Demand Response! Power FROM the People
by Dave Loucks
John J. White
March 1, 2010
Summary:
“We won! People were undocking computers,
turning off lights, unplugging coffee pots and tuning in to see how they were
making a difference!” said Brandon Ekberg, business unit manager for Eaton’s
Metering and Software business. However, at the very end, and two floors below,
the folks in the machine shop were watching, too, and they had a few ideas of
their own to ultimately claim victory in Eaton’s Demand Response (DR) Event.
With technology leading the way in the form of building
automation and lighting controls, it was the human element that made this DR
Event unique. And Ekberg’s Power Xpert Metering led the way — allowing the more
than 700 employees to take part as they could see in real-time: (1) how their
section/floor was contributing to the power reduction (2) how they compared to
other sections/floors of the building and (3) how the load reduction was
surpassing a 200kW goal. John White, Eaton’s energy management and
environmental solutions manager shared, “It’s amazing what our own employees
were able to do by having a dashboard with key energy information…this was truly
power from the people.”
At its peak, Eaton reduced power demand by more than 900kW
or more than 50 percent (from over 1,500 down below 700 kW). In the second year
of participating in the PJM (Pennsylvania, New Jersey and Maryland)
Demand Response program with Comverge Corporation, Eaton’s commitment to the
grid stood at 200kW, so this event was a huge success. Eaton facilities
manager, Todd Pfeifer, was determined to provide the 200kW through true-load
curtailment rather than rely on emergency power generation. Pfeifer has assumed
responsibility for Eaton’s recently completed LEED Gold building expansion and
has worked to provide a completely compostable cafeteria, so it was no surprise
that Todd wanted to participate in Demand Response in a “green” way.
Key Demand Response
Elements
According to Dave Loucks, Eaton’s solutions manager,
who specified the functionality of the Demand Response dashboard for Eaton’s
Power Xpert Metering and Software, “The data we got from Power Xpert also
allowed for extensive diagnostics to better understand key contributors to the
power curtailment.” With the event taking place on August 18, outdoor air
temperatures were rising to over 82 degrees Fahrenheit (F) when by 1 p.m. In an
effort to balance comfort with power curtailment, Eaton relied on Trane Company
to maintain space temperatures below 76 F (up from 72 F) and reduce fan speeds
utilizing Eaton’s Variable Frequency Drives. Starting at 12:45 p.m.,
temperature setpoints were elevated, duct static pressure was lowered (slowing
down fan speeds) and Eaton’s Pow-R-Command lighting control reduced ambient
lighting levels down to 35 percent (from a setpoint of 65 percent of maximum
output) and/or completely shut down lighting in some areas. Personnel from
facilities also identified areas that did not have centralized lighting control
and powered down areas such as Eaton’s Executive Briefing
Center.
After about 40 minutes, the air conditioning (HVAC) system
began to draw more power just to maintain a 76 F temperature setpoint and
eroded the initial 900kW peak power reduction. At the end of the hour, power
demand was still down by more 572kW with a large part of on-going and
sustainable power curtailment coming from lighting reductions (159 kW),
employee actions with plug loads (317 kW) and the remaining HVAC load (249 kW).
|
Load
|
Prior to DR (Baseline)
|
Start of DR
|
End of DR
|
|
|
Power
|
kW Saved
|
Percent Saved
|
Power
|
kW Saved
|
Percent Saved
|
Power
|
kW Saved
|
Percent Saved
|
|
Total Load
|
1542 kW
|
0
|
0
|
668kW (_43__ percent)
|
873 kW
|
57 %
|
970 kW (_63 percent_
percent)
|
573 kW
|
37 %
|
|
HVAC (South & West)
|
590 kW
|
0
|
0
|
53 kW
|
537 kW
|
91 %
|
356 kW
|
234 kW
|
40 %
|
|
HVAC (North & East)
|
254 kW
|
0
|
0
|
49 kW
|
205 kW
|
81 %
|
130 kW
|
124 kW
|
49 %
|
|
Total Lighting
|
250 kW
|
0
|
0
|
108 kW
|
141 kW
|
57 %
|
91 kW
|
159 kW
|
64 %
|
|
Employee Actions (Plug Load)
|
396 kW
|
0
|
0
|
296 kW
|
101 kW
|
25 %
|
241 kW
|
154 kW
|
39 %
|
|
Total Curtailed
|
|
0
|
0
|
|
873 kW
|
57 %
|
|
573 kW
|
37 %
|
|
|
|
|
|
|
|
|
|
|
|
After months of planning to curtail load, there were several
keys elements that made this event different. Most customers respond to a DR
event by either running their emergency power generation or simply raising the
space temperature setpoint (HVAC). Although Eaton did see some short-term power
reduction from the HVAC system, it appears empowering people to participate,
coupled with central lighting control offer the most sustainable curtailment
potential. Keys to our success in these areas were:
· Creation of an email campaign led by Mark
Schnirel, marketing specialist. Schnirel prepared building occupants with an
email the day before the event, a follow-up email 15 minutes before the event
and a post-event email thanking participants and sharing results.
· Providing a real-time dashboard with
sub-metering systems that allowed our employees to see how their actions were
contributing to the power curtailment success of their wing/floor, which in
turn established competition between wings (some areas had reduced 80 percent
of load).
· Centralized lighting control was a must — with
the click of a mouse, lighting in entire wings was driven down. Unlike daylight
harvesting at an individual fixture level, this method maintained a consistent
setting and then allowed for prompt return to “normal” once event was complete. Ultimately, the Eaton
DR event leveraged both technology personal
commitment from employees. As a precursor to the Smart Grid, Demand Response
events offer a hint of what is truly possible if you use the ”power from
your people.”
Supplemental
Materials:
Note that instrumentation failure prevented the capture of
actual watts load for HVAC roof top units 7 and 8. To compensate we examined
the ratio of amperes (which we did capture for all RTUs) with watts for RTUs 9
and 10. We then took that ratio (approximately 0.65 kW per ampere) and
multiplied it times the average of the three phase currents for RTU 7 and 8 and
estimated their instantaneous power (kW).
Lessons Learned
1. The Foreseer system allows the user to press a button
and select when it should capture a baseline energy reading. Savings are
computed from that point. Since the DR event was publicized in email and
through an employee meeting (that included about ½ of the building’s
employees), and since these announcements occurred prior to the beginning of
the DR event, it is likely that some demand reduction occurred prior to the
Noon time frame of these graphs. For example, when zooming in at the total
building load, it appears that it is already on a downward path at Noon.
However, prior to this time, the outside air temperature was continuing to
climb. The challenge was to select a “starting point” for Foreseer to declare
this as the highest energy consumption point while employees were turning off
loads but while the HVAC system was increasing its cooling load. The compromise
was reached by activating the system at approximately 12:15 p.m. This resulted
in a maximum power of 1451 kW, or 91 kW lower than the values shown on the
Excel graphs. The reason is that the Excel graphs used peak power not when the
Foreseer system was switched into DR mode, but from visually examining the peak
value prior to the DR event time window. As mentioned earlier, these graphs
were made with an export of the Foreseer data over the time range of 12 noon to
2:30 PM. This gave a viewing window of 1 hour prior to the start of the DR
event, and 15 minutes further back in time than how the Cherrington Foreseer system
calculated savings (since it began with a later starting point 91 kW lower).
2. Outside temperature was recorded as ranging from 90.5
to 81 degrees over the interval monitored, with a time-weighted average outside
temperature of 85 degrees F. Using 68 degrees F as the cooling-to-heating base
temperature, we averaged 17 degrees above this value. Using the 735 kW peak
just prior to the event, that works out to 43.2 kW per degree above our base
temperature.
Once we raised the temperature of the building and reduced the airflow, the
HVAC load stabilized at a value of 487 kW, or a 298 kW reduction. Dividing this
savings by the number of degrees the outside temperature exceed our base
temperature, that works out to 28.6 kW per degree (298/17 = 28.6) above base
temperature which works out to a reduction of 14.6 kW per degree (or a
reduction of 33.7 percent).
Extrapolating this 33.7 percent savings over different sized buildings, we can
estimate the likely savings per area per degree. For example, normalizing these
results to our 280,000 square feet (160,000 existing plus 120,000 new)
Cherrington HQ building, our savings 298 kW in a 280,000 ft2
building operating at 17 degrees above our base temperatures works out to be
0.0521 Watts / degree / square foot.
Normalizing based on temperature factors in the effect of temperature on the
day of the DR event. Normalizing based on building size allows this data to be
extrapolated to buildings other sizes.
3. Data center loading was interesting. On first
examination, the data center load appeared to remain constant.
Average data center load dropped from 157.9 kW
prior to the DR event to 156.7 kW during the DR event, a drop of only 0.8
percent. Also, the average rate of change of power consumption remained
relatively flat prior to DR event. This is consistent with our assumption that
data center loading remain relatively constant before, during and after a DR
event. And as we assumed, power consumed prior to the DR event remained
relatively constant.
The trend line drawn through the middle of the
chart is a linear extrapolation of the trend over time. It is essentially flat.
However, when we look at the trend of the data center loading during the
interval, a surprising new trend line appears.
While the combined average power over the DR
interval was 156.7 kW, and this was a modest drop from the average just prior
to the interval, we notice that the trend of this consumption is decreasing
over the interval. Using a linear curve fit, the trend line of kW loading is
found to be decreasing at a rate of 6.9 kW per hour, or dropping from an
initial average of 160.2 kW down to 153.3 kW, or a 4.3 percent reduction over
the course of the interval.
At the conclusion of the interval, energy consumption again increased, this
time at a rate of 5.1 kW per hour.
1. The question we have is, “What would contribute to this decline during the DR event and subsequent increase of power consumption in the data center at the conclusion of the DR event?” Apparently the servers (and associated computer room air conditioners) used less and less energy over the course of the demand response interval. Was less work done by the building causing less load on the servers? 2. Similar to our work in item 1, we next normalized this data into energy savings per unit space. This data forms a baseline for what we would expect other commercial buildings of similar construction and function to be able to provide with respect to demand response savings. Using these baseline values, customers could estimate what like reductions would be possible at their facilities and therefore how much money would be available from the demand response provider. Using this funding, a return on investment calculation could be made that looks at the instrumentation, audits or other costs necessary to insure a successful DR event, balanced with the estimated funding available from the savings.
| | Prior to Event | | | | | | | | | | | | | | | | reduction | | | | Total Demand | kW | W/ft2 | percent of FL | percent saved | kW | | HVAC | 734.6 | 2.6237 | 100.00 percent | 0.00 percent | 0.0 | | Lighting | 249.6 | 0.8914 | 100.00 percent | 0.00 percent | 0.0 | | Data Center | 170.0 | 0.6071 | 100.00 percent | 0.00 percent | 0.0 | | Plug/Unknown | 396.2 | 1.4151 | 100.00 percent | 0.00 percent | 0.0 | | Total | 1542.3 | 5.5082 | 100.00 percent | 0.00 percent | 0.0 | | | | | | | | | | | | | | | | | | | reduction | | | | Existing Building (S & W) | kW | per sq. ft | percent of FL | percent saved | kW | | HVAC | 590.0 | 2.1071 | 100.00 percent | 0.00 percent | 0.00 | | Lighting | 182.3 | 0.6512 | 100.00 percent | 0.00 percent | 0.00 | | Data Center | 170.0 | 0.6071 | 100.00 percent | 0.00 percent | 0.00 | | Plug/Unknown | 146.8 | 0.5242 | 100.00 percent | 0.00 percent | 0.00 | | Total | 1036.4 | 3.7016 | 100.00 percent | 0.00 percent | 0.00 | | | | | | | | | | | | reduction | | | | New Building (N & E) | kW | per sq. ft | percent of FL | percent saved | kW | | HVAC | 253.83 | 0.9065 | 100.00 percent | 0.00 percent | 0.00 | | Lighting | 67.31 | 0.2404 | 100.00 percent | 0.00 percent | 0.00 | | | | | | | | | Plug/Unknown | 258.92 | 0.9247 | 100.00 percent | 0.00 percent | 0.00 | | Total | 579.4 | 2.0693 | 100.00 percent | 0.00 percent | 0.00 |
| | Start of Event | | | | | | | | | | | | | | | reduction | | | | Total Demand | kW | W/ft2 | percent of FL | percent saved | kW | | HVAC | 101.3 | 0.3617 | 13.78 percent | 86.22 percent | 633.4 | | Lighting | 108.3 | 0.3868 | 43.39 percent | 56.61 percent | 141.3 | | Data Center | 163.6 | 0.5843 | 96.24 percent | 3.76 percent | 6.4 | | Plug/Unknown | 295.7 | 1.0561 | 74.63 percent | 25.37 percent | 100.5 | | Total | 668.9 | 2.3888 | 43.37 percent | 56.63 percent | 873.4 | | | | | | | | | | | | | | | | | | | reduction | | | | Existing Building (S & W) | kW | per sq. ft | percent of FL | percent saved | kW | | HVAC | 52.6 | 0.1879 | 8.92 percent | 91.08 percent | 537.40 | | Lighting | 85.2 | 0.3043 | 46.73 percent | 53.27 percent | 97.13 | | Data Center | 163.6 | 0.5843 | 96.24 percent | 3.76 percent | 6.40 | | Plug/Unknown | 129.6 | 0.4628 | 88.27 percent | 11.73 percent | 17.22 | | Total | 431.0 | 1.5392 | 41.58 percent | 58.42 percent | 605.47 | | | | | | | | | | | | | | | | | | | | | | | | | | reduction | | | | New Building (N & E) | kW | per sq. ft | percent of FL | percent saved | kW | | HVAC | 48.66 | 0.1738 | 19.17 percent | 80.83 percent | 205.16 | | Lighting | 23.10 | 0.0825 | 34.31 percent | 65.69 percent | 44.22 | | | | | | | | | Plug/Unknown | 166.14 | 0.5933 | 64.17 percent | 35.83 percent | 92.78 | | Total | 237.9 | 0.8496 | 41.06 percent | 58.94 percent | 341.50 |
| | | | | | | | | End of Event | | | | | | | | | | | | | | | | reduction | | | | Total Demand | kW | per sq. ft | percent of FL | percent saved | kW | | HVAC | 485.7 | 1.7345 | 66.11 percent | 33.89 percent | 249.0 | | Lighting | 91.0 | 0.3251 | 36.47 percent | 63.53 percent | 158.6 | | Data Center | 151.2 | 0.5400 | 88.94 percent | 11.06 percent | 18.8 | | Plug/Unknown | 241.9 | 0.8638 | 61.04 percent | 38.96 percent | 154.4 | | Total | 969.7 | 3.4634 | 62.88 percent | 37.12 percent | 572.5 | | | | | | | | | | | | | | | | | | | reduction | | | | Existing Building (S & W) | kW | per sq. ft | percent of FL | percent saved | kW | | HVAC | 355.6 | 1.2700 | 60.27 percent | 39.73 percent | 234.40 | | Lighting | 73.2 | 0.2613 | 40.12 percent | 59.88 percent | 109.18 | | Data Center | 151.2 | 0.5400 | 88.94 percent | 11.06 percent | 18.80 | | Plug/Unknown | 106.3 | 0.3795 | 72.39 percent | 27.61 percent | 40.52 | | Total | 686.2 | 2.4508 | 66.21 percent | 33.79 percent | 350.23 | | | | | | | | | | | | | | | | | | | | | | | | | | reduction | | | | New Building (N & E) | kW | per sq. ft | percent of FL | percent saved | kW | | HVAC | 130.07 | 0.4645 | 51.24 percent | 48.76 percent | 123.76 | | Lighting | 17.87 | 0.0638 | 26.54 percent | 73.46 percent | 49.44 | | | | | | | | | Plug/Unknown | 135.59 | 0.4843 | 52.37 percent | 47.63 percent | 123.33 | | Total | 283.5 | 1.0126 | 48.94 percent | 51.06 percent | 295.87 |
Sidebar: What is demand response and how does a curtailment service provider work?
The supply of
electricity must precisely increase and decrease to perfectly match the demand
for electricity on the regional power grid. A Demand Response (DR) program
enables businesses to earn financial incentives by temporarily reducing
(curtailing) the use of electricity (or by running emergency generators) at
times when electricity is in short supply.
Typically, during the
hottest summer days, there are periods when electricity demand is forecasted to
exceed supply. During these peak demand hours, DR programs are activated across
the region to help relieve the power grid and keep the power flowing. Such
programs are designed to minimize the occurrence of rotating outages and are a
result of a joint effort between the Federal Energy Regulatory Commission
(FERC), state utility commissions, and various state agencies.
Sidebar: Eaton Fast Facts
· Eaton has
received over $40K during the past two years through the Demand Response
program. The company plans to reinvest the funds into building upgrades to
expand on energy-efficiency.
· Eaton
recently received LEED Gold status for its 3,875 square-foot Executive
Briefing Center
(EBC)- completed as part of its Electrical Sector America’s headquarters’ expansion
in September 2008. The EBC is a venue for Eaton’s business and technology
leaders to meet with global customers and partners, collaborate on solving
business challenges and demonstrate Eaton’s electrical solutions.
|
By: Jasmine
Posted: March 15, 2010 11:51 AM