Embed Size (px)
Citation preview
D. Meulendijks
Preventing unstable operation of a
synthesis compressor using
Dynamic System Simulation
* Corresponding author
Mr. D. Meulendijks, M.Sc.
TNO Science & Industry
T: +31 15 269 2263
F: +31 15 269 2111
Mr. J.P.M. Smeulers, M.Sc.*
TNO Science & Industry
T: +31 15 269 2121
F: +31 15 269 2111
February 13 - 16, 20111st Middle East TurboMachinery Symposium2
Unstable operation of a synthesis gas compressor
• An example of unexpected dynamic response
• Results of an analysis using modeling of the complete
compressor installation
1. Lay-out of the compressor system
2. Description of the problem
3. Measurements
4. Analysis by means of a computer model
5. Conclusions
February 13 - 16, 20111st Middle East TurboMachinery Symposium3
Lay-out of the compressor system
1 2 3
ASC valve
Recyclecompressor
nitrogen
Hydrogen plant
Reactor
FC
PCRPM
control
• 3-stage compressor, with all stages on a single shaft
• Suction pressure speed controlled
• Suction pressure 23 bar; Discharge pressure 133 bar
• Nitrogen flow is controlled to have a constant 3:1 hydrogen-
nitrogen ratio
February 13 - 16, 20111st Middle East TurboMachinery Symposium4
Problem description
• Compressor operated stable for rated capacity
• Compressor started surging far away from the surge point,
limiting the operating envelope
• No operation possible at part load
• At 90% capacity pressure variations occurred that increase until
surge occurs
• Temporary solution was to apply a large surge margin, resulting
in a considerable reduction of part load efficiency
February 13 - 16, 20111st Middle East TurboMachinery Symposium5
Analysis - Measurementssuction
discharge 1st
stage
discharge 2nd
stage
discharge 3rd
stage
0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0- 0 . 8- 0 . 6
- 0 . 4- 0 . 2
0
0 . 2
M e a s u r e d d y n a m i c p r e s s u r e s
0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0
- 0 . 5
0
0 . 5
0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0
- 0 . 5
0
0 . 5
0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0- 1
0
1
T i m e [ s ]
P d
ischarg
e [bar]
surge
2 0 0 2 0 2 2 0 4 2 0 6 2 0 8 2 1 0 2 1 2 2 1 4 2 1 6 2 1 8
- 0 . 1
0
0 . 1
0 . 2
P s
uction [bar]
M e a s u r e d d y n a m i c p r e s s u r e s
2 0 0 2 0 2 2 0 4 2 0 6 2 0 8 2 1 0 2 1 2 2 1 4 2 1 6 2 1 8
- 0 . 4
- 0 . 2
0
0 . 2
0 . 4
P inte
rsta
ge1 [bar]
2 0 0 2 0 2 2 0 4 2 0 6 2 0 8 2 1 0 2 1 2 2 1 4 2 1 6 2 1 8
- 0 . 2
0
0 . 2
P inte
rsta
ge2 [bar]
2 0 0 2 0 2 2 0 4 2 0 6 2 0 8 2 1 0 2 1 2 2 1 4 2 1 6 2 1 8
- 0 . 5
0
0 . 5
T i m e [ s ]
P d
ischarg
e [bar]
suction
discharge 1st
stage
discharge 2nd
stage
discharge 3rd
stage
Pressure variations start to grow autonomously
→ not a classical surge phenomenon, but a system instability
February 13 - 16, 20111st Middle East TurboMachinery Symposium6
Analysis – Measurements interpretation
• Each stage has a phase shift of approximately 60 degrees
• Consumed power varies for each stage
• The sum of the power variations even out to a large extent:
• When stage 1 requires more power the other stages require less
T im e [s ]
∆P [ba r]
1s t s tage2nd s tage3 rd s tage
February 13 - 16, 20111st Middle East TurboMachinery Symposium7
Analysis – Computer model
A model has been built to analyse this using PULSIM, which has building blocks for:
• Fluid machinery: compressor stage, rotor, driver
• Pipe system: pipes, volumes, …
• Controllers
dP - flow 1st stage
10
12
14
16
18
20
22
8 10 12 14 16 18 20 22
Flow [kg/s]
Pressure difference [bar]
surge
point 11000 RPM
10500 RPM
30 m70 m70 m70 m100 m
~ 8 m3 ~6 m3
Rotor
PC
steamturbine
artificial
excitation
30 m70 m70 m70 m100 m
~ 8 m3 ~6 m3
Rotor
PC
steamturbine
artificial
excitation
CQBQAP ++=∆ .. 2
Approximated dP - flow curve
A negative, B~ RPM and C~ RPM2
February 13 - 16, 20111st Middle East TurboMachinery Symposium8
Analysis - Response to artificial excitation
Response curve flow pulsation @ suction
0
0,2
0,4
0,6
0,8
1
1,2
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Frequency [Hz]
Flow [%pp]
100%load
95% load
90% load
85%load
unstable range
Conclusion: system is marginally stable for 85% load
February 13 - 16, 20111st Middle East TurboMachinery Symposium9
Analysis - Causes of unstable behavior
1. Reduced compressor impedance at lower flows
dP - flow 1st stage
y = -0.101x2 + 2.3302x + 7.1155y = -0.101x
2 + 2.2256x + 6.4908
15
16
17
18
19
20
21
22
8 9 10 11 12 13 14 15 16 17 18
Flow [kg/s]
Pressure difference [bar]
surge
pointoverall
surge
point
10500 RPM
11000 RPM
• Idealised curve
∆P=A.Q2+B.Q+C
• Compressor impedance
Rc= -d ∆P/dQ= -2.A.Q-B
• Surge for Rc<0 or
Q<-B/(2.A)
• @90% load the
compressor impedance is a
factor of 2 smaller than
@100% load
February 13 - 16, 20111st Middle East TurboMachinery Symposium10
Analysis - Causes of unstable behavior
1. Reduced compressor impedance at lower flows
2. Exchange of power between stages via rotor
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
Time [s]
Pow
er [M
W]
Power 1, 2, 3 and total
1st stage
total
2nd stage3rd stage
February 13 - 16, 20111st Middle East TurboMachinery Symposium11
Analysis - Causes of unstable behavior
1. Reduced compressor impedance at lower flows
2. Exchange of power between stages via rotor
3. Poor controllability of suction pressure control: small changes in
suction pressure lead to large variations in flow
4. Variation of molecular weight caused by pulsations
Flow variations at mixing point cause change of composition that
travels with the flow velocity to the compressor. Travel time 7
seconds or ca. 3.5 period of instability cycle.
70 m100 m
hydrogen
nitrogen
February 13 - 16, 20111st Middle East TurboMachinery Symposium12
Solution - Stabilising the system
1. Improve characteristic of
first stage (steeper)
2. Flow control instead of suction pressure control (cascade with suction pressure control)
3. Reduce variation of molecular weight (mixing vessel or hydrogen flow control)
4. Dampen the resonance: orifice plate or control valve in hydrogen line
Difficult to achieve in an existing system
The orifice was actually implemented,
and the system operated stable0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Frequency [Hz]
Flow amplitude
Original
With orifice
February 13 - 16, 20111st Middle East TurboMachinery Symposium13
Concluding remarks
1. Unexpected instabilities can occur in a compressor installation as a result of
pipe system and compressor dynamics
2. Low pressure loss systems with multi-stage compressors are susceptible for
instabilities
3. Compressor head-flow curves are not as smooth as presented here. Especially
‘multiple wheel stages’ show deviations from the idealised quadratic curves.
Differences exist between ‘design’, ‘test’ and ‘real’ curves.
4. Analysis of the compressor control during the design stage is recommendable.
Powerful simulation tools are available to analyse the dynamics and control.
However, this is no replacement for understanding what is going on in the
system.
5. Surge cycles itself can also be simulated nowadays, i.e. the unstable behavior.
