1. Recommendations for Isolation Monitor Ground Fault Detectors on Residential and Utility-Scale PV Systems
Jack Flicker
Jay Johnson
Mark Albers
Greg Ball

2. Ground Fault Detection Options
Solar ABCs steering committee investigated
ground faults and the detection blind spot
Concluded that fuse-based GFDI designs vulnerable to
faults in the grounded current-carrying conductor
GFDI fuse impedance decreases detection window
for all faults
Number of alternatives identified
Current Sense Monitor (CSM)
Residual Current Detection (RCD)
Isolation Resistance Monitor (Riso)
Also called “Morning Check”

3. Riso Research Goals and Findings
Current Riso standards vary widely from extremely conservative (UL 1741 CRD 2010) to very liberal (UL 1741 CRD 2012)
Standards take into account array/inverter V, but not module/inverter isolation
Project Goal Find Riso thresholds to:
maximize the number of ground faults that are detected
minimize unwanted tripping
Balance unwanted tripping with safety
Appropriate trip point can be picked for given maximum fault power
Need knowledge of the array
Finding There is no a priori one-size fits all solution to completely fulfill both safety and unwanted tripping concerns
It may be necessary to go to a range of values
 Part of commissioning process?

4. Isolation Resistance (Riso)V I
Inject voltage pulse (~50 V)
using an external power source
isolation calculated from current draw + -
Rfault
Rmod
Rinv
Vpulse
If Riso < threshold (kW)
isolation monitor trips x x
Riso measurements on ungrounded systems
grounded systems which temporarily unground
continuous monitoring possible by superimposed
Vpulse on nominal DC bus
Often before the inverter begins to export power
“morning check”

5. Isolation Resistance (Riso)V I
Rmod modeled as S*M resistor array
of single module leakage (Rleak)
S = # strings
M = modules/string
Parasitic pathway to ground via inverter or other BOS components
~1 MW for residential systems
100’s of kW for utility systems
some measured as low as 20 kW (!!)
Riso modeled as equivalent resistance of parallel resistors + -
Rfault
Rmod
Vpulse x x

6. Riso Model Riso modeled as S*M resistor array S = # strings
M = modules/string
Slight deviations due to voltage drops across parasitic resistors, bypass diodes, and/or module photodiodes
Validated by measurements on 750 and 3 kW systems (more information in proceedings)
Riso dominated by REGC
Riso dominated by Rfault
Riso dominated by Leakage
(module and inverter) 6
J. D. Flicker, and J. Johnson. “Electrical Simulations of Series and Parallel PV Arc-Faults” Proceedings Photovoltaics Specialists Conference (PVSC), pp , 2013.

7. Riso Models Ungrounded Arrays Grounded Arrays Worst-case
In general, the fault current is not analytically solvable
Luckily for us, the worst case scenario (fault at a current carrying conductor) has analytical form
Ungrounded Arrays
Adrian Haering
Grounded Arrays I
Rfault
Rinv
Ifault
I-Ifault
Relatively low current for high Riso values because of “clamping” effect
Ungrounded array ceases to float and becomes referenced to ground
turns into grounded array
Inverter Acts as Current Divider
Worst-case
Rinv << Rfault

8. Riso Models Fault Current Ungrounded Arrays Grounded Arrays
Fault Power
Grounded arrays provide significantly more fault current
Dump much more power into fault
 Significantly more dangerous
Will consider grounded array with fault on the current carrying conductor to be the worst case

9. Riso Trip Point Recommendations
We have developed equations that predict fault current as a function of Rfault
How does this translate to Riso trip point?
Start with allowable energy
#1 Divide by Voc to get allowable Ifault trip times
#2 Use Ifault equation to get allowable Rfault
Need system Voc
~14.5 kW for 70 W Pmax
Voc=1000 V
Imp=2089 A
Rmp=0.34 W

10. Riso Trip Point Recommendations
Use Riso equation to get trip point
Need system information
# strings, # modules in string, Module isolation, Inverter isolation
Riso = 5.66 kW
S,M are well known according to system design (S=100, M=10)
Conservative Rleak value would be minimum IEC value (40 MWm2)
Inverter isolation can vary widely, conservative value 20 kW

11. Riso Trip Point Recommendations
Use Riso equation to get trip point
But what if we don’t have system information…such as an inverter manufacturer?
IEC Minimum Isolation
= 40 MWm2
Measured or
Assume worst-case
15 kW ?? 2
Assume worst-case
5% ??

12. Riso Trip Point Recommendations
Allows us derive an Riso trip threshold vs. system size for maximum allowable fault power dissipation (Pfault)
Riso has a dependence on system size (Pinv)  find Riso trip points vs. system size
Assume worst case= leakiest modules/inverter?
Rleak = 40 MWm2, Rinv = 15kW, h = 5%, Voc = 1000 V
What if someone puts on better modules?
Rleak = 20 MWm2, Rinv = 15kW
h = 20%, Voc = 1000 V
More overhead  less unwanted tripping
But Riso trip point  unsafe condition (Pfault > 70 W)
What if you assume worst case
 least leaky modules/inverter?
Riso ≈ Rfault
Safe if installer puts on leaky modules/inverter, but unwanted tripping issues

13. Riso Trip Point Recommendations
Current Riso standards vary widely from extremely conservative (UL 1741 CRD 2010) at 100 kW to very liberal (UL 1741 CRD 2012) at 700 W for a 30 kW system
Take into account array/inverter V, but not module/inverter isolation
h = 5%
Inverter isolation = 15 kW
PDC/PAC = 1
Voc = 1000 V
Appropriate trip times for array size can be altered according to what the appetite for maximum fault power is
But to truly balance unwanted tripping with safety you have to have some knowledge of the array and there is NO a priori one-size fits all solution to completely fulfill both safety and unwanted tripping concerns
It may be necessary to go to a range of values which can be chosen by array operator