A Private Plan

By Leonard W. White, P.E., RCDD, Senior Principal, Stanford White Assocs. Consulting Engineers, Raleigh, N.C. August 1, 2006

Your client is a major university hospital: a campus of nearly 60 buildings consuming almost 7,500,000 sq. ft. It wants a new fire alarm system. Where do you start?

This was exactly the dilemma our firm faced in designing a new fire alarm system for Duke University Medical Center campus in Durham, N.C. To begin, it’s necessary to identify the most important buildings and then create a reasonable scope. In this case, target No. 1 was Duke Hospital North, or simply Duke Hospital, a 70’s-era high-rise hospital that had evolved into a complex of five major buildings — including a children’s hospital — that took up more than 1,500,000 sq. ft.

As the hospital grew, each new building was provided with an independent fire alarm system, all of which reported to a 24/7 central building automation system. However, with the aging of the buildings and their embedded fire alarm systems, it became apparent that the original systems could not easily be modified or expanded to meet the ever-increasing alarm-notification requirements of codes, nor would the existing systems provide the pinpoint event-location ability that is a built-in advantage of a modern intelligent device-addressable fire alarm system. Furthermore, repair parts were becoming more difficult to obtain. Thus, the decision was made to replace the entire system in all the hospital buildings with a new system that would protect the facilities, serve as a central monitor point for other existing medical center fire alarm systems and form a stable base for additional fire alarm system upgrades on the campus.

Plan of action

To get a handle on such a massive undertaking, it’s best to focus on a few key areas and at the onset of the construction process, take a critical look at specifications and contract documents.

With constant, unrelated renovations scheduled throughout a 2.5-year construction period, a mechanism was necessary to contain the cost of change orders due to the changing scope of work defined by the contract documents. In this case, the M/E team was in a favorable position in that this was a private project. Thus, the only restrictions on equipment specification were the suitability of the equipment for the application and the usual life-cycle cost evaluations.

This proved a boon for this particular job. If you’ll pardon a digression, the state of North Carolina, like most jurisdictions, has a set of requirements intended to give all equipment manufacturers equal access to the bidding process. In fact, North Carolina’s requirements [GS: 133.3] are very open in that they require designers to specify products by using performance-acceptance criterion, where practical definitions are provided as opposed to generating a set of acceptable manufacturers. The concept is well-intentioned, but arguably allows products into the construction process that would not otherwise be allowed by the designer. On this job, fortunately, the free-access requirements did not apply. Therefore, one of the most critical decisions of the project was the choice, early on, to go with an almost closed spec and design the replacement system around a single manufacturer and entertain proposals only from contractors that would provide the desired equipment.

Thus, project documents were very broad:

1) The system must be state-of-the-art, capable of supporting intelligent, addressable devices and be easily expanded to monitor multiple facilities over a wide physical area.

2) In general, fire alarm systems, equipment and communications protocols are proprietary. While it is possible, especially at the system level, to monitor and/or control one brandname system with another for maximum interoperability, it is best if all systems are a single brand. The selected brand name is intended to become the system of choice for all protected buildings.

3) The major brand-name fire alarm equipment manufacturers are relatively stable in the marketplace; distribution agents, dealers and installers are much less stable. A major requirement for the system of choice is that the manufacturer be stable and that it provides its products to multiple dealers and installers.

4) The system installer should be a well-established local company, staffed to meet the construction constraints of the project.

As a result, more attention was given to installation and coordination issues. This was especially true for documents relating to the requirements for maintaining the existing system in full operation until the new system was installed, 100% tested and approved by the local authority having jurisdiction. Fortunately, the AHJ inspectors recognized the problems of such a complex system and the necessity of installing it in sections, and were extremely cooperative in scheduling testing at times that would minimize disruptions to hospital activities. This cooperation, especially when testing was done in the surgical suite areas, was extremely important in helping make the project a success.

After significant market research and product evaluations, it was decided to design the replacement system around the Edwards Systems Technology brand name. While this clarified certain matters, contract documents could not be as broad as the specifications. Again, to control change orders, it was decided the team would employ a method where bidders were required to provide unit prices for selected project items, such as spot smoke detectors, duct detectors, pull stations, door hold-open magnets and similar items that were likely to be required as various areas were renovated during the 2.5-year construction interval.

The details of the project specifications for unit prices required a fixed price for the particular item when it was to be added within a certain distance of an existing similar item. This scheme allowed the bidders to provide a considered price in a competitive bid environment for work that was reasonably likely to be required. In practice, it worked well in maintaining control over change order prices.

In making contract provisions for unit prices, it was necessary to provide a method to evaluate, and perhaps adjust, the individual unit prices prior to awarding the contract for construction. Generally, the construction contract is awarded on the basis of lowest base bid plus owner-selected alternates; without the ability to negotiate the individual unit prices with the apparent low bidder, the unit price figures offered by the apparent low-bid contractor may be unreasonable. To avoid this possibility, appropriate wording was inserted into the Instructions to Bidders and the Supplementary General Conditions to allow the hospital the ability to negotiate individual unit prices or outright reject the entire contract and move on to the next lowest bidder should negotiations be unproductive.

The unit price methodology, as described above, works well, in practice, for private clients. It should be noted, however, that the approach might not be appropriate for certain public owners as they may be bound by statute or other mandates that rigidly control the award process. If such is the case, then the entire issue of dealing with reasonably expected change becomes more problematic—and expensive.

Design issues

With the equipment plan outlined, it’s helpful next to examine the major issues that would significantly affect the design and construction of the fire alarm system itself and point out lessons learned. Two challenges were power and wiring. For the most part, the size of an average fire alarm system requires a limited number of power connection points and total power. But with a system of this size and topology, there were not enough life-safety panel boards to provide branch circuits to support the new network panels and supplemental notification alarm circuit (SNAC) panels. Furthermore, for a hospital fire alarm system, connection to the life-safety branch of the essential electrical system is mandated both by the National Electrical Code and NFPA-99, Standard for Health Care Facilities. Consequently, the life-safety branch had to be expanded with additional risers and panel boards to support the new fire alarm equipment.

As far as wiring, a significant part of the old system in the main building was installed in communications cable tray routed in unusually large interstitial spaces that run between floors of the hospital—a total main-floor-to-floor distance of about 25 ft. The general accessibility of the cable tray in this space simplified the removal of the old wiring, and new wiring was installed in raceway, typically EMT. In other sections of the facility, where there was no interstitial space, the original system was installed in EMT, the accessible parts of which were removed as a part of the demolition process.

Another issue in dealing with large buildings with multiple fire compartments is alarm annunciation. The configuration used at Duke Hospital provides a voice alarm message to alert the occupants of the incident fire zone, with an advisory message going to all adjacent zones. Other specialized pre-recorded messages are also provided depending on the exact nature of the alarm condition. With close to 150 separate zones, the programming and scheduling of initial and subsequent messages is complex and requires careful coordination and testing.

On the subject of coordination, the necessity of having the new and existing fire alarm systems operating simultaneously definitely caused some problems. In many locations, the original system devices and alarm appliances did not meet the current requirements of NFPA 72 and the ADA. But these were the “easy” places because new equipment could be installed in new locations. On the other hand, in many locations, it was necessary to exchange an old item for new. Locations where the wall or ceiling in the area was “hard”—plaster, masonry or finished wood—proved difficult. Fortunately, it was possible, in most cases, to remove the existing device and relocate it temporarily on the wall with an exposed cable connection. The existing raceway to the existing location was reused for the new item and after the new system was accepted by the AHJ, the old item was removed.

Surprisingly, one of the more persistent difficulties in the actual construction process arose from the simplest initiation devices: pull stations. In some cases, new device boxes were installed for the new items, especially in locations where the original location was not technically in compliance with the revised codes. In other cases, a double-wide surface box—where the location was appropriate—was provided with both pull stations being located side-by-side. One of the pull stations was labeled as “not in service” so that there would be no confusion should the need to activate the station arise. For reasons that were never understood, the building occupants, or visitors, tended to remove the labels. This required constant attention to be sure all non-operational devices were clearly identified.

Coordination issues

Originally, the project was scheduled for completion within a year. However, this timetable changed in the review process when it was observed that a one-year construction interval translated into the need to complete more than 4,000 sq. ft. of work—that would also have to be tested—per day, based on a schedule where everyone worked every day of the year. The timetable was then stretched to 2.5 years, which, with a five-day work week, allowed for a more reasonable goal of 2,500 sq. ft. per day.

Of course, working in sensitive patient areas also proved tricky. As expected, access to these areas had to be scheduled around patient care activities and was often delayed by patient care emergencies. This was especially true in critical-care areas where patients may remain in the same location for long periods of time and where patient movement for non-medical reasons is not acceptable. Such delays, while unavoidable, led to additional construction expenses. Fortunately, due to the large area of the overall project, there was little downtime due to accessibility issues. Thus, the construction was simply shifted to other areas until the critical areas became available.

The situation was further complicated by the hospital’s requirement of barriers to provide improved infection-control measures. This was another expense but was ultimately resolved by constructing an adjustable-height, dust-tight, roll-around containment cart that could be fitted tight against the ceiling so that there would be minimal dust generated in the patient-care areas. This cart worked so well that it is being used in continued maintenance efforts in the facility.

A job well done

Despite the complexity and numerous challenges, construction was completed in the spring of 2006, with final demolition being completed about two months later. Today, Duke Hospital North is fully outfitted with a technologically top-of-the-line fire alarm protection system, ready to serve the hospital as it continues to serve its multitudes of patients.

Anatomy of a Fire Alarm Retrofit

Following is a summary of the total number of initiation devices, alarm appliances and various other system components used in the recent upgrade to the Duke University Medical Center fire alarm system:

Base components

Firefighter command center: 1

Network panels: 24

Power supply modules: 64

Dual loop drivers: 33

Single-loop drivers: 13

Addressable initiation devices

Smoke detectors: 2,644

Heat detectors: 17

Manual pull stations: 524

Duct detectors: 310

Life-safety control points: 728

Addressable sprinkler system initiation devices

Tamper switches: 170

Flow switches: 111

Auxiliary monitor points—halon panels, preaction systems, hoods, and the like: 13

Alarm appliances

8-in. ceiling speakers: 2,180

Strobes: 2,100


Remote annunciators: 20

Zone amplifiers (total power: 6,420 watts): 146

Backup amplifiers: 23

Total system capacity at maximum build-out

Addressable devices: 19,750

Signal-line circuits: 79

A Big Deal

Working on a project the size and scope of Duke Hospital is in and of itself a design challenge. But this project had a number of unique issues. Like most hospitals, it is in a constant state of change. At any given moment, there are several ongoing projects in the buildings, ranging from simple office modifications to full-scale renovations of entire departments. In addition to more than 980 licensed patient beds, the buildings have a number of unusual occupancies that require special attention. Some of these areas include a complex kitchen and cafeteria with a dedicated chemical-suppression system, a data-processing area with separate chemical-suppression system, MRI suites with high-magnetic fields, more than 40 operating rooms, elevator systems, several dedicated pre-action sprinkler systems and the patient rapid transport system. In addition, designers had to deal with the usual requirements for high-rise buildings such as two-way voice communications and smoke control/evacuation systems.

Applying Elevator Standards

While the Duke Hospital buildings look fairly typical from the outside, there are several unusual features that have allowed the facility to keep up with the constantly changing demands placed on a hospital physical plant, as well as pose challenges to the fire alarm project.

For example, the hospital features a unique transportation hub that enables easy movement of patients, visitors and staff between the Duke Hospital complex and other buildings on the medical center campus. As originally conceived, the main hospital building was intended to be the center of a sophisticated medical center campus with a “transportation” level that provided direct, environmentally controlled access to parking areas and other medical center facilities. In addition to communicating with outlying buildings, the transportation level is part of Duke Hospital’s central elevator core, housing eight elevators that provide vertical access to the inpatient housing and treatment areas of the complex.

The transportation level is one floor below ground level and houses two automated patient transport systems—one that moves between one of the larger parking decks and Duke Hospital, and the other that moves between the hospital and related medical campus clinics. The transport system, called “patient rapid transport” and generally referred to as PRT, moves approximately 5,000 people per day.