Engineers must consider a number of issues when designing a world-class research facility. Getting power to the right location is only the beginning. Flexibility in design is critical to limit downtime during renovation of lab space as research teams are switched out. Massive amounts of computing capability located within the facility can also be desirable to ensure the security of crunching billions of calculations. Specialized research lighting — as well as generator and uninterruptible power supplies — must all be considered.
Delving deeper into the electrical system, engineers must also address electrical interference, reliability, and redundancy levels. One extended power outage could cost a facility millions of dollars in lost research. Electromagnetic waves that originate at nearby transformers, power lines, and other sources might inadvertently alter a scientist's results and render months of work useless. Noise from nearby electrical or mechanical equipment could seriously affect the normal functions of animals or rodents.
How do electrical engineers support the research that takes place in these facilities? At any research site, the “silent assistant” is the building's electrical system. Electrical engineering experience has shown that the best lab utility systems are those that are rarely noticed. When design engineers provide the appropriate amount of electrical power where it's needed and effectively manage interruptions, scientists can focus their attention on research — not the health of their building systems.
Capacity. Research laboratories require more power than the average office building. Depending on the research equipment it uses, the average facility can draw about 16W to 20W per square foot, compared to about 6W per square foot at a standard computer-filled office space. To determine power requirements for new labs, the University of Washington's design team uses a graph based on data assembled from existing labs. It allows an engineer to design the proper amount of power with acceptable spare capacity for each lab. As the lab space increases in square footage, the electrical requirements in watts per square foot decreases. Larger labs use the extra space for added support functions, which require less electrical load. These labs also tend to be more diverse as they typically support simultaneous operation of all the equipment.
Many modern research labs require computational server rooms, which have extremely dense power requirements; high-density server racks can accommodate 130 servers each. And power demands will only increase. IBM recently developed a new supercomputer that despite taking up only as much space as a 30-inch television (or about half a rack) can complete 2 trillion mathematical operations per second. This level of computing power, along with the associated individual cooling requirements, can easily push design electrical loads within these spaces to 150W per square foot.
Providing access to a nearby panelboard is especially advantageous. The typical facility is only useful for 10 to 12 years, and within that timeframe the type of research it hosts may change significantly. Because of this frequent turnover and the related evolution of power needs, 25% to 40% spare capacity must be available in nearby panelboards so adjacent labs aren't interrupted.
Reliability. Ask a research scientist what timeframe is acceptable for his lab to be without power and the answer will surely be “zero.” Accommodating this request is neither realistic nor possible. In addition, the game of chasing reliability 9s is an expensive one. Telling users the implications — and the associated costs — of their request enlightens them very quickly.
Most research facility owners want their electrical system to be 99.9% to 99.999% reliable. The difference between an annual 9-hour outage and a 3-minute outage is a large and potentially expensive jump. If the facility is located in an area where local power is commonly interrupted, the investment in backup systems can be costly.
Other facilities, such as the University of Washington Health Sciences, have built a strong initial base of reliability into the electrical distribution system. Using a three-transformer spot network to serve the new Bioengineering and Genomic Science buildings from the campus primary feeders allows the university to limit the actual downtime of the system. In fact, only one 4-hour outage has occurred in the last 10 years. This equates to a 99.995% reliability record — and this is just within the normal power system. Add to that the 1.5MW back-up generator for certain critical loads and equipment, and 99.999% is readily available.
Many electric utility companies in large cities use a spot network to serve the downtown core, but labs located in suburbs and outlying areas will experience longer, more numerous outages. Therefore, adding either additional medium-voltage feeders from separate substations or installing multiple on-site generators is recommended to achieve reliability to support expensive research functions.
Interstitial space. For building owners with a bigger budget, interstitial space provides a level of storage beyond what an accessible ceiling can offer. These storage spaces are virtual “floors between floors,” complete with poured concrete flooring and access to the lab area or floor below. Although perceived as mostly a convenience for service personnel, it can save a facility thousands of dollars in labor and remodeling costs when installed above labs. Speed to market is vital, so making adjustments to labs while research is underway is most efficient with interstitial construction.
Panelboards, transformers, bus ducts, and disconnects can all be located in the interstitial space of research facilities. This maximizes space on the lab floor for lab equipment and bench space.
In many cases, the overwhelming majority of electrical and mechanical alterations can occur in interstitial spaces while research continues below. To prevent research disruption, final connections to any new equipment can be performed during off hours. This can be highly desirable for a lab that's rushing to get a new discovery to the market.
Interstitial space is more expensive to construct at the outset, as Washington State University (WSU) discovered after analyzing the initial design of the new Biotechnology building. A 5% to 10% increase in the initial capital cost of construction can be difficult to overcome in the public arena.
“The major problem when working in the state budgeting process is that the capital and operating budgets are separate,” said Ryan Ruffcorn, a project manager in capital planning and development at WSU. “Therefore, increases in the first-costs of a building are never directly credited to the project budget. The increased construction costs would then require that program space be reduced in an effort to keep the project on budget.”
Rather than construct interstitial space, most publicly funded lab buildings opt for a more traditional design that uses pathways within accessible ceilings.
It's also becoming popular for design and construction firms to build generic labs on spec. In this situation, the tenants aren't defined at the time the building is designed, allowing the building owner to cater to a variety of biotech clients with different needs. However, engineers can't design a building to support every conceivable condition. Savvy developers of these spec labs understand that additional space in the electrical rooms and risers is advantageous to accommodate growth or unique tenant requirements.
Occasionally, a facility will find a middle ground and use interstitial space above the most crucial or sensitive lab areas like animal holding areas. By doing so, electrical personnel don't need to scrub down every time they work on the space. A researcher's main concern is contamination from outside sources. Controlling building systems workers' access to animal spaces dramatically reduces the potential for human contamination.
Lighting. The most commonly overlooked component of research lab design is the lighting system. Fine instrument bench work requires not only plenty of horizontal footcandles but a strong vertical contribution. A common design in the past still used today involves illuminating lab areas with rows of recessed fluorescent fixtures over the benches. This configuration offers a significant number of horizontal footcandles (70 to 100 in most cases) but introduces annoying shadows.
To support the significant amount of bench work that takes place in the vertical plane, IESNA recommends only 50 footcandles in the horizontal plane and 30 footcandles in the vertical plane. Many researchers prefer to have more than 50 footcandles in the horizontal plane, which is easily accomplished by using 3-lamp T-5 HO fixtures. Obtaining the vertical requirement, however, requires an indirect contribution.
A recent mock-up at the University of Washington Bioengineering department averaged 80 to 90 horizontal footcandles on the bench top, while eliminating most of the shadow effect with about 60 footcandles in the vertical plane. This was accomplished using a single row of “drop-basket” fixtures recessed in a 9-foot ceiling and located between benches spaced at 10.5 feet apart.
Installing energy-saving lighting controls contributes value to a building's overall sustainable design. A separate switch for the center lamp allows a stepped approach to savings in areas where researchers don't require added light levels. Daylight dimming should also be considered where available to maximize exterior light contribution.
Lastly, lighting of animal areas also requires special attention. Dimming and override fixture control is important to ensure the animals aren't disturbed during research. Fire alarm strobes shouldn't be installed in these areas — assuming the AHJ accepts the non-public occupancy — because fire alarm system tests can greatly disturb rodents and animals in cages.
By thoroughly examining and documenting researchers' needs, design engineers can plan electrical systems that are both highly effective and rarely noticed. Limited or planned power outages, easy and quick lab renovations, capacity for future equipment, and shadow-free work benches are the benchmark of quality electrical design.
Thrun is principal of Sparling, an electrical engineering and technology-consulting firm based in Seattle.
Sidebar: Make Use of Interstitial Space
The interstitial space method of construction in laboratory buildings can benefit construction in several ways:
Workers and contractors aren't required to use ladders because most electrical equipment and systems can be accessed and maintained from the interstitial space.
Construction schedules can be compressed, resulting in cost savings. This is mainly due to the fact that lab spaces can be fit-up at the same time building systems are being installed.
Ongoing operation and maintenance of equipment and infrastructure systems is simplified. The result is direct operational cost savings and reduced life-cycle costs for the building.
Maintenance of equipment and systems can be completed without disrupting laboratory services.