Typical building tertiary pumps (one in operation and one standby).
Harvard University is serious about energy and environmentally responsible building systems. In last month's issue, Harvard's School of Public Health demonstrated just how serious this commitment is as actions speak louder than words. Harvard Business School (HBS) took a somewhat different engineering/environmental course when the firm responsible for facility management at HBS, GCA Services Group, Inc., initiated their innovative approach to energy conservation: hydraulic modeling of the campus chilled water distribution system. Doug Scatterday, director, facilities operation; Chuck Stronach, assistant director, facilities operation; Jason Munro, assistant manager of buildings and trades; and member of GCA Services Group, Inc., contracted Richard D. Kimball Inc. (RDK) to study the campus chilled water system for potential energy conservation and operating savings.

The timing for the study was just the next step by HBS to enhance its central plant chilled water management plan. At the same time, this latest action went beyond typical master planning procedures, using hydraulic modeling to begin the process of analytically assessing the system's existing conditions, potential retrofit opportunities, and possible future system expansion needs.

Hydraulic modeling provided the facilities operations department with the ability to troubleshoot the campus' chronic chilled water problem with the existing primary-secondary-tertiary pumping system.

Inherently, this type of system is difficult to operate efficiently or cost effectively while meeting space comfort. Anticipating the hydraulic model solution would provide them with valuable information on how the system functioned systemwide, the HBS/RDK team was anxious to consider the possibility of eliminating 26 tertiary pumps without compromising chilled water distribution or capacity performance. If successful, a retrofit could allow them to dramatically improve the system configuration, with a focus on bypassing existing tertiary pumps and avoiding the cost of adding new tertiary pumps in the face of future building growth at the 1.5 million-sq-ft campus.

A Benchmark

A telltale sign that energy consumption could be saved with this 24/7 distribution system was to benchmark current pumping horsepower divided by tons cooling. The rule of thumb used by RDK was 0.05 hp/ton to 0.1 hp/ton for peak pumping performance. At HBS, the system was operating at 0.175 hp/ton or approximately twice the pump energy estimated to operate this 30-building site. This rule of thumb was the first technical indicator to reinforce the HBS' belief that pumping capabilities were excessive (pressure and horsepower) and that energy consumption could be reduced. How much of a horsepower reduction remained to be assessed.

A second sign of inefficiency at HBS was the primary-secondary-tertiary distribution's inability to maximize chilled water differential of 11°F (supply temperature at 45° and return temperature of 56°). With the current installation, the primary pumping is constant flow, secondary pumping is variable speed/variable flow, 60% of the tertiary pumps at constant flow, and 40% of the tertiary pumps with variable speed/variable flow. At peak chilled water performance, the energy profile in 2004 was:

  • 110 hp (2 to 40 hp and 1 to 30 hp) constant flow, primary pumping;
  • 400 hp (2 to 200 hp) variable speed at 65%~ secondary pumping;
  • 100 hp (six pumps) variable speed at 65%~ tertiary pumping; and
  • 160 hp (20 pumps) constant flow, tertiary pumping.

This system configuration, like so many other primary-secondary-tertiary pump arrangements, is analogous to keeping a Grand Prix race car finely tuned and operating efficiently and effectively 24/7. Continuous success is limited at best, and inefficiency is inherent. When HBS considered building expansion, the existing chilled water system configuration could compound this dilemma of primary-secondary-tertiary pumping. A budget estimate for adding a new tertiary pump is to figure a furnished and installed cost of approximately $10,000 for a 4-in. piping system and $15,000 for a larger piping system. Double this price when a standby pump is included, along with the building program scope of work. Another issue associated with tertiary pumping is the space these pumps take up within a building (approximately 100 sq ft) that can add another $10,000 (at $100/ sq ft) in construction cost to a new project.

Annual maintenance costs associated with preventive maintenance can be estimated at approximately $400 per installation. Over a 20-yr period, these tertiary pump installations can be quite costly, not to mention the inefficiencies associated with primary-secondary-tertiary energy consumption.

For new construction, the value engineering benefit of studying pumping configuration options using hydraulic model software can be a three-month ROI. At HBS, and more specifically at the facility operation group, you could say they inherited the primary-secondary-tertiary configuration as the chilled water system has grown over the past 20 years. As a course correction, HBS chose to seek out potential alternatives to the existing conditions that could potentially be cost effective enough to abandon the primary-secondary-tertiary configuration. Their motivation for hydraulic modeling software was to correct past engineered solutions so that the campus chilled water system would be easier to operate, maintain, and, most importantly, save money and energy in future years.

HBS Identifies The Opportunities

Prior to contracting RDK, HBS facility operation initiated operating parameter adjustments to increase the chilled water ΔT (temperature differential between supply and return water), which would increase the chiller equipment energy performance. With a larger differential, the chillers would improve/lower their kW/ton consumption. In addition, they also knew that larger Δ equated to reduced flow (gpm) that would equate to reduced pump energy consumption. Their dilemma was that these changes had little influence on reducing tertiary pumping with 60% of those pumps operating at constant flow.

While concentrating on central plant performance, the HBS team began to be suspicious of the need to have tertiary pumps. Through a trial and error effort of installing a bypass line around a tertiary pump, they found that it might be possible to eliminate the tertiary pump configuration. What was not known was whether this approach could be implemented throughout the campus installation. Also unknown was whether there were any negative affects to this approach, and whether the facility group could continue to provide adequate cooling capacity to each building.

While striving to improve system performance, HBS had a much larger issue on the horizon: proactively anticipating any possible future growth. As a result, the school's master plan will have an influence on the existing central chiller plant and eliminating tertiary questions as to whether these pumps would be needed again in the future. The ability to forecast what would be needed from the central plant consisted of the following:

  • New chiller(s) will be needed based on current capacity demands.
  • What will the optimum future chilled water capacity be in both summer and winter?
  • Increase in chiller capacity will influence a need to increase chilled water flow.
  • What will be the potential to increase chilled water ΔT to offset and/or compensate for increased flow?
  • Can existing building equipment and their original chilled water ΔT accommodate larger ΔT?
  • What are the limits on the chilled water system infrastructure existing pipe sizes?
  • How will chilled water management strategy change?
  • Operating cost associated with central plant growth will increase in energy consumption and in PM.
  • Electrical power infrastructure will be impacted by central plant growth.

What was needed by HBS was a means to assess various central plant and system distribution configuration scenarios. This would mean an engineering study using life-cycle analysis software capable of effectively incorporating "what if" scenarios to consider numerous alternatives and design parameters.

While this portion of the master planning is yet to be completed, the trial and error chilled water distribution efforts to date were worth the time, but a better, more analytical approach to piping design was needed. Hydraulic modeling software would be the answer for this portion of the central plant master planning questions. Once the modeling is completed, HBS will be able to refocus on the central chiller plant design to come up with solutions while studying and deciding on the optimum chiller configuration:

  • Continue with parallel chiller configuration;
  • Consider chiller(s) in series with parallel;
  • Continue with electric centrifugal chillers;
  • Consider combination of electric with absorption; and
  • Consider thermal storage.

Whatever the chiller configuration, the pumping configuration will be the equipment delivering the cooling capacity. Collectively, the optimum solution will provide a sustainable chilled water management process. Without the optimum pumping strategy, the optimum chiller strategy will be compromised and energy and the environment will also be compromised.

Baker Hall Library, overlooking the Charles River. Potential future growth of the campus caused officials to overhaul its chilled water system.

Hydraulic Modeling Software

Hydraulic modeling software provides the designer with the ability to take a snapshot of system performance at any pre-selected point in time. The software is a computerized representation of pipe distribution that can capture the system at peak performance, as well as at partial load performance. The program can simulate a single pump operating up and down its pump curve at constant speed or variable speed, as well as capture all the pertinent engineering data and system pressure drop along the way, with pumps in parallel and/or in series based on system configuration.

This computerized business tool can assist a design engineer in predicting system behavior and evaluating capacity additions. The model would be able to assist in making some technical decisions in the coming years as they pertain to possible capacity additions.

For the facility manager, this is a troubleshooting program to assess potential problematic issues including low system Δ or high system pressure. At HBS, both of these issues were known and an immediate concern as well as a long-term problem. RDK, working with the HBS team, would implement a software simulation that would clear up what was happening throughout the campus and building distribution.

The other benefit of this software modeling is its ability to accommodate potential reductions in operating costs. The HBS/RDK team effort would reap significant operating, economic, and energy information through the use of hydraulic modeling, as the results would confirm the HBS belief that the tertiary pumps could be taken offline without compromising chilled water capacity. The simulation printout documented the elimination of tertiary pumps equating to 260 hp. Based on 24/7 operation in the air conditioning season, and an energy cost of approximately $0.09/ kWh, the savings will total approximately $38,000 in annual utility costs. The estimated savings should be further improved upon once the system has been retrofitted to bypass the tertiary pumps as the two 200-hp secondary pumps, with their VSDs, carry the secondary and tertiary load, and further reduce the estimated total energy consumption. An additional benefit from the program results was the confirmation that future construction projects would not require the addition of tertiary pumps.

The team also simulated a second alternative/option, which was to remove the primary chilled water pumps, making the chilled water distribution a primary variable flow option. Based on the software model, the results showed that the 110-hp primary pumping could be eliminated in favor of the secondary pumps doing the work, with the assistance of one 25-hp tertiary pump at one of the furthest points in the chilled water system to provide a boost in the overall pumping performance.

While this second option of the software simulation appeared to be of interest, based on the future growth of the central plant capacity, adding chillers will require additional pumping performance. It was determined that the cost to retrofit the primary-secondary to primary variable flow would not be a value-added decision, knowing the system will be retrofitted again in the next couple of years to add one or more chillers.

The Test of Efficiency

With the knowledge that the chilled water distribution system can function without the tertiary pumps and the system pump pressure will be less, phase two of this process will be to retrofit the distribution to accommodate primary-secondary pumping. The scope of work to complete this effort will include pump bypass piping and rebalancing the water flows to each building followed by a retrocommissioning effort to tune up the chilled water management strategy. Once implemented, energy metering can be used to monitor, measure, and benchmark current energy consumption to past utility bills. Monitoring and measuring the chilled water ΔT will also be part of the enhanced chilled water management plan.

Once the water system is rebalanced, the software model can be updated to reflect current operating conditions in flow, temperature, and pressure drop. With real operating data, HBS will be able to continue to look for opportunities to increase system performance while reducing energy consumption. At the same time, HBS will have far more accurate operating data to make the annual operating decisions, as well as contribute to the school's master plan.

The existing chilled water system with primary pumps at constant volume, secondary pumps with VSDs in parallel, and campus tertiary pumps: one in operation and one is standby.


With 20/20 hindsight and new technology, what was considered a great design several years ago is not necessarily a good design today. Technology is not the only thing that has changed over the years. At HBS, the campus has changed dramatically over the past 10 to 15 years. It also requires 24/7 operation to accommodate its Executive Education program, where as just 10 years ago, the campus operation had the summer to make changes, improve system performance, and replace equipment without interfering with the academic program.

An immediate benefit of the hydraulic model results has been to eliminate the proposed new tertiary pumping configuration for the school's current LEED™ Silver certification Hamilton Hall renovation project. Still in the design phase, this chilled water pump/standby pump deletion will save the project approximately $20,000. In turn, this savings will more than pay the hydraulic model project cost, a real win-win investment for HBS.

HBS continues to look to the future, with both near and long-range master planning demands. For the operation group, this vision is just part of the challenge they face with day-to-day O&M needs. This can be a balancing act between efficient management and operation while continuing to provide building occupancy space comfort. It is this commitment to do the very best job possible that has pushed HBS to think "outside-the-box" for creative solutions.

Hydraulic modeling, which is becoming better known as a computer technology, continues to grow in the building industry. This computerized process can assist the design engineer and the facility engineer. At HBS, the software simulation confirmed what was speculated, which was that the campus distribution didn't require 26 tertiary pumps totaling 260 hp. It also helped to raise awareness about what else could be done to improve system performance while contributing to energy conservation and the environment. Hydraulic modeling is a computerized business tool that, as a rule will probably pay for itself in less than one year. At HBS, the payback was immediate, with the influence it had on the Hamilton Hall LEED project. ES