Task 15 – Case Study – LIPAEDGE Direct Load Control Program – USA

LIPAedge Direct Load Control Program – USA

This is the 8th article in a series highlighting the case studies of DSM Task 15, Network Driven DSM. This Task demonstrated that DSM can be successfully used to support electricity networks in two main ways:

1) by relieving constraints on distribution and/ or transmission networks at lower costs than building ‘poles and wires’ solutions, and

2) by providing services for electricity network system operators, achieving peak load reductions with various response times for network operational support.


The LIPAedge Direct Load Control Program was developed by the Long Island Power Authority (LIPA) to use central control of residential and small commercial air-conditioning thermostats to achieve peak load reduction.  The program commenced in early 2001.  However, from 31 July 2003 the program was closed to new participants because LIPA had enough air-conditioner
load under direct load control.


The LIPAedge program uses the programmable ComfortChoice thermostat (see Figure 1).  This was designed by the Carrier Corporation with associated communication infrastructure provided by Silicon Energy to provide emergency peak reduction for utilities.


Figure 1.  ComfortChoice Thermostat Used in the LIPAedge Program

The system operator uses an internet-based system provided by Silicon Energy to control a
demand-side resource comprising about 20,000 thermostat‑controlled air-conditioners. Skytel two-way pagers are used to transmit a curtailment order to the thermostat and to receive acknowledgment and monitoring information.  One or more pager signals are generated and transferred to the SkyTel pager network (see Figure 2).  Commands go via satellite to
pager towers, where they are broadcast to the thermostats.  The thermostats take immediate action or adjust their schedules for future action, depending on what the system operator
ordered.  The thermostats log the order and respond via pager, enabling LIPA to monitor the response to the event.  The thermostats also collect data every minute on temperature, set point, and power consumption (hourly duty cycle).  They retain this information as hourly averages and report it to the utility.  The thermostat itself holds 7 days of hourly data.


Figure 2.  The Carrier/Silicon Energy Direct Load Control System

For a summer load curtailment, the system operator might send a command at 9:00 am directing all thermostats to move their set points up four degrees, starting at 2:00 pm and ending at 6:00 pm.  Alternatively, the system operator could send a command directing all thermostats to completely curtail immediately.  The command would be received and acted upon by all loads, providing full response within about 90 seconds.  This is far faster than generator response, which typically requires a 10-minute ramp time.

Thermostats can be addressed individually, in groups, or in total.  This important advantage provides both flexibility and speed.  System operator commands that are addressed to the entire resource are implemented through a single page that all thermostats receive.  Similarly, 15 subgroups can be addressed if response is required in a specific area to alleviate a
transmission constraint.  Thermostats can be addressed individually as well.  This capability is useful for monitoring the performance of the system (each thermostat is checked weekly for a “heartbeat”).

The customer also receives benefits.  The thermostat is fully programmable and remotely accessible, with all of the associated energy savings and convenience benefits.  A web-based remote interface is provided for customer interaction.  Customers can also override curtailment events. The override feature appears to be important to gain customer acceptance and it probably increases the reliability benefit.  The system operator can block overrides if necessary.  Typically. this is not done for demand curtailment events, but it may be useful for spinning
reserve events.

Two-way paging communication enables the utility to monitor load performance both during response events and under normal conditions.  Response from the thermostats is staggered over a time period set by the utility to avoid overwhelming the paging system.  It typically requires 90 minutes for 20,000 thermostats to respond.  Thus the system provides for performance monitoring but not in the 2 to 8-second intervals typical for large generators.

Communication is more reliable from the system operator to the thermostat than from the thermostat to the system operator.  The pager tower has a 500-watt transmitter, while the thermostat’s transmitter is only one watt.  The thermostat makes four attempts to report
back if the pager tower fails to receive any of its signals.  The thermostat continues to take control actions and respond to new commands even if return communication is lost.  Hence the system is more reliable than would be indicated by the list of “failed” units generated by the
“heartbeat” report.  About 4% to 5% of the thermostats fail to report back.

The LIPAedge Program

The LIPAedge program is the largest residential direct load control program using two-way communication in the United States.  Two-way communication allows LIPA to monitor capability and response.  It also enables customers to control their individual thermostats via the internet, a benefit that motivates participation.

The LIPAedge program is available for residential customers with central air-conditioning and for small business customers, though the program is now closed to new participants.  There are about 20,000 residential customers and 3,000 small business customers participating in the program.  Customers who sign up to the LIPAedge program receive a ComfortChoice thermostat and installation free of charge.  Customers also receive a one-time bonus payment of USD 25 (residential customers) or USD 50 (small business customers).

During 2001, when the program commenced, LIPAedge customers were offered an opportunity to earn a USD 20 cash reward for each LIPA customer referral they provided who installed a LIPAedge thermostat.

LiPAedge customers agree to have their
central air-conditioning system adjusted between the hours of 2 pm and 6 pm for a maximum of seven days throughout the four month summer season.  Customers have access to a dedicated web page for their thermostat and are able to remotely change the set point of their air-conditioner whenever they want.

LIPA initiates curtailment events by either increasing the set point on LIPAedge thermostats by 3 to 4 degrees, or by cycling air-conditioner compressors off for a portion of each hour (see
Figure 3).

Customers can override curtailment messages sent to their thermostat, though LIPA encourages its customers not to override during a curtailment event.  If the customer decides to override the curtailment, the change is recorded by the thermostat and a wireless message is then sent back to the central server.


Figure 3.  The Curtailment Process in the LIPAedge Program


Results Achieved

LIPA collected name-plate power consumption information on the air-conditioning equipment being controlled when it installed the ComfortChoice thermostats for the LIPAedge program.  It also directly measured the power consumption of a subset of those loads to estimate the actual load of the aggregation.  LIPA determined that the average capacity of residential air-conditioning units being controlled was 3.84 kW, while the average capacity of small commercial units was 6.38 kW.  The total 23,400 individual loads had a peak capacity of 97.4 MW if all the units were on at 100% duty cycle.

LIPA monitored the performance of 400 units from 1 May 2002 through 29 September 2002.  Hourly data were collected from each unit for duty cycle and facility temperature.  Those
data were used to estimate the performance of all 23,400 responsive loads.  LIPA found that each controlled load provided an average of 1.06 kW of demand reduction (1.03 kW per residential air-conditioner and 1.35 kW per small commercial air-conditioner).  LIPA expected 24.9 MW of peak reduction response from the full 23,400 controlled air-conditioners.

LIPA tested the actual performance of the system to reduce energy demand during peak hours on three days during the summer of 2002.  It also monitored performance on seven other days to provide baseline data. The results are shown in Tables 1 and 2.  Table 2 shows that an
increasing number of residential thermostats were overridden as the 14 August curtailment event continued; the proportion of units overridden increased from 5.7% at 3 pm to 20.8% at 6 pm.



The LIPAedge program cost was USD 515 per residential customer and USD 545 per commercial customer. This yielded a combined average cost of USD 487/kW of demand reduction. LIPA paid all costs.

Download case study here

This article was contributed by David Crossley, Managing Director of Energy Futures Australia Pty. Ltd and Senior Advisor at The Regulatory Assistance Project.  For more information on this case study and others, visit Task 15, Network Driven DSM at:

Task 15 Network Driven DSM