Draft of Proposed Commentary Statements
1st Draft by Ernst W. Kiesling
08/09/07
Chapter 3
Commentary Committee: Tezak (Chair), Waller, Kiesling: Review; Levitan
C FOREWORD
Overarching considerations in shelter design
Storm shelters are designed to provide a very high degree of occupant protection in a
windstorm even if the surrounding structure or building is totally destroyed. Because life
safety (occupant protection) is at stake, greater reliability and larger margins of safety are
appropriate than when designing other types of buildings or structures. Whether built outdoors
or within host buildings, shelters should be designed to withstand wind induced pressures and
debris impacts as if totally exposed to wind and windborne debris.
Clarity and integrity are important in making claims and establishing expectations in the
levels of protection that storm shelters offer. Peace of mind is one of the major benefits to be
derived from purchasing a storm shelter for the home or in going to a community shelter. This
standard establishes design criteria to protect occupants from the worst-case event that can
reasonably be expected at the shelter’s location. The shelter’s value is greatly diminished if
the occupant feels unsafe or uncertain when a severe storm is approaching or when a severe
weather warning is issued.
C 304
Need to consider gust and impact together
Tornadoes generate high winds and debris. Yet we do not expect peak winds and maximum
debris impacts simultaneously.
1. Criteria for design wind speeds and debris impacts are extremely high and conservative,
such that two maximum effects are unlikely to occur simultaneously. There is precedence
though that you can have maximum for one and reduced value for the other; e.g. full wind
load and reduced impact load or vice versa..
2. The mechanism of debris generation and missile travel is such that the most critical
missile is not likely to be at the location of maximum wind speed.
C304.8
Response time of storm shelters to pressure changes
The response time of a volume V0 to atmospheric pressure changes as occurs when a tornado
passes over, can be calculated in a similar way to the response time to wind-induced pressure
changes when a broken window occurs.
The response time is given by :

2  .p .V0
kAp0
where  is the density of air
p is the difference in pressure
V0 is the volume
 is the ratio of specific heats of air
k is the orifice constant for the air flowing in or out of the building
A is the total opening area for the flow
p0 is the atmospheric pressure before the change occurs
(This equation is given in a different form for the wind-induced pressure change case, in
‘Wind Loading of Structures’ by J.D. Holmes - Equation 6.5)
Example.
Take a storm shelter with a internal volume of 30 m3 , and a venting area of 0.10 m2 (this is
equivalent to the venting area of 1ft2 per 1000 ft3 given in the NSSA Standard).
Take p as 20,000 Pa ( 3 p.s.i.) and p0 as 100,000 Pa (sea level conditions)
The orifice constant, k, can be taken as 0.6 (potential flow theory), and the ratio of specific
heats for air is, , 1.4. The air density, , can be taken as 1.20 Kg/m3 (average value for sea
level conditions).
Then the equation above gives a value of time constant equal to 0.78 seconds.
In fact this value applies to all cases of the venting ratio of 1 ft2 per 1000 ft3 specified by the
NSSA Standard, as the volume and venting area appear in the numerator and denominator,
respectively.
This time constant is clearly much smaller than the time taken for a tornado to pass over a
structure the size of a storm shelter – typically 10 to 20 seconds. Thus the internal pressure
will change nearly simultaneously with the external atmospheric pressure, giving negligible
pressure differences across the building walls and roof.
A smaller venting area than that assumed above will give a longer time constant. In this case
the internal pressure change, and resulting pressure differences, can be calculated by a simple
numerical solution, given a typical atmospheric pressure change caused by a moving tornado.
NOTE
These paragraphs do not belong in the commentary but represent thinking that should
be countered or overcome by positive statements that present why we did what we did.
In adopting a 10,000 year MRI for hurricanes we are dealing with hurricane wind
speeds comparable to tornado wind speeds (higher in So. Florida and Gulf Coast) that
will destroy buildings and generate debris. Furthermore hurricanes generate
tornadoes, albeit not as intense as Mid-American tornadoes. Hurricane missiles do not
travel as far as tornado missiles so a lower ratio of missile speed to wind speed is
appropriate. We are inconsistent in adopting a much lighter missile for hurricanes
while adopting a ratio of 0.40 in both cases.
We can achieve greater consistency by going to a 15 lb missile for hurricanes and
lowering the ratio of missile speed to wind speed to, say 0.30 (which was my first
suggestion a couple of years ago). We can meet the concern of Coulbourne, Plisich,
Ingargiola, et. al. by simply requiring that community shelters in hurricane regions be
designed for both hurricanes and tornadoes.
I think it is important that we write a very thorough discussion for the commentary
explaining how we arrived at the criteria shown in the Standard. The following is a
hlf-hearted attempt by Ernie Kiesling
C305.1.1
C305.1.2
Design missile – type
The design missile chosen for much of the work done in protective structures and in storm
shelters is a 2 by 4 board. Surely much of the debris generated by extreme winds consists of
boards which came from buildings being torn apart by wind-induced pressures. Furthermore,
the 2 by 4 board has been shown to have more perforation potential then other common types
of debris including 2 x 6 boards of the same length and traveling at the same speed. Hence a 2
x 4 board is chosen as the design missile for shelter design. The speed with which the missile
travels is a function of the type of wind – straight wind, tornado, or hurricane – as well as the
wind speed. The design missile will vary with the type and intensity of extreme winds
expected at a given location.
While it is recognized that this is not the only type of debris that is carried by extreme winds,
it is considered a reasonable representative missile to be used for design and testing purposes.
In considering perforation of a structure or wall section, worst case conditions are assumed.
The missile is assumed to strike perpendicular to the surface of contact and it is assumed that
there is no pitch or yaw in the missile in flight. Testing at Texas Tech University determined
that blunt (square-faced) boards are more likely than pointed ones to perforate shelter
surfaces. Furthermore, in numerous post-storm damage documentation studies, it was
observed that 2 x 4 boards are the most-often found missiles to have perforated building
surfaces. While beams, bar joists, concrete blocks, and heavier objects are sometimes found,
they are most often found on the ground close to the point of origin.
For tornadoes a 15 lb 2 x 4 has been used as the missile criterion for several decades.
Study of many debris fields created by tornadoes and missile penetrations of buildings
observed in post-storm documentation studies tend to validate the criterion. For hurricanes,
few loose boards become wind-borne and travel distances are very short compared to
tornadoes. While arguments might be made to use 15 lb 2 x 4’s as the design missile for
hurricane testing, little field date in post-storm inspections supported such criteria.
Furthermore, considerable testing using a 9 lb 2 x 4 board has been done on building envelope
materials in Florida following ASTM test procedures specifying such missiles. Hence test
methods and a limited data base of test results is available. These considerations led to the
selection of the 9 lb 2 x 4 as the test missile for hurricanes.
Design missile – Speed
The speed attained by a 2 by 4 board driven by extreme winds depends on several factors
including the height of release, the density of the board, and the time in flight. These
parameters, in turn, determine the ratio between maximum wind speed and the speed attained
by the missile. Several publications address this issue. Wills, Lee, and Wyatt ( ) performed
wind tunnel tests on three-dimensional cubes for which they found the ratio of missile speed
to straight line wind speed to be approximately 0.36. For sheet structures they found the ratio
to be 0.64. They state that the 2 by 4 board is a pole structure that would have flight
characteristics somewhere between the cube and the sheet. They recommend the ratio of
missile speed to wind speed to be 0.5.
In a consulting report for building design in the Camen Islands, Joseph E. Minor uses a value
of 0.33, concluding that a 15 lb 2 x 4 board traveling at 50 mph in a 150 mi. per hour (3sec
gust) wind speed is appropriate for building design. He points out that a higher ratio should be
used when designing for occupant protection: a range of 0.4 to 0.6 is considered appropriate.
The missile criteria for tornadoes specified in FEMA 320 is a 15 lb 2 x 4 striking vertical
surfaces at 0.40 times the design wind speed and two-thirds that on horizontal surfaces. The
same criteria are used in this standard.
Because debris generated from failed structures is likely to be released from height less than
40 ft., and is not likely to be lifted higher in the air or carried by straight winds or hurricane
winds, we may concluded that they will be in the air for a very short period of time, likely less
than three seconds. Hence it is appropriate to consider 3-second gust wind speeds in
calculations. But John Holmes points out that significant time in flight is required for drag
forces to accelerate pole-type missiles to 0.40 times the wind speed, longer than the missile
will likely be in flight in a hurricane or straight wind. Hence it seems reasonable to reduce the
ratio from 0.5 recommended by Wills and colleagues to 0.4 or even lower for hurricane or
straight winds. For this standard a 9 lb 2 x 4 traveling at 0.4 times the design wind speed on
vertical surfaces and 0.10 times on horizontal surfaces were chosen as the hurricane missile
criteria. Hence a lighter missile than some advocated was chosen but a higher ratio of missile
speed to wind speed was chosen.
Addendum: 8_09_07
C 101.3 (Not sure where this should go)
The importance of integrating sound storm shelter design with building codes is emphasized
in a paper:
Waller, James E. and Kiesling, Ernst W., “Building Codes and Storm Shelter Safety”,
Building Safety Journal, August 2003, P.19. The publication may be accessed on the
National Storm Shelter Association Web page at http://www.nssa.cc/Publications.php
C 102
Prescriptive designs for residential shelters are presented in FEMA 320. Modifications to
those designs that permit larger shelter sizes are described in a publication:
Zain, Mohammed, Budek, Andrew, and Kiesling, Ernst, “Size Limits for AboveGround Safe Rooms”, Journal of Architectural Engineering, American Society of
Civil Engineers, Date ___, Volume ___, Number ___, p. ___ (Accepted for
Publication). The publication may be accessed on the National Storm Shelter
Association Web page at http://www.nssa.cc/Publications.php
C 309.1.2
Guidance on design of concrete slabs to anchor storm shelters is available in publications
including:
Galani, Navin K., Budek, Andrew, Kiesling, Ernst W., and Zain, Mohammed, “OnGrade Reinforced Concrete Floor Slabs for Storm Shelters, Building Safety Journal,
April 2006, p.23. The publication may be accessed on the National Storm Shelter
Association Web page at http://www.nssa.cc/Publications.php