Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: Planning: P1547 series and P2030.2 year 2014 dates (most tentative)



WG members, especially utility members:   I am soliciting feedback on how other utilities are handling, or will handle, the following high penetration voltage issues:

 

·         DG intermittency:

o   PV cloud cover issue

§  PV output may drop precipitously within seconds due to fast moving clouds.

·         We have been applying flicker criteria to this condition.  In other words, we study for maximum voltage fluctuation for full on and full off conditions.  If the primary voltage change by more than 3 volts, on 120V base, for urban feeders and 5 volts for rural feeders, we require mitigation.  This appeared to work pretty well and kept the voltage within Rule 2 limits after interconnection.  But many proposed installations dropped out due to the high cost of reconductoring.  Hence, there are concerns that this may be overly conservative .  So, what is your utility practice on this issue? 

§  When you do studies, do you look at the cumulative DG on the feeder or just the unit under study?  The concern is that a cloud cover may cover the entire feeder area and cause all of the PV on the feeder to drop in output.  Currently, PG&E just study the single unit and assumed the rest stay constant.  This is a carry-over practice of when we study individual DGs that did not have common response modes.

·         Feeder reclose after a momentary fault. 

o   All DGs are required to trip during a feeder fault.  However, distribution feeders are typically designed to reclose to recover from a momentary fault.  But, during reclose, the DGs will be off line and the load that was supplied by the DGs before the fault will be seen by the feeder breaker relay and the LTC/voltage regulators. 

§  The feeder breaker relay should be set for the full load level, without DG,  to avoid potential false trip under this condition.  This should not be a problem if the relay setting can accommodate the full load current without tripping.

§  Our electromechanical step voltage regulators are currently set for about 15-30 second time coordination delay between each regulating stages to avoid hunting.  On the PG&E system, we have some pretty long skinny feeders with up to 5 stages of voltage regulation in series to provide proper steady state voltage to all feeder customers.  So, the total uncorrected voltage drop from the substation to the end of the line may be over 20% on some feeders.  Our distribution line conductors are sized to carry steady state radial load current only and are subject to significant voltage drop that are currently corrected by radial voltage regulators.  Most of the distribution conductors are small conductors that are subjected to significant voltage drops.  This is not an issue when there is no DG since the regulators are set to correct for the voltage drop due to the load current already.  But the DGs can mask the load before the trip, and the regulator is not providing the full regulation needed for the full load due to the presence of the DGs.  After the feeder and DG triped, the feeder load without DG may be much higher and with much higher voltage drop.  So, the feeder customers may be subjected to undervoltage beyond Rule 2 limits until the regulators are able to respond and correct for the new load level. 

·         Adjacent feeder fault: 

o   All DGs feed off the same substation bus will sense the undervoltage due to a close in fault.  Due to the current tight IEEE-1547 UV settings and even the proposed IEEE-1547a settings,  DGs may trip for a close-in adjacent feeder fault.  This may be a worse event than the above feeder fault condition since this will happen more frequently and may dump all of the DG on the bus if the feeder relay was set for time coordinated tripping that will typically take longer than 0.16 sec.

§  Under this condition, how do you deal with the undervoltage condition before the voltage regulators have a chance to respond?  CA Rule 2 does give us some flexibility for feeder fault conditions.  But this is an out-of-section adjacent feeder fault condition that normally the units should be able to ride through.

§  How long would you tolerate an undervoltage event down to 90%?, 85%?, 80%?

§  In your opinion, what is the maximum allowable undervoltage and duration to avoid damage to customer equipment?

§  How long would you tolerate 110%, 120% voltage?

§  What is the maximum allowable overvoltage and duration to avoid damage to customer equipment?

·         Potential mitigation measure that we may want to consider:

o   Speed up the re-energization time, and possibly keeping the DG controls hot, but with no export, during these momentary events, to minimize the DG restoration time and associated undervoltage duration.

 

Also, I do want to point out that the distribution system is regulated to +/-5% of nominal voltage normally by radial voltage regulators and capacitors but the radial distribution regulators do not work well in a loop configuration with multiple sources and hence the step voltage regulators are not used on the transmission system.  The transmission conductors are typically sized for both local load and loop flow.  So, they tend to be uniform and much bigger conductors.  The transmission voltage where we typically connect generators are also allowed much wider voltage fluctuation since the voltage is regulated on the distribution system where the end use load equipment are located.    But now we are putting DG power sources next to load customers and load equipment and potentially subjecting the customer load equipment to abnormal voltages.

 

Hope to see you all at the IEEE-1547 full revision meeting in Las Vegas in a couple of weeks.

 

 

Chase Sun
Principal Engineer - DG
PG&E-Distribution Planning

Rm 939B, 245 Market St,

SF, CA 94105
415-973-5665 (Work)
510-295-7236 (Cell)

 

 



PG&E is committed to protecting our customers' privacy.
To learn more, please visit http://www.pge.com/about/company/privacy/customer/