =3.14 and =4.01), as shown in the table below.
Boundaries 3.14 4.01

Built-in radial pump 1.288 0.35
0.45 0.42
25 25
Built-in mixed pump 1.97 0.348
1.49 0.389
147 147
The calculated pump 1.67 0.33
93.33 70.54
0.79 0.42
64.88 70.54
As an example, the pseudo-specific speed for the head curve ( ) at the boundary of
=3.14 can be calculated as follows.
However, at the boundary of = 4.01, the built-in head curve is too close for interpolation
to be used. In this case, there is a large error, as is indicated below.
This value is ridiculously large. Therefore, we recommend that the actual specific speed
(70.54) be used in this case.
At this point, we have obtained four pseudo-specific speeds at two Suter Curve
boundaries; i.e., the head curve ( ) and torque curve ( ). In the next stage, we shall
deduce a whole Suter Curve for the pump.
5.3 A Whole Suter Curve
Firstly, we assume that the specific speeds at the unknown zone linearly shift between the
two above-mentioned boundaries (i.e., the phase angle is from = 4.01 to = 3.14+2 ).
Here, 2 means that pump curve reaches the next circle. For instance, the pseudo-
specific speed for head curve at the phase angle of 0 can be calculated as follows.
For instance again, the pseudo specific speed for head curve at the phase angle of 1.5
can be calculated by:
After obtaining the pseudo-specific speeds at the full-range phase angle (except the
known zone), we can deduce the unknown part of Suter curve using interpolation. The
process for calculating values at phase angles of 0 and 1.5 are illustrated in the following
table.
Boundaries 0 1.5
25 25
Built-in radial pump
0.634 -0.556
147 147
Built-in mixed pump
-0.69 -1.5
84.16 90.5
The calculated pump
-0.008 -1.063
The calculation for the head curve at =0 is as follows.
6. Program
In this appendix, we have shown how our spreadsheet program can be used for
extrapolating a built-in Suter Curve to obtain a curve for representing the actual pump.
Although the above-mentioned method is not difficult to understand, the actual calculation
is quite involved. Of course, this simple task can easily be carried out using a computer. In
fact, a small program was written in Excel. The full-range Suter curve can be obtained
merely by inputting the boundary conditions; i.e., the phase angles and the pseudo-
specific speeds at the start and end points of the known zone. However, the input pseudo-
specific speeds must lie between 25 and 261, or else an unacceptable error may result
from the interpolation. The table below the final section of this appendix contains the
calculated results for the above example. All of the required input data is coloured in red,
and all known portions of the Suter Curve are marked as “-“.
7. Summary
The steps for obtaining the Suter Curves are as follows.
1. Obtain the necessary information from the manufacturer, including the P-Q
curve, pump speed curve, efficiency curve (or power curve), rated flow rate,
head, speed, and the total moment of inertia of the pump.
2. Calculate the known part of Suter Curve.
3. Deduce the specific speed of the pump.
4. Interpolate the pseudo-specific speed at the ends of the known Suter Curve.
5. Input the calculated data (i.e., the pseudo-specific speed and the corresponding
phase angle) into the Excel program to obtain a Suter Curve for the full range.
APPENDIX C. THE CALCULATION OF HYDRAULIC TRANSIENT

FORCES IN A PIPE NETWORK
1. Introduction
The PIPENET VISION Transient Module can be used to accurately calculate dynamic
forces in a network.
Our own studies have shown that, even in simple systems, the manual approach is risky,
and could potentially cause a serious problem to be underestimated. There are two
aspects to this.
In a case that we studied, we found that, with a simplified manual method, the maximum
force was underestimated by up to 52%. (Please see Table 2 below.)
It is also well known that using the maximum force as a static force, instead of using
dynamic forces, could also cause the extent of the problem to be underestimated.
2. Basic Concepts and Principles
2.1 Control volumes and surfaces
It is not possible to discuss dynamic forces without addressing control volumes. A control
volume is a part of a network where a force is to act. A control volume can be enclosed by
any number of (or a part of) network elements. The control surfaces are the surfaces
through which fluid flows, and are defined by their normal vectors.
An example of a control volume about a straight pipe is shown below (in Figure 1).
Fe
m2, V2 m1, V1
a
P2 P1
L2 L1
+X
L
Figure 1. Control volumes
2.2 Dynamic Forces

In Figure 1, it is assumed that the fluid flows from the right to the left; i.e., in the negative-X
direction. A valve at the downstream end is closed suddenly. Suppose that the resulting
pressure wave propagates through the network at a speed of “a”, say, and that the wave
surface reaches the control volume that is indicated by the broken circle. The force in the
direction of the X Axis that acts on the control volume is given by
(1)
where
is the synthetic force,
is the friction between fluid and pipe wall,
is the static force due to the pressure difference at the both control surfaces,
is the kinetic force due to the change of fluid momentum.
The calculation of and are discussed in the following sections.
2.3 The Static Force

The mechanism of static force in a transient process has been discussed in several
papers. The basic equations include
(2)
and
(3)
Therefore,
(4)
The above equations are applicable to either compressible or incompressible fluids. For
the incompressible liquid, or the less compressible gas (such as that in a high-pressure
steam system, in which the pressure fluctuation is relatively small, so that the gas pressure
and the gas density can be considered as constants), Equation 4 can be simplified as
(5)
2.4 The Kinetic Force

The kinetic force relates only to the inflow momentum and outflow momentum:
(6)
For the same reason as mentioned above, Equation 6 can be simplified for an
incompressible liquid or a less-compressible gas to
(7)
3. Example: A Simple Steam Network
In this example, we consider the following simple stream network.
Figure 2. The schematic of a simple steam network
3.1 Specifications
The following data is used.
Pressure: 260 bar G,

Pipe size: 569 mm (OD) X 419 mm (ID),
Temperature: 569 ˚C,
Wave speed: 654.65 m/s,
Valve type: linear (i.e., valve is linear to the valve setting),
Valve full-closing time: 0.2 s (see Figure 3).
Figure 3. The operation of the valve
3.2 Calculation Method

The wave is created between (a) the time that the valve starts to close and (b) the time at
which the valve is fully closed. Thereafter, the wave propagates throughout the network,
during which time it becomes damped. The length from the front to the back of the wave is
given by
(8)
where “ ” is the wave speed and “ ” is the valve operating time. In some papers, “ ” is
referred to as the critical length. In this example, the critical length is 130.93 m, so Pipes 3
and 4 can contain a whole section of the wave, but Pipes 1 and 2 can only contain a part
of it. Figures 4 and 5 contain profiles of the pressure and flow rate during the time that the
wave moves through the network. It is assumed that the valve and the flow rate are
linear with respect to the valve setting.
Pressure
wave back end
A
P0+P B
a
C
D
P0
wave front end

Lw
O a (t-tc) at Distance
Figure 4. Pressure profile in the network
Flow Rate
D
m
C wave front end

wave back end
a
B
Figure 5. Flow-rate profile in the network
Based on the linear assumption for the valve, it is very clear in the above figures that the
equations used to calculate the maximum forces are different, and depend on the pipe
length. The corresponding equations are summarized below.
Static Force:
A. When the pipe is longer than the critical length (e.g., between Points A and D),
(9)
B. When the pipe is shorter than the critical length (i.e., between Points B and C),
(10)
Kinetic Force:
A. When the pipe is longer than the critical length (e.g., between Points A and D),
(11)
B. When the pipe is shorter than the critical length (e.g., between Points B and C),
(12)
The synthetic forces that act on the pipe are therefore
(13)
It can be seen from Figure 5 that the flow-rate difference between the points A and D (or B
and C) is a constant, irrespective of whether the pipe length is longer or shorter than the
critical length. For the shorter pipe, the expression “ ” attains its minimum value
when the Point B is located at the back end of the wave; i.e., . However, for the
longer pipe, “ ” is always constant between the Points A and D. Therefore,
A. When the pipe is longer than the critical length (i.e., the forces attain a maximum
value when and ),
(14)
B. When the pipe is shorter than the critical length,
(15)
The forces attain their maximum value when and :
(16)
From the above equations, it appears that the maximum forces are not affected by the
value of the valve coefficient, . However, according to the results of the simulation
carried out using the PIPENET VISION Transient Module, the maximum forces do depend
on the value of . The larger values of result in greater dynamic forces. The main
reason is that the above equations are based on a precondition; namely, the flow rate is
linear with respect to the valve setting. However, the resistance of the pipes is not taken
into account. If the valve is large enough, the main pressure drop occurs in the pipes,
and not at the valve. So, changing the valve setting when the valve is either at or near its
fully open position has a lesser effect on the flow rate and the pressure drop in the
network. A greater change occurs when the valve is almost closed. Please refer to Figures
6 and 7 below.
Pressure
wave back end
A
P0+P
B a
C
D
P0
wave front end

Lw
Figure 6. The actual pressure profile in the network

Flow Rate
D
m
C wave front end

m1
a
m1’ wave back end

A B
m2
Figure 7. The actual flow rate profile in the network

At this stage, Equations 5 and 7 are still valid. Therefore, Equation 14 is still applicable to a
pipe that is longer than the critical length. However, Equation 16 is no longer valid for a
pipe that is shorter than the critical length. When , the equation should be replaced
by
(17)
It is obvious that the actual flow-rate value for m1 is much larger than the assumed value,
m1’, which is based on a linear relationship. However, an accurate calculation of the flow
rate is very difficult to obtain. To achieve this goal, it may be necessary to use a quality
program, like the PIPENET VISION Transient Module. With reference to Equations 16 and
17, it is very clear that Equation 16 does not result in a conservative estimate. The use of
Equation 16 may, therefore, result in a much lower value for the dynamic forces than is
actually present. In this case, it may, therefore, be quite dangerous to use Equation 16.
4. Results and Discussion
4.1 Case 1: Small Value of
Table 1 below contains a summary of the calculation results for the simple steam network,
based on either (a) Equations 14 and 16 or (b) the PIPENET VISION Transient Module.
The value of the valve is 225.7 m3/hr.bar. The pressure drop when the valve is fully
open is 99% of the overall pressure loss in the network. Although Equations 14 and 16
could produce a greater accuracy, the difference is less than 10%. The reason may be that
Equations 14 and 16 do not take into account the resistance of the pipes.
Pipe number 1 2 3 4
Pipe size (ID) 419 419 419 419 mm
Pipe section area 0.138 0.138 0.138 0.138 m2
Pipe length 5 20 150 300 m
Pressure 260 260 260 260 bar G
Temperature 569 569 569 569 C
Density 79.208 79.208 79.208 79.208 kg/m3
Wave speed 654.65 654.65 654.65 654.65 m/s
Valve operating time 0.2 0.2 0.2 0.2 s
Critical length 130.93 130.93 130.93 130.93 m
Flow rate 265 265 265 265 tons/hr
73.61 73.61 73.61 73.61 kg/s
If longer than critical length No No Yes Yes
Static force 1840 7361 48190 48190 N
Kinetic force -1 -12 -496 -496 N
Synthetic force (Equations
14 and 16) 1840 7350 47693 47693 N
Synthetic force (PIPENET) 2050 8100 48500 49500 N
Difference -10% -9% -2% -4%
Table 1. The data used in the calculations and the comparison of the forces
Please note that the graphs in this appendix were not obtained using the most recent
version of the PIPENET VISION Graph Viewer.
Figure 8. The force in Pipe 1 as calculated using PIPENET VISION

4.2 Case 2: Normal Valve
Table 2 below contains a summary of the calculation results for the simple steam network,
based on either (a) Equations 14 and 16 or (b) the PIPENET VISION Transient Module.
The value of the valve is 2257 m3/hr.bar, and the pressure drop at the valve (8.15 bars)
is 45% of overall pressure loss in the network (18 bars). The use of Equations 14 and 16
results in a maximum error of 52%, as it can be seen in Table 2. In the actual main steam
system or the preheating system, the pressure drop at the pipes and the valve are half to
half. The length of the single straight pipe is about 5-25m; i.e., the results obtained using
Equations 14 and 16 are only 40% to 50% of the actual value. This error will increase
greatly for pipes shorter than 5m. Note that the difference for Pipe 4 is much larger than
that for Pipe 3. The reason is that the frictional pressure drop in the pipes is not considered
in the equations. The pressure profile at the inlet of Pipe 4 is not flat during the wave pass,
as is illustrated in Figure 17.
Pipe number 1 2 3 4
Pipe size (ID) 419 419 419 419 mm
Pipe section area 0.138 0.138 0.138 0.138 m2
Pipe length 5 20 150 300 m
Pressure 260 260 260 260 bar G
Temperature 569 569 569 569 C
Density 79.208 79.208 79.208 79.208 kg/m3
Wave speed 654.65 654.65 654.65 654.65 m/s
Valve operating time 0.2 0.2 0.2 0.2 s
Critical length 130.93 130.93 130.93 130.93 m
Flow rate 1800 1800 1800 1800 tons/hr
500 500 500 500 kg/s
If longer than critical lengthNo No Yes Yes
Static force 12500 50000 327325 327325 N
Kinetic force -33 -534 -22890 -22890 N
Synthetic force (Equations
14 and 16) 12467 49466 304435 304435 N
Synthetic force (PIPENET) 26000 95000 315000 370000 N
Difference -52% -48% -3% -18%
Table 2. The data used in the calculations and the comparison of the forces
Figure 12. The pressure drop in the valve

Figure 13. The flow rate in Pipe 4


5. Conclusions
From the above calculations, comparisons and discussion, we can draw the following
conclusions.
1. The simplified equations, such as Equation 14 and 16, which are possibly widely
used, are based on the linear relationship between the valve setting and the flow
rate in the network. However, in most cases, the pressure drop in the pipes could
not be neglected. With a larger value of , the calculation error increases
dramatically in the shorter pipe and the valve.
2. The use of the simplified equations may result in the dynamic forces being greatly
underestimated. The calculation results normally attain only 50% (or even 20%) of
the actual value, depending on the pipe length. Therefore, it is imperative that the
PIPENET VISION Transient Module is used to obtain an accurate solution.
PIPENET VISION Transient Module has been in use for more than ten years, and
has been validated in many projects.

Transient - Chapter 5 - Cooling Water System Part 2

Uploaded by

Copyright:

Available Formats

Transient - Chapter 5 - Cooling Water System Part 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Transient - Chapter 5 - Cooling Water System Part 2

Uploaded by

Copyright:

Available Formats

PIPENET VISION

A COOLING WATER SYSTEM INCLUDING A

Revision 2.5, September 2017

1.1 Introduction to Control Systems

1.1.1 Proportional Term

1.1.2 Integral Term

1.1.3 Derivative Term

1.2 Cascade Control

1.3 Concepts of Turbo Pumps

Consider the relationships

is the flow rate,

1.3.1 Operating Zones for the Pump

The eight operating zones are illustrated in the following diagram.

1.3.2 Characterisation of Turbo Pumps by Suter Curves

is the power (kW),

1.4 Turbo Pumps in PIPENET VISION

Appendix 2 of this chapter of the Training Manual contains a description of how to

3.1 Initialisation Data

The units that are to be used are tabulated below.

3.3 Fluid Properties

3.4 Library Data

3.4.1 Pipe Schedule Data

Enter the following data for GRP 1.

3.4.2 Turbo Pump Data

3.4.3 Fittings data

45BE Type K-factor, K-factor=0.11

3.5 Initialisation Data Revisited

4.1 Short Pipes

The data for these short pipes is tabulated below.

Label Input Output Diameter Length Net Height Roughness

As an example, a completed dialog box for Pipe 1 is shown below.

4.2 Elastic Pipes

The elastic pipes of the model are highlighted below.

A completed dialog box for Pipe 36 is shown below.

The numbers of these nodes can be seen on the following diagram.

The specifications for the nodes are as follows.

Node Specification Type Time Function Constant Value

4.3.1 Turbo Pumps

A sample Properties Dialog Box is as follows.

4.3.2 Receiving Vessels

The two receiving vessels are highlighted in the diagram below.

4.3.3 Flow Sensors

The node numbers for the sensors are as follows.

Label Input Node Output Node Information Node

The Properties Dialog Box for Flow Sensor 1 is as follows.

4.3.4 Pressure Sensors

The properties of the first sensor are indicated below.

4.3.5 Transfer Functions

The data for these transfer functions is tabulated below.

It is important to note that a transfer function represents the dynamics of measuring

A sample Properties Dialog Box is shown below.

4.3.6 PID Controller Data

The data for the controllers is as follows.

Below is a sample Properties Dialog Box.

4.3.7 Operating Valves and Control Valves

The values for the DN1500 Valve are tabulated below.

s-value Flow Coefficient

Label Input Node Output Node Information Node Valve Characteristic

5. The Simulation Cases