To: "'seaint(--nospam--at)seaint.org'" <seaint(--nospam--at)seaint.org>
Subject: RE: storage racks with partially restrained moment connections - wind frame analysis in high seismic zones
From: "Sprague, Harold O." <SpragueHO(--nospam--at)bv.com>
Date: Mon, 19 Mar 2001 17:41:16 -0600
The 1997 NEHRP, ASCE 7-98, and the IBC simply states that the storage rack
installation shall accommodate seismic displacements, and that the assumed
displacements are not less than 5% of the height.
Harold O. Sprague
> -----Original Message-----
> From: Peter Higgins [SMTP:76573.2107(--nospam--at)compuserve.com]
> Sent: Monday, March 19, 2001 5:36 PM
> To: INTERNET:seaint(--nospam--at)seaint.org
> Subject: RE: storage racks with partially restrained moment
> connections - wind frame analysis in high seismic zones
> Message text written by INTERNET:seaint(--nospam--at)seaint.org
> >The finite element programs we have at my work do not have rotational
> elements for beam to column connections. I have seen ways that beam
> of inertia can be adjusted and ways to put in flimsy dummy members to
> account for the rotational stiffness but these do not account for a
> plateau when the connection does not provide any additional rotational
> This is common to all buildings too. They plateau out, but our design is
> based on the initial tangent modulus anyway. Given that, why should be
> treat racks any differently?
> >I have heard 500 kip*in/rad is a typical rotational stiffness for most
> partially restrained storage rack moment connections. Assuming the
> connectors do not fail what is a typical limiting rotation when a joint
> plastic and does not provide any additional rotational stiffness? <
> 500 is actually a bit high for many types, but depends on the type of
> connector. However the initial tangent stiffness (i.e. before any
> yielding) is typically over 1000. Our office uses an average stiffness
> which is curve fit to the hysteretic curve which we determine by test. By
> the way, the ductility of these connectors is superb, usually exceeding
> 0.3-0.4 radians (!). If we had buildings like these we'd be happy indeed.
> This behavior is very likely the reason for the excellent performance of
> properly designed storage racks.
> >1997 UBC 2222.5 requires racks to be installed with a maximum tolerance
> 1" in 10 feet. There are no storage rack drift limits in the UBC for high
> seismic areas, has this been discussed for future codes? This would make
> wholesale retail areas safer. Nobody may have been squished by storage
> related damage in Northridge, but when did the earthquake happen in the
> morning and were the wholesale retail warehouses open yet?<
> No. and all the code changes in the world wouldn't have saved the store
> which had the wrong rack in it. It never had a chance. This is an
> enforcement, not design issue. On the other hand all the properly designed
> stores had NO damage in Northridge. Whether or not they were open is not
> important. No one was or would have been hurt (by the racks at least).
> >I would like to suggest a code change to limit the calculated
> period of vibration for racks, require the fundamental period to be
> calculated with the Rayleigh method or a finite element model, and have
> reasonable drift limits. <
> 1) This is throwing the baby out with the bath water. The storage racks do
> have these drifts and periods. They have been experimentally verified.
> Changing the code to give a different period is flying in the face of
> 2) The large drifts and superb ductility are what make racks work.
> Tightening up the drift limits would destroy this behavior. What's wrong
> with high drifts? They're good for the safety of the structure. Think of
> as base isolation taken a bit further. No one lives in them, so they can
> have unrestricted drifts without difficulty and enjoy all the benefits
> associated with it.
> 3) Agreed, and a code change is not required. Calculation methods are
> already restricted to the Rayleigh (or other rational) methods. You can't
> use the empirical Method A stuff at all.
> >It seems crazy to try to get fundamental periods out past 1.5, 2 or 3
> seconds to lower the design lateral load on a heavily loaded non-building
> structure in occupancies that people are congregating in. It does not
> matter if the rack does not collapse when people are squished by
> refrigerators or are squashed by pallets full of 5 gallon laundry soap
> buckets flying off the top of 16 foot storage racks. I don't run when I
> walk down an aisle in a wholesale retail warehouse, but I watch for ground
> Period if fundamentally related to the inverse square root of the period.
> You can't have one without the other, We aren't "trying" to get it out
> there, it simply is what it is. If you don't like stuff flying off the
> racks, you certainly don't want to tighten up the drift limits which
> increases the accelerations in the racks. The longer the period, the
> it is for the stuff staying on the racks. For what it's worth, we shook
> actual merchandise. Very little of anything fell off, including 5 gallon
> paint cans (the 1 gallons did come off).
> I think we're in fair agreement. It's just that racks require a shift in
> gears. Building engineers are taught to restrict drifts because high
> are "bad" in the sense that they make people seasick and crack walls and
> windows in little shakers. The idea of allowing very high drifts is
> to their thinking. However, drift is actually a good thing. That's why all
> drift limits are removed from the non building structures if they can be
> accomodated. The ideal is the stationary pendulum. An earthquake doesn't
> shake it at all, it simply stays put and the supports move. That's why we
> took out the restraints on many suspended wlements in the 97 UBC.
> Peter Higgins, SE