The core of Corgan’s data center design effort is protecting the servers. Sophisticated electrical and mechanical infrastructure provides continuous power and cooling to ensure continuous operation. The shell building is expected to protect both the servers and this critical infrastructure. The threats to consider are legion, but not all are equal. They range from high to low probability and from serious impact to mere inconvenience. Recent weather events have shown that the task of avoiding these impacts is becoming more difficult, not less.
But one type of risk stands out as a pervasive threat that can be a near-term menace in virtually any location. That threat is wind. And unlike some threats, which may never come to pass, the question regarding a wind event isn’t “if” but “when.”
Wind can arise from several meteorological events, all of which are caused by differential heating or cooling of the atmosphere—usually through evaporation or condensation involving “latent heat.” Hurricanes can cover wide areas with winds up to 185 mph. Tornados are more local, but they can be more intense, packing winds of 300 mph. Downbursts associated with thunderstorms can create winds up to 100 mph, and more local and intense microbursts can generate winds up to 168 mph. These straight-line winds can occur over storm fronts stretching for hundreds of miles.
The damage from such events depends on several factors, certainly including wind speed but also the duration of the event, the size of the affected area, the speed at which it’s moving, and the amount and type of wind-borne debris. Although published wind speeds are associated with each tornado category in the Enhanced Fujita Scale (EF Scale), the EF value assignment takes place after the fact by evaluating the level of damage.
For other short-term events, wind speeds are usually estimated rather than measured. For this reason, Corgan avoids characterizing our designs as tornado “proof” or resistant to a particular EF tornado designation, instead defining the wind speed they can be resist. As usual, however, the devil is in the details. Wind is a direct threat, but it works in collusion with other weather components to make life difficult.
Once we’ve established a design wind speed with the owner, the direct wind force exerting lateral pressure on the building walls is relatively simple to calculate and resist through structural-moment connections, frame bracing or both. The effect of wind passing over a roof is less straightforward. It creates uplift pressure that reverses the stresses in roof structures and tries to pull the roof membrane up and off the roof. High parapets and building forms can create additional wind patterns that increase uplift pressure. The roof perimeter, particularly corners, is the worst and requires additional fasteners. Before we complete the building-envelope design, we look at how the openings are protected. Depending on the mechanical system and other components, these openings may be large. If the building openings can be sealed with construction that resists the design wind speed, the building is designed as a “closed” structure. If not, the assumption is that the wind will enter the building and try to blow it apart from the inside, adding to the forces pulling on it from the outside.
Additionally, this determination must consider wind-borne projectiles, as they increase the risk of breaching the building enclosure at openings where it tends to be weakest. Often, Corgan designs a strong envelope for the critical-function areas that are less robust, with more glazing in areas that the customer intends to abandon during a weather event and that may suffer some damage. Staff can retreat to the more-secure building areas, and the structure is considered “sacrificial.”
Outside the building, electrical components, such as transformers, are generally constructed with robust enclosures designed to resist weather events. Other components, such as generators, are protected by enclosures that can be designed to resist wind events. Usually these components are shielded by screen walls providing additional protection. Mechanical systems must release huge amounts of heat outside the building, generally into the atmosphere. This process requires large areas of exchange that may include openings, louvers, fans, evaporative media or other components. In our most weather-resistant facilities, the cooling towers are made of reinforced poured-in-place concrete.
As Corgan evaluates threats and threat responses, we are careful to note and take advantage of the collateral benefits of different solutions. For example, facades designed to resist blast forces are inherently wind and projectile resistant. Building structures designed to resist high wind speeds are inherently resistant to earthquakes. But taking full advantage of these overlapping benefits requires careful review of the secondary system design to ensure the entire system is uniformly resilient and is free of weak spots that still require scrutiny. Careful attention to these issues allows owners to legitimately claim all the benefits that their structures have to offer.
As Corgan considers the risk/cost/reward balance with our clients, historical data is useful, but it may contain a trap. Historical data has been relatively stable during the currently “recorded history,” but whether it will continue to be so is an open question. Weather conditions and events, at their historic extremes, are expressed in statistical terms. The “100-year flood plain,” for example, defines the boundary that a water level has a 1% chance of exceeding in any given year. The ASHRAE “1% design day” values for a particular location represent the characteristics that are historically experienced 99% of the time. Actual conditions are only higher or lower for 1% of the hours in the historical record—or “about as bad as it gets.” If we experience historic weather events with increasing frequency and record-breaking magnitude, these historic norms will seem less predictive of future events.
When weather events affect larger geographic areas, the response, recovery and emergency services we rely on become more stressed and less immediately reliable. What our experience tells us about past-event durations may be misleading us regarding future events. After all, when it rains, it pours.
About the Author
Brian George is a Principal of Corgan’s Critical Facilities Architecture Group and leads a team of talented architects who are experts in the design of mission-critical facilities. His 39 years of experience and in-depth knowledge of critical electrical and mechanical systems brings extraordinary value to the planning and execution of data center projects nationwide and internationally. Brian is highly sought after as an expert in his field, speaking annually at industry conferences around the world. As a registered architect, is a member of the American Institute of Architects, is LEED BD+C accredited and has served on Corgan’s Management Committee since 2010.