advanced inverter functions for distributed solar … inverter functions for distributed solar...
TRANSCRIPT
2/23/2016
Pgh PELS: Advanced Inverter Functions 1
Advanced Inverter Functions for Distributed Solar Integration
Tom McDermott, [email protected]
Pittsburgh IEEE PELS Chapter
February 23, 2016
1. DOE Sunshot and state‐wide incentives for distributed solar.2. Managing voltage fluctuations with smart inverters3. The need for PV to provide ride‐through and grid support4. Evolution of IEEE Standard 15475. On‐going research into PV inverter functions and models
Sponsors:
State Net Metering Policies put the focus on distributed resources rather than transmission‐connected plants.
2/23/2016 2Pgh PELS: Advanced Inverter Functions
2/23/2016
Pgh PELS: Advanced Inverter Functions 2
Net metering tiers can distort the interconnection; a 3 MW generator becomes three 1‐MW generators.
2/23/2016 3
Source: Neil LaBrake, Jr., National GridPgh PELS: Advanced Inverter Functions
2/23/2016 4Pgh PELS: Advanced Inverter Functions
2/23/2016
Pgh PELS: Advanced Inverter Functions 3
PV Angle Effects may enhance (or degrade) the economics of a PV interconnection.
• Best elevation depends on latitude and season
• Best azimuth depends on time of day
2/23/2016 5Pgh PELS: Advanced Inverter Functions
The inverter optimally matches the photovoltaic DC output voltage to the AC load via Maximum Power Point Tracking.
2/23/2016 Pgh PELS: Advanced Inverter Functions 6
0 100 200 300 400 500 6000
4
8
12
16
20
24
100%80%60%40%20%
Vpv [Volts]
Ipv
[Am
ps]
2/23/2016
Pgh PELS: Advanced Inverter Functions 4
There are three main concerns with integrating VG on distribution systems.
1. Voltage Fluctuationsa) Inverters can operate at fixed power factor
2. Fault Currents and Overvoltagesa) Inverters contribute only 1.1 – 2.0 times rated
b) Inverters shut down quickly upon open‐circuit or short‐circuit conditions
3. Unintended Islandinga) Utility‐interactive inverters have built‐in anti‐
islanding detection schemes to operate in 2 s
72/23/2016 Pgh PELS: Advanced Inverter Functions
Sandia, NREL and EPRI have promoted updates to the 15% screening criteria for no‐study PV interconnects.
2/23/2016 Pgh PELS: Advanced Inverter Functions 8
DG > 2/3 of Min Load?
Q Match within 1%?
Synch Gen > 25% of DG?
1 Product > 2/3 of DG?
Yes
Yes
Anti‐Islanding Screen
Detailed Study
YesYes
No Study
No
No
No
No
Coddington et. al.: Updating Interconnection Screens for PV System Integration
2/23/2016
Pgh PELS: Advanced Inverter Functions 5
Optimal (fixed) power factor dispatch alleviates voltage fluctuations, until vendors implement IEEE 1547a
2/23/2016 9
1 12
2 2
100( )( )
(100 Re ) (Im ) 100
drop n nn
drop dropn
V R jX P jQU
dVV V
U
Estimated % Voltage Change:
(R, X in ohms, Sn in MVA, Un in kV)
Pgh PELS: Advanced Inverter Functions
Basic Distribution Software Tasks have been limited to load flow and short circuit analyses.
• Current Flow and Voltage Drop
2/23/2016 10
Conductor Overload? Voltage Range?
• Overcurrent Protective Device CoordinationRS
Detected?R Trip First?
Fuse Saving?
Pgh PELS: Advanced Inverter Functions
2/23/2016
Pgh PELS: Advanced Inverter Functions 6
Quasi‐static Time Series (QSTS) Simulation addresses load and generation variability.
• Time‐stepping through a variable power profile
• Tap changers are active, with time delays
• Capacitor controls with time delays
• Controls remember state at next time
• But it’s not a dynamic simulation:• No inertia
• No numerical integration
• Simple RMS load flow solution at each time step
0
100
200
300
400
500
600
700
0 300 600 900 1200 1500 1800 2100 2400
Power Output [kW]
Time [s]
0
10
20
30
40
50
60
70
80
90
100
7 9 11 13 15 17 19
Percent Output
Hour of Day
0 180 360 540
On/
Off
Sta
tus
Time [s]
Combustion Turbine Profiles
OffOnOff
OnOffOn
Daily PV (50 kW Unit)Conventional Generator40‐minute Wind
2/23/2016 Pgh PELS: Advanced Inverter Functions 11
0:00 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 0:00
[Hour of Day]
8‐Aug
Peak
Avg
Daily Load Variations)
Intelligent Volt‐Var function attempts to hold a constant voltage using percent available VARs.
2/23/2016 Pgh PELS: Advanced Inverter Functions 12
Source: “Common Functions for Smart Inverters, v3”, EPRI 3002002233, February 2014
2/23/2016
Pgh PELS: Advanced Inverter Functions 7
“Percent available” output means the physical slope [VARs/volt] varies.
2/23/2016 Pgh PELS: Advanced Inverter Functions 13
What was the intent? “’Available VARs’” implies whatever the DER is capable of providing at the moment, without compromising Watt output. In other words, Watt output takes precedence over VARs in the context of this function”.Source: “Common Functions for Smart Inverters, v3”, EPRI 3002002233, February 2014, p. 9‐3.
Dynamic Reactive Current function provides an event‐based response to changes in voltage.
2/23/2016 Pgh PELS: Advanced Inverter Functions 14
Source: “Common Functions for Smart Inverters, v3”, EPRI 3002002233, February 2014
Tested with a three‐phase high‐impedance fault in EPRI 3002002271
2/23/2016
Pgh PELS: Advanced Inverter Functions 8
Dynamic Reactive Current function provides reactive power in percent of rating.
2/23/2016 Pgh PELS: Advanced Inverter Functions 15
Source: “Common Functions for Smart Inverters, v3”, EPRI 3002002233, February 2014
Dynamic Reactive Current function uses a moving average to define the voltage set‐point.
2/23/2016 Pgh PELS: Advanced Inverter Functions 16
Source: “Common Functions for Smart Inverters, v3”, EPRI 3002002233, February 2014
2/23/2016
Pgh PELS: Advanced Inverter Functions 9
A voltage‐control test circuit includes both grid voltage and solar output variability.
2/23/2016 Pgh PELS: Advanced Inverter Functions 17
Unity Power Factor
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Intelligent Volt‐VAR mitigates fluctuation, but constrains operation (e.g. voltage reduction requires communication).
2/23/2016 Pgh PELS: Advanced Inverter Functions 18
Unity Power Factor
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Vreg=1.01, Slope=50
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
2/23/2016
Pgh PELS: Advanced Inverter Functions 10
The burden of choosing settings; detailed studies needed; the wrong choice makes things worse.
2/23/2016 Pgh PELS: Advanced Inverter Functions 19
0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1-1
-0.5
0
0.5
1
Voltage (p.u.)
vars
(p.
u. a
vaila
ble) Voltvar Response
0.94 0.96 0.98 1 1.02 1.04 1.06 1.08 1.1-1
-0.5
0
0.5
1
Voltage (p.u.)
vars
(p.
u. a
vaila
ble) Voltvar Response
Source: Abate, McDermott, Rylander and Smith, “Smart Inverter Settings for ImprovingDistribution Feeder Performance”, IEEE PES General Meeting, Denver, CO, July 2015.
Adaptive Voltage Regulation provides the best of both intelligent volt‐VAR and dynamic reactive current.
2/23/2016 Pgh PELS: Advanced Inverter Functions 20
2 2-rated outS P
2 2-rated outS P
1 tregV V e
2/23/2016
Pgh PELS: Advanced Inverter Functions 11
Adaptive Voltage Regulation mitigates fast voltage fluctuation while tracking slower grid voltage trends.
2/23/2016 Pgh PELS: Advanced Inverter Functions 21
Vreg=1.01, Slope=50
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Adaptive Vreg, Slope=50
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
Vreg
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Dynamic Reactive Current performance may be close to Adaptive Voltage Regulation; Slope = 50.
2/23/2016 Pgh PELS: Advanced Inverter Functions 22
Dynamic Reactive
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Adaptive Vreg
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
2/23/2016
Pgh PELS: Advanced Inverter Functions 12
Dynamic Reactive Current and Adaptive Vreg tested with high‐impedance fault on clear day.
2/23/2016 Pgh PELS: Advanced Inverter Functions 23
Unity Power Factor
4 8 12 16 200.9
0.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-150
-100
-50
0
50
100
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Dynamic Reactive
4 8 12 16 200.9
0.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-150
-100
-50
0
50
100
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Adaptive Vreg
4 8 12 16 200.9
0.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
4 8 12 16 20-200
-150
-100
-50
0
50
100
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Voltage regulation set‐point and reactive power can be limited, if desired.
2/23/2016 Pgh PELS: Advanced Inverter Functions 24
Q limited to 20 kVAR
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
Vreg
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
1.0 < Vreg < 1.03
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
Vreg
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
2/23/2016
Pgh PELS: Advanced Inverter Functions 13
Instead of attempting to regulate the grid voltage, dispatch reactive power (i.e. the bias point is not zero).
2/23/2016 Pgh PELS: Advanced Inverter Functions 25
Absorb 40 kVAR
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
Vreg
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Supply 40 kVAR
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
Vreg
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Shorter Vreg time constants don’t necessarily work better; the optimal value seems to be 1200 seconds.
2/23/2016 Pgh PELS: Advanced Inverter Functions 26
Vreg Tau = 300 s
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
Vreg
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
Vreg Tau = 1200 s
4 8 12 16 200.95
1
1.05
Time [hr]
Vol
ts [
pu]
Vsys
Vreg
4 8 12 16 20-200
-100
0
100
200
Time [hr]
[kW
] or
[kV
AR
]
P
Q
2/23/2016
Pgh PELS: Advanced Inverter Functions 14
Results from an actual feeder with 1.7 MW of PV show that Adaptive Voltage Regulation performs well.
2/23/2016 Pgh PELS: Advanced Inverter Functions 27
Volt‐Watt function can limit steady‐state voltage rise, especially with low X/R ratios (Hawaii uses this).
2/23/2016 Pgh PELS: Advanced Inverter Functions 28
Simplified Formula: V R P X Q But do we really want to control V this way, even if P has more leverage than Q?
2/23/2016
Pgh PELS: Advanced Inverter Functions 15
A Frequency‐Watt function can help dampen system frequency swings.
2/23/2016 Pgh PELS: Advanced Inverter Functions 29
(60 )
60pre
f dbP P
k
• P is the real power output [pu]• Ppre is the pre‐disturbance real power output [pu]• f is the disturbed system frequency [Hz]• db is a single‐sided deadband (default to 0.1 Hz)• k is the per‐unit frequency change corresponding
to 1 per‐unit power output change (defaults to 0.05)
Conclusion – Adaptive Voltage Regulation should be the default behavior of PV inverters.
• The default settings work well (so far…):– Slope = 30– Tau = 1200 s– 0.95 ≤ Vreg ≤ 1.05 [pu]
• Without smart‐grid communications– No need for detailed coordination studies– No wake‐up time for inverter’s voltage response
• With smart‐grid communications– Fail‐safe behavior– Dispatch reactive power, just like shunt capacitors
• Default high‐frequency roll‐off as well
2/23/2016 Pgh PELS: Advanced Inverter Functions 30
2/23/2016
Pgh PELS: Advanced Inverter Functions 16
Bulk system planners have been concerned with losing large amounts of DG after a fault.
2/23/2016 Pgh PELS: Advanced Inverter Functions 31
Source: Mahendra Patel, PJM(shaded region has V < 50%)
Source: Nick Miller, GE
Loss‐of‐DG events have caused problems in the European bulk power system.
2/23/2016 Pgh PELS: Advanced Inverter Functions 32
NERC IVGTF Task 1‐7 Report
2/23/2016
Pgh PELS: Advanced Inverter Functions 17
IEEE 1547a‐2014 allows voltage regulation and requires more adjustability in voltage and frequency trip settings.
2/23/2016 Pgh PELS: Advanced Inverter Functions 33
After more than 10 years, IEEE 1547 is in a full revision process to more fully address the issues.
2/23/2016 Pgh PELS: Advanced Inverter Functions 34
75% Ballot Return and 75% Yes
Before 2012: a 5‐year Life Cycle (if “Reaffirmed”)
Since 2012: a 10‐year Life Cycle
2/23/2016
Pgh PELS: Advanced Inverter Functions 18
An early draft of P1547 clause 4.2 opens the door to technology‐dependent ride‐through requirements.
2/23/2016 Pgh PELS: Advanced Inverter Functions 35
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
110%
120%
130%
140%
150%
0.01 0.1 1 10 100
Voltage (%
of base)
Cumulative Time (sec)
B
A
C
D
EMandatory Overvoltage TripClearing Time (default values)
Mandatory Undervoltage TripClearing Time (Class I default values)
Class II Low‐Voltage Ride‐through and Clearing Times (triangle markers)
One of “shall do” standards (1547.1‐2005) supports UL 1741, which defines how single inverters are tested.
2/23/2016 Pgh PELS: Advanced Inverter Functions 36
2/23/2016
Pgh PELS: Advanced Inverter Functions 19
How to strengthen the provisions for intentional islanding (aka micro‐gridding) with new functions?
2/23/2016 Pgh PELS: Advanced Inverter Functions 37IEEE 1547.4‐2011
What islanding detection methods will be reliable in multi‐inverter and high‐penetration scenarios?
2/23/2016 Pgh PELS: Advanced Inverter Functions 38
• Xu et. al.: A Power Line Signaling Based Technique for Anti‐Islanding Protection of Distributed Generators—Part I: Scheme and Analysis
• Wang et. al.: A Power Line Signaling Based Scheme for Anti‐Islanding Protection of Distributed Generators—Part II: Field Test Results
2/23/2016
Pgh PELS: Advanced Inverter Functions 20
Pitt and Duquesne Light are monitoring PV output variability at multiple sites, on a 1‐second time scale.
2/23/2016 Pgh PELS: Advanced Inverter Functions 39
June 7-8, 2014
00:00 06:00 12:00 18:00 00:00 06:00 12:00285
290
295
Time of Day
Vo
lts R
MS
00:00 06:00 12:00 18:00 00:00 06:00 12:0050
100
150
200
250
Time of Day
Am
ps R
MS
00:00 06:00 12:00 18:00 00:00 06:00 12:00-60
-40
-20
0
20
Time of Day
kW @
60
Hz
00:00 06:00 12:00 18:00 00:00 06:00 12:00-25
-20
-15
-10
Time of Day
kVA
R @
60
Hz
2/23/2016 40
User Local Machine
‐ GIS extract files‐ Model adjustments‐ Simulation settings
‐ Simulation results‐ Visualizations‐ Model export
Web Server
Feeder Simulations(OpenDSS)
Browser(HTML5,Javascript)
Model Creation(PERL)
Server(Apache or IIS, PHP)
GIS Feeder Extract(SQL)
GIS Data Server
Pgh PELS: Advanced Inverter Functions
OpenDSS running on a Web Browser enables engineers at FirstEnergy to simulate voltage fluctuations.
2/23/2016
Pgh PELS: Advanced Inverter Functions 21
The user can extract and analyze OpenDSS feeder models from the Geographic Information System.
2/23/2016 Pgh PELS: Advanced Inverter Functions 41
Symmetrical components can work with PV inverter sources, but results may be pessimistic?
2/23/2016 Pgh PELS: Advanced Inverter Functions 42
12 0LZ
12 0LZ 0SZ
02 YZ 02 YZ
1 0TZ
12 1LZ
12 1LZ 1SZ
12 YZ 12 YZ
1 1TZ
12 1LZ
12 1LZ 1SZ 1 1TZ
1.05puSV
0 1 2F F FI I I
XI
YI
12 YZ 12 YZ
0THZ
1THZ
DGV
2THZ
1.1puDGI
NORTZ
2 0TZ
3 NZ
2 1TZ
2 1TZ
1mB
2mB
0mB
2/23/2016
Pgh PELS: Advanced Inverter Functions 22
Detailed PV transient models present issues with software licenses, proprietary data & learning curves.
2/23/2016 Pgh PELS: Advanced Inverter Functions 43
SLGF TOV, 30 KVA Grounding, 0.16s Delay
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 10
2
4
Vpc
c [p
u]
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 10
5
10
Ipcc
[pu
]
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 10
50
100
150
Ifdr
[pu
]
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 158
60
62
64
Fre
quen
cy [Hz]
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1-1
0
1
2
Time [s]
Spc
c [p
u]
P
Q
Simplified forOpenDSS
Inverter open‐circuit and short‐circuit behavior testing to identify models for application studies.
2/23/2016 Pgh PELS: Advanced Inverter Functions 44
DC
Out
pu
t DC
Inpu
t
AC
Ou
tput
1ph
s A
C1p
hs
AC
L3
L2L1
3p
hs
AC
1ph
s A
C1p
hs
AC
GN
DN
2/23/2016
Pgh PELS: Advanced Inverter Functions 23
Creating a better PV inverter model for OpenDSS, using system identification.
2/23/2016 Pgh PELS: Advanced Inverter Functions 45
Step Response of Linear Block
Piecewise Linear Function
representing output
nonlinearity
Piecewise Linear function representing
input nonlinearity
Piecewise Linear function representing
input nonlinearity
Hammerstein‐Weiner Modeling Framework
45
What can you do? The next P1547 meeting will be in Juno Beach, FL, March 8‐9. We need consistent participation!
2/23/2016 Pgh PELS: Advanced Inverter Functions 46
http://grouper.ieee.org/groups/scc21/