Derecho Outage Survey







Mark B. Lively

Consulting Engineer, Utility Economic Engineers (Gaithersburg, Maryland)

MbeLively@aol.com

 

The Washington, DC, area was hit by a derecho on Friday, 2012 June 29.  The derecho had strong wind and many lightning strikes, which felled many trees, which in turn knocked out power to perhaps a million customers in the Washington, DC, area.  The local utilities hired crews from hundreds of miles away to assist with the restoration efforts.

The triage process for deploying the various crews working on restoring service gives primacy to health and safety facilities, such as hospitals and fire stations.  Then crews are assigned based on doing the most good, fastest, essentially seeking to achieve the most “bang for the buck.”  For instance, a single tree branch on a distribution line, taking a half hour to fix, might be preventing service to a thousand customers in a densely packed neighborhood.  Removing the tree branch restores many customers to service quickly.  Politically, this quick action should reduce the complaints against the utility.  Economically, this quick action resumes the flow of revenue to the utility earlier.  Thus, there are many reasons to select the “big bang” items for quick action.  But sometimes it is hard to determine which items provide the biggest bang, as will be discussed later.

This report is based on a survey of people who participate in the activities of the National Capital Area Chapter of the US Association for Energy Economics (NCAC-USAEE or just NCAC).  Some 1200 people were surveyed and 42 people responded, a response rate of about 3%.  For those who had already paid NCAC dues for 2012, the response rate was about 9%.  For other people in the NCAC data base, the response rate was about 1%.  The survey sought the number of hours the participants were out of service, the name of the electric utility, and the political jurisdiction in which the participant resided.  The following table presents the data collected in the survey.

 

 

 

The most respondents were for PEPCo in its Montgomery County, Maryland, service area, the wealthier of the two Maryland counties adjacent to Washington, DC.  That there were no respondents from Prince George’s County, the other Maryland county adjacent to Washington, DC, suggests that there may have been bias in the survey, such as where the NCAC participants lived relative to the general population of the Washington, DC, area.  PEPCo also serves Washington, DC, from where only 6 people responded, 4 of whom said they did not have an outage or the outage was insignificant.  The few number of responses from Washington, D.C., also suggests a survey bias.

The graphs in this paper are cumulative distribution functions.  The horizontal axis is the reported length of the outages.  The vertical axis is the fraction of customers who reported an outage that lasted for a time that was less than or equal to the indicated length of the outage.  Figure 1 is for PEPCo, including separate lines for Montgomery County, for DC, and for a composite of the two jurisdictions.  Figure 2 presents the cumulative distribution functions for BG&E, for VEPCo, and for a composite of the two utilities.  Figure 3 presents the cumulative distribution functions for BG&E, for VEPCo, for a composite of PEPCo, and for a composite of all three utilities.  Figure 4 provides a head to head comparison between PEPCo and VEPCo the two utilities with the largest presences in the DC area.

 

 

 

 

 

 

 

 

The cumulative distribution function for PEPCo-Montgomery County in Figure 1 provides an indication of the results of the triage process, as does the cumulative distribution function for VEPCo in Figure 2.  The line slopes up rapidly during the first day, suggesting that PEPCo crews had identified optimal tasks for restoring many customers quickly, the one branch that had been affecting hundreds of customers.  Other surges occur during the second and third day.  These secondary surges are other forms of the “big bang” theory.

PEPCo had crews that came in from out of state, driving hundreds of miles to arrive in the DC area, perhaps arriving late on the first day or sometime during the second or third day.  The surges on the second and third days thus represent additional waves of optimized restoration of service.  After the third day, the restoration process slowed.  The crews were no longer restoring hundreds of customers at a time, but were restoring only ten customers at a time, or perhaps only one.  The cumulative distribution function grows less rapidly.  The day two surge is especially noticeable for VEPCo.

I have previously mentioned four biases that could skew the curves shown in Figures 1-3.

  • The most obvious bias is the sample selection.  NCAC dues-paying members had a response rate of 9% versus the response rate of 1% for non-members.  NCAC dues paying members could have a different geographic dispersion than do non-members, and both could be different from the geographic dispersion of the population in the DC area, as was mentioned previously in regard to the absence of any responses from Prince George’s County.
  • There were relatively few responses from DC itself, and none from Prince George’s County.  The housing density, tree density, etc., could be significantly different for these jurisdictions than for the area as a whole.
  • I postulated “the most bang for the buck” based on the number of customers.  The survey responses were for residential customers.  Utilities earn more money by restoring commercial customers and/or very large residential customers.  The triage process would be most economical by restoring first those customers who are the largest consumers of electricity and pay the most for electricity.
  • There can be a non-egalitarian political aspect to the restoration process.  There is the thought that the more vocal consumers are richer and live in more remote areas, areas that are restored one customer at a time.  Egalitarian and economic concepts suggest these odds and sods would be restored last.  Minimizing political fallout suggests that such customers would be restored earlier instead of later.

Thus, though the curves are drawn based on customer numbers, the issues of customer size and political impacts can also be important.

India has a severe shortage of electric power, with some areas normally receiving electricity for only a short period each day.  Many stories suggest 8 hours of electric service is typical in some locations.  The blackouts that occurred in India in July 2012 produced stories of India’s utilities responding to political pressure as to which neighborhoods to serve when there is a shortage.  Thus, one neighborhood might receive service for the cited 8 hours a day while another neighborhood, one that is more politically connected, might receive electric service for 24 hours a day, on most days, with the blackout being the great equalizer.   That concept could translate into the US as to which neighborhoods should be restored first politically versus on an egalitarian basis.

The Golden Triangle is a 43-block neighborhood in Washington, D.C. that is served by a 501(c)6 non-profit organization.  The Golden Triangle Business Improvement District (BID) has a mission that includes “street improvements” and “keeping the streets clean.”  These cited services, which are nominally provided by a city government, show one way that the triage for power restoration could be accomplished on a non-egalitarian, economic basis.  Private citizens could hire crews to restore power to their homes ahead of the normal triage schedule.  To an extent, the fact that we have different utility service areas in one locality accomplishes this concept.

Click to view a printable version of this article.