hstory of fault contribution solar inverter

Impact and Sensitivity Studies of PV Inverters Contribution toFaults based on Generic PV Inverter Models

Ontario Grid Connection Study

Prepared for: CanSIAProject Manager: Wesley Johnston

Director of Policy and Research

Prepared by: Quanta Technology, LLCProject Manager: Farid Katiraei, Ph.D.Contributors: Farid Katiraei, Juergen Holbach, Tim Chang

May 2, 2012

Ontario Connection Study Summary of Impact & Sensitivity Studies Final (May 02, 2012) Page 2

CONFIDENTIAL/PROPRIETARY: This document contains trade secrets and/or proprietary, commercial, orfinancial information not generally available to the public. It is considered privileged and proprietary to theOfferor, and is submitted by Quanta Technology LLC. in confidence with the understanding that its contents arespecifically exempted from disclosure under the Freedom of Information Act [5 USC Section 552 (b) (4)] andshall not be disclosed by the recipient [whether it be Government (local, state, federal, or foreign), privateindustry, or non-profit organization] and shall not be duplicated, used, or disclosed, in whole or in part, for anypurpose except to the extent in which portions of the information contained in this document are required topermit evaluation of this document, without the expressed written consent of the Offeror. If a contract isawarded to this Offeror as a result of, or in connection with, the submission of this data, the right to duplicate,use, or disclose the data is granted to the extent provided in the contract.


Contact and Information:For any question and/or further information about this study and results, please contact:

Wesley JohnstonDirector of Policy and ResearchCanadian Solar Industries Association/L'Association des Industries Solaires du Canada150 Isabella Street, Suite 605, Ottawa, ON, CANADA, K1S 1V7tel: 613-736-9077 ext.224 fax: 613-736-8938toll free: 866-522-6742 [email protected]

Farid KatiraeiPrincipal Advisor, Protection/Automation & Distributed GenerationQuanta Technology, LLC1031 McNicoll Avenue, Suite 102, Toronto, Ontario, Canada, M1W 2L8Ph: 416-477-5813Cell: [email protected]

Acknowledgments:The enclosed document reflects outcome of the work, discussions, and comments by several members of theproject team and PV inverter manufacturers. Quanta Technology would like to thank all the participants andtheir great effort in reviewing and enhancing the document.

The present Study Working Group includes: Wesley Johnston and David Wills (CanSIA), Luis Marti, BingYoung, Andrew Yan, and Arun Narang (Hydro One), Gary Thompson and Peter Baroutis (Toronto Hydro),Janos Rajda (SMA), Farid Katiraei, Juergen Holbach, Tim Chang, Julio Romero Aguero and Carl Wilkins(Quanta Technology).

The PV inverter vendor participants in alphabetical order are: Fronios, Santerno, SatCon, Schneider Electric,SMA, and Sustainable Energy Technologies. Quanta Technology appreciates their outstanding efforts todevelop and assist with verification of PV inverter transient models.


Contents1 Executive Summary........................................................................................................................................ 6

1.1 Project Background .................................................................................................................................. 61.2 Study Scope.............................................................................................................................................. 71.3 Findings and Conclusions ........................................................................................................................ 81.4 Summary of Recommendations ............................................................................................................. 10

2 Introduction................................................................................................................................................... 113 Benchmark Study System............................................................................................................................. 13

3.1 Resistive Load ........................................................................................................................................ 134 PV Inverter Overview................................................................................................................................... 145 Impact Study Cases....................................................................................................................................... 17

5.1 Fault Studies for PV Inverters at Location A......................................................................................... 195.1.1 Effect on voltages and Inverter tripping ......................................................................................... 195.1.2 Effect on fault currents.................................................................................................................... 20

5.2 Components of Fault Currents ............................................................................................................... 225.3 Effect of PV Inverter PLL design on current contribution..................................................................... 25

6 Impact of PV Inverter Fault Contribution on CB ratings ............................................................................. 296.1 Influence of PV inverter current on Symmetrical Circuit breaker rating............................................... 296.2 Influence of PV inverter current on asymmetrical Circuit breaker rating.............................................. 306.3 Influence of PV inverter current on close and latch rating .................................................................... 316.4 Effect of change in PV inverter frequency............................................................................................. 34

7 Summary of variations in the PCC voltages and fault detection time for different PV inverter locations... 368 Summary of Fault Current differences ......................................................................................................... 439 Summary of Sensitivity Study Cases............................................................................................................ 48

9.1 Fault levels for no-generation conditions of benchmarks ...................................................................... 489.2 Fault current comparison........................................................................................................................ 50

10 Summary of Observations and Findings....................................................................................................... 5210.1 Summary of main observations from the studies ............................................................................... 5210.2 Summary of main findings ................................................................................................................. 55

11 Recommendations and Future Work ............................................................................................................ 5712 References..................................................................................................................................................... 5913 Appendix A Benchmark Specifications..................................................................................................... 60

13.1 Circuit Breakers and Reclosers........................................................................................................... 6114 Appendix B PV Inverters Response to Voltage Drop and Faults .............................................................. 62

14.1 Voltage step changes .......................................................................................................................... 6214.2 Fault Current Summary ...................................................................................................................... 63

15 Appendix C Simulation results for various faults and PV inverter locations ............................................ 6515.1 PV Inverters at Location A................................................................................................................. 66

15.1.1 Three Phase Fault at location A...................................................................................................... 6615.1.2 Single Phase Fault at location A ..................................................................................................... 7015.1.3 Three Phase Fault at location E ...................................................................................................... 7415.1.4 Single Phase Fault at location E...................................................................................................... 7515.1.5 Three Phase Fault at location C ...................................................................................................... 7715.1.6 Single Phase Fault at location C ..................................................................................................... 81

15.2 PV Inverters at Location B ................................................................................................................. 8515.2.1 Three Phase Fault at location C ...................................................................................................... 8715.2.2 Single Phase Fault at location C ..................................................................................................... 9115.2.3 Three Phase Fault at location A...................................................................................................... 9515.2.4 Single Phase Fault at location A ..................................................................................................... 99


15.3 PV Inverter at Location C................................................................................................................. 10315.3.1 Three Phase Fault at location D.................................................................................................... 10415.3.2 Single Phase Fault at location D ................................................................................................... 108

15.4 PV Inverter at Location D ................................................................................................................ 11215.4.1 Three Phase Fault at location A.................................................................................................... 11415.4.2 Single Phase Fault at location A ................................................................................................... 118

15.5 Four PV Inverters at Location B....................................................................................................... 12215.5.1 Three Phase Fault at location C .................................................................................................... 12415.5.2 Single Phase Fault at location C ................................................................................................... 12715.5.3 Three Phase Fault at location E .................................................................................................... 13015.5.4 Single Phase Fault at location E.................................................................................................... 132

15.6 Two PV Inverters at Location D....................................................................................................... 13415.6.1 Three Phase Fault at location A.................................................................................................... 13615.6.2 Single Phase Fault at location A ................................................................................................... 140


1 Executive Summary

1.1 Project Background

In early 2011, solar industry participants were having high level conversations with the Ontario Ministry ofEnergy and Hydro One regarding the challenges of connecting solar photovoltaics (PV) projects equal to or lessthan 500 kW to the Ontario electrical grid system. Many solar industry participants were experiencing delaysregarding the connection of their projects. The reason for these delays and the inability to connect to theelectrical grid were due to technical reasons as stated by Hydro One and some local distribution companies.

As solar PV is now playing a larger role in the Ontario electricity grid system, there was a clear need for greaterinformation and analysis concerning the connecting of solar photovoltaics (PV) projects equal to or less than500 kW and short circuit impacts. A jointly-funded study approach between the solar industry and Hydro Onewas proposed which become known as the Ontario Connection Study.

The Canadian Solar Industries Association (CanSIA) took on the leadership role of organizing and seeking solarindustry participants to partake in this joint solar industry and Hydro One study. A number of solar industrycompanies participated in the funding of this project (See Solar Connection Champions list). A Project SteeringCommittee (PSC), co-chaired by both Hydro One and CanSIA, was formed to guide the project. Additionally, aTechnical Working Group was also created to advise the PSC and to take on many of the technical elements ofguiding this study.

The objective of this project was to conduct a third party study to explore the impact of connecting solar PVgeneration equal to or less than 500 kW per installation to the electrical grid, to understand the technicalchallenges for both utilities and developers, and to explore possible solutions related to the technical challenges.

The PSC held its first official meeting on April 14, 2011 in Toronto, Ontario. At that time a Terms of Referencefor both PSC and the TWG were created. A process was put in place to create a scope of work for this importantand comprehensive research project and to create an RFP to seek a qualified research consultant to complete thetask. A project manager was hired to assist the TWG in structuring the RFP and selecting the right researchconsultant. On June 17, 2012 Quanta Technology was selected as the research consultant to perform thisresearch task as they demonstrated the qualifications required to perform this research and put together a workplan that addressed all the elements of study scope of work as defined by the TWG.

Solar Connection ChampionsAdvanced Energy AMP Solar GroupCanadian Solar Inc. CanSIACelestica Deivendran Consulting Inc.Fronius Hydro OneJCM Capital OZZ SolarPhotowatt Ontario Potentia SolarSanterno SatconSchneider SharpSMA America Great Circle Solar Income CorporationSolar Power Products SunEdisonSunRise Power Sustainable Energy TechnologiesToronto Hydro


1.2 Study Scope

The specific focus of this joint CanSIA and Hydro One study was to determine characteristics of commercialPV inverters under short circuit conditions and to investigate the potential contribution of the PV inverters toshort circuit current levels when there are faults on the distribution and transmission systems. This study took asystematic approach to examining the PV inverter transient short circuit characteristics and their short circuitcontributions to the electrical grid that a distribution or transmission utility needs to reflect in impactassessments.

The overall approach of the study was to obtain detailed modeling information from manufacturers of PVinverters that were likely to be utilized for solar PV projects equal to or less than 500 kW in Ontario, developthe necessary mathematical and software simulation models that provide the representative response of suchinverters and then conduct simulations of these devices on an electrical system that is reflective of thedistribution and transmission network where currently short circuit constraints exist. A number of realisticscenarios were simulated and evaluated to assess the short circuit impact of various parameters from the PVinverter and the system aspects. In addition, a literature survey was also conducted to further inform thedevelopment of the study approach, identify issues identified by others and provide insight on the state-of-artwith respect to modeling of PV inverters and their potential impacts on system short circuit levels.

The short circuit phenomenon as they apply to the operation of power system equipment takes place typicallyfrom a few to ten cycles (within the PV inverter maximum interruption timeframe), which translates into atimeframe of up to 170 milliseconds. During this period, power system protection devices such as circuitbreakers must operate to interrupt the short circuit current which can be in the order of thousands to tens ofthousands of Amps and represent tremendous levels of energy. Both the electrical behavior of generationdevices and their control devices and the electrical phenomenon seen by protective devices during this transienttimeframe are extremely complex.

As a result, advanced simulation tools to conduct switching transient analysis, such as PSCAD/EMTDC, wererequired to look at the complex control systems of PV inverters and their impacts to the system. Moreconventional programs used for production level connection impact studies by utilities such as PSS/E andCYMDIST are referred to as steady state programs.. Both the analysis capability and the associatedsimulation models of such programs are not suitable to properly assess the dynamic three-phase analysisrequired in this study.

Significant effort was required to consult with PV manufacturers to obtain and assess models that simulateinverters in the 60 Hz to 5 kHz time frame and their corresponding data. A total of 8 PV manufacturers wereconsulted in an effort to cover as much of the Ontario market share as possible. Ultimately, 6 manufacturermodels were obtained, verified and their short circuit characteristics analyzed. These 6 models represent morethan 80% of the PV inverters being installed in Ontario.

Following further analysis and verification of the transient behavior of the 6 models under short circuitconditions, it was determined that these models could be further grouped into two categories:

1. PV model with fast disconnection (i.e. in less than one cycle)2. PV model with continued operation for up to 10 cycles


There was one model in the first category and this was referred to as Model 1. Five of the six models fall inthe second category and it was determined that this group could be presented by a single Generic Model.

Once the inverter models were established, a PSCAD/EMTDC model of a benchmark system was developed inconsultation with Hydro One. The benchmark model reflects an 115kV transmission system supplying typicalstep-down transformer stations (TS) connecting a distribution system. The distribution system includes the TS,downstream distribution stations (DS), reclosers and lower voltage distribution feeders to represent the systemlevels that supply typical end-user customers and where PV installations may be found. The transmission anddistribution system includes modeling details such as transformer winding configurations and groundingdevices that may impact short circuit levels. This system represents an actual portion of the Hydro Onedistribution and transmission system where short circuit constraints presently exist.

A large number of simulations were then performed looking at the short circuit impact with variouspermutations of:

1. PV inverters at four different locations,2. One, two and four inverters at each location,3. Applying faults at five locations along the feeders,4. Applying three types of faults: line-to-ground, line-to-line-ground and 3-phase

Over 840 simulation case studies were conducted to cover a wide range of conditions and scenarios. Both theinverter behavior and the impact of the inverters on the system short circuit levels were reviewed. Furthersensitivity analysis was conducted looking at the short circuit impact of PV inverters to various systemparameters including line/feeder length, transformer reactance, X/R ratios and system strength.

An assessment of the impact of PV inverters on the close and latch capability of circuit breakers was alsoconducted. The IEEE standard C37.04-1999 establishes the close and latch requirement for circuit breakers andmaximum instantaneous current levels must not exceed these requirements to ensure the integrity of the breakeroperation. Simulations were conducted for different scenarios and fault conditions to assess the impact of thePV inverters during the short circuit period.

1.3 Findings and Conclusions

The main findings and conclusions from the study are outlined below:

1. The PV inverter current contribution during a fault is not zero and varies by design.

a) It was observed that, for most fault conditions, several PV inverters continued supplying current to thefeeder subsequent to a fault for a period ranging from 4 to 10 cycles.

b) The current contribution level is a function of the voltage at the terminal of the PV inverter during afault, which is determined by the type and location of a fault.

c) For most vendor-specific PV inverter models, the PV inverter current during the fault were above therated inverter current and may reach up to 120% of the inverter rated current.

d) For a few PV inverter models, the inverter current during faults was maintained at the pre-fault invertercurrent.


e) In one inverter model, the inverter current was dropped to zero and the inverter was disconnected in lessthan 0.5 cycles for a fault case in which the terminal voltage reached below 50% on any phase.

2. In computing the short circuit level seen by a circuit breaker, the PV inverter rated current (aggregate size)with an adjustment factor of 1.2 (maximum) may be used and arithmetically added to the system faultcurrent, as calculated from conventional distribution short circuit study tools (without incorporating PVinverters).

a) Applying an arithmetic summation is a generalization approach for the worst case scenario. The accuratecalculation by using the phase angle difference between the two current vectors contributing to the faultat the contact parting time of a circuit breaker requires a complex transient study, but can result insmaller values.

b) Depending on the vendor-specific inverter models, the adjustment factor may vary in the range of 1.0 to1.2 per unit. However, the aggregate aspects of multiple inverters from various vendors with potentiallydifferent adjustment factors complicate the consideration of a lower factor.

c) In one specific vendor model and only for the faults causing a voltage drop more than 50% at the PVinverter terminal, there was no current contribution (i.e. adjustment factor of zero).

3. In PV Project site cases where system fault current levels are close to upstream breakersymmetrical/interrupt rating the computation of the fault current level increment due to PV inverter unitsaddition/connection can be better estimated through performing electromagnetic transient simulationstudies.

a) In order to apply vector summation of the two contributing current components, one needs to determinethe phase angle difference between the two currents at the contact parting time of a circuit breaker.

b) Typically, there is an initial phase angle difference of 72 to 90 degrees between the two currents.However, under some fault conditions, the phase angle of the inverter current may vary with time.

c) Five out of the six selected inverter models showed changes in the phase angle of the inverter currentduring the fault due to specific inverter control behavior associated with design of the Phase LockedLoop block. The frequency of the PV inverter current during a three-phase bolted fault at the PV inverterterminal may be different than the pre-fault grid frequency. This introduces a time-varying phase anglewith the potential of two current vectors gradually becoming in phase at some point of time.

d) A relatively complex transient simulation study will be needed to model PV inverters control behaviorand properly investigate the PV inverter current contribution to the system fault and impact on thesymmetrical rating of a circuit breaker.

4. From the circuit breaker close and latch viewpoint, the PV inverter current contribution during the fault isnot a limiting factor, if the symmetrical rating requirements are properly met. The close and latch rating canbe estimated based on the findings stated below:

a) The PV inverter current contribution to the peak of the first cycle fault (base for the calculation of theCB close and latch rating) is about 20% of the aggregated PV inverter rated current.

b) The PV inverter current does not have any direct current component (dc offset) that would cause anyshift in the output current.


c) Based on the IEEE standard C37.04-1999, the close and latch rating of a circuit breaker is 2.6 times theasymmetrical rating.

1.4 Summary of Recommendations

The following recommendations are derived from the studies:

Develop and propose an industry best practice guideline for the design of the PV inverter behaviorduring faults.

Share the findings of the report with various Standard Development organizations (e.g. CSA, ANSI,IEEE, etc.).

Develop new study tools and/or enhance the existing utility engineering software tools to provide theindustry with a proper method for analysis of system faults in presence of multiple PV inverters ofvarious sizes.

Investigate effect of vendor-specific active anti-islanding schemes on fault detection time of PVinverters for cases that the voltage drop is not large enough.

Perform field testing of various three-phase PV inverters to further examine the characteristics duringfaults and to validate and enhance generic PV inverter models.

Investigate other impacts of the PV inverters on distribution system voltages and operation of the feederequipment that may be considered as limiting factors in achieving medium to high penetration of PVinverters.

It is also highly recommended to review the findings of the study in three to five year timeframe, due to dealingwith a fast changing and evolving industry. Because of the industry needs and trend toward enabling high PVpenetration, new control and protection capabilities are being introduced by vendors for consideration. Someaspects such as the maximum current contribution factor of 1.2 pu can be lowered for the PV inverters with newgeneration of controls.


2 IntroductionThis report outlines a summary of the impact studies and sensitivity analysis that were performed on arepresentative benchmark system with PV inverters to investigate effect of PV inverters on short circuitcapacity and fault current characteristics of distribution systems in Ontario. Two inverter models were selectedfor the purpose of studies following a series of simulation model evaluations for commercial PV inverters fromsix vendors. The list of vendor models represents products that are presently utilized in the Ontario Marketcovering over 80% of market participation. The individual PV inverter size was limited to 500 kW maximumrating, while incorporating cases of multiple PV inverters in parallel.

A representative system was introduced and used as the benchmark for the impact studies. In order to accountfor the variations in the design and characteristics of a wide range of distribution systems, applicable ranges ofchanges in the benchmark parameters were determined to vary system specifications. The impact study caseswere repeated and the effect of fault contribution was examined for each parameter change.

The impact studies outlined in this report is based on the following approach:

First, the base-case fault currents for faults at specified points on the feeder without any generation wereexamined and reported in table format for three-phase and single phase faults,

A PV inverter unit based on Model 1, in one iteration, and Generic Model, in the second iteration wasadded to the benchmark model at the first location (point A). The fault cases at multiple locations weresimulated to capture transient voltage and current waveforms at PCC and each of the circuit breakers orreclosers.

The PCC voltage was analyzed to determine level of voltage drop and the ability of the PV Inverter todetect a fault condition,

The fault current waveforms were examined to determine any change in the level of fault current basedon measuring the change in the first cycle peak, the second cycle peak, up to the fifth cycle peak, as wellas the rms of fault currents per phase,

The simulation results were documented and plots generated to illustrate transient currents, The fault current differences for the corresponding cases with and without a PV inverter was calculated

and reported to determine change in the fault level, In the next step, the PV inverter location or model was changed and fault analysis repeated.

After completing the impact study cases for the base benchmark system, the benchmark parameters wereadjusted to reflect a range of system characteristics according to the approach described for the sensitivitystudies. Fault analyses were repeated for two new benchmarks.

A summary of the impact studies and sensitivity analysis cases are highlighted in this report based on selectedsimulation results. The comprehensive set of illustrations for relevant measurement quantities (voltages,currents, active and reactive power) from the studies are documented as individual plots to cover all thesimulation cases.

The remaining part of this report is organized as follows. Sections 3 and 4 will provide a brief description of thestudy benchmark and the PV inverter models. A summary of impact studies for selected cases is provided inSection 5. Impact of faults on the short circuit levels seen by the CB is investigated and discussed in Section 6.Summary tables of variations in PCC voltages, fault currents, and the expected PV inverter tripping conditions


are given in Sections 7 and 8. A brief summary of sensitivity study results are outlined in Section 9. Section 10represents the major observations and conclusions from the studies.


3 Benchmark Study SystemA schematic diagram of the impact study benchmark system is shown in Figure 1. The benchmark system isprovided by Hydro One to represent typical distribution system topologies in Ontario. The system is modeled inPSCAD/EMTDC simulation software for this study. The benchmark parameters are given in Appendix A.Detailed description is provided in Reference 1.

Figure 1: PV Impact study system

The study system comprises of part of a 115 kV transmission system, identified with a 115 kV source that feedsa 115 kV / 27.6 kV distribution substation (DS) through two 14.4 km long lines (Line 1 and Line 2). The 115kV / 27.6 kV substation includes two 83 MVA transformers (Y/) and two grounding bank transformers (GT1 & GT2). The grounding banks are 4.78 MVA each and connected to the 27.6 kV bus. The distributionsubstation typically feeds ten 27.6 kV feeders. For the purpose of this study, one of the feeders (Feeder 1) isconsidered in detail. A second feeder is also considered with characteristics similar to the first feeder toprovide a few points for applying faults on an adjacent feeder. The remaining adjacent feeders (about 8 feeders)are represented as an aggregate load connected to the 27.6 kV bus as shown in Figure 1.

Feeder 1 in Figure 1 is about 15 km long and supplies several 27.6 kV customers. This feeder ends on atransformer station (TS). An 8.32 kV feeder (Feeder 2), about 10 km long, connects the rest of the customers onthis circuit. The total customer loads for each of Feeder 1 and Feeder 2 are modeled as two spot loads, Load 1and Load 2, respectively.

3.1 Resistive LoadMost of the fault simulations were performed with the original loadings of the feeder as outlined in AppendixA. However, during the simulation studies, it was noted that, even without any generation source, a dc flow andcurrent contribution to a fault from downstream load can be observed. This was primarily associated with theinductive nature of the loads on the feeders that create a slowly damping discharging current from thedownstream load toward an upstream fault. The discharge may last for several cycles and the discharge currentis in the order of PV inverter contribution. To avoid any erroneous interpretation of the simulation results, it wasdecided to repeat and/or continue studies with purely resistive loads.


4 PV Inverter OverviewBased on extensive simulation studies and examination of six vendor-specific PV inverter models, from thefault characteristic viewpoint, two inverter models were selected and used for the impact studies. The modelsare: Model 1 that interrupts the inverter current contribution immediately during a fault event if the PCCvoltage drop is beyond 50%, and Generic Model that continues the inverter current for a few cycles(maximum of 10 cycles) after a fault incidence, even if the voltage reduction at the terminal of the generic PVinverter model is below 50%. The differences between the current characteristics of the two PV inverter modelsduring a fault are shown in Figure 2 and Figure 3.

a) Model 1 (time in cycles)

b) Generic Model (time in cycles)

Figure 2: Inverter currents during a three phase fault close to PCC


a) Model 1 (time in cycles)

b) Generic Model (time in cycles)

Figure 3: Inverter currents for a single phase fault close to PCC

The PV inverter represented by the model is a three-phase 500kW 200V unit with the inverter bridge introducedas detailed model. The block diagram is shown in Figure 4.

Figure 4: Model 1 block diagram


The nominal characteristics of the PV inverter model are shown in Table 1. It should be noted that the internalcurrent limit specifies the continuous kVA rating of the inverter. The maximum rms and instantaneous peakcurrents are associated with transient rating of an inverter that should be provided by a vendor.

Table 1 - PV Inverter characteristics for the simulation modelsReal PowerRating (Pn)

kvarating(Sn)

Lowestpf

InverterVoltage

BaseCurrent

(rms)

MaximumCurrent(% rms)

InstantaneousCurrent(% peak)

M1 500 kW, 3Ph 500 kVA 0.9 200 Vrms 1443 A 102 % 180%

The PV inverter represented by the Generic Model is a three-phase 500kW 208V unit with the inverter bridgeintroduced as current controlled voltage source equivalent. The block diagram is shown in Figure 5.

Figure 5: Generic Model block diagram (current controlled voltage source)

The nominal characteristics of the PV inverter model are shown in Table 2.

Table 2 - PV Inverter characteristics for the simulation modelsReal PowerRating (Pn)

kVArating(Sn)

Lowestpf

InverterVoltage

BaseCurrent

(rms)

MaximumCurrent(% rms)

InstantaneousCurrent(% peak)

Generic 500 kW, 3Ph 500 kVA 0.9 208 Vrms 1450 A 120 % 180%

Both PV inverter models include typical under/over voltage and frequency protection as prescribed by IEEE1547. The voltage measurement and protection is performed per phase. The default settings are reported inTable 3.

Table 3: Inverter protection for modelsThreshold Delay Notes

Fast ac over-voltage >1.2 p.u. 160 ms Generic modelSlow ac over-voltage >1.1 p.u. 1 sFast ac under-voltage 60.5 Hz 160 ms Generic modelFast ac under-frequency


5 Impact Study CasesThe objective of the impact assessment is to investigate and demonstrate (through simulations) some of thetechnical challenges specific to the connection of PV inverter based generation to the electrical grid. The focusof the study is to primarily evaluate fault current contribution of a PV inverter based generation with specificfault characteristic on a benchmark system which represents a typical distribution system in Ontario.

The base case simulation model of the benchmark system as described in Section 2 is used for the impactstudies. The base case uses the minimum loading of the feeders and typical configurations/parameters for all theassociated components. Several fault cases are simulated to investigate inverter behavior during various faultsand quantify the impact on the benchmark system. The impact study of this task is performed for a multitude offault conditions to cover feasible scenarios that may occur in the field. Eight (8) fault locations are selected,with two fault types (single-line-to-ground and three-phase bolted faults) investigated for each location. Thestudy system with fault locations without any PV Inverter is shown Figure 6..

Figure 6: Study system with selected fault locations

As references for fault comparison, Table 4 and Table 5 summarize the fault currents and the minimum voltagemeasurements at selected locations on the feeders for the no generation condition. Green background identifiesthe cases that fault is at or downstream of a PV location. The maximum fault current (first cycle peak) of thebase case benchmark is about 34.3 kA at 27.4 kV bus for a three phase fault at point A or E. The single phasefault current is about 20.6 kA peak for fault either at point A or E.


Table 4 - Currents and voltages during fault at main buses for No generation condition (3Ph fault, base case)

peak (1stcycle)

RMS(steady-

state)

peak (1stcycle)

RMS(steady-

state)

peak (1stcycle)

RMS(steady-

state)

peak (1stcycle)

RMS(steady-

state)

HV I 90.80 34.86 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

A 34.33 13.13 34.33 13.13 0.00 0.00 0.00 0.00 0.00 0.00 0.00

B 4.99 2.31 4.99 2.31 0.00 0.00 0.03 0.02 0.91 0.00 0.00

C 8.54 3.70 2.73 1.12 8.54 3.70 0.03 0.02 1.01 0.00 0.00

D 1.91 0.92 0.61 0.28 1.91 0.92 0.03 0.00 1.08 0.83 0.00

E 34.33 13.98 0.00 0.00 0.00 0.00 34.33 13.99 0.00 0.00 0.00

G 8.54 3.70 0.03 0.02 0.04 0.03 2.73 1.12 1.01 1.01 1.00

H 1.91 0.92 0.03 0.02 0.04 0.03 0.61 0.28 1.08 1.08 1.07Fee

der2

8.32 kVBus

(feeder 1)

End offeeder

(point D)

Fault Currents (kA) Minimum RMS Voltage (pu)

FaultLocation

Fault Current CB1 Rec1 CB2

Feed

er1

27.6 kVBus

Table 5 - Currents and voltages during fault at main buses for No generation condition (SLG fault)

peak (1stcycle)

RMS(steady-

state)

peak (1stcycle)

RMS(steady-

state)

peak (1stcycle)

RMS(steady-

state)

peak (1stcycle)

RMS(steady-

state)

HV I 95.67 39.46 0.02 0.01 0.01 0.01 0.02 0.01 0.60 0.27 0.27

A 20.55 8.02 20.55 8.04 0.03 0.02 0.01 0.01 0.00 0.75 0.75

B 3.15 1.46 3.16 1.46 0.03 0.02 0.03 0.02 0.90 0.77 0.77

C 10.42 4.40 1.83 0.78 10.42 4.40 0.03 0.02 1.04 0.00 0.00

D 1.25 0.65 0.24 0.13 1.25 0.65 0.03 0.02 1.09 0.94 0.00

E 20.55 8.04 0.01 0.01 0.03 0.02 20.55 8.05 0.00 0.75 0.75

G 10.42 4.40 0.03 0.02 0.04 0.03 1.83 0.78 1.04 1.03 1.02

H 1.25 0.65 0.03 0.02 0.04 0.03 0.24 0.13 1.09 1.09 1.08

Fault Currents (kA) Minimum RMS Voltage (pu)

FaultLocation

Fault Current CB1 Rec1 CB227.6 kV

Bus

8.32 kVBus

(feeder 1)

End offeeder

(point D)

Feed

er1

Feed

er2

In the following subsections multiple fault locations are introduced and simulated for each PV Inverter locationor a combination of multiple PV inverters. Selected detail simulation results are presented in Appendix C.


5.1 Fault Studies for PV Inverters at Location AFor this study case, PV inverter is located at point A which is downstream of the 27.6 kV bus on feeder 1. Thefaults are applied to several points along the feeders and on the HV line.

Detailed simulation results and plots for faults at selected locations are shown in Appendix C, including thefaults at points: A and E (27.6 kV bus), C (8.32kV bus) and I (115kV bus). Faults at three points (A, E and I)are examined closely in the following sections for the cases of single or multiple PV inverters at location A.

Figure 7: Study system with a PV Inverter at location A and multiple fault locations

5.1.1 Effect on voltages and Inverter trippingTable 6 and Table 7 summarize the minimum voltage measurement at PCC and the PV inverter tripping timefor various fault types and locations. For simplicity, the table structure reflects the configuration of the studybenchmark by the order of fault locations with respect to the PV location. It should be noted that the reportedPV inverter trip time is based on detection time of conventional under/over voltage and frequency protectionschemes and does not include any active anti-islanding scheme.

Table 6 Trip time and PCC voltages for the case of the Generic PV Inverter at Location A

Vpu Phase Vpu Vpu Vpu Phase Vpu PhasePCC PCC PCC PCC PCC

SLG 2s 0.55 AB 10.5c 0.00 A NA 0.86 A NA 0.97 A NA 1.02 ABTPH 10c 0.00 ABC 10c 0.00 ABC 2s 0.85 ABC NA 0.94 ABC NA 1.00 ABC

Vpu Vpu Phase Vpu PhasePCC PCC PCC

10.5c 0.00 A NA 0.97 AB NA 1.02 AB

10c 0.00 ABC NA 0.94 ABC NA 1.00 ABCPoint E Point G Point H

Trip

Trip Trip Trip

Trip Trip Trip

115kV 27.6kV 8.32kV

1x PV

Point I Point A Point B Point C Point D

Trip

Phase

Phase Phase


Table 7 Trip time and PCC voltages for the case of the model1 PV Inverter at Location A


SLG 0.3c 0.55 AB 0.3c 0.00 A NA 0.86 A NA 0.98 AB NA 1.02 ABTPH 0c 0.00 ABC 0c 0.00 ABC NA 0.86 ABC NA 0.94 ABC NA 1.01 ABC


0.3c 0.00 A NA 0.98 AB NA 1.02 AB

0c 0.00 ABC NA 0.94 ABC NA 1.01 ABC

Trip

Trip Phase Trip Trip

Point E Point G Point H

Trip Trip Phase Trip Phase Trip

115kV 27.6kV 8.32kV

1x PV


According to Table 6 and Table 7, when the PV inverter is at location A, it may not detect faults on feeder 2 orfeeder 4. Generally, there was not enough voltage drop for any fault downstream of the 8.32 kV bus to trip thePV inverter based on an under-voltage protection scheme.

5.1.2 Effect on fault currentsThe following tables provide the fault current values and comparison with corresponding fault currents for no-generation condition. Table 8 shows a summary of the peak and rms values of the fault currents and circuitbreaker currents for each of the first 5 cycles subsequent to the fault. Green background identifies the cases thatthe fault is at or downstream of the PV location.

Table 8 Peak and rms values of fault currents and CB currents for a Generic PV Inverter at A and resistive load

1 90.80 51.59 0.02 0.01 0.00 0.002 83.46 49.17 0.02 0.01 0.00 0.003 77.42 45.04 0.02 0.01 0.00 0.004 72.45 42.02 0.02 0.01 0.00 0.005 68.36 39.84 0.02 0.01 0.00 0.001 34.33 19.89 34.33 19.90 0.00 0.002 30.12 17.49 30.14 17.50 0.00 0.003 27.06 15.63 27.06 15.63 0.00 0.004 24.82 14.54 24.81 14.53 0.00 0.005 23.16 13.91 23.15 13.90 0.00 0.001 34.33 19.89 0.02 0.01 34.33 19.902 30.12 17.49 0.02 0.01 30.12 17.493 27.06 15.63 0.02 0.01 27.06 15.634 24.82 14.54 0.02 0.01 24.82 14.545 23.16 13.91 0.02 0.10 23.16 13.91

E

HV I

Feeder 1 A

Feeder 2

Currents during fault (kA)Fault

LocationFaultcycle

Fault Current CB1 CB2peak RMS peak RMS peak RMS


For further evaluation of impact of multiple inverters and large aggregate size, the fault simulations wererepeated for the cases of four PV inverters (2MW), eight PV inverters (4MW) and sixteen PV inverters (8MW)at location A. The fault locations of interest were primarily points A, E, and I. For reference, Table 9 providesthe peak and rms values of the fault currents and CB currents for the case of sixteen PV inverters.

Table 9 - Peak and rms values of fault currents and CB currents for 16 Generic PV Inverters at A and resistive load

1 90.70 51.50 0.26 0.16 0.00 0.002 83.32 49.09 0.25 0.19 0.00 0.003 77.33 44.99 0.28 0.20 0.00 0.004 72.45 42.02 0.28 0.20 0.00 0.005 68.36 39.85 0.28 0.20 0.00 0.001 34.30 19.88 34.36 19.92 0.00 0.002 29.88 17.39 30.16 17.51 0.00 0.003 27.00 15.62 27.08 15.64 0.00 0.004 25.06 14.68 24.82 14.54 0.00 0.005 23.27 14.02 23.16 13.91 0.00 0.001 34.30 19.88 0.30 0.16 34.30 19.882 29.88 17.39 0.25 0.19 29.88 17.393 27.00 15.62 0.28 0.20 27.00 15.614 25.06 14.68 0.28 0.20 25.06 14.685 23.37 14.02 0.28 0.20 23.37 14.02

Feeder 2 E

HV I

Feeder 1 A

Currents during fault (kA)Fault

LocationFaultcycle

Fault Current CB1 CB2peak RMS peak RMS peak RMS

Table 10 shows the percentage of changes in the fault currents with respects to No PV fault currents for faults atvarious points on the feeder. The corresponding fault currents for No PV case is given in Appendix C. Table 10includes the changes in the fault currents for a case of one PV inverter versus sixteen PV inverters (8 MW) atlocation A. For the comparison, the variations in the peak and rms values per cycle for each of the first 5 cyclesafter the fault are also reported.


Table 10 Changes in fault currents with & without PV Inverters at location A

1 0.00% 0.00% -0.11% -0.17%2 0.00% 0.00% -0.17% -0.16%3 0.00% 0.00% -0.11% -0.11%4 0.01% 0.01% 0.00% 0.00%5 0.01% 0.01% 0.02% 0.03%1 0.00% 0.00% -0.09% -0.06%2 -0.05% -0.04% -0.84% -0.64%3 -0.02% -0.01% -0.23% -0.11%4 0.13% 0.06% 1.11% 1.04%5 0.06% 0.06% 0.52% 0.87%1 0.00% 0.00% -0.09% -0.06%2 -0.05% -0.05% -0.84% -0.64%3 -0.02% -0.01% -0.23% -0.11%4 0.06% 0.06% 1.03% 1.05%5 0.06% 0.06% 0.96% 0.88%

I

A

E

16 PV InvertersFault Current Difference (%)

peak RMS1 PV Inverter

peak RMSFault

LocationFaultcycle

The comparison in Table 10 shows that negligible changes in the fault currents were observed for the first threecycles after the fault. In some cases, the fault current was slightly reduced. Table 10 also suggests that the levelof changes in the fault currents is not proportional to the number or aggregate size of PV inverters. Hence, thePV inverters contribution to the fault is not scalable by the numbers of multiple PV inverters.

5.2 Components of Fault CurrentsIn order to examine the changes in the fault current, the PV inverter current contribution (IPCC) and the CB1current (ICB1) are overlaid in one plot as shown in Figure 8 for a three phase fault at point A. The invertercurrent is scaled up by a factor of 1000 (A vs kA) to become observable as compared with the CB1 current. Thefault current (IFL) which is the vector sum of the CB1 current and the PV inverter current is also shown on thesame plot. Figure 8 clearly shows a phase angle difference between the peak of CB1 current and the PV invertercurrent. In addition, due to extremely low current contribution from PV inverter (20A vs 31 kA), the CB1current has not changed. It was also observed that the CB1 current is the same as fault current.


Main : Graphs

sec 1.990 2.000 2.010 2.020 2.030 2.040 2.050

-20

-10

0

10

20

30

40IPCC1A IFLA ICB1A

-40

-30

-20

-10

0

10

20IPCC1B IFLB ICB1B

-20.0

-15.0

-10.0

-5.0

0.0

5.0

10.0

15.0

20.0

25.0IPCC1C IFLC ICB1C

Figure 8 Fault current components (3Ph fault at location A, Generic PV Inverter model at A)

A major observation from Figure 8 is that the phase angle difference between the PV inverter currents and thesystem currents plays an important role in the contribution to the total fault current. However, this phase angleis changing with time. In the first three cycles, the instantaneous values of the inverter currents are negative atthe instances of the peak currents for the system fault contribution (or CB1 currents). In the subsequent cycles,the currents slowly become in phase with the system fault current (e.g. the fourth and fifth cycles after thefault). The time-varying phase angle difference suggests a change in the frequency of the PV inverter currentsduring the fault, since the frequency of the system fault contribution is constant at 60Hz.

Further investigations of the behavior of the vendor-specific PV inverter models during faults revealed that thefrequency of the PV inverter currents is highly dependent on the design of the phased locked loop (PLL) blockof a specific inverter. This block is used to determine phase angle and frequency of the grid voltage (at the lowside of the interconnection transformer) to synchronize the PV inverter operation with the grid and export powerto the grid at a pre-specified power factor (mainly unity power factor). Detailed description and variousapproaches in design of a PLL block can be found in several literature, including Reference 3.

Inverter current (A) CB1 & Fault currents (kA)


For a three-phase fault at the terminal of a PV Inverter, the voltage measurements are too low for a PLL blockto estimate grid frequency properly. Due to the fault transients and an insufficient voltage measurement, thePLL frequency may undergo some frequency transients. Typically, the frequency excursions are limited byinternal maximum/minimum thresholds built into PLL design. In case of a three-phase bolted fault, due to lackof inertia, the frequency of PV inverter currents suddenly jump to the upper threshold subsequent to the fault,which causes the gradual phase angle variations.

Re-examination of the vendor-specific PV inverter models showed that the PLL design varies among thevendors. As shown in Figure 9 and Figure 10, it was determined that Models 3 and 6 have an upper frequencythreshold of about 70 Hz. The source currents and the PV inverter currents during the fault will become in phaseafter 5 cycles from the fault inception, before they move apart again. Model 1 uses a threshold of 61 Hz, or 1Hz maximum deviations from the base frequency. Among all, applying a three-phase fault to Model 1 showedthat the phase angle of the PV inverter current would not change during fault, as shown in Figure 11. In thisModel, the PLL frequency showed short transients, subsequent to the fault. The frequency was immediatelyrestored and maintained afterward at pre-fault value (60 Hz).

For any other fault types with at least one healthy phase (LG and LLG), or a three phase fault at some distancefrom the PV inverter location that results in enough residual voltage, the inverter frequency during the fault willbe maintained at 60 Hz and there will be a fixed phase angle difference of 72 to 90 degrees between thecontributing current components.

0 2 4 6

-5

0

5

curr

ent

Phase A

from Sourcefrom PCC

0 2 4 6-5

0

5

curr

ent

Phase B

0 2 4 6-5

0

5

cycles

curr

ent

Phase C

Figure 9 - Response of the PLL for Model 3


Figure 10 - PLL response for Model 6

0 2 4 6

-5

0

5

curr

ent

Phase A

from Sourcefrom PCC

0 2 4 6-5

0

5

curr

ent

Phase B

0 2 4 6-5

0

5

cycles

curr

ent

Phase C

Figure 11 - PLL response for Model 2

5.3 Effect of PV Inverter PLL design on current contribution


To investigate the effect of PLL design on the PV inverter current contribution, the PLL upper threshold for thegeneric model was modified to 61 Hz and fault studies with PV inverters at location A were repeated. Thestudies with this change in the PLL threshold are referenced as Modified PLL as compared to Original PLL.

The simulation results of the short circuit studies with the modified and original PLL are shown in the followingfigures for various aggregate sizes of PV inverters (up to 16 MW). The peak and rms values of the fault currentsfor the first and the third cycles after the fault are compared in the figures by varying the PV generation capacity(multiple inverters at location A). For all cases, the fault was applied at zero crossing of phase A voltage toensure maximum fault current. Faults were applied at point A, E and I.

0 2 4 6 834.28

34.3

34.32

34.34

34.36

Peak

Cur

rent

(kA

)

1 cycle Fault Current (Peak) Fault A

Original PLLModified PLL

0 2 4 6 834.28

34.3

34.32

34.34

34.361 cycle Fault Current (Peak) Fault E

Peak

Cur

rent

(kA

)

0 2 4 6 890.6

90.7

90.8

90.91 cycle Fault Current (Peak) Fault I

Peak

Cur

rent

(kA

)

PV Capacity (MW)

Figure 12 Change in the first cycle peak of fault current with increase in PV generation capacity


0 2 4 6 819.86

19.88

19.9

19.92

19.94

RM

SC

urre

nt(k

A)

1 cycle Fault Current (RMS) Fault A


0 2 4 6 819.86

19.88

19.9

19.92

19.941 cycle Fault Current (RMS) Fault E

RM

SC

urre

nt(k

A)

0 2 4 6 851.45

51.5

51.55

51.61 cycle Fault Current (RMS) Fault I

RM

SC

urre

nt(k

A)

PV Capacity (MW)

Figure 13 - Change in the first cycle rms of fault current with increase in the PV generation capacity

0 2 4 6 826.8

26.9

27

27.1

27.2

Peak

Cur

rent

(kA

)

3 cycle Fault Current (Peak) Fault A


0 2 4 6 826.8

26.9

27

27.1

27.23 cycle Fault Current (Peak) Fault E

Peak

Cur

rent

(kA

)

0 2 4 6 8

77.35

77.4

77.45

77.53 cycle Fault Current (Peak) Fault I

Peak

Cur

rent

(kA

)

PV Capacity (MW)

Figure 14 - Change in the third cycle peak of fault current with increase in PV generation capacity


0 2 4 6 815.5

15.55

15.6

15.65

15.7R

MS

Cur

rent

(kA

)

3 cycle Fault Current (RMS) Fault A


0 2 4 6 815.5

15.55

15.6

15.65

15.73 cycle Fault Current (RMS) Fault E

RM

SC

urre

nt(k

A)

0 2 4 6 844.95

45

45.05

45.13 cycle Fault Current (RMS) Fault I

RM

SC

urre

nt(k

A)

PV Capacity (MW)

Figure 15 - Change in the third cycle rms of fault current with increase in the PV generation capacity

The conclusions from the above simulations are as follows:

For the Original PLL setting (70 Hz threshold), it can be noted that, up to the third cycle after the fault, thecombined effect of the inverter current and the system fault contribution did not change the fault level orslightly lowered the system fault current, as compared to a no-PV fault case. The primary reason is due to thefact that there was a phase angle difference between the inverter current and the fault current contribution fromthe source - the two current components contributing to a fault. The phase angle difference between twocontributing currents means that the total current is not the arithmetic summation, but the vector summation ofthe components. In other words, the instantaneous value of inverter current may be negative at the instance thatthe fault current from the source has its peak.

For the Modified PLL setting (61 Hz threshold), the total fault currents for the faults on the 27.6 kV bus wereslightly increased. However, the PV inverters contribution to a fault on 115 kV bus did not change the faultlevel or slightly reduced the fault current. This was associated with a larger phase angle difference between the faultcurrent components at 115 kV bus.

In general, there was no observation of significant change in the fault currents with addition and/or increase inPV generation capacity.


6 Impact of PV Inverter Fault Contribution on CB ratingsThe objective of the studies reported in this section is to examine impact of PV inverters integrated intodistribution systems on fault current carrying ratings and interruption capacity of circuit breakers and switches.The study benchmark includes circuit breakers at the beginning of 27.6 kV feeders and reclosers on 8.32 kVfeeders.

IEEE standards C37.04-1999 (Reference 4) and C37.06-2009 (Reference 5) define the fault currentcarrying/interruption capacity of ac high voltage (1000 Vac) circuit breakers. The standards also provide criteriato relate the symmetrical and asymmetrical ratings of a CB at specific operating time such as contact partingtime. The contact parting time is defined by the standard as the sum of circuit breaker operating time and a cycle relay detection time.

6.1 Influence of PV inverter current on Symmetrical Circuit breakerrating

During the study it was found that the PV inverter fault current will be in phase with the voltage measured on itsterminals before the fault occurred. This behavior becomes important as it will introduce a phase shift inrelationship to the system fault current, which has a phase shift of approx. 80-90 degrees with respect to thesystem source voltage. Hence, the total symmetrical fault current amplitude has to be calculated from the vectorsummation of the PV inverter current and the system fault current that will therefore result in smaller valuescompared to an arithmetic summation. The notion of vector summation (phasors) has to be applied at each timeinstance, due to time varying phase angle of the inverter currents.

To determine the phase shift between the two currents, it is useful to relate both to the system source voltage.As stated above, the phase angle relationship between the system fault current and the system source voltage isgiven by the total fault impedance between the source and the fault location. In the standard C37.04-1999 a faultimpedance X/R ratio of 17 is used for the calculation of the standard dc offset. This ratio translates to a phaseshift of 86.6 degrees (arctan(17) =- 86.6 ) between the source voltage and the system fault current. For systemsusing cables, the X/R ratio can be much lower. The lower angle results in a faster decline of the dc offset. Thebase case benchmark system selected for the study had an X/R ratio of approx. 20 (phase angle = 87.1 degree).

Figure 16 - Phase angle between fault current and the system voltage for the benchmark system

The PV inverter current vector will be in phase with the voltage at the PV inverter terminal during the loadcondition. The phase angle to the system source voltage is hereby determined by the load and the impedance


between the source and the PV inverter location. In the benchmark system, the measured phase angle wasapprox. 10 degrees, but for high load situations of the feeders, it can be assumed that the phase angle can be ashigh as 15 degrees.

Based on the above assumptions, it can be concluded that the initial phase angle between the system faultcurrent and the PV inverter current is always greater than 72 degrees as shown in . In a no load situation the PVinverter voltage (and fault current) would be in phase with the system source voltage what would result to themaximum phase angle between this two current.

Figure 17 - Determination of the initial phase angle of inverter current

The vector summation of these two current vectors at each time instance can be calculated with the followingformula:

The total current (I total) should be used to evaluate the circuit breaker symmetrical rating, if the PV invertermaintains the system frequency during a fault.

6.2 Influence of PV inverter current on asymmetrical Circuit breakerrating

In calculation of the asymmetrical rating of a circuit breaker the effect of dc offset in the fault current has to beconsidered. Therefore, the circuit breaker contact parting time is important, as it specifies the dc current value.In the study, it was observed that none of the PV inverter models induced any dc component in the fault current.

The symmetrical rms current rating of a CB is related to asymmetrical rms current rating (It) with a factordefined by the level of dc offset as described by equation (2) of IEEE standard C37.04-1999.

Because the PV inverter current contribution is relatively small (< 10% ) compared to the system fault current,the above formula given in the IEEE standard can still be used to calculate the asymmetrical current. Although,this approach may introduce a small error (


However, the above formula can be modified to reflect the vector summation of the two current components (asintroduced in section 6.1) to achieve high accuracy:

This modification is needed as only the system fault current includes a dc component.

6.3 Influence of PV inverter current on close and latch ratingThe close and latch rating of a circuit break is evaluated based on the first peak of the fault current. The faultcurrent generated by a synchronous generator is well understood and can be described by the following formula:

)tcos()cos(eI2)t(i PTt

f

Where:

= Angle of switching on the voltage waveform,

22

mf

LR

VI

is the r.m.s. value of the primary symmetrical fault current,Tp = L / R is the dc time constant of the primary current,

)RLarctan(

is the phase angle difference between the voltage and the current.

Substituting for ( -) the expression above can be written as:

)tcos()cos(eI2i PTt

f)t(

The second term in this expression is the steady state sinusoidal variation and the first term is the transient partof the fault current, which vanishes theoretically a few cycles.

At t=0, it can be seen that the transient component equals the steady state component and since both haveopposing polarities the net current is zero at t=0.

The transient component will be zero when = 2

and will have maximum value when = 0o. Making an

assumption that the network is predominantly inductive this corresponds to switching taking place on thevoltage waveform, when it is passing through the peak and zero crossing, respectively.

22

100%2

DCIII Isystemtotalasym


Figure 18 - Fault current with the maximum dc offset

As mentioned earlier, the PV inverter current contribution during a fault does not have any dc component and isin phase with the voltage measured at the PV inverter terminal before the fault. The inverter current can bedescribed by the following formula:

)angleloadsin()( nominal_ tIFti Inverterinverter

The factor F applies the fact that the PV inverter current in the first cycle may have a higher transientmagnitude. The maximum value was found to be less than 1.8 pu.

The load angle is the angle between the system voltage and the voltage measured at the PV inverter terminalbefore the fault and is mainly caused by the feeder load.

Figure 19 - PV current during fault with a factor of F=1.8 and angle of 10 degrees


From the simulation results and the previous discussions, it can be concluded that both the PV inverter currentand the system contribution to the fault have their maximum values during the first cycle, but at different timeinstances. Figure 20 shows how the contribution of the PV inverter current will change the location and thevalue of the first cycle peak of the system fault current. For demonstration purposes both currents are assumedat 1 PU.

Figure 20 - Example of total fault current

The peak value of the total fault current including the PV inverter current during the fault can be calculated as:

The first term in the above equation is identical to the vector summation of the two current vectors asintroduced before; however, it is now multiplied by a sqrt(2) factor to get the peak value. The second termexpresses the dc offset at the time of the maximum and is associated with the system fault current.

For ease of calculations, we can assume that the time when the maximum current is reached will be at the samemoment as the first maximum of the total current without a dc offset. This is a simplification as the dc offsetwill move the actual maximum to a slightly earlier moment. The system current is assumed with a maximum dcoffset and is basically a sinusoidal waveform with a 90 degree phase shift. The phase shift of the PV invertercurrent is the same as explained earlier. Based on this assumption, the time for the first cycle peak can becalculated as:

sree

reeII

III

ttotalsystem

Invertertotalsystem

deg21600

deg2

arccos180222

maximum

Figure 21 shows the increase of the peak value of the first cycle fault current in relationship with the level of thePV penetration and different phase angles between the system fault current and the PV inverter current during

100%

)cos(22 )(22 peaktsystemInvertersystemInvertersystempeakDC

IangleinverteranglefaultIIIIi


the fault. It can be concluded that the increase in the peak value due to the PV inverter current contribution isnot critical for the evaluation of the circuit breaker rating as long as the symmetrical current requirements aremet as described earlier.

Figure 21 - Peak value of the total system fault as a function of the PV penetration (in per unit of the 500 MVA short circuitcapacity) for two different phase angles between the contribution currents

6.4 Effect of change in PV inverter frequencyThe previous considerations are correct as far as the PV inverter maintains the system frequency during thefault. However, for situation in which the PV inverter cannot measure proper level of voltages at the terminal todetermine frequency and phase angle (e.g. a three phase bolted fault close to the PV inverter), the PV inverterfrequency can suddenly change to a value as high as 70 Hz. This higher frequency will cause the vector of thePV inverter current to rotate with respect to the system fault current. The speed at which the phase angle shiftsper cycle can be calculated as:

For example, if the inverter frequency changes to 70 Hz, it will result in a 60 degree phase shift per cycle.

]/[60][deg360]/)[(

scyclesreescyclesff systeminverter


Figure 22 Rotation of the PV inverter current vector as a result of higher frequency than the grid frequency

The actual phase angle at the contact parting time of the circuit breaker must be calculated and considered if theasymmetrical rating is a limiting factor. For 3 and 5 cycle circuit breakers with the associated 2 and 3 cyclecontact parting time, the considerations above show that the frequency shift effect can be ignored in calculationof the asymmetrical ratting as it will only improve the requirements. Therefore, it will not be considered as theworst case scenario. However, for breaker failure conditions and backup protection operations, the contactparting time of breakers can occur beyond 5 cycles. For such situations and for symmetrical short circuit level,higher PV inverter frequency means that the arithmetic summation of the two currents should be used, sincethere is the possibility that at the moment of contact parting time both currents become in phase.

For the close and latch rating a higher PV inverter frequency than 60 Hz will always result in a lower value forthe vector summation, as the additional phase shift during the first half cycle will always lower the total current.

7 Summary of variations in the PCC voltages and fault detection timefor different PV inverter locations

This section provides an overall summary of the minimum voltage measurement at PCC and the PV invertertripping time for various fault type and locations. Each table is for a specific PV inverter(s) location, for eithergeneric or model1 PV inverters. For simplicity, the table structure reflects the configuration of the studybenchmark by the order of fault locations with respect to PV location. It should be noted that the reported PVinverter trip time is based on detection time of conventional under/over voltage and frequency protectionschemes and does not include any active anti-islanding scheme.

Based on the simulation results and protection schemes, if there is not enough voltage drop at PCC (more thanspecified 15% level in IEEE 1547), a PV inverter equipped with only conventional protection schemes may notdetect a fault condition.

Table 11 Generic PV Inverter at Location A


SLG 2s 0.55 AB 10.5c 0.00 A NA 0.86 A NA 0.97 A NA 1.02 ABTPH 10c 0.00 ABC 10c 0.00 ABC 2s 0.85 ABC NA 0.94 ABC NA 1.00 ABC


10.5c 0.00 A NA 0.97 AB NA 1.02 AB


Trip

Trip Trip Trip

Trip Trip Trip

115kV 27.6kV 8.32kV

1x PV


Trip

Phase

Phase Phase

Table 12 Model1 PV Inverter at Location A


SLG 0.3c 0.55 AB 0.3c 0.00 A NA 0.86 A NA 0.98 AB NA 1.02 ABTPH 0c 0.00 ABC 0c 0.00 ABC NA 0.86 ABC NA 0.94 ABC NA 1.01 ABC


0.3c 0.00 A NA 0.98 AB NA 1.02 AB


Trip




115kV 27.6kV 8.32kV

1x PV



Table 13 Generic PV Inverter at Location B


SLG 2s 0.55 AB 10.5c 0.00 A 10.5c 0.00 A NA 0.72 A NA 0.98 ABTPH 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 2s 0.54 ABC NA 0.90 ABC

V Vpu Phase Vpu PhasePCC PCC PCC

10.5c 0.00 ABC NA 0.97 AB NA 1.01 AB


Trip Trip Trip Trip

Trip Trip Trip

Trip

Phase

Phase Phase

1x PV


115kV 27.6kV 8.32kV

Table 14 Model1 PV Inverter at Location B


SLG 0.3c 0.55 AB 0.3c 0.00 A 0.3c 0.00 A 1.1c 0.72 A NA 0.98 ABTPH 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC 0c 0.55 ABC NA 0.90 ABC


0.3c 0.00 A NA 0.97 AB NA 1.01 AB


Trip




115kV 27.6kV 8.32kV

1x PV



Table 15 Generic PV Inverter at Location C


SLG 10.5c 0.25 A 2s 0.69 AC 2s 0.72 AC 10.5c 0.00 A NA 0.87 ATPH 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 2s 0.77 ABC


2s 0.79 AC NA 0.96 A NA 1.01 A

10c 0.00 ABC NA 0.94 ABC NA 1.00 ABC

115kV 27.6kV 8.32kV

1x PV



Trip Trip Phase Trip Phase Trip Trip


Table 16 Generic PV Inverter at Location D


SLG 10.5c 0.25 A 2s 0.70 A 2s 0.70 AC 10.5c 0.00 A 10.5c 0.00 ATPH 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC


2s 0.70 A NA 0.95 A NA 1.00 A


Point D

Trip Trip Trip Trip

Trip Trip Trip

115kV 27.6kV 8.32kV

1x PV

Point I Point A Point B Point C

Trip

Phase

Phase Phase

Table 17 Model1 PV Inverter at Location D


SLG 0.3c 0.25 A 2s 0.70 AC 2s 0.70 AC 0.3c 0.00 A 0.3c 0.00 ATPH 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC


2s 0.70 AC NA 0.94 A NA 1.00 A


Trip




115kV 27.6kV 8.32kV

1x PV



Table 18 Two Generic PV Inverters at Location A and two Generic PV Inverters at Location B

Vpu Phase Vpu Vpu Vpu VpuPCC PCC PCC PCC PCC

SLG 2s 0.55 AB 10.5c 0.00 A NA 0.86 A NA 0.97 AB NA 1.02 ABTPH 10c 0.00 ABC 10c 0.00 ABC 2s 0.85 ABC NA 0.94 ABC NA 1.00 ABC

V Vpu VpuPCC PCC PCC

10.5c 0.00 A NA 0.97 AB NA 1.02 AB

10c 0.00 ABC NA 0.94 ABC NA 1.00 ABC


SLG 2s 0.55 AB 10.5c 0.00 A 10.5c 0.00 A 2s 0.75 A NA 0.98 ABTPH 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 2s 0.54 ABC NA 0.90 ABC


10.5c 0.00 A NA 0.97 AB NA 1.01 AB

10c 0.00 ABC NA 0.94 ABC NA 1.00 ABC

2x PV

Trip Trip Trip Trip

Trip Trip Trip

2x PV



Point B Point C Point D

Trip Trip Trip TripTrip

Phase

Phase Phase

Phase Phase

PhasePhase

Trip Trip Trip

115kV 27.6kV 8.32kV

2x PV

Point I Point A

2x PV

Trip

Phase

Phase

Point E

Phase

Phase Phase

PhasePhase

Point G Point H


Table 19 Two Model1 PV Inverters at Location A and two Generic PV Inverters at Location B


SLG 0.3c 0.55 AB 0.4c 0.00 A 10c 0.85 A NA 0.97 AB NA 1.02 ABTPH 0c 0.00 ABC 0c 0.00 ABC 10c 0.85 ABC NA 0.94 ABC NA 1.00 ABC


0.4c 0.00 A NA 0.97 AB NA 1.02 AB



SLG 0.3c 0.55 AB 0.3c 0.00 A 0.3c 0.00 A 1.1c 0.70 A NA 0.98 ABTPH 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC 0.3c 0.55 ABC NA 0.90 ABC


0.3c 0.00 A NA 0.97 AB NA 1.01 AB


Phase

Trip Phase Trip Phase Trip Phase





2x PV 2x PV


Phase Trip


115kV 27.6kV 8.32kV

2x PV 2x PV


Phase Trip Phase


Table 20 Four Generic PV Inverters at Location B


SLG 2s 0.55 AB 10.5c 0.00 A 10.5c 0.00 A 2s 0.72 A NA 0.98 ABTPH 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 2s 0.54 ABC NA 0.90 ABC


10.5c 0.00 A NA 0.97 A NA 1.02 AB


Trip Trip Trip Trip

Trip Trip Trip


115kV 27.6kV 8.32kV

Trip

Phase

Phase

4x PV

Phase

Phase Phase

PhasePhase

Table 21 Four Model1 PV Inverters at Location B


SLG 0.3c 0.55 AB 0.3c 0.00 A 0.3c 0.00 A 1.2c 0.70 A NA 0.98 ABTPH 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC 0.3c 0.55 ABC NA 0.90 ABC


0.3c 0.00 A NA 0.97 AB NA 1.02 AB


Phase Trip Phase



4x PV


115kV 27.6kV 8.32kV


Table 22 Two Generic PV Inverter at Location D


SLG 10.5c 0.25 A 2s 0.70 AC 2s 0.70 AC 10.5c 0.00 A 10.5c 0.00 ATPH 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC 10c 0.00 ABC


2s 0.70 AC NA 0.95 A NA 1.00 A


Point C Point D

Trip Trip Trip Trip

115kV 27.6kV 8.32kV

2x PV

Point I Point A Point B

Trip

Phase

Phase

Trip

Phase

Phase Phase

PhasePhase

Trip Trip

Table 23 Two Model1 PV Inverter at Location D


SLG 0.3c 0.25 A 10c 0.70 AC 10c 0.70 AB 0.3c 0.00 A NA??? 0.00 ATPH 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC 0c 0.00 ABC


10c 0.70 AC NA 0.94 A NA 1.00 A


Phase Trip Phase



2x PV


115kV 27.6kV 8.32kV

8 Summary of Fault Current differencesThe following tables provide summary of the maximum (first cycle peak) fault current differences as seen bythe circuit breakers and reclosers of the benchmark system for two cases: with and without PV Inverters. Theresults are provided for the original loading of the feeder that considers active and reactive power of the load.

Table 24 - Fault current differences of a Generic PV Inverter at A with respect to a No PV Inverter case

CB1 CB2 Rec1A 0.014 0.000 0.000 0.070 A 0.000 0.000B 0.014 0.000 0.000 0.015 0.000 0.000C 0.015 0.000 0.006 0.015 0.000 0.007 AD 0.015 0.000 0.000 0.015 0.000 0.002 AE 0.020 0.030 0.000 0.018 0.080 A 0.000G 0.015 0.000 0.000 0.015 0.001 AC 0.000H 0.015 0.001 0.000 0.015 0.000 0.000I 0.020 0.005 0.015 0.020 B 0.000 0.000

Maximum fault current difference (kA)

Rec1

A1

TPH SLGCB1 CB2

PVLocation

Fault

Table 25 - Fault current differences of a Model1 PV Inverter at A with respect to a No PV Inverter case

PV Fault CB1 CB2 Rec1A 0.025 0.007 0.008 0.040 A 0.001 0.000B 0.016 0.000 0.000 0.018 0.000 0.000C 0.015 0.000 0.025 0.015 0.000 0.020 AD 0.015 0.000 0.006 0.015 0.000 0.003 AE 0.030 0.030 0.008 0.020 0.060 A 0.000G 0.015 0.008 0.000 0.015 0.004 AB 0.000H 0.015 0.002 0.000 0.015 0.000 0.000I 0.020 0.000 0.000 0.020 0.000 0.000

current difference (kA)TPH SLG

CB1 CB2 Rec1

A1


Table 26 - Fault current differences of a Generic PV Inverter at B with respect to a No PV Inverter case

CB1 CB2 Rec1A 0.050 0.000 0.006 0.070 A 0.000 0.000B 0.012 0.000 0.000 0.018 B 0.000 0.000C 0.010 0.000 0.040 0.016 C 0.000 0.038 AD 0.015 0.000 0.008 0.015 0.000 0.005 AE 0.020 0.050 0.007 0.018 0.070 A 0.000G 0.015 0.002 0.000 0.015 0.001 AB 0.000H 0.015 0.000 0.000 0.015 0.000 0.000I 0.020 0.005 0.007 0.020 B 0.000 0.000


B1

TPH SLGCB1 CB2 Rec1

PVLocation

Fault

Table 27 - Fault current differences of a Model1 PV Inverter at B with respect to a No PV Inverter case

PV Fault CB1 CB2 Rec1A 0.050 0.008 0.010 0.050 A 0.001 0.002B 0.015 0.000 0.000 0.015 0.000 0.001C 0.016 0.000 0.040 0.020 0.000 0.050 AD 0.015 0.000 0.015 0.015 0.000 0.005 AE 0.020 0.050 0.100 0.020 0.050 A 0.002G 0.015 0.008 0.000 0.015 0.004 AB 0.000H 0.015 0.002 0.000 0.015 0.000 0.000I 0.020 0.002 0.002 0.020 0.000 0.002


CB1 CB2 Rec1

B1


Table 28 - Fault current differences of a Generic PV Inverter at D with respect to a No PV Inverter case

CB1 CB2 Rec1A 0.050 0.000 0.060 0.090 A 0.000 0.060B 0.012 0.000 0.060 0.018 BC 0.000 0.060C 0.013 0.000 0.045 0.015 C 0.000 0.060 BD 0.013 0.000 0.045 0.015 C 0.000 0.060 BE 0.020 0.050 0.060 0.018 0.110 A 0.060G 0.015 0.002 0.050 0.015 0.001 AB 0.050H 0.015 0.000 0.050 0.015 0.000 0.050I 0.020 0.005 0.060 0.020 B 0.000 0.060


CB1 CB2 Rec1

D1

TPH SLGPVLocation

Fault

Table 29 - Fault current differences of a Model1 PV Inverter at D with respect to a No PV Inverter case

PV Fault CB1 CB2 Rec1A 0.004 0.006 0.080 0.040 A 0.000 0.070B 0.015 0.000 0.080 0.020 0.000 0.070C 0.015 0.000 0.050 0.015 0.000 0.050D 0.015 0.000 0.005 0.015 0.000 0.050E 0.020 0.050 0.080 0.020 0.050 A 0.060G 0.015 0.008 0.005 0.015 0.004 AB 0.050H 0.015 0.002 0.050 0.015 0.000 0.050I 0.020 0.000 0.080 0.020 0.000 0.070


CB1 CB2 Rec1

D1

Table 30 - Fault current differences of two Generic PV Inverters at A and two at B with respect to a No PV Inverter case

CB1 CB2 Rec1A 0.080 0.000 0.004 0.100 A 0.000 0.003B 0.030 0.000 0.000 0.060 BC 0.000 0.003C 0.050 0.000 0.090 0.060 0.000 0.080 AD 0.060 0.000 0.020 0.060 0.000 0.010 AE 0.080 0.140 0.004 0.070 0.100 A 0.003G 0.060 0.007 0.000 0.060 0.004 AB 0.000H 0.060 0.002 0.000 0.060 0.000 0.000I 0.070 0.002 0.003 0.080 BC 0.000 0.000

A2B2

TPH SLGMaximum fault current difference (kA)

CB1 CB2 Rec1PV

LocationFault


Table 31 - Fault current differences of two Model1 PV Inverters at A and two at B with respect to a No PV Inverter case

PV Fault CB1 CB2 Rec1A 0.070 0.004 0.012 0.060 0.000 0.000B 0.050 0.000 0.000 0.050 0.000 0.000C 0.060 0.000 0.080 0.060 0.000 0.120 AD 0.050 0.000 0.025 0.060 0.000 0.120 AE 0.100 0.070 0.012 0.080 0.120 A 0.002G 0.060 0.015 0.000 0.060 0.006 AB 0.000H 0.060 0.003 0.000 0.060 0.001 AB 0.000I 0.090 0.006 0.010 0.080 0.000 0.002


CB1 CB2 Rec1

A2B2

Table 32 - Fault current differences of four Generic PV Inverters at B with respect to a No PV Inverter case

CB1 CB2 Rec1A 0.070 0.000 0.008 0.100 A 0.000 0.005B 0.050 0.000 0.000 0.070 BC 0.000 0.009C 0.060 0.000 0.150 0.040 0.000 0.140 AD 0.060 0.000 0.035 0.060 0.000 0.018 AE 0.070 0.130 0.008 0.070 0.090 A 0.009G 0.060 0.007 0.000 0.060 0.004 AB 0.000H 0.060 0.002 0.000 0.060 0.001 AB 0.001I 0.070 0.002 0.007 0.080 B 0.000 0.001

B4

TPH SLGMaximum fault current difference (kA)

CB1 CB2 Rec1PV

LocationFault

Table 33 - Fault current differences of four Model1 PV Inverters at B with respect to a No PV Inverter case

PV Fault CB1 CB2 Rec1A 0.070 0.004 0.012 0.050 0.000 0.005B 0.050 0.000 0.001 0.050 0.000 0.003C 0.060 0.000 0.120 0.070 0.000 0.200 AD 0.050 0.000 0.040 0.050 0.000 0.200 AE 0.100 0.070 0.012 0.080 0.120 A 0.004G 0.070 0.013 0.000 0.050 0.006 AB 0.000H 0.050 0.003 0.000 0.050 0.000 0.000I 0.080 0.001 0.008 0.080 0.000 0.005


CB1 CB2 Rec1

B4


Table 34 - Fault current differences of two Generic PV Inverters at D with respect to a No PV Inverter case

CB1 CB2 Rec1A 0.060 0.000 0.120 0.060 A 0.000 0.110B 0.020 0.000 0.120 0.035 B 0.000 0.120C 0.030 0.000 0.080 0.030 C 0.000 0.110 BD 0.025 0.000 0.090 0.030 C 0.000 0.110 BE 0.040 0.080 0.120 0.035 0.070 A 0.120G 0.030 0.004 0.100 0.030 0.002 AB 0.100H 0.030 0.001 0.100 0.030 0.000 0.100I 0.040 0.004 0.120 0.040 B 0.000 0.120

CB1 CB2 Rec1TPH SLG


Fault

D2

PVLocation

It should be noted that the load displacement by PV inverters are considered in the above currents. For instance,2 PV inverters at Location D, displace about 1 MW of load (about 70A). After fault, the recloser current maychange up to twice load current (120 A).

Table 35 - Fault current differences of two Model1 PV Inverters at D with respect to a No PV Inverter case

PV Fault CB1 CB2 Rec1A 0.050 0.007 0.120 0.040 A 0.000 0.120B 0.030 0.000 0.120 0.040 BC 0.000 0.120C 0.030 0.000 0.080 0

hstory of fault contribution solar inverter

Documents