Types of Chiller:
4-Refrigerant Loss Rates Reduced
6-Microprocessor-Based Control Systems
7- Selection Criteria
THERE ARE TWO basic types of reciprocating chillers:
1-1 hermetically sealed units and
1-2 units of open construction. Of the two types, hermetically sealed units are the most common.
In hermetically sealed units, the motor and the compressor are direct-coupled and housed in a single casing that is sealed to the atmosphere. In open construction units, the motor and the compressor are in separate housings and are connected by a direct drive shaft or by a V-belt. In general, open construction units have a longer service life, lower maintenance requirements and higher operating efficiencies.
Reciprocating chillers offer several advantages. Compared with other types of chillers, their initial cost is low for units of 100 tons or less. Reciprocating chillers have a higher condensing temperature than other chillers, making them more suited for applications where air-cooled condensers must be used. It is easy to closely match the capacity of the chiller to the building load by installing multiple machines. Multiple machines also allow the facility manager to stage operation for part-load conditions, increasing operating efficiency.
A major drawback of reciprocating chillers is their high level of maintenance requirements in comparison with other chiller types. Reciprocating chillers have more moving parts than centrifugal or rotary chillers, resulting in an increased need for wear-related maintenance activities.
Reciprocating chillers also generate high levels of noise and vibration. Special precautions must be taken to isolate the chillers from the facility to prevent transmission of machine-generated vibrations and noise.
Finally, reciprocating chillers are not well suited for applications with cooling loads in excess of 200 tons. As the units grow in capacity, their space requirements and first costs exceed those of other chiller types. In addition, the energy requirements for larger units exceed that of other chillers types.
Centrifugal chillers are variable volume displacement units. Typically, an electric drive powers one or more rotating impellers that use centrifugal force to compress the refrigerant vapor. The cooling capacity is controlled through the use of inlet vanes on the impellers that restrict refrigerant flow.
Centrifugal chillers are generally manufactured in capacities from 90 to 1,000 tons, with most units falling in the range of 150 to 300 tons.
Like reciprocating chillers, centrifugal units are available in both hermetically sealed and open construction. Despite its lower operating efficiency, the hermetically sealed unit is more widely used.
Centrifugal chillers are popular in part because of their low energy cost per ton of cooling produced relative to other chiller types. Centrifugal chillers also are small in comparison with reciprocating units, and do not produce as many vibrations. However, the high-speed drive unit centrifugal chillers use produces a high-frequency noise that can disrupt operations in adjacent spaces.
Typical full-load efficiencies for new chillers, rated at standard ARI operating conditions, range between 0.50 and 0.62 kW/ton. These efficiencies offer a significant improvement over those of the chillers they are replacing. Chillers made during the 1970s averaged 0.90 kW/ton for standard units and 0.80 kW/ton for high-efficiency units. And these values were for chillers when they were first installed. Over the years, normal wear and tear and fouling of the heat transfer surfaces lowered these efficiencies. Today's replacement chiller uses 25 percent to 30 percent less energy to produce the same cooling.
Chiller efficiencies haven't topped out, and industry experts expect them to continue improving over the next few years. At least one company is producing large centrifugal chillers with an operating efficiency below 0.50 kW/ton.
New chillers are larger than the ones they are replacing, even with no change in capacity. Efficiency improvements are primarily the result of three design changes: larger heat exchangers, redesigned refrigerant flow aerodynamics and more efficient motor designs. All of these changes increased chiller size. With mechanical rooms historically undersized, there simply may not be enough room to install the new, larger chiller.
A serious drawback to centrifugal chillers has been their part load performance. When the building load decreases, the chiller responds by partially closing its inlet vanes to restrict refrigerant flow. While this control method is effective down to about 20 percent of the chiller's rated output, it results in decreased operating efficiency. For example, a chiller rated at 0.60 kW per ton at full load might require as much as 0.90 kW per ton when lightly loaded. Since chillers typically operate at or near full load less than 10 percent of the time, part load operating characteristics significantly impact annual energy requirements.
Centrifugal chillers also can be difficult to operate at low cooling loads. When the cooling load falls below 25 percent of the chiller's rated output, the chiller is prone to a condition known as surging. Unrestricted, surging can lead to serious chiller damage. To reduce the chances of damage from surging, manufacturers add special controls, but most of these controls further reduce the part load efficiency of the units.
The problems with part load performance and low load operation to a great extent have been eliminated through the addition of variable frequency drives (VFDs) to centrifugal chillers. The VFD alters the line voltage and frequency powering the chiller based on the cooling load. As the cooling load decreases, the VFD decreases its output voltage and frequency, slowing the chiller. The result: A chiller that operates at near full load efficiency for almost the entire operating range. Additionally, the VFD units allow the chiller to operate as low as 10 percent of rated output without experiencing surging.
One VFD system drawback: It increases the chiller's full-load current draw by 5 percent to 10 percent due to internal losses and conversion inefficiencies. While this penalty is more than offset by savings produced under part-load conditions, it can be significant, particularly in facilities with high electrical demand charges.
Gas-engine-driven centrifugal chillers are an alternative to electrically driven chillers in areas where electrical demand charges are high and natural gas rates are low. A switch to natural gas can offset a significant portion of their facility's electrical demand.
Another recent development in chiller design improves the efficiency of HFC-134a centrifugal chillers by replacing the expansion valve in the refrigerant system with a small, two-phase turbine. With expansion valves, the energy potential available across the expansion valve due to the high pressure differential is lost. By replacing the valve with a turbine, some of the energy can be recovered and used to help drive the chiller's compressor, reducing energy the motor must supply.
Rotary or screw chillers, like reciprocating chillers, are positive-displacement compressors. An electric motor drives two machined rotors that compress refrigerant gas between their lobes as they mesh. Units are available in both hermetically sealed and open construction.
Rotary chillers are available in capacities ranging from 20 to 2,000 tons, with most installations falling in the range of 175 to 750 tons. Typical chiller efficiencies are between 0.70 and 0.80 kW per ton, making rotary chillers more efficient than comparably sized reciprocating chillers, but less efficient than centrifugal chillers.
Two major advantages of a rotary chiller are its compact size and light weight. With a relatively high compression ratio and few moving parts, rotary chillers are smaller and lighter than reciprocating and centrifugal chillers of the same cooling capacity. Rotary chillers also offer quieter, vibration-free operation.
The major drawback of rotary chillers is their high first cost. For small cooling loads, reciprocating chillers are less expensive to purchase and install; for large loads, centrifugal chillers cost less.
4-Refrigerant Loss Rates Reduced
New chiller designs include features to significantly reduce refrigerant losses. Before chlorofluorocarbons (CFCs) were an issue, a loss rate of 15 percent per year was considered acceptable. Many older installations have even higher loss rates. But now that CFC refrigerants are both an environmental and an economic concern, building managers cannot continue to replace refrigerants at that rate.
A new system can limit refrigerant losses to less than 0.5 percent per year, including the losses that occur during system maintenance. These low loss rates are the result of changes in the design and installation of the chiller, including minimizing the number of shaft seals required, installing isolation valves around such items as filter assemblies, and using brazed instead of flared fittings.
New systems benefit from the use of high-efficiency purge units. And monitoring the run time of the purge unit allows operators an early warning of system leaks.
Reciprocating, centrifugal and rotary chillers use mechanical energy in the form of a motor to drive the cooling cycle. Absorption chillers use heat as the energy source to drive the process.
There are two basic types of absorption chillers: direct- and indirect-fired units. Direct-fired absorption chillers typically burn natural gas to generate heat to drive the cycle. Indirect-fired units use low pressure steam, hot water or waste process heat. The most common refrigerants used include water and ammonia.
Absorption chillers range in capacity from 100 to 5,000 tons, with most of the currently installed machines operating in the range of 300 to 500 tons. Thermal efficiencies typically are between 11,000 to 19,000 Btu of heat input per ton-hour of cooling produced.
Absorption chillers offer the advantage of using an energy source other than electricity to power the air conditioning system. The electrical energy used in an absorption chiller is typically less than 10 percent of the electricity required by other chillers. This low demand for electricity makes the units well suited for applications where there is insufficient electrical capacity for motor driven chillers, or where the local utility's electrical demand charges that would be incurred would be excessive.
Indirect-fired absorption chillers also offer the flexibility of being powered by a range of heat sources, including low pressure process steam, hot water, solar energy and waste heat. If a facility has a source of waste heat that must be disposed of, absorption chillers can provide low-cost chilled water while reducing the waste heat discharge temperature.
In the United States, the future looks bright for absorption chillers. Once widely used in facilities with sources of waste heat, or underused central steam systems, the chillers were for the most part replaced with lower maintenance, electrically driven, centrifugal units. Ten years ago, the U.S. absorption market was all but dead even though other countries, particularly Japan, continued to use and refine absorption technology.
The situation has changed. First, the phasing out of CFC-based refrigerants increased the demand for refrigeration cycles that were CFC free. Absorption chillers, with water as their refrigerant, offer a good alternative.
Another factor that helped to increase interest is economics. In areas of the country that have high electrical demand rates, facilities professionals can reduce energy costs by investing in air-conditioning systems that are not driven by electricity. Because absorption machines are driven by steam, waste heat or natural gas, they won't contribute to the facility's electrical demand charges. The savings produced through avoided electrical demand charges can be high enough to justify the system purchase. Utility rebate programs have made the installation of absorption machines even more economical.
Finally, manufacturers have improved the design and operation of absorption machines to the point where today's systems are economical and reliable. New technologies, including the use of microprocessor-based control systems, have eliminated nearly all of the operational problems experienced with absorption units in the past. Look for additional improvements in the near term as manufacturers continue to refine absorption chiller designs.
6-Microprocessor-Based Control Systems
One of the most significant changes in new chiller design is the control system. Gone are the electro-mechanical systems of the past. Today's chiller control systems are almost exclusively microprocessor-based electronic controls.
Microprocessor-based controls offer five major advantages over older generation control systems:
Precision. Microprocessor controls and their electronic components offer far more accuracy and durability than their electro-mechanical counterparts. Although older systems that were in calibration and good working condition could hold temperatures within a few degrees of the desired setpoint, microprocessor-based control systems are accurate to within 0.1 degree. Equally important, the systems can remain this accurate for years. Electro-mechanical controls drift with time and must constantly be adjusted to stay within desired limits.
Point Density. Point density refers to the number of data elements that feed the control systems information. That data then is used to initiate control actions based on a programmed control strategy. The more data elements available, the greater the level of control that can be implemented. In electro-mechanical control systems, the number of data elements is severely limited. Microprocessor-based systems use five to 10 times as many data elements to initiate control actions. The result: Microprocessor-based control systems can initiate more complicated and sophisticated control strategies. Having additional data elements also allows the control system to more closely monitor the operation of the chiller.
Flexibility. Electro-mechanical control systems are limited in their ability to regulate chiller operation. Control actions are based on a limited number of sensor inputs and must be hard wired. Microprocessor-based control systems, with their higher point density and programmable control sequences, allow nearly unlimited, complex control actions. Additionally, changes to control actions require only that the software be modified. Similar changes in electro-mechanical systems require hardware modifications.
Reliability. One of the most serious problems with the older electro-mechanical control systems was their need for maintenance. Controllers were prone to go out of calibration. Relay contact pitted and burned. Pneumatic elements were destroyed by water and dirt in the supply air. Microprocessor-based control systems have eliminated most of these problems simply by reducing the number of mechanical elements in the control system.
Information Feedback. With electro-mechanical control systems, the amount of information the facility staff could get out of the system was limited. Systems generally provided readouts of chilled water supply and return temperatures, condenser water supply and return temperatures, and current draw. But there was no way of determining if the readings were "normal" or if they were changing with time.
With microprocessor-based controls, data tables can be built into the control system. Monitored values are constantly compared to expected ranges. Values falling outside the normal range automatically trigger alarms. Values also can be tracked over time to determine trends and anticipate load changes. Monitoring multiple sensors also provides an accurate depiction of chiller health. Maintenance problems can be diagnosed and corrected early, before they develop into costly problems.
When selecting a chiller for a new or retrofit application, install sufficient capacity to meet the imposed cooling loads. But more than peak loading must be evaluated in selecting a chiller. One must also look at how those loads vary on a daily and seasonal basis. Remember, a typical chiller operates at less than peak loading more than 90 percent of the time.
For applications where the chiller must operate at 50 percent or less of capacity for more than 50 percent of the time, staged reciprocating units or a centrifugal unit equipped with a variable frequency drive (VFD) offer the most energy-efficient solution. If the chiller is to operate at higher loads for longer periods of time, centrifugal or rotary units may be best.
The cooling load profile will help to determine the type of chiller to use, and if single or multiple chillers should be installed. Multiple chiller installations allow facilities professionals to stage their operation to match building loads while keeping the chillers operating at energy-efficient loading.
To a great extent, space availability will dictate the type of chiller to be installed. If space is tight, the rotary chiller might be the best solution. If the space is adjacent to an area where noise is a major concern, reciprocating chillers aren't the best option. Even the high-pitched whine of centrifugal chillers can be a serious problem.
Facilities professionals must take into consideration the maintenance requirements of the different types of chillers and balance them with the capabilities of the maintenance staff. Few organizations have the in-house expertise required to perform much beyond routine chiller maintenance. Most elect to use maintenance contracts for major items. But it's the routine maintenance, or lack of it, that will determine how well the chiller operates. And routine maintenance requirements vary by chiller type.
In general, reciprocating chillers require more routine maintenance than centrifugal chillers; centrifugal chillers require more than rotary chillers. Absorption chiller requirements fall somewhere between those for reciprocating and centrifugal machines.
Finally, facilities professionals must consider the level of service and support that can be provided by the manufacturer in that location. Unfortunately, neither are consistent and can be determined only by trial and error, or by talking with owners of similar systems located close by.
These notes are Written by:
James E. Piper, P.E., Ph.D., is an automation manager, plant maintenance and engineering, at the University of Maryland-College Park. Information was compiled from articles published in May 1993 and January 1995 issues of Building Operating Management.