Answering the Central Question

Examples of centralized and distributed strategies permeate our environment. Even our own bodies are arranged with a central cardiopulmonary apparatus, widely distributed sweat glands and regional lymph nodes.

By George P. Karidis, P.E. and John D. Richards, P.E. SmithGroup Detroit November 1, 2000

Examples of centralized and distributed strategies permeate our environment. Even our own bodies are arranged with a central cardiopulmonary apparatus, widely distributed sweat glands and regional lymph nodes. In a broad sense, why choose one approach over the other?

Narrowing the query to our built environment, when is a set of buildings better served by a central plant for heating, cooling or other services than by distributed equipment , and why ? The answer must consider the influence of loads, operations and costs on a particular project, as well as the various system options and planning strategies.

Load-related influences

One of the biggest influences favoring a central plant is a site with a high load density , as is often seen in campus settings such as universities, government facilities and hospitals, or in office and industrial complexes and high-rise urban districts. A high load density means a less extensive distribution system and shorter piping runs. This lowers the initial cost of the site distribution system, which often represents the lion’s share of construction costs. Shorter runs also minimize thermal and pressure losses and maintenance costs.

A desirable companion to high load density is a favorable load factor . Generally, this means that the aggregate load over time tends to approach the peak block-load condition. A favorable load factor can aid central-plant feasibility by leveraging identified operational advantages over a longer period of time or a higher percentage of full load, or both. An important special case of favorable load factor is a daily load profile with a short, prominent peak period, which may be suited for a thermal-storage strategy.

It is the interaction, or product, of these two load-based influences that best favors a central plant arrangement. For example, Detroit’s Cobo Center houses a 4,000-ton river water-cooled chiller plant that serves the Center’s convention space and 10,000-seat multiuse arena, as well as Joe Louis Arena, home of the Detroit Red Wings; Hart Plaza, a summer festival venue; and the UAW-Ford National Programs Center, an office/training building. This set of adjacent buildings offers a high chilled-water load density. Though the load factor for most of them is very poor, combining the diverse set matches off-peak arena profiles with on-peak convention/office profiles, dramatically improving the load factor while reducing the required chiller capacity.

Operation and first-cost influences

In addition to the load-density/load-factor pairing, a second set of major influences is the interplay of operational advantages and the first cost of construction.

Operational advantages involve energy, maintenance, reliability, redundancy, peak shifting or any other aspect of operating a set of buildings.

The first-cost outlook for building a central plant anticipates the necessary level of initial investment within the context of an owner’s funding availability. This investment takes credit for all the costs avoided by not installing distributed equipment. It does not include future plant-expansion financing.

Since funding the initial construction can be the greatest hurdle to a central plant’s financial viability, an owner’s set of values and financial perspective can have a pervasive influence in assessing operational advantages and the first-cost outlook. A building developer may consider only strategies within a close investment horizon, while a government agency may consider solutions projected over a 40-year life cycle. A large corporation may require a minimum internal rate of return, whereas a university may seek opportunities to finance the first step.

As with load influences, it is the interaction, or product, of operational advantages and first-cost outlook which favors a central plant over a distributed approach. This interaction can make even a site with moderate load density and load factor suitable for a central plant. The U.S. Navy’s Bachelor’s Enlisted Quarters at the Great Lakes Naval Station, Ill.-a multiple-building residential complex-uses centrally-located, high-efficiency centrifugal chillers in lieu of less efficient air-cooled machines at each building. The results are significantly improved energy consumption, a lower total installed capacity and reduced maintenance-characteristics in line with the owner’s perspective.

Central-system alternatives

The utility services most often suited to central generation include space or process heating and cooling, electrical service and, in some industrial applications, compressed air or domestic hot water. However, all potential utility services can initially be explored.

For heating and cooling systems, the design choices to consider include the heat-transfer medium, how it is distributed and what generation equipment will produce it.

High-temperature heating hot-water systems allow greater temperature differences and thus smaller piping than low-temperature systems. In addition, the upper portion of the differential temperature can generate steam. At Evans Army Community Hospital in Fort Carson, Colo., high-temperature water from a central plant is cascaded in series, first to 60-pounds-per-square-inch (psi) steam generators, then to 15-psi generators and finally to all low-temperature needs-thus managing the highest quality of heat for its highest use.

Condensing steam conveys more heat per pound of mass flow than hot water, though at lower density and higher velocity. By increasing the steam pressure, density and temperature are increased, allowing a smaller pipe size to be used. Steam systems typically require more maintenance than heating hot-water systems, due to the requirements for proper steam trapping and condensate collection and return. In fact, some manufacturing owners, faced with the cost of labor and escalating maintenance costs for aging steam systems, have opted to decommission steam plants in favor of direct-fired gas equipment where appropriately high ventilation rates exist.

Available chilled-water mediums include water and either brine or inhibited glycol solutions-which offer freeze protection but increase fluid viscosity and decrease heat-transfer performance. Central-plant chilled-water system design should maximize the differential temperature to minimize pipe size and pump energy. However, this requires diligence in minimizing user flow under all conditions, and in maximizing installed heat-transfer capacity at all user locations.

Piping-distribution options include direct buried, direct buried in double-wall piping, shallow concrete trenches, fully accessible tunnels and aboveground trestles. The considerations when selecting the piping location include first cost, aesthetics, insulation type, system life and ease of maintenance. Both chilled- and hot-water systems should engage primary-secondary variable-flow pumping with consideration for distributed tertiary pumping. Individual building users may be served directly or indirectly through a heat exchanger-typically plate-and-frame type for chilled-water systems and shell-and-tube type for steam or high-temperature heating hot-water systems. Local metering of total flow and British thermal units (Btus) offers monitoring and the ability to allocate energy costs.

Generation equipment

The selection of generation equipment requires careful consideration of first costs as well as current and projected utility rates. Other essential factors include maintainability, durability and operational requirements such as load matching and fuel-source flexibility. Boilers may be arranged for dual-fuel or tri-fuel firing, including natural gas, multiple grades of oil, coal or even landfill gas. Central-plant centrifugal chillers may be driven by electric motors or natural-gas or diesel engines, and may employ alternate refrigerants such as ammonia. Absorption chiller options include steam or natural-gas firing. Employing a combination of chiller types may offer significant operational benefits and greater fuel-source flexibility.

In addition to multiple fuel sources, many plants employ thermal storage and cogeneration to help respond to an unpredictable energy market. Ice or chilled-water storage offers operating savings by generating chilled storage during the utility’s off-peak rates. Due to greater space requirements and complex piping arrangements, chilled-water storage systems are well suited for central plants. At the Veterans’ Administration Hospital in Detroit (pictured on page 29), an installed chiller capacity of 5,500 tons would have been substantially larger if chilled-water storage were not incorporated into the central-plant design. Besides energy cost savings, this resulted in significant first-cost savings for the chiller capacity and attendant electrical service.

Cogeneration alternatives include collecting waste heat from electrical generators or engine-driven chillers. This may provide low-pressure steam or low-temperature heating hot water for district distribution and for preheating boiler feed-water or combustion air in the central plant.

Funding strategies

Getting funding for the initial infrastructure construction is critical for many owners. Ideally, the initial plant construction would include sufficient infrastructure to support at least two generators for each utility provided. The distribution system would cover a substantial portion of the intended site, allowing flexibility for adjustments in the master site plan as development progressed.

In North Carolina, Research Triangle Park’s initial build for chilled water service included three miles of distribution, 36-inch-diameter mains and a 5,000-ton chiller plant expandable to 30,000 tons (see diagram at right). While a substantial initial phase carries a high price tag, it can provide a catalyst for robust development.

If low first cost is an owner priority-but strong load and operational influences favor a central plant-then third-party ownership or financing of a central plant may make sense. Many companies will build and operate a central utility generation facility for an owner. These facilities provide chilled water or steam to the owner’s buildings, similar to the manner in which electricity or natural gas is brought to the facility. Sometimes the generation facility is located on the owner’s site, though often-especially in urban areas-the generation facility is located on the service provider’s property. When a third party operates the generation facility, the first cost of the plant is saved, but the advantages of the central plant are not sacrificed. The first cost of the plant is generally paid for over a period of time as part of the provider’s contract with the owner. Many urban generation systems operate in this manner.

Once an initial phase establishes plant and distribution infrastructure to serve a number of buildings, funding for subsequent phases can be allocated more easily to subsequent building projects. This may involve incremental growth of plant bays and equipment, possibly with distribution line extensions as incremental steps. Eventually, a second plant site may be planned at the opposite end of a site loop (assuming constant-diameter piping), possibly doubling the capacity of a given loop-pipe size. Adding thermal storage can be one of the incremental phases-perhaps as early as the second-and possibly at another location on the site, static head and site topography permitting.

Planning strategies

Ultimate size: Central plants can reach a maximum feasible size, limited by site constraints, piping size, static pressure considerations and dynamic losses. At Michigan State University in East Lansing, a regional chilled-water plant was evaluated for ways to expand capacity and determine how best to serve the new Biophysical Sciences Building. While various options of extending and enhancing the plant were discovered, it proved to be in the university’s best interest to build a new plant with the new building. This plant also picked up some load from other existing buildings, thus preserving and increasing the untapped capacity of the existing chilled-water plant for other regional expansion.

Equipment interconnection: By extending piping between two or more equipment installations, each can back up the other, with more efficient operation under various load levels. Such interconnection efforts may face hydraulic, static pressure or pumping arrangement issues. With the construction of St. Jude Children’s Hospital’s Integrated Research Center, Memphis, Tenn.-which was too tall to be served directly by an existing chilled-water system-so new chiller equipment was connected with an existing site system to allow two-way back up. One of the new chillers may be valved to serve the lower pressure site system directly, or the site system may supplement the new building’s load via a plate-and frame heat exchanger. Interconnection may also be favorable if schedule or cash- flow constraints prevent building a central plant with the first buildings on a site. With one chiller or boiler in each building instead of two, savings from simpler installations should help offset site piping costs.

Ultrahigh reliability: With certain mission-critical corporate or government facilities, reliability rules. Even “N+2” redundancy in equipment counts may not be enough, for there are certain risks inherent in a central plant system that may not exist with numerous independent pieces of equipment locally backing each other up. In very critical instances, redundant headers with alternate routes may be employed to achieve a better level of service availability (e.g. #0.9999). On-site generation may also be considered, with or without cogeneration. A variation of this strategy would apply large gas-turbine generators to back up an entire site-increasing reliability while creating a valuable peak-shifting or load-shedding opportunity in a deregulated environment.

Space availability: Many campus or urban settings impose draconian site constraints on the location of new plant equipment. On the other hand, finding suitable locations in each building for equipment and its intake/discharge and access/replacement may be just as challenging. The University of Michigan’s Palmer Drive Development in Ann Arbor found an opportunity to build the initial phase of a regional chiller plant beneath one of the new buildings in the project. By providing space in the plant for additional chillers, it will be able to serve additional buildings in the future. Another space availability issue may exist with utility right-of-way and site-disruption issues, particularly in urban district systems.

Evaluation considerations

Many factors converge in the evaluation of central plant versus distributed equipment approaches, including economic analysis, plant ownership, timing constraints, organizational leadership, politics and the environment.

The economic analysis requires an accurate determination of first cost and operational costs for each proposed option. This is critical to making a proper evaluation and final decision, both for the overall approach and system choices. Items necessary for a life-cycle cost evaluation include first costs, utility costs, maintenance costs, operational costs, utility escalation rates and the owner’s cost of money.

Utility costs generally play a large role in determining the outcome of the study. Local utilities should be contacted, and all rate options understood. An advanced computer energy simulation can determine utility usage for many different equipment and utility combinations, matching operating parameters as closely as possible. With the deregulation of natural gas and electricity and the proliferation of utility service options, purchasing energy from an alternate provider and paying local transportation charges may prove to be a benefit.

Maintenance costs should be included for all of the equipment options, including annual and periodic maintenance, such as the rebuilding of an engine drive for a chiller.

Operational expenses include the cost of personnel to run the central plant, insurance, chemicals and other annual recurring costs. In some areas, local or state codes may require licensed boiler and/or chiller operators to be on duty 24 hours a day when the machines are operating. Most codes have minimum requirements that require operators for equipment over a given Btu output or square footage of heat transfer area. This requirement may burden a central plant with a very large annual operating expense that distributed plants may avoid.

Balancing the decision

Both centralized plant and distributed equipment strategies prove their viability in various building examples. Central plants can offer operational advantages for multiple building projects having a high load density, a favorable load factor and a matching first-cost outlook. The analysis preceding an owner’s final decision may be lengthy, but should address system alternatives, planning strategies and evaluation considerations.

So why does the human body have a centralized cardiopulmonary system? A look at its load influences reveals that it handles the largest volumetric processes in the human body-nonstop. For operational advantages, circulating one fluid accomplishes oxygenation and a host of other functions. And the first-cost outlook is feasible via the much-admired embryological project delivery system. In contrast, the body’s evaporative cooling system is simple and effective just as it is-covering our widely distributed epidermis with a widely distributed approach. As to why humans have regionally arranged lymph nodes.consult a physician.

About the authors

Serving clients out of SmithGroup’s Detroit office, George P. Karidis, P.E., is a vice president and director of mechanical engineering, and John D. Richards, P.E., is a principal and chief mechanical engineer of the technologies practice group. SmithGroup, not included in Consulting-Specifying Engineer’s August 2000 Giants survey, billed $35.4 million last year in mechanical and electrical engineering design work, which would have ranked the firm in the top 20 of the survey.