Transcript NEPTUNE Power System: Transient Analysis and Circuit Studies

North East Pacific Time-series
Underwater Networked
Experiment (NEPTUNE):
Power System Design, Modeling
and Analysis
Aditya Upadhye
Outline
NEPTUNE
Power system requirements
Two design alternatives
o
o
o


o
o
o
o
Version 1
Version 2
Cable analysis
Models
Simulation results
Conclusions and future work
NEPTUNE
Explorer Plate
Juan de Fuca Ridge
Pacific Plate
Junction Box
Juan de Fuca
Plate
Cable
study area
Gorda Plate
Nedonna Beach
North
American
Plate
Science requirements






Communication bandwidth - Gb/s
Power – 200kW
Reliability
Robustness of design
Thirty year lifetime
Maintenance and support
Power System Design

Basic tradeoffs







Frequency: ac versus dc
Network: radial versus interconnected
Loads: series versus parallel
Shore station supply at 10kV, 200kW
Max. current-carrying capacity = 10A
User voltage = 400V / 48V
Max. power at each node = 10kW
Power System Design

Protection



Monitoring and control



Sectionalizing circuit breaker
Breaker control
Current – voltage measurements
State estimation
Shore station control hardware /
software
Power System Design:
Version 1
Version 1 Circuit
Backbone cable
10kV
Spur
cable
Circuit
Breaker
Back-biased
diodes
Node dc-dc converter
and control
electronics
NODE
400V
Science Loads
DC Circuit Breaker
Need


During initial energization
For fault isolation
Required features
To force a current zero and minimize arcing
To prevent breaker restrikes

DC Circuit Breaker
Open Circuit
S2
S3
R1
S1
R2
S4
C
DC Circuit Breaker
Soft Closing
S2
S3
S4
R1
R2
S1
C
DC Circuit Breaker
Closed circuit
S2
S1
S3
S4
R1
R2
C
DC Circuit Breaker
Capacitor charging
S3
S2
S4
R1
R2
S1
C
DC Circuit Breaker
Capacitor discharging
S2
S3
S4
S1
R1
R2
C
DC Circuit Breaker
Hardware prototype





125V, 5A breaker circuit
Breaker control
MOSFETs drive the switch solenoids
Opto-isolator between logic circuit and driver
circuit
Control logic has a counter, which
continuously cycles through the breaker
operations
DC Circuit Breaker
Hardware prototype test results







Continuous Voltage: 125V
Continuous Current: 4.5A
Total Breaker Cycles: 125,000
Normal cycle switching frequency: 20Hz
Maximum cycle switching frequency: 100Hz
Maximum tested voltage: 200V
Maximum tested current: 5A
Power System Design:
Version 2
Version 2 Circuit
10kV
Circuit
Breaker
Back-biased
diodes
BRANCHING
UNIT
Node dc-dc converter
and control
electronics
400V
Science Loads
NODE
Branching Unit
Solenoid of S1
Solenoid of S2
Solenoid of S3
S1
S5
S6
L1
L2
I1
Z1
Z3
S2
Z2
12V
12V
BU
Controller
1
BU
Controller
2
I4
I3
I1 I2 I3 I4
I1 I2 I3 I4
S3
Science Load
Dummy Load
I2
Series Power Supply
•Indigenous power supply for each BU
•Less reliance on node converter
•Use of zener diodes in reverse region
•Back-to-back zener diodes
Solenoid # 1
0.5A
Solenoid # 2
0.5A
2-5A
12V
12V
0V
2-4A
Modes of Operation




Normal
Fault
Fault-locating
Restoration
Special case

System startup
Normal mode
1. Current measurement at
shore
2. Communication from nodes
experiencing voltage collapse
3. PMACS determining nodes
that drop-out.
Fault?
No
Yes
System
shutdown
System startup
PMACS determining
presence of fault and its
location
Fault locating
mode
State
Estimation
Is fault
located?
No
Yes
System
shutdown
Restoration
mode
Raise
voltage to
10kV
Comparison of
Version 1 and Version 2
Version 1
Version 2
Conventional approach to
power system design
Based on the philosophy that
cable faults are rare but possible
Response to a fault is at the
local level by the nearest circuit
breaker
Response to a fault is at the
system level by the shore station
controls
Circuit breaker is complicated
with many components
Complexity of circuit breaker is
greatly reduced
Fault current is interrupted;
arcing and restrikes are
possible
Fault current is not
interrupted; arcing and
restrikes are not possible
Single node failure can cause
failure in a large section of the
network
Single node failure is not
catastrophic for the system as
that node only will be out of
service
Reliability is low
Reliability is increased
Electromagnetic Transients
Program (EMTP)
Alternate Transients Program
ATP Theory
o
o
o
ATP is a universal program system for digital
simulation of transient phenomena of
electromagnetic as well as electromechanical
nature
With this digital program, complex networks
and control systems of arbitrary structure
can be simulated
Trapezoidal rule of integration
Cable Parameters
ALCATEL OALC4 Cable
Insulating sheath Ø 17
mm
Steel wires
strand
Composite
conductor
Opticafibers
l
Thixotropi Jelly
c
Steel
Ø: 2.3 mm
tube
Inductance Calculations



The generalized formulae were applied to the OALC4
cable
The core (steel) current caused flux linkages within
a) the core
b) the sheath
c) the insulation
The sheath (copper) current caused magnetic flux
linkages within:
a) the sheath
b) the insulation
Inductance Calculations
T  i  e
Where T is the total flux linkage associated
with the conductor, i is the flux linkage
internal to the conductor, and e is the flux
linkage external to the conductor
L
T
icable
Where icable is the total current in the cable
Results
Theoretical ATP values ALCATEL
values
values
R (/km)
1.03
1.03
1.00
L (mH/km)
0.3947
0.3948
0.4
C (F/km)
0.179
0.179
0.2
Simulation Models
Version 1: Opening of Circuit Breaker
t = (topen-t)
Switch closed
t = topen
Switch open:
initial arcing
t =( topen +t)
Capacitor
charging
Simulation of Restrikes
RESTRIKE!!!
Vmax
topen
Initial Arcing
Period
Restrikes: Simulation Circuit
NODE A
S2
S3
S1
250 km
Cable
10kV
DC
S4
Rd
250 km
Cable
NODE B
Rs
C
Sf
Load Z2
Load Z1
Capacitor Current
2
Restrike
1
3
No
Restrike
2
1
Capacitor Voltage
2
Restrike
1
No
Restrike
2
1
3
Simulation Results
Maximum voltage across
switch
Travel time of
switch
15 kV
5 ms
Minimum value of
capacitor to prevent
restrikes (F)
2
15 kV
10ms
5
15 kV
18 ms
10
25 kV
15 ms
1
25 kV
18 ms
1
25 kV
20 ms
1
Current Limiting Operation



The shore station power supplies are
rated at 200kW, 10kV
The steady-state system current = 10A
Under certain conditions, the system
current may increase due to



Cable faults
Topology changes
Load fluctuations
Current Limiting Operation



The system current is limited to a value below
10A using the control circuitry in the shore
station
This is done by dropping the shore voltage
which in turn reduces the current
The control action is initiated only for steadystate overcurrents and not transient
overcurrents.
Fault Analysis
Version1: Simulation Circuit
100
km
NODE A
100
km
NODE B
50
km
50
km
Shore
current
limiting
LOADZ1
DC
LOADZ2
Sf
LOADZ3
Results of Current Limiting: Shore
Output voltage and Current
1
Voltage
3
2
2
Current
3
1
Voltage and Current at Node 2:
No Current Limiting
3
Voltage
1
2
1
Current
2
3
Capacitor Current of Node 2
1
2
3
Version 2: Fault Studies



A pre-insertion resistance may be placed at
the shore station to limit the fault current
This resistance will limit the fault current
before the shore controls take the
appropriate mode-dependant control action
Three controllable parameters in
simulations:
A.
B.
C.
Value of pre-insertion resistance
Response time of control circuitry
Distance of fault from the shore station
Simulation Circuit
'X'
km
BU A
Preinsersion
resistance
100
km
BU B
900
km
Sf
LOAD Z
LOAD Z
DC
DC
V1
X=100km/1200km
V2
Results: Vary Response Time
Fault
distance=100km
Fault
distance=1200k
m
600
I^2t values
500
400
300
200
100
0
0
10
20
30
40
50
Response time of shore (ms)
60
Results: Vary Fault Distance
Peak Current Transient (A)
250
200
150
100
50
0
0
200
400
600
800
1000
1200
1400
Distance of fault from shore (km)
600
I^2t values.
500
400
300
200
100
0
0
200
400
600
800
1000
Fault distance from shore.
1200
1400
Conclusions






A sub sea observatory NEPTUNE is the first of its kind and will
open up new and exciting areas of scientific research
The NEPTUNE power system implements a ‘dc network’
Version 1 dc breaker is designed and a hardware prototype was
built in lab
Version 2, the preferred design choice is philosophically different
from conventional terrestrial power systems
Transient studies of the system is performed using EMTP for
worst-case scenarios from the point of view of component
design and fault analysis
Theoretical analysis of the cable was performed and EMTP
models were developed for the above
Future Work



DC breaker prototype for Version 2
Control and monitoring systems for the above using
microcontroller and/or array logic
A comprehensive transient model for the entire
NEPTUNE network which is generic enough to
simulate any fault type and any operating scenario