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