DC Systems Working Group EFCOG ESTG Fermi National Laboratory July 18-22, 2016

DC - History On the EFCOG ESTG website On the EFCOG ESTG website See DC Working Group e.PPTX for: See DC Working Group e.PPTX for: DC Arc Flash WG Phase I – 10/2010 DC Arc Flash WG Phase I – 10/2010 DC Systems WG Phase II – 10/2012 DC Systems WG Phase II – 10/2012 DC Systems WG Phase III – 7/2014 DC Systems WG Phase III – 7/2014 See DC Working Group d1.PPTX for: See DC Working Group d1.PPTX for: DC Systems WG Phase IV – 7/2015 DC Systems WG Phase IV – 7/2015

Advertised Agenda This working group continues efforts from last year to quantify the degree of hazard posed by DC sources, especially where it differs from AC system hazards, and safe work procedures for DC systems, such as battery and photovoltaic systems, where de-energization methods are not obvious or practical. It provides input to standards-making bodies where justified. This working group continues efforts from last year to quantify the degree of hazard posed by DC sources, especially where it differs from AC system hazards, and safe work procedures for DC systems, such as battery and photovoltaic systems, where de-energization methods are not obvious or practical. It provides input to standards-making bodies where justified.

Detailed Agenda Prepare EFCOG Best Practices from 2015 ESW Battery Risk Assessment Flowchart DC Arc Hazard Evaluation Spreadsheets (verified) Excel calculator based on Doan’s NFPA 70E 2012 Excel calculator based on Doan’s NFPA 70E 2012 Two Excel calculators based on Ammerman by INL and LLNL Two Excel calculators based on Ammerman by INL and LLNL High Energy Ground Stick Standard for Pulsed Power Applications, from LBNL. New issues concerning ground faulting in large PV systems that require 2x voltage protection for workers New DC Arc Hazard calculations for PV systems that correct the 4x underreporting of IE for workers.

DC Arc Flash WG 2016 Members new members this year Stan Berry, Gary Dreifuerst, Adam Green, Timothy Kuneli, Adam Martinez, Peter McNutt, Lynn Ribaud, Melkie A. Tega, Justin Tokash, Donald Bourcier, Kyle Carr, David Florez, Thomas Florez, Martin Iedema, Eugene Ormand, Jason Sempsrott, Tim Snow, Gedeon Teame, Michael Utes Stan Berry, Gary Dreifuerst, Adam Green, Timothy Kuneli, Adam Martinez, Peter McNutt, Lynn Ribaud, Melkie A. Tega, Justin Tokash, Donald Bourcier, Kyle Carr, David Florez, Thomas Florez, Martin Iedema, Eugene Ormand, Jason Sempsrott, Tim Snow, Gedeon Teame, Michael Utes Facilities-ANL, DOE, FNAL, LANL, LBNL, LLNL, Navy, NREL, NSTech, PNNL, SNL Facilities-ANL, DOE, FNAL, LANL, LBNL, LLNL, Navy, NREL, NSTech, PNNL, SNL

Battery Risk Assessment Flowchart (2014) START MIN PPE REQ: Safety Glasses, No Metal/Jewelry, Insulated Tools >3kW <100V Electrolyte Thermal Chemical Perform ARC FLASH Calc No SHOCK or ARC FLASH Rated PPE REQ Y N N Y Chemical PPE REQ No Chemical PPE REQ NY Additional Hand/Face PPE REQ No Thermal PPE REQ <1.2 cal/cm² No SHOCK or ARC FLASH Rated PPE REQ (note 1) Can it be segmented <100V N Y N <40 cal/cm² SHOCK PPE REQ, no ARC FLASH Rated PPE REQ Y SHOCK AND ARC FLASH Rated PPE REQ Y Electrical Revise scope/work plan to get Arc Flash calculation <40 cal/cm 2 N Note 1: ARC FLASH Rated PPE may be required during segmentation. Perform ARC FLASH calculation.

Battery Risk Assessment Flowchart 70E 2015 START MIN PPE REQ: Safety Glasses, No Metal/Jewelry, Insulated Tools <100V Perform ARC FLASH Calc No SHOCK or ARC FLASH Rated PPE REQ Y N <1.2 cal/cm² No SHOCK or ARC FLASH Rated PPE REQ (note 1) Can it be segmented <100V N Y N <40 cal/cm² SHOCK PPE REQ, no ARC FLASH Rated PPE REQ Y SHOCK AND ARC FLASH Rated PPE REQ Y Electrical Revise scope/work plan to get Arc Flash calculation <40 cal/cm 2 N Note 1: ARC FLASH Rated PPE may be required during segmentation. Perform ARC FLASH calculation. Other Hazards to Consider 1.Thermal (>3kW) 2.Chemical (electrolyte) 3.Gas (explosion) 4.Pressure (case rupture) 5.Weight (lifting, rigging)

DC Arc Hazard Evaluation Methods Observations: Doan and Ammerman result in similar (within 8%) incident energies for a battery system, using constant clearing time of 2 seconds, 130 – 260 V, and 0.5 – 2 inch gaps. Estimates were significantly higher than measured test data over these conditions (Kinectrics Bruce Power data – comparative results presented in 2014 Cantor Doble paper) Doan and Ammerman result in similar (within 8%) incident energies for a battery system, using constant clearing time of 2 seconds, 130 – 260 V, and 0.5 – 2 inch gaps. Estimates were significantly higher than measured test data over these conditions (Kinectrics Bruce Power data – comparative results presented in 2014 Cantor Doble paper) Ammerman may result in more accurate incident energies than Doan in cases where clearing time is dependent on Iarc (e.g. determined from TCC). Ammerman may result in more accurate incident energies than Doan in cases where clearing time is dependent on Iarc (e.g. determined from TCC). INL Mathcad / EXCEL tool is useful for computing Iarc and resulting incident energy for the Ammerman method INL Mathcad / EXCEL tool is useful for computing Iarc and resulting incident energy for the Ammerman methodRecommendations: Recommend more testing to evaluate accuracy of existing models and/or develop additional empirical models. Insufficient data points currently available to validate models. Recommend more testing to evaluate accuracy of existing models and/or develop additional empirical models. Insufficient data points currently available to validate models. Need to determine applicability of models to PV and other DC sources Need to determine applicability of models to PV and other DC sources

Excel calculator based on Doan’s NFPA 70E 2012

INL Excel calculator based on NFPA 70E Annex D8.1.2 Stokes, Oppenlander and Ammerman papers

LLNL Excel calculator based on NFPA 70E Annex D8.1.2 Stokes, Oppenlander and Ammerman papers, page 1

LLNL Excel calculator based on NFPA 70E Annex D8.1.2 Stokes, Oppenlander and Ammerman papers, page 2

High Energy Ground Stick Standard for Pulsed Power Applications, from ESW 2015

LBNL Build Standard for Ground Hook Safety

LBNL High Energy Ground Hook Safety for Pulsed Power Applications, from ESW Merge with previous 2 and incorporate comments by deHope to create a BP for EFCOG

Where are Utility DC systems going? Increased by 20% since 2015 to reach 2 GW

Existing code structure Existing code structure 690 Solar Photovoltaic (PV) Systems 690 Solar Photovoltaic (PV) Systems 692 Fuel Cell Systems 692 Fuel Cell Systems 694 Wind Electric Systems 694 Wind Electric Systems 705 Interconnected Electric Power Production Sources 705 Interconnected Electric Power Production Sources 708 Critical Operations Power Systems 708 Critical Operations Power Systems New for 2017 proposed New for 2017 proposed Where is the NEC going for Energy?

PV Systems Voltage PPE, p1 Grounded PV systems are allowed by the UL 1741 Inverter standard to attach to ground through only a 5A fuse. Grounded PV systems are allowed by the UL 1741 Inverter standard to attach to ground through only a 5A fuse. UL 1741 defines the “DC Ground Fault Detector/Interrupter” in section 31 UL 1741 defines the “DC Ground Fault Detector/Interrupter” in section The ground-fault detector/interrupter (GFDI) shall sense a ground fault, interrupt the ground-fault current path and provide an indication of the fault when the ground-fault currents exceed the limits shown in Table The ground-fault detector/interrupter (GFDI) shall sense a ground fault, interrupt the ground-fault current path and provide an indication of the fault when the ground-fault currents exceed the limits shown in Table For >250kWDC the current level is 5A For >250kWDC the current level is 5A 31.3 A ground fault detector/interrupter that has tripped in accordance with 31.2 shall not be capable of automatic reclosure A ground fault detector/interrupter that has tripped in accordance with 31.2 shall not be capable of automatic reclosure.

PV Systems Voltage PPE, p2 Grounded PV systems are allowed by the UL 1741 Inverter standard to attach to ground through only a 5A fuse. Grounded PV systems are allowed by the UL 1741 Inverter standard to attach to ground through only a 5A fuse. This fuse clears when a fault occurs in any panel string (rated at ~9Adc). This sets an error bit, blocks the inverter and alerts the operators. The array will continue to generate voltage. This fuse clears when a fault occurs in any panel string (rated at ~9Adc). This sets an error bit, blocks the inverter and alerts the operators. The array will continue to generate voltage. When repair is started, the first isolation switch (DCB1) will be opened. At this time the ground fault is somewhere in a PV string. With the DCB1 switch open, there exists 2x the normal strin voltage across the switch. When repair is started, the first isolation switch (DCB1) will be opened. At this time the ground fault is somewhere in a PV string. With the DCB1 switch open, there exists 2x the normal strin voltage across the switch. For a 1kVdc system design that means 2kVdc across the switch terminals (positive to negative terminals). The maximum DC rating for Class O gloves is 1500 Vdc. This must be considered when designing for maintenance. A single pole switch would not cause the Class O glove to be underrated. For a 1kVdc system design that means 2kVdc across the switch terminals (positive to negative terminals). The maximum DC rating for Class O gloves is 1500 Vdc. This must be considered when designing for maintenance. A single pole switch would not cause the Class O glove to be underrated.

PV Systems Voltage PPE, p3 All closed DPST switch

ground fault PV Systems Voltage PPE, p4 ground fault All closed Fuse opens due to fault All closed GRD fault DPST switch

PV ground fault-repair 5A GFI fuse PV Systems Voltage PPE, p5 PV ground fault-repair 5A GFI fuse All closed Fuse repaired closed GRD fault closed open -1kVdc +1kVdc 0V DPST switch Now compare the DPST switch with a SPST (negative rail is always bonded together)

PV Systems DC Arc Flash calcs, p1 Eduardo Enrique, “DC Arc Flash Calculations for Solar Farms”, st IEEE Conference on Technologies for Sustainability Eduardo Enrique, “DC Arc Flash Calculations for Solar Farms”, st IEEE Conference on Technologies for Sustainability The methods in 70E2015 Annex D.5 use the maximum power method which can underestimate the incident energy by as much as 3x. The methods in 70E2015 Annex D.5 use the maximum power method which can underestimate the incident energy by as much as 3x. In Enrique’s survey of Utility-grade PV Modules, he found that the arcing voltage is max power voltage and that the same holds true for the current. Both values are on the panel’s nameplate. In Enrique’s survey of Utility-grade PV Modules, he found that the arcing voltage is max power voltage and that the same holds true for the current. Both values are on the panel’s nameplate.

PV Systems DC Arc Flash calcs, p2 Ratio of Electrical Characteristics for Utility- Grade PV Modules Ratio of Electrical Characteristics for Utility- Grade PV Modules VI for 300W PV VI for 300W PV V oc-pv I sc-pv

PV Systems DC Arc Flash calcs, p3 For the standard example in his paper, the IE pv /IE max was 3.24, when the temperature compensation factor (raises the PV output voltage for cold temperatures) was included the factor was That is basically 4x the 70E method for calculating AF PPE. For the standard example in his paper, the IE pv /IE max was 3.24, when the temperature compensation factor (raises the PV output voltage for cold temperatures) was included the factor was That is basically 4x the 70E method for calculating AF PPE.

PV Systems DC Arc Flash calcs, p4 Need new Excel Calculator for PV styled like Doan calculator – TBD in working group Need new Excel Calculator for PV styled like Doan calculator – TBD in working group