a study of ssi effects incorpora ng seismic wave ......2014 u.s. doe natural phenomena hazards mee...

A  Study  of  SSI  Effects  Incorpora�ng  Seismic  Wave  Incoherence  within  

the  DOE  Complex    

2014  U.S.  DOE  Natural  Phenomena  Hazards  Mee�ng  

 Carl  J.  Costan�no  and  Associates  

www.cjcassoc.com  

Introduc�on  

  CJC&A  was  commissioned  to  evaluate  the  effects  of  seismic  wave  incoherency  on  the  SSI  response  at  a  DOE  complex  in  the  eastern  US.    The  site  is  characterized  as  a  rela�vely  s�ff  rock  site  with  high  frequency  input  mo�ons.  –  Input  mo�ons  peak  near  30  Hz.  –  Structure  has  a  large  spa�al  area  (160,000  sq.  �.)  –  Site/structure  is  an  excellent  candidate  for  reduc�on  of  high  frequency  mo�ons  due  to  spa�al  varia�on  of  ground  mo�ons.  

2  cjcassoc.com  

Background    Seismic  wave  incoherence  is  the  spa�al  varia�on  of  the  amplitude  and  phasing  of  seismic  ground  mo�ons  over  an  extended  area.  

  In  terms  of  its  impact  on  SSI,  incoherence  generally  tends  to  reduce  founda�on  mo�ons  rela�ve  to  the  free  field  input  mo�on.  –  Generally,  if  the  founda�on  is  rela�vely  large  and  s�ff,  results  in  reduced  seismic  demands  within  the  structure.    

–  The  amplitude  of  founda�on  mo�ons  generally  decrease  as  the  frequency  of  the  mo�on  increases.  

  Incorpora�on  of  incoherency  in  SSI  has  been  implemented  within  the  nuclear  power  industry  and  is  accepted  by  regulators.  

  Incorpora�on  of  incoherency  in  SSI  within  the  DOE  complex  has  NOT  yet  been  implemented  to  date.  

cjcassoc.com   3  

Incoherency  Implementa�on    EPRI  has  performed  valida�on  of  the  implementa�on  of  the  U.S.  NRC  approved  methodologies  for  considering  seismic  wave  incoherence  in  CLASSI  and  SASSI.  –  EPRI  report  1015111,  “Program  on  Technology  Innova�on:  Valida�on  of  CLASSI  and  SASSI  Codes  to  Treat  Seismic  Wave  Incoherence  in  Soil-­‐Structure  Interac�on  (SSI)  Analysis  of  Nuclear  Power  Plant  Structures.”  

  This  study  uses  the  SASSI-­‐SRSS  methodology  implemented  in  SASSI2010.  –  Decomposes  the  coherency  matrix  into  spa�al  modes.  –  Transfer  func�ons  are  modified  for  each  spa�al  mode  and  combined  by  the  SRSS  method.  

cjcassoc.com   4  

Incoherency  Implementa�on  (2)  

cjcassoc.com   5  

  The  Abrahamson  rock  coherency  func�on  was  used  for  this  study.    This  rock  coherency  model  is  accepted  by  the  NRC  for  both  rock  and  soil  sites.  

Surface  Input  Design  Spectrum  at  the  Site  

cjcassoc.com   6  

Descrip�on  of  the  SSI  Problem  

  The  SSI  problem  is  characterized  by  the  following:  –  Large  footprint,  squat  shear  wall  structure  –  Soil  profile  is  based  on  excava�on  of  residuum  and  “so�”  weathered  shale  at  the  site,  and  replaced  with  mass  concrete  fill  from  the  top  of  shale  to  the  grade  level.  

– Mass  concrete  is  founded  on  layers  of  weathered  and  unweathered  shale  (3000  fps  <  Vs  <  6000  fps)  

  The  original  SSI  model  incorporates  the  mass  concrete  fill  substructure.  

cjcassoc.com   7  

Schema�c  of  the  Original  Coherent  SSI  Model  used  for  Design  

cjcassoc.com   8  

Grade  

Infinite  Halfspace  

Free  Field  Soil  Layers  

Superstructure  

Mass  Fill  Concrete  Wea.  Shale  

Unweathered  Shale  Unweathered  Shale  

Excava�on  Limit  

Substructure  Soil  Model  

Incoherency  Analysis  Models  

cjcassoc.com   9  

Superstructure  

Mass  Fill  Concrete  Wea.  Shale  

Unweathered  Shale  Unweathered  Shale  

Superstructure  

Free  Field  Soil  Layers  

Superstructure  Founded  on  a  Layered  Halfspace      This  is  the  typical  implementa�on  for  

incorpora�on  of  incoherency  into  SSI.    Interac�on  nodes  are  at  the  surface.  

Interac�on    Nodes  (typ.)  

Superstructure  Founded  on  Embedded  Substructure      This  is  an  embedded  SSI  model.    Interac�on  nodes  are  at  the  sides  and  bo�om  of  

excava�on.  

Incoherency  Analysis  Models  (2)  

  A  simplified  FE  model  of  the  facility  was  generated.    The  FE  model  has  the  following  characteris�cs:  –  Same  dominant  natural  frequencies  of  the  target  facility.  

–  Seismic  mass  and  building  footprint  were  consistent  with  the  target  facility.  

–  Localized  slab/roof  proper�es  around  the  structure  were  varied  to  produce  a  broad  frequency  range  of  response  around  the  structure.  

cjcassoc.com   10  

Incoherency  Analysis  Parameters    As  previously  shown,  coherent  and  incoherent  SSI  analyses  were  performed  on  the  surface  and  embedded  FE  models.    A  single  determinis�c  SSI  analysis  is  performed  on  each  using  the  BE  soil  profile  for  the  site.    Other  sensi�vity  studies  were  performed  which  are  not  all  covered  here:  – Number  of  spa�al  modes  required.  –  Effect  of  basemat  s�ffness.  –  Rocking  and  torsion  response.  – Other  building  specific  features.  

cjcassoc.com   11  

Observa�ons  from  the  Surface  Founded  Model  (Superstructure  on  a  Halfspace)  

  The  results  are  consistent  with  industry  experience  and  expecta�ons.    The  large  footprint  and  high  frequency  input  results  in  reduc�ons  in  ISRS  of  up  to  40%  (system  dependent).  –  ZPA  reduces  up  to  30%.  

  No  surprises  or  unusual  behaviors  were  observed.    Addi�onal  rocking  and  torsional  responses  were  observed,  but  were  very  low  magnitude.  –  This  structure  is  highly  symmetric.  

cjcassoc.com   12  

cjcassoc.com   13  

Coherent

Incoh. 10 Modes

Incoh. 20 Modes

Incoh. 30 Modes

0.5 1.0 5.0 10.0 50.0 100.00.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Frequency �Hz⇥

SpectralAcc.�g⇥

Incoh. 10 Modes

Incoh. 20 Modes

Incoh. 30 Modes

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

Frequency �Hz⇥

MeanSpectralRatioofIncoh.:Coherent

Coherent

Incoh. 10 Modes

Incoh. 20 Modes

Incoh. 30 Modes

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

Frequency �Hz⇥

SpectralAcc.�g⇥

Incoh. 10 Modes

Incoh. 20 Modes

Incoh. 30 Modes

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

Frequency �Hz⇥

MeanSpectralRatioofIncoh.:Coherent

Typical  Basemat  Level  ISRS  –  Surface  Founded  Structure  

5%  Damped  ISRS   Mean  Spectral  Ra�o  Horizontal  

Ver�cal  

cjcassoc.com   14  

Coherent

Incoh.

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Frequency HHzL

SpectralAcc.HgL

Incoh.:Coh. Mean

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

Frequency HHzL

MeanSpectralRatioofIncoh.:Coherent

Coherent

Incoh.

0.5 1.0 5.0 10.0 50.0 100.00

1

2

3

4

Frequency HHzL

SpectralAcc.HgL

Incoh.:Coh. Mean

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

Frequency HHzL

MeanSpectralRatioofIncoh.:Coherent

Typical  ISRS  at  Intermediate  Floor  –  Surface  Founded  Structure  

5%  Damped  ISRS   Mean  Spectral  Ra�o  Horizontal  

Ver�cal  

cjcassoc.com   15  

Example  Results  for  Rocking  and  Torsion  Responses  

Observa�ons  from  the  Embedded  Model  (Superstructure  on  a  Substructure)  

  In  addi�on  to  the  surface  founded  case  (typical  implementa�on),  the  embedded  substructure  model  was  analyzed.  

  The  purpose  was  to  inves�gate  the  impact  of  embedment  effects  on  incoherent  response  rela�ve  to  the  surface  model.  

  In  general,  high  frequency  ISRS  amplitude  reduce  up  to  ~30%.  

  However,  at  several  loca�ons  on  the  basemat  there  are  points  of  increased  ISRS  rela�ve  to  the  coherent  analysis.  

cjcassoc.com   16  

Superstructure  

Mass  Fill  Concrete  Wea.  Shale  

Unweathered  Shale  Unweathered  Shale  

cjcassoc.com   17  

Coherent

Incoh.

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Frequency HHzL

SpectralAcc.HgL

Incoh.:Coh. Mean

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

Frequency HHzL

MeanSpectralRatioofIncoh.:Coherent

Coherent

Incoh.

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Frequency HHzL

SpectralAcc.HgL

Incoh.:Coh. Mean

0.5 1.0 5.0 10.0 50.0 100.00.0

0.2

0.4

0.6

0.8

1.0

Frequency HHzL

MeanSpectralRatioofIncoh.:Coherent

Typical  ISRS  at  Intermediate  Floor  –  Surface  Founded  Structure  (X  Direc�on)  

5%  Damped  ISRS   Mean  Spectral  Ra�o  Surface  Structure  

Embedded  Model  

cjcassoc.com   18  

Basemat  ISRS  –  Basemat  Level  

5%  Damped  ISRS   Mean  Spectral  Ra�o  X  Direc�on  

Y  Direc�on  

Overall  Observa�on  

  This  large  footprint  structure  on  a  s�ff  site  is  an  excellent  candidate  for  the  applica�on  of  seismic  wave  incoherency.  

  High  frequency  ISRS  reduc�ons  are  observed,  consistent  with  industry  experience.  

  The  method  of  modeling  the  SSI  problem  can  impact  the  results:  –  The  standard  surface  implementa�on  of  incoherence  behaves  consistent  with  expecta�ons.  

–  The  applica�on  to  an  embedded  problem  can  introduce  ar�facts  in  the  response  not  observed  in  the  surface  case.  

cjcassoc.com   19