Energy Manager

Taking advantage of demand response to cut building operating expenses: how a hybrid chiller system can help

Canada has abundant resources for energy independence; from the Alberta oil patch to the hydro plants in BC and Newfoundland, and the 3,000 MW of wind farms spread across the provinces. Nevertheless, the distribution of that power isn’t always handled as efficiently as it could.

March 10, 2010  By Joe Zwers

Implementing a smart grid and demand response programs can help ensure reliable power at an affordable cost. In March 2008, the Independent Electricity System Operator (IESO) started the Ontario Smart Grid Forum ( which brings together representatives from the Ministry of Energy, Hydro One, Hydro Ottawa, Burlington Hydro, Toronto Hydro and other industry groups to develop ways to create a more efficient and reliable grid. All the provinces now have some sort of demand response (DR) system in place, and Natural Resources Canada (NRCan) is conducting projects to make it easier for campuses and commercial buildings to participate in DR programs.

Ontario Options
Demand Response systems, where electrical customers reduce their electrical load in response to a request from the utility, are still under development. The most advanced is in Ontario.

“The only province that has a response program for commercial buildings is Ontario,” says Meli Stylianou, a project leader with the Intelligent Buildings Technologies Team at NRCan’s Canmet Energy Technology Center in Varennes, QC, just outside of Montreal. “The other provinces have different versions, generally interruptible power programs for industrial facilities.”

Ontario’s program came out of the report Blueprint for Demand Response in Ontario that Navigant Consulting prepared for the Independent Electricity Market Operator (now the Ontario Power Authority) in April 2003.


Residential and small business customers can participate in a program called peaksaver, which is funded by OPA and administered through the local utilities. A small device is installed near the water heater or central air conditioner. During the DR periods, typically between 2 and 6 pm on summer weekdays, the device will turn off the water heater or will cycle the AC off for 15 minutes out of every 30. DR periods never exceed four hours. In early July 2008, OPA had its first province wide peaksaver demand response call. The program was estimated to shave 40MW off the peak power demand.

For businesses, OPA currently has a three stage DR program ( DR1 is a voluntary program where large customers have the choice of either reducing their consumption, or not, when the price exceeds an agreed upon level. There is a minimum participation level of 500kW for one hour per year. Participants are paid based on the difference between their baseline electricity use and the amount used during the DR period.

While DR1 is voluntary, DR2 requires a firm commitment from the commercial or industrial facility to shift load every business day from the DR period between 7 am and 7 pm to a time at night or on the weekends. The company can choose to reduce electricity use anywhere from 4 to 12 consecutive hours during the DR period.

DR3 requires participants to select one of two possible schedules in which they agree to reduce electricity on demand for a four hour period. The property must have a peak demand of no less than 50kW and sign up through a DR aggregator. If the property has 5MW DR capability or more, it can contract directly with OPA. The DR can be provided by reducing usage or by operating a generator. (See below for a possible generator configuration.)

Todd Williams, the Navigant Consulting director who was the primary author on the DR Blueprint report says that the key to effectively participating in the market is to already have a good building control system and demand reduction procedures in place.

“It all revolves around good monitoring and control,” he says. “If I know what is operating and I have flexibility in terms of control, I can provide that demand response.”

Participants have the option of either manually reducing demand or going with a completely automated version, where the signal comes in from OPA and the building management system takes preprogrammed actions.

“Most people don’t want to go that far,” Williams says. “They have a sense of what they can do and when the call comes in, there may be some choices of what they can shed to achieve the reduction target.”

Hybrid Chillers
One way to reduce electrical demand is to use a hybrid chiller system. The most common approach is to use a mix of two different drivers — such as a steam turbine and an electric motor or reciprocating gas engine — to power two centrifugal or absorption chillers. The building manager would select which chiller to use depending on the current price of electricity, gas or steam. In most cases, the motor driven chiller would operate during the cooler hours of the day and be supplemented or replaced by the engine or steam turbine during the peak hours or when there is a DR call,

The downside of a hybrid chiller system is that it roughly doubles the initial capital expenditure, as two separate compressors are required plus the piping and valving to connect to the evaporator. This can, however, be quickly offset by the reduced operating costs. York International Corporation compared the capital and operating costs of two 500-ton electric centrifugal chillers and six different hybrid designs which could meet the same cooling requirements. It found that the different hybrid systems would pay for themselves over a period of six months to five years. Those payback calculations were based on particular temperature ranges and certain pricing structures for gas and electricity and may not be applicable to one’s own building’s conditions.

In general, the temperatures used for the calculations are warmer than what is seen in Canada, so those savings may not be possible. A way to cut down on the capital costs, while still achieving high energy savings, is to connect two drivers to a single compressor using self-synchronizing over running clutches such as an SSS Clutch. This approach, while new to building HVAC systems, has been successfully used for years with natural gas pipeline compressors, allowing them to operate off either natural gas or electricity, depending on the relative cost of the two. It may also meet the OPA DR3 option of switching to on site generation.

Dual driving a chiller compressor with an electric motor and reciprocating engine or steam turbine can provide demand side management capability, emergency or standby electricity and/or chiller power, while reducing energy costs. All of this is accomplished at a considerable first cost savings compared to separate engine generators and electric driven chillers.

Hospitals, campus facilities and large office complexes often see huge differences in day to night energy charges. Many of these same facilities require emergency electric power capability, oftentimes including keeping at least some of their chillers running to provide cooling for critical computer equipment. This particular system, then, might also serve this purpose.

The mechanical system arrangement for the dual driven chiller is engine-clutch-motor-clutch-compressor. The motor would be the synchronous type designed for both motor and generator operation. For peak demand shaving operation with the motor already driving the chiller compressor, the engine would be started, brought up to speed under governor control, and the SSS Clutch would automatically synchronize the engine to the motor shaft. The engine load would be adjusted to pick up as much of the motor load as desired. Or, if the engine is sized appropriately, and allowed by the local utility, the engine could handle the entire chiller load and operate the motor/generator in the generator mode feeding power back into the grid. This same mode of operation could be used if the price of fuel for the engine is less than the cost of electric power for the motor, which often occurs during peak load times on the power grid.

In the event of a power failure the system could be operated in one of 3 modes:  1. Engine driving motor/generator generating emergency electrical power with the clutch to the chiller in the locked out position. 2. Engine powering the chiller only with both clutches engaged but the motor/generator breaker open.  3. The engine driving both the motor/generator and chiller producing chilled water and electrical power. 

Driving the chiller compressor directly with the engine during emergency operation allows a considerably smaller engine compared to sizing an engine to drive a generator sized to start the electric motor driven chiller. Starting reliability is also enhanced. 

In the peak shaving mode the motor/generator does not have to act as a generator paralleled with the grid but can remain in the motor mode. The engine load can be automatically adjusted to pick up the greater part of the chiller load but not pump power back into the grid, thus avoiding the complex engineering studies and special electrical equipment required by the utility for a generator paralleled with their system (a reverse current relay would be needed). The system would still be able to generate emergency power by including a transfer switch.  A separate engine generator used for peak shaving or demand management would need to be paralleled with the grid.   

The dual driven chiller could be incorporated into a CHP system, particularly for an operation that has a need for heat in the winter but not in the warmer months. The engine would be used continuously in the winter but in the warmer months only operate for peak shaving and demand management, using the motor most of the time. Such an arrangement would be well suited for parts of Canada, such as Ontario, which have their peak demand in the summer months. If using an engine, however, be sure that its operation meets with any environmental permits or regulations.

Making it Easier to Respond
Natural Resources Canada is working on two projects to make it easier for buildings to participate in DR programs. Both of these take advantage of earlier and ongoing work done at Lawrence Berkeley National Laboratory’s Demand Response Research Center (DRRC).

One project involves testing different control strategies. Currently there are two buildings in the test, one at NRCan’s Varennes research center and the other at the British Columbia Institute of Technology in Burnaby. OPA will identify a third building in Ontario.

“Our objective is to link these three buildings to a server located in Varennes and to simulate different demand response activities,” says Stylianou. “We are also in the process of modifying the controls of the buildings to receive this information and to react appropriately to reduce the electricity demand.”

The DRRC has created software and a set of standards called OpenADR ( that automates a range of different DR programs, and can be adapted to serve others.  More than sixty vendors have implemented the OpenADR client software.

“Here in California, we all three investor owned utilities have automation servers and we have over 200 facilities that are automated using OpenADR,” says DRRC director Mary Ann Piette. “It is a very flexible, open-based model to allow utilities to send signals to the buildings.”

But even more important is what steps the building management system takes in response to the automated signal. The California model, where the peak demand is on hot summer afternoons, calls for such actions as precooling the building during the night and changing the settings on the air conditioners. It also may involve lowering the lighting, since lights are a heat source as well as a power draw. Such a strategy would not be appropriate in colder areas, where peak demand comes during winter mornings and evenings, when both light and heat are needed. In Canada, Ontario follows a similar Summer afternoon peak pattern to California, but the other provinces have winter peaks.

“It is important to understand the electric load shape of the building and what is on at those early morning hours, and you may need to preheat the building,” says Piette.

In areas that rely heavily on hydro, such as BC, the DR strategy may not be related to cutting peak demand, since hydropower can easily follow load shifts, but be used to cut power usage during droughts. Wind power and solar also have an effect on demand response strategies. In Canada, there was been a tenfold increase in wind power capacity over the last decade and more is on the way.

“As we target more renewables on the grid, we have intermittent loads and demand response becomes even more important, she says. “So utilities are not just looking at hot summer afternoons, but times when the wind dies down or the solar loads are coming down.”

Stylianou’s tests will determine the best strategies buildings in different parts of Canada should employ in implementing DR. His other project is to modify the DRRC’s software that models the impact of DR on a building. The DRRC software is designed for the California’s climate and buildings.

“We are modifying it to so users can model climactic conditions and office buildings in Halifax, Montreal, Toronto, Winnipeg, Calgary and Vancouver,” he says. “We will have a tool to estimate, for different sizes of office buildings, what their potential for demand response would be.”

Print this page


Stories continue below