Using chiller plants in humid zones
Mechanical engineers have always been faced with balancing functionality, health/safety codes, reliability, maintainability, and flexibility. Now the race to achieve net-zero facilities results in a variety of energy and sustainable scenarios that change based on each owner’s individual goals and preferences.
This article will review the use of dual chiller plants to achieve energy efficiency for federal office buildings located in humid climates, specifically in Washington, D.C. While humidity is a relative term, the fact remains that this area of the country enjoys four full seasons of varying weather conditions, along with fairly balanced energy price structures. These conditions can extend the payback of renewable energy investments such as wind or solar to well over 25 years. There are few silver bullets available to achieve enhanced energy goals. Thus, an integrated design approach by the team is necessary.
The facility under discussion represents approximately 360,000 gross sq ft of podium (including entries, a breakfast/lunch kitchen, conference facilities, and fitness center); approximately 1.4 million gross sq ft of tenant-specific office space throughout 15- and 17-story towers connected through the 10th floor; and approximately 1.3 million gross sq ft of parking structures.
Energy efficiency for the project began during the proposal phase in 2008. The building request for proposal (RFP) indicated that ASHRAE 90.1, LEED Silver, and the Energy Policy Act of 2005 were part of the contract requirements. The team reviewed energy model results that incorporated various options for exceeding baseline energy performance for the facility. As a strategy the design-build team chose a combination that not only met the facility’s needs but, more importantly, provided the future ability to exceed the minimum energy goals. This decision resulted in a cost-effective upgrade to LEED Gold for the owner during the construction phase of the project.
Southland Industries, as the mechanical engineer of record, developed initial energy models based on a variety of similar RFP criteria. The modeling showed a progression of energy savings with each added feature. For instance, multiple wall R-values were modeled to investigate the point of diminishing return. This was compared to an R-value of virtual infinity, as a super wall. The process continued adding features based on the anticipated level of impact such as windows, roof, and lighting. After minimizing the loads on the facility, multiple central plant options were explored. Typically any option with less than 1% savings or an extended payback of over 20 years was not necessary to meet the energy goals for this project. This process showed that series counter flow chillers (although successful on a previous project) were not beneficial for this facility. We also knew that decoupling the outdoor air for dehumidification and internal loads was a plus, and found that two separate chiller plants maximized savings, creating a 55 F sensible plant and 42 F standard chiller plant.
The energy strategy for the overall project was conveyed through an energy responsibility matrix. This document defined the minimum building envelope values, window assemblies (taking color rendition into account), proposed lighting reduction (including task lighting selection and control), data center plug load management, elevator motor design, and mechanical system selection. This matrix included the owner, general contractor, mechanical and electrical subcontractor responsibilities and goals, along with a “decision date” representing a passing window of opportunity. The “decision date” was important to meet the aggressive project completion schedule. The chillers, for instance, needed to be selected early in the process to be available for rigging with a 17-story crane. The performance criteria for each team member were critical because the success of the whole team depended on all of the interrelated systems in the building. A team member missing a goal early in the process meant that the subsequent systems had to pick up the remaining slack, or jeopardize the desired energy performance goals. The team leaders recognized the importance of each window of opportunity and hosted biweekly meetings to verify conformance.
Owners are encouraged to indicate LEED credits that they are willing to support within the RFP documents; otherwise, the decision date indicated above has passed before the design-build team can read the documents. For instance, daylighting is impossible to achieve without the owner’s support during the tenant fit-out process. Tenants that are accustomed to private offices along perimeter walls and multiple enclosed conference rooms prevent the use of this valuable energy-saving tool. After the building footprint and shell have been established in the design process, it is difficult to capture a perimeter window design that can support day lighting for a high percentage of occupants.
The RFP set the bar for energy performance through the referenced standards and design requirements indicated within the document. After the goals were met in the RFP process and the team was approximately 90% into design, the customer asked why more energy saving options were not incorporated. A simple answer is that additional features typically make a team’s proposal noncompetitive unless it achieves the next level of LEED or a milestone that is desirable. We suspect the reason for waiting until design is substantially complete to ask this type of question is that some projects take years to receive funding so energy rebates and opportunities are not readily apparent, or the building energy manager may not be identified early in the process. Either way, viable energy saving options can be requested and priced as part of the RFP process. Fishing for remote options can be time consuming for the design-build team. It is more cost effective to ask for two additional energy points and allow the team to pick the most
efficient means to achieve the goal.
Chiller plant design
The design team based its energy strategy on the concept of providing a latent and sensible (dual) chiller plant design with the ability to de-energize both chiller plants at ambient temperatures of approximately 50 F and below, and increasing the availability of the full water side economizer.
This direction was chosen because the baseline building used in ASHRAE 90.1 Appendix G is an all-air variable air volume system with water cooled chiller plant. If this is the baseline, then this design needed to exceed the two most dominant energy consumers under that system. Fan energy was addressed first by reducing the facility’s supply air to appropriate ventilation levels. Reducing the distribution airflow saves fan energy but significantly limits the effectiveness of the air side economizer. For the Washington, D.C., climate, an air-side economizer is not the most efficient option and until recently was not required for the entire region under ASHRAE 90.1. The team chose to focus on a water-side economizer, which delivers higher energy savings in office buildings. As a result, the team turned to the second baseline building issue, which is overall
chiller plant performance. By looking at elevated loop temperatures for sensible loop chillers, the economizer hours of operation were increased in conjunction with a reduction in chiller energy consumption. An unexpected reduction in energy efficiency of the latent loop occurred but adjustments to the condenser water temperatures alleviated the problem. This success gave the team confidence to move forward with the decoupled design for the project.
The team initially explored the use of Type II and Type III desiccant wheels to remove moisture and operate with a single sensible loop from the chiller plant. This addressed the baseline building latent chiller efficiency indicated above. Without a continuous supply of recovered steam or hot water to maintain a constant discharge air dew point, the heating requirements to reactivate the wheel were a significant energy burden. Most of the equipment literature is geared toward schools or small office buildings that are looking for load neutral ventilation air and can allow fluctuations in final space relative humidity. For this facility, space dew point drives the sensible chilled water temperature setpoint and resulting terminal equipment capacity. Using a regenerative wheel that cannot guarantee a consistent discharge air dew point temperature will impact the facility’s ability to remove sensible heat from each space. A second limitation is that of wheel capacity at approximately 30,000 cfm based on static pressure, physical
shipping dimensions, and efficiency. With an anticipated DOAS design of approximately 200,000 cfm, multiple wheels would be required for each DOAS unit, resulting in a complex penthouse configuration.
To make the proposed dual plant design possible, all zones and sensible loads (plug, lighting, data center, telecom, and solar) had to be combined with the sensible loop. Remaining loads that need direct and frequent humidity control, such as the cafeteria, loading dock, and kitchen, received standard HVAC equipment with full air side economizer capability. The most demanding solution was for the data and telecom spaces. CRAH unit manufacturers typically do not rate their equipment for nonstandard conditions. Eventually the team found a unit manufacturer that offered to work with them to develop 55 F equipment, using variable speed electronically commutated motor (ECM) direct drive plug fans.
The resulting CRAH units were energy efficient, designed for sensible temperatures, and required an increased footprint to reduce coil static pressure drops. This physical dimension change made sense, since historically CRAH unit casings were designed with a primary purpose of creating a reduced footprint (to save expensive data center space) and the ability to be removed through standard-width doorways. Energy efficiency was not high on the original list of priorities. The maintenance and service department confirmed that 95% of water cooled CRAH units are repaired in place, so size and footprint were not a major concern if service access and clearances are addressed. Even though the energy goals were met, the remaining design team members continued to develop telecom floor plans based on years of layout experience around the less efficient equipment. This increase in the CRAH unit footprint had to be addressed early when space planning was underway before each team member pulled a standard 20 ton CRAH unit footprint from their favorite catalog. With coordination completed and confirmation that the CRAH units could operate with the 55 F sensible loop parameters, we moved forward with the sensible loop design.
Chiller plant equipment selection followed with multiple chiller and tower combinations. The lead engineer, energy engineer, chiller supplier, and tower supplier came together in a collaborative effort (looking at various tower/chiller combinations, various ΔT’s and approach temperatures, different tube bundles, variable frequency drive (VFD) opportunities, etc.). A hyperbola of performance was developed to select the most effective central plant design for this specific facility. The preferred chiller manufacturer provided an option for water cooled VFDs, which were accepted as part of the energy-saving strategy. Knowing the operating goals early in the process helped set parameters for selection during this phase of the design.
The team began with a concept of five 1350 ton, ARI 0.550 kW/ton chillers and five 80 bhp towers (total 400 hp) in 2008 and by early 2009 had refined the plant to:
- Three 1250 ton centrifugal R123 (HCFC) chillers: ARI 0.585; NPLV 0.380
- Two 1250 ton centrifugal R123 (HCFC) chillers: ARI 0.434; NPLV 0.229
- One 2433 ton plate and frame HX for water side economizer
- Eight 40 bhp tower cells with VFD drives (total hp of 320).
With the dual chilled water loop design well under way, systems requiring emergency power still needed special attention. In this case, placing a single sensible or latent plant chiller on emergency power may not be adequate. Small areas intended for emergency operation may require the addition of a dedicated DX DOAS unit and would use the sensible loop as a condenser source for water source heat pumps, or use the latent loop with a blending station for sensible loads. Similar to all emergency power designs, this takes attention to the size of areas served, time frame of operation, and available power for equipment.
Low-approach cooling towers
The cooling tower selection was made using a 4 F approach verses the 7 F, which is typically found in the Washington, D.C., area. This is based on a 78 F design wet bulb temperature and not only maximized the water side economizer and chiller selections, but reduced the overall horsepower by approximately 20%. To achieve this efficiency, both the tower surface area and footprint were increased by approximately 56%.
To support the anticipated tower temperatures, each centrifugal chiller was provided with a head pressure control valve for proper operation in conjunction with the colder water side economizer requirements. The purpose of the valve is to modulate condenser flow in order to maintain the head pressure on the chiller’s compressor, keeping it operational at high efficiency during periods of low condenser temperatures. The preferred solution was to let each chiller modulate its associated control valve to avoid issues with a BAS failure or override. The valve receives signals from the condenser barrel refrigerant gas pressure (head pressure) to keep the chiller operating at peak performance with a variety of condenser water supply temperatures. The valve can also support a cold start after the condenser water has been used during evening hours to generate chilled water via the economizer heat exchanger. Each manufacturer should be consulted for specifics about its equipment and control preferences.
The dedicated outside air equipment was designed to operate without chilled water, when ambient temperatures are approximately 50 F and below. This allows the chilled water system to be de-energized during most of the spring, fall, and winter months, especially during off-peak evening hours. The sensible chilled water loop will supplement the water side economizer as necessary. Both central plant systems are configured with a primary pumping arrangement.
Zone dew point is continuously calculated to analyze the need for latent cooling. Older induction designs worked on these principles under control of pneumatic systems. The worst-case scenario for the building operator is to maintain the latent loop and DOAS, leaving air conditions without reset. With the addition of demand control ventilation, the latent loads can be reduced, which has a dramatic impact on overall energy efficiency for the facility. For this type of design, peak season refers to peak dehumidification cycles such as a humid/foggy August day and not a hot July summer.
The design-build team presented a goal of using the water side economizer to generate sensible chilled water, and allow early balancing of the associated risers and CRAH equipment. This could have worked in winter conditions with lower dew points, but warmer spring weather requires the use of the latent loop to reduce space humidity. If early use of a sensible loop is desired, the design team may need to upgrade the insulation for all distribution piping. Condensation at the equipment must be reviewed and addressed because it is not anticipated for normal operation. Isolation valves can be provided to protect sections of the building, and overall start-up sequences required greater attention. This will require coordination with the general contractor to define areas that are open to ambient conditions during the normal construction process. Temporary construction barriers are not vapor barriers and cannot be expected to deliver humidity control.
Chemical-free condenser water treatment
The condenser water treatment system was first in line to be commissioned because the towers needed time for passivation. Under normal operation, blow-down and drains can discharge to the storm system, in lieu of the sanitary system, thanks to the chemical-free operation. However, the design team provided a temporary sanitary connection so chemicals could be used to accelerate the passivation process. It reduced the time for passivation but still resulted in a four-week process using the water treatment chemical Pre-Q, due to local water quality. The tower fans were not used over the four-week passivation period, which limited the available capacity to 10% of the duty load or approximately 400 tons of cooling. Specific decisions were made regarding where to use this available capacity during the early commissioning process.
Different configurations are often the result of designing and installing chemical-free systems. For this facility, two units in parallel were installed for the entire system. Even though the arrangement functions as designed, individual treatment units at each tower or chiller provide fewer issues with pressure drop in the standard condenser plant configuration.
It is worth noting that the chemical-free system results in different operating sequences for each design. For systems with chemicals, the chillers can remain idle for weeks before the treatment may no longer be effective (i.e., standby equipment, swing seasons when peak loads do not exist, or N+1 equipment). Chemical-free systems only provide protection and treatment when water is moving and evaporating. Having a tower or chiller sit idle for weeks or months will result in little to no protection. As a result, automatic rotation of the equipment on a weekly basis, or a shutdown by removing condenser water from the equipment is required. For this facility, the BAS will rotate a chiller/pump/tower daily. A similar operation was used to treat the plate and frame heat exchanger.
Sensible terminal equipment is not typically a major energy-saving feature for the dollars spent, but can support sensible chilled water systems or demand control ventilation, which significantly enhance the building’s energy efficiency. A variety of methods should be explored to address building latent loads and create a sensible chilled water loop for terminal equipment. The building size, percentage of conference facilities, data center loads, and tenant-specific design parameters typically make the difference between each of the available options. Comparisons to ASHRAE 90.1 Appendix G help target areas that will lead to success. A number of projects have achieved building certifications under previous LEED rating system criteria, with little or no energy credits. The current rating system has closed that loophole and should result in more efficient facilities. The next challenge is to compare initial energy models to the final building results. Despite the fact that the model is not an exact representation of the final metered data, enhanced metering should bring clarity to this issue. As a result, design teams that can deliver a facility at the initial budget and provide an educated view of future energy performance will provide real savings to the end user. The use of dual chilled water loops associated with sensible space cooling is another tool in this effort.
To summarize the design team’s results, anticipated LEED energy savings for the facility are calculated to be 24.66% resulting in 5 credits. This includes loads associated with the full tenant fit-out for the facility.
Winkler is managing principal engineer for the mid-Atlantic division at Southland Industries. He is responsible for planning and design of mechanical systems for educational, office, healthcare, lab, and industrial facilities. He maintains certifications in plumbing engineering (CIPE) and energy management (CEM).
Sustainable features checklist
This project was awarded based on a design-build competition, indicating that a dual plant design did not impact the team’s ability to generate a competitive proposal. The following sustainable features were part of the proposal for the facility:
- Sensible loop (55 LWT) centrifugal chillers
- Latent loop (42 LWT) centrifugal chillers
- A sensible/chilled water stand-by chiller
- Primary variable flow pumping for each chilled water loop
- Induced draft counter flow cooling towers based on a 4 F approach and oversized box to reduce fan energy
- Chemical-free water treatment system
- A plate and frame water side economizer heat exchanger
- Dedicated outdoor air units (DOAS), with total enthalpy recovery wheels
- Sensible chilled water cooling to each space via terminal units (similar design parameters as chilled ceilings or chilled beams)
- Sensible chilled water vertical computer room air handling (CRAH) units
- Demand control ventilation (DCV) throughout the building.
For this project, “appropriate ventilation” levels needed to be evaluated in three ways:
1. The applicable IMC code requires 20 cfm/p for office space.
2. LEED references ASHRAE 62.1, which uses a combination of people and area outdoor air requirements, and is typically less than 20 cfm/p.
3. Finally, our limiting factor, which was humidity removal from the space and could be met with 20 cfm/p. This selection drove our final DOAS coil and latent chilled water design parameters.
Condenser blowdown and discharge into the storm system sounds straightforward to the mechanical design engineer. For this facility, we are using a siphonic roof drainage system for the 15- and 17-story towers. Point loads are not easily accepted by the typical siphonic design and need to be coordinated with the plumbing designer early in the process. Express risers are typically required, which can be easily added during the design phase of the project.