Designing an optimum heating-cooling scheme for the new Molson Centre in Montreal, Canada certainly had some very interesting diversity and load profiles to consider. Hockey games in the winter and concerts in the summer provided a definite challenge for the optimization of equipment used in the heating ventilating and air conditioning (HVAC) systems.

With relatively large equipment required to handle peak loads, the engineers needed to focus on the comfort, maintenance and operating costs during the significant number of hours the building was influenced by partial load conditions. All components used to deliver heating and cooling to the building were closely scrutinized to provide the best and most cost effective overall solution.

The philosophy chosen for the three heating and one cooling distribution systems employed the popular primary-secondary principle (shown in Figure 1).

This layout allowed the "splitting" of the primary boiler and chiller equipment, in addition to the pumping systems expected to distribute this energy to the building. Splitting the equipment naturally allows a more accurate match of "production source" to building "load," with smaller components operating at higher efficiency levels.

Constant vs. Variable

An important factor in using this principle, especially where chilled water is concerned, is the provision of an environment whereby the production of chilled water can use constant volume with the distribution system using varying volume. In the case of the Molson Centre, five chillers were used, each with its own primary pump. As a chiller was staged on, its primary pump was started in unison.

This operation guaranteed that the chiller, once operating, received its full design flow. Using the primary-secondary principle allowed the use of two-way valve controlled, varying volume pumping in the distribution circuit. Here lies another opportunity to ‘split’ demand into a number of parallel running pumps. The chilled water system flow rate in this case totaled 6,900 US GPM.

How many pumps should be run in parallel? The answer, in part, depends on the diversity expected in the building load demand. Using Bell & Gossett’s ESP-PLUS software program to calculate annual operating costs for multiple pumping scenarios drastically reduced the complexity of this task.

Dividing the secondary system design flow rate between two-through-six pumps and then modeling operating costs against the building load profile is simple. The idea here is to minimize pumping costs by running fewer pumps to match the load conditions at any given time.

After all, the pump that is going to save the most money is the one not running at all.

While the focus is on saving operating costs, why not consider the option of taking advantage of the pump affinity laws, accounting for speed reduction. Reducing the speed of a centrifugal pump greatly reduces the cost of its operation. The relationship between pump speed and energy (BHP) consumed is a cubed function (as shown in Figure 2).

Figure 2

Speed Reduction

If the speed of a pump can be reduced to 50% to respond to half load conditions, the power consumption can be reduced to approximately one-eighth of a pump running at 100%. With this new dimension included in the process, it is now possible to incorporate speed reduction into the multiple parallel running pump analysis.

The goal is to optimize operating costs further by reducing the speed accordingly of the pumps that are on. An added feature of ESP-PLUS produces very comprehensive operating costs for this parallel variable speed pumping scenario. Data derived from the exercise includes speed, staging and wire-to-water efficiency calculations, among others.

In the case of the Molson Centre, it was calculated, based on five cents per kilowatt-hour (5c/kW hr), that approximately $18,000 per year of operating cost savings could be realized by running three, staged, equal-sized, variable speed pumps in parallel.

Besides the distinct economical reasons for variable speed pumping, other positive consequences result:

1. Two-way control valves do not have to oppose the force of increased differential pressure imposed by the ‘rising to shut-off head’ characteristics of secondary pumps.

2. Improved control of flow through two-way valves and terminal units.

3. Rotating components of pumps running at reduced speed prolong their service life.

4. Ramping up of pump speed gradually provides an inherent ‘soft start’ for the system, both electrically and hydraulically.

Variable Speed Design

The three variable speed parallel pumps proved to be a viable solution for the large secondary distribution systems in this building. Running three variable speed pumps in parallel is not a simple process. One must be fully aware as to how a centrifugal pump reacts to ‘system’ changes. Speed changes made to a centrifugal pump in an environment of non-compressible fluid and extremely fast rates of change in system pressure can result in unpredictable outcomes.

It is very important to associate the operation of the varying speed pumps with the "system resistance curve."

This curve represents the varying system friction losses that become the very reason for our speed reduction control philosophy. In addition to optimizing pump speed, the control mechanism must also prevent potentially damaging conditions, such as hunting, surging and operation of pumps past their "end of curve" point.

An option available to the design group for the Molson Centre was the Bell & Gossett Technologic Pump Controller. This application-specific controller provided an instant answer to the complex control needs for the secondary distribution pumping systems. This controller not only maximizes the potential of pump speed reduction, but also eliminates the potentially damaging conditions indicated above.

The new schematic layout, incorporating all control mechanisms, is shown in figure 3. This typical layout indicates the major components of a variable speed pumping system, a sensing element, feedback signal, pump controller, adjustable frequency drive and, of course, the heart of the system - the pump. It is critical that all of these components be matched precisely to allow the full benefits of variable speed pumping. Obviously, the pump saving the most energy is the one turned "off." The Technologic Controller runs only as many variable speed pumps as necessary to match the load conditions at any point in time. This means the algorithm used must maintain adequate system flow in addition to eliminating over -and under- pumping during dynamically changing pump-staging points. It must also scan all sensing points (six in the case of the chilled water system) at an adequately fast rate to ensure that system information is updated almost immediately.

To ensure that the building management system was always informed, status points were connected to the controller to send back needed information. All of these features arrived at the jobsite ready to perform. The supply of matched components eliminates "finger pointing" and guarantees the instant, maximum benefits of variable-speed pump operation. System piping layout was simplified with the use of Bell and Gossett VSC Series pumps with their vertical suction and discharge connections. These pumps minimized the installed footprint and allowed maximum optimization of floor space in the mechanical rooms. The fact that all of these system components came from a single source helped ensure the benefits of pumping-strategy optimization requested by the design team.

This article originally appeared in the March 1998 issue of Mechanical Buyer & Specifier, and will appear in a future issue of Consulting-Specifying Engineer and Energy User News.

Reprinted from TechTalk September 1998
Copyright 1998 by ITT Industries