In the past few decades radiant cooling systems have progressed and are becoming recognized as one of the most comfortable and, from the standpoint of energy use and materials, highly efficient HVAC systems available on the market today.(1) Their efficiency includes not only energy efficiency but materials efficiency as well. As a hydronic system they use approximately 1/2 the horsepower and 1/2 the materials to move heating and cooling energy around a building compared to air systems.
They have developed in Europe during the past several decades and are now just starting to be looked at closely - and initially specified - here in America. A new low flow injection pumping system is now available to make a radiant cooling system even more efficient. The first system of its kind in North America was recently installed in Milton, Ontario, outside of Toronto. The system delivers heating and cooling energy to a variety of terminal units including chilled ceiling panels, chilled beams, fan coils and heat pumps - all in the same piping distribution system despite each of the terminal units requiring different temperatures.
The European Experience
Like many hydronic-based system developments, radiant cooling had its origin in commercial HVAC systems in Western Europe stretching back to the 1980s. Traditionally, European commercial buildings were typically supplied only with heating systems. However, with the introduction of personal computers these buildings now required a measure of sensible cooling.
The Europeans developed chilled radiant ceilings to satisfy this need for indoor cooling. Radiant chilled ceilings consist of metal panels with hydronic tubing attached to the top side of the panels. Chilled water is circulated through the panels to produce radiant and convective cooling. (2) While many European commercial buildings have limited ceiling space, the system requires minimum space since small pipes instead of large ducts are used to transport the cooling energy. Newer chilled panels are now available that use plastic pipe embedded in ceilings, walls or floors.
Approximately 50-60 percent of the heat transfer from a radiant chilled panel is radiant and 40-50 percent is convective. The chilled water temperature must be above dew point to prevent condensation from forming on the underside of the panels. This is in the range of 55º F to 60º F. The driving force or temperature difference between the chilled water and the room is therefore reduced, falling within the range of 15º F to 20º F as opposed to a conventional chilled water system of 30º F to 35º F, using 40º F to 45º F chilled water.
As a result, higher chilled water flow rates are required to achieve reasonable capacities. These flow rates are in the range of 4.5 to 6 gpm per ton using chilled water delta T's of 4º F to 5º F as opposed to conventional chilled water systems of 2 to 3 gpm per ton using delta T's of 8º F to 12º F.
The chilled water flow rate for chilled panels and ceilings is approximately double that of conventional chilled water systems, as shown in Figure 1.
Even with higher flow rates the capacity of radiant chilled panels and ceilings is relatively low, in the range of 20 to 40 btuh/sq. ft. of panel area. While this is within the range of cooling loads for interior spaces, it may not be adequate for exterior spaces. For the European experience in the 1980s some cooling was better than none.
Exterior spaces with larger glass areas approach 60 btuh/sq. ft. of floor area or higher. For full sensible cooling in these exterior spaces either the cooling load must be reduced or the ceiling panels supplemented with other cooling sources. The solar load can be reduced by using shading devices on the windows. This can take the form of interior blinds or exterior sun shades that close when the windows are exposed to direct sunlight. Additional cooling sources could be chilled wall or floor panels. (3)
Because the chilled water supply temperature supplied to a radiant chilled ceiling is above dew point, radiant chilled panels cannot provide latent cooling capacity. However, providing a 100 percent dedicated outside air system (DOAS) to accomplish latent cooling allows the combination decoupled system to provide both sensible and latent capacity. (4) Again, this was not important for the European experience two decades ago, but for full comfort cooling in the United States latent cooling capability is a necessity.
The biggest advantage of decoupling the sensible and latent loads is the substantial reduction in airflows. A typical air-based cooling system will require 7 to 10 air changes per hour of airflow (recirculated and outside air). A radiant cooling system employing a DOAS will only require 1 to 2 air changes per hour (outside air only). This makes for a substantial reduction in the horsepower and materials required to move cooling energy.
Passive and Active Chilled Beams
The Europeans discovered from their experience that by lowering the chilled panel below the ceiling that the convection cooling component of the individual panels could be increased; this satisfied the increased cooling loads from increased use of computers experienced in the 1990s. In addition there was a desire to provide higher cooling capacities for exterior zones to provide better overall comfort.
By lowering the panel below the ceiling the capacity of the chilled panel can be increased to approximately 120 to 150 btuh/sq. ft. of the beam or coil area of the beam. This configuration has been designated as a passive chilled beam by the industry. It resembles a beam when mounted below the ceiling. It is passive since the convective cooling component is natural convection.
The higher cooling capacity of a passive chilled beam could now satisfy increased equipment loads in interior spaces and solar loads in exterior spaces. To increase the cooling capacity of a passive chilled beam even further the conditioned ventilation air from the DOAS can be used to flow air through the chilled coil. This further increases the convective component of the beam. The ventilation air is introduced to the chilled beam through a venturi, generating a higher velocity and subsequently lower pressure region inside the chilled beam. This low pressure region induces room air to flow up through the chilled coil and mix with the primary air from the DOAS. The airflow over the chilled coil is reversed for an active chilled beam vs. a passive chilled beam, with the induced room air now flowing up through the coil.
Active chilled beams are sometimes referred to as "induction diffusers." The air from the active chilled beam is introduced into the space through a slot diffuser, creating a Coanda effect. Inducing warm room air to blow through the chilled coil substantially increases the capacity of the chilled beam. Active chilled beam capacities are in the range of 350 to 600 btuh/sq. ft. of the beam or coil area of the beam. Added to this is the capacity of the primary air from the DOAS. Depending on the temperature and quantity of this primary supply air, this can add up to 300 btuh/sq. ft. of the beam or coil area of the beam. An active chilled beam can deliver from 500 to 900 btuh/sq. ft. of the beam or coil area between the chilled coil and the primary air. An active chilled beam configuration is shown below.
Any of the radiant or chilled beam systems have one thing in common: substantially lower air volumes (1 to 2 air changes per hour vs. 7 to 10). This results in substantially lower horsepower and fewer materials needed to achieve the same or better levels of comfort than air systems such as VAV, rooftop, VVT or CAV.
Figure below shows the reduction in horsepower for a typical all-air VAV system vs. a radiant chilled ceiling or chilled beam system.
The total peak power demand for the radiant cooling system, including the transport as well as generation systems, is almost 25 percent or 1/4 less than an all-air system, a substantial reduction. This includes the 50 percent or 1/2 reduction in energy transport. This is due to the reduction in air quantities needed to accomplish the task. There is, however, an increase in pump horsepower, the result of the radiant cooling system requiring approximately twice as much chilled water flow as a conventional system.
Lowering Peak Power Demand with Injection Pumping
If the chilled water flow for a radiant cooling system could be reduced to that of a conventional system then the peak power demand could be reduced even further. This can be accomplished by the use of injection pumping. Injection pumping has been used for a number of years in radiant heating systems by mixing down the higher temperature boiler water (at 180º F) to that needed for a radiant floor panel (100º to 120º F). This same principlecan be applied to a radiant cooling system only in reverse - to mix up low temperature chilled water (40º F to 45º F) to that required by a chilled ceiling panel or beam (55º F to 60º F).
Shown below is a schematic piping layout for a radiant cooling low flow, low temperature injection piping system.
In this system the primary chiller flow is actually lower than a conventional system. The primary chilled water temperature difference is 16º F (58º - 42º F) or 1.5 gpm per ton. A conventional system is in the range of 8º F to 12º F or 2 to 3 gpm per ton. This is a flow rate 1/2 that of a conventional system and 1/4 of a typical radiant cooling system. This results in a corresponding decrease in pump horsepower and less materials for smaller pipe. This system lowers the transport horsepower to move heating and cooling energy around a building to a minimum.
Another advantage of using low temperature chilled water is the ability to spot dehumidify. The use of a 100 percent DOAS pressurizes the building to negate infiltration of outside air. At building entranceways the amount of natural infiltration will temporarily overwhelm the amount of outside conditioned air delivered by the DOAS when an outside door is opened, especially in humid climates.
To overcome this temporary infiltration overload local dehumidification is needed, and is typically provided by a fan coil installed at building entrances. The fan coil will need chilled water in the range of 50º F maximum to achieve adequate dehumidification. This cannot be supplied by a distribution system using 55º F to 60º F chilled water to radiant panels and chilled beams. Use of an injection pumping system will enable chilled water in this range to be available to the building entranceways by proper layout of the low flow piping system. In addition, operating a chiller at lower temperatures (e.g., 40º F to 45º F) allows the DOAS to use chilled water rather than a DX unit.
A comparison of the peak power demand for the low flow injection pumping system is shown below.
The peak power demand for the low flow, injection pumping radiant cooling system is 33 percent or 1/3 less than an all-air system, again a substantial reduction. This system reduces the transport energy to only 20 percent of the total HVAC system and only 1/3 of the transport energy of an air system. This system, presently under development by Taco, Inc., promises to be one of the most efficient HVAC systems to become available on the U.S. market. It combines hydronic heating and cooling energy transport with injection radiant heating and cooling energy delivery in the conditioned space.
Posko, Chilled Beams in Chicago, Engineered Systems, October 2008.
Mumma and Conroy, Ceiling Radiant Cooling Panels, ASHRAE Transactions, 01-7-5.
Oleson, Radiant Floor Cooling Systems, ASHRAE Journal, September 2008.
Mumma and Jeong, Designing a Dedicated Outside Air System, ASHRAE Journal, October 2006.
Author Info - Greg Cunniff, PE, is Applications Engineering Manager for Taco, Inc. He previously was a partner in a consulting engineering firm, Drapes Engineering, a partner in a manufacturers' representative firm, Vemco, established a temperature control contracting firm, Electro Controls, and his own design-build firm, the Summit Group.