Variable frequency drives (VFDs), also known as variable speed drives (VSDs), are used to modulate the speed of motors in various applications, from hydronic water pumping to VAV fan units to vacuum systems.

VFDs allow machinery to match the demand of a system by modulating speeds and therefore capacity, but the benefits hardly end there. They allow engineers and designers to exploit what may seem like a loophole in physics. The fan laws and pump laws, also known as affinity laws, describe how VFDs can modulate capacity non-linearly relative to power input. For example, a fan motor moving 1,000 cfm at full speed will move 500 cfm at 50% but will only consume approximately 12.5% of the power used at full speed. The power use is proportional to the cube of the speed, as expressed below:

power use

For example, if a motor draws 10 kilowatts (kW) at full speed, what would the power draw be at half speed?

motor draws 10 kilowatts

In HVAC applications, VFDs really shine during part-load periods when cooling or heating loads may be minimal. During part-load conditions, heating and cooling loads drop, allowing fan speeds to drop proportionally with those loads. When those fan speeds drop, power consumption decreases non-linearly. In a constant volume fan system, a fan system without a VFD-driven motor, those savings are nowhere to be found. The relationship between motor power and motor speed or system load are illustrated in Figure 1.

figure 1


VFD use is becoming ubiquitous on air-side systems

VSDs are used on a variety of systems in buildings. Engineers and designers typically specify their use on all new hydronic air-handling-system fan motors. More recently, VFD retrofits on existing systems like these have become very economical, and one may see payback periods spanning less than three years. VFD use on new multistage, direct-expansion package unit fan motors is becoming more common, and there are even VFD retrofit solutions for package unit supply fan motors.


VFDs in chiller plants — work in progress

VFD adoption in central plants has taken a little longer to catch on. A minimum flow must typically be maintained through chillers and boilers to avoid scale buildup and prevent laminar flow, which decreases heat transfer. This minimum flow problem can make a primary chilled water or condenser water variable flow retrofit with VFDs complicated and fussy. A proper VFD retrofit of a chilled water plant should start with converting a constant volume cooling tower fan to variable volume, since a varying cooling tower fan speed based on wet-bulb and condenser water set points has no negative impacts on chiller performance.

In primary-secondary chilled water pumping systems, the secondary chilled water loop should be converted to variable flow since the primary loop dictates the flow through the chillers and will remain as constant flow pumps. In primary-only chilled water pumping arrangements, the chilled water distribution system may be converted to variable flow if the chiller can handle reduced water flow through the evaporator barrel. The minimum flow must be determined based on the manufacturer’s specifications. Testing and balancing will be required as well as devices to measure and verify minimum flow through each evaporator barrel.

Recently, VFD-driven centrifugal and rotary-screw compressors have made their way into the chiller market. These chillers are capable of efficiently unloading and pair well with variable primary flow pumping systems. One might ask, if variable primary flow can be achieved safely with chillers, why has variable condenser water flow lagged behind?


Variable condenser water flow — the final frontier of VFD adoption

Variable condenser water flow is rarely the first retrofit of choice for chiller plant upgrades. In fact, many new construction projects skip out on the large savings that can be captured with the implementation of variable condenser water flow, even when fully modulating chillers are installed.

A primary deterrent to implementing variable condenser water flow is the concern of scale buildup due to low fluid velocities as mentioned earlier. Fast-moving water inhibits scale buildup in the heat transfer surfaces within the chiller condenser barrel. When water slows down, it has the opportunity to hang around and drop off mineral deposits. These deposits buildup over time and can penalize performance as the chiller ages. To prevent this, chillers have low-flow prevention systems. Chillers need a minimum flow of condenser water or else the chiller will cycle off; however, most modern chillers have a minimum condenser water flow rate of 30%-50% of the maximum condenser water flow.

A lesser-known concern is that reducing condenser water flow can cause areas of the cooling tower fill material to dry up. This drying effect of cooling tower fill is known as “dry air disease” and should be avoided at all costs because it will actually increase energy consumption over the long run. When sections of the fill material dry up, they begin to build up mineral deposits at the interface between wet and dry fill material, meaning the fill may require replacement earlier than anticipated. Additionally, the air that is drafted into the cooling tower fill will take the path of least resistance: the dry spots. No heat transfer occurs at these sections because there is no water to evaporate, and thus the fan or condenser water pump motor will compensate by increasing speed. This dry air disease can enter into a downward spiral of lost cooling tower efficiency and continual scale buildup.

Testing should be done to determine the minimum condenser water volume needed to keep fill material wet. This can vary depending on outdoor weather conditions so consider testing in shoulder months and under different wet bulb conditions.

Many cooling tower manufacturers now offer towers that can adapt to changes in condenser water flow for a small marginal upfront cost. For example, a well-known cooling tower manufacturer has a new feature that allows the cooling tower fill to be divided into “inboard” and “outboard” sections by installing elevated nozzle cups on the inboard section of the hot water basin. During a low-flow condenser water scenario, water will only travel through the outboard nozzles, as shown in Figure 2. However, the humidity in the air passing over the inboard section will keep the inboard section of fill material wet. During higher flow scenarios, the water in the hot water basin will fill up to the point that water will flow through the elevated nozzle cups and begin flowing down the inboard section of the cooling tower fill, as illustrated in Figure 3.

figure 2


figure 3


Imagine a facility with a chilled water plant that includes one chiller and one cooling tower on a mild day. The chiller is loaded to 30% and the cooling tower fan is off; however, the condenser water pump is running at 100% because the tower must not be allowed to “dry-up.” If elevated nozzle cups are installed, the chiller can reduce its condenser water pump speed to 30%, provided no low-flow switches are tripped and the media will remain wetted. One must consider, with reduced condenser water flow, the cooling tower fan may click on depending on the fan control sequence. There is considerable optimization that could be had here on the condenser side with variable condenser water flow.

On a multi-chiller, multi-tower system with elevated nozzle cups installed, all three towers could pass the condenser water of just one chiller with each fan running at 33% speed. This is effectively trading fan energy for wetted fill, and it’s a really good bargain thanks to the affinity laws.

A three-tower system under part-load conditions can either distribute condenser water to one tower as illustrated in Figure 4 or distribute water to all three towers, reducing total fan energy, as illustrated in Figure 5.

figure 4


figure 5



Control strategies for variable condenser water systems

When variable condenser water flow and variable speed cooling tower fans are implemented on the same system, care must be taken in the development of control strategies. Typically, the best course of action is to establish a condenser water minimum flow for each pump and vary the condenser water flow based on chiller loading. Then, have the cooling tower fan speed respond to the condenser water return temperature or condenser water delta T. This prevents the condenser water flow and cooling tower fan speed from continually modulating  (imagine a cat chasing its tail) like they would if the systems responded to the same control point.

Another control strategy is to have condenser water flow respond to the condenser water return temperature or delta T. When condenser water flow reaches 100%, the cooling tower fans can cycle on and track condenser water return temperature or delta T. In this scenario, there would be two stages of condenser water cooling determined by whether condenser water flow is 100% or not.


Real-world implementation of variable condenser water flow

Wood Environment and Infrastructure Solutions has implemented variable condenser water flow on a chilled water plant located in New Jersey as part of a larger energy conservation project for a national client. The building is a large distribution center that primarily uses chilled water for cooling. The existing chilled water system consisted of two variable flow primary chilled water pumps, three variable speed cooling fans, and three constant volume condenser water pumps. The three Trane CVHE450 centrifugal chillers also had variable speed compressors. The system was ideal for a variable condenser water conversion project, and the mild but variable cooling load meant part-load efficiencies could be exploited with a fully modulating chilled water system.

The building runs a year-round sorting operation and has a relatively high internal heat load. In addition, trailer docks around the building contribute to a significant infiltration load. Despite having a relatively short cooling season with an average of 1,740 cooling degree days, chillers were typically disabled only for a few months in winter. The three chillers had a combined capacity of 1,080 tons, although only two can run simultaneously.

The HVAC retrofit design intent was to reduce energy consumption while minimizing the impact on  comfort. This was done by installing VSDs on heating hot water pumps and condenser water pumps, reducing VFD minimum speed set points, adjusting economizer set points and minimum outdoor air damper positions, testing and balancing, programming energy efficient set points, setbacks and schedules, installing a premium-quality weather station, and revising controls sequences. One thing to note is the cooling tower was replaced with a like-and-kind cooling tower and was deemed to have little effect on energy consumption.

In the design of the condenser water-side sequence of operation, the chiller minimum flow rate was determined to be 720 gpm based on chiller product data. In the condenser water pumping system, this corresponded to a specific pressure differential across the condenser barrel. This pressure differential became the minimum chilled water flow rate lower limit. The condenser water speed was then set to modulate based on the chiller’s running load amps (RLA).

Cooling tower fan speed, however, is modulated to meet a condenser water supply temperature set point. When condenser water supply temperature drops below the minimum set point, which is typical upon initial startup and low ambient conditions, the cooling tower bypass control valve will modulate open to allow condenser water to glow directly into the return.

In addition, the automation system was programmed to determine the optimal condenser water cooling tower supply temperature set point at any chiller load and ambient wet bulb temperature.

The entire HVAC retrofit project resulted in a modeled annual savings of 1,400,000 kWh, 30,700 therms, and 570 kilogallons of water. Although the impact of condenser water pump energy savings for this project cannot be determined from overall energy savings, in the next section several scenarios are presented that estimate the energy impact of variable condenser water pumps.


Modeled scenarios of variable condenser water flow

Using energy modeling software, a fictional building was modeled in four locations around the U.S. The building is a 100,000-square-foot light manufacturing facility. Table 1 shows the results of the energy model and expected savings of variable condenser water flow. Condenser water VFD retrofits can reduce total facility electric consumption between 4-9%. The retrofit has the highest impact in hot and humid climates as seen in Table 1, but savings are still substantial in all climate zones where water-cooled chillers are used.

table 1

> TABLE 1.



Most utility companies have offered incentives for retrofitting pump and fan systems with variable frequency drives, often paying out between $50-200 per horsepower installed. However, with incentive budgets drying up as VFD retrofit projects become more economical, Wood has found that variable flow condenser water retrofit applications to be one of the few HVAC-related VFD incentives available at many utilities across the country. Even the notoriously fussy incentive program at Southern California Edison still lists “variable-speed drives on condenser water pump control” as their only HVAC-related VFD express solution.



A properly implemented variable condenser water pumping system is an affordable addition to any energy retrofit, central plant renovation, or new construction project. In a retrofit project, existing conditions, such as chiller unloading capability, condenser water, and cooling tower control, need to be carefully examined to ensure the system will function properly following the retrofit. In new construction, care must be taken during chiller and cooling tower equipment selection. Finally, the implementation of variable condenser water flow should be considered only if variable chilled water flow and variable speed cooling tower fans are present or are in the pipeline for construction.

Back To Top