ying pile caps in seismic regions

huangyhg

tying pile caps in seismic regions.
hi,
recieved some conflicting advice lately regarding whether or not is necessary to tie pile caps together when constructing in a siesmic area such as middle east.. the pile caps in question are not supporting the lateral framing structure, however they are supporting the gravity loaded columns that support a diaphragm between shear cores - any comments greatly appreciated.
cheers

if you are designing per ubc, then 1807.2 states that the pile caps should be interconnected unless you can prove that the soil has enough passive resistance to provide restraint for a force equal to 10% of larger column load of the caps you are trying to connect
ibc section 1808.2.23.1 states that if you are in sdc "c" or higher, then you need to interconnect the piles by using ties (or) slab on grade (or) grade beams. however, it seems like there may be an exception for "interior posts which are not subject to lateral load".
can this be interpreted as follows? (maybe others can chime in)
any building that has been detailed such that the column is not part of a lateral frame can be detailed such that you don't have to connect the pile caps?

no. in my opinion, earthquake movement is a random phenomenon in all directions. during such event, the earth vibravates horizontally and vertically. unless the structural element is isolated completely, it needs to be tied together at all levels.  
consider that the lateral system for a building on deep foundations extends deep into the ground.  for most areas requiring piles, the top soils are soft and do not prevent the piles from deflecting relative to a stiff concrete floor system or tie beam.
therefore, the energy gets applied to the foundation, not necessarily at grade, but some distance below.
now a typical piled foundation has all different pile layouts for each pile cap and may include battered piles for lateral resisting walls from the building.  the mode shape for each of these pile groups may be considerably different.  the deflection att the top of the caps may be in different directions during a seismic event.  this would increase the delta at the basse of a column versus the building overhead and add significantly to the p-delta effect on a column just needed for axial load.  
to prevent this, the code requires us to tie the foundations together so that all the foundation elements move as one.
@teguci, thanks for the explanation.