IDAHO NATIONAL LABORATORY Structural Dynamics Research and Development

IDAHO NATIONAL LABORATORY Structural Dynamics Research and Development NLSSI: Exploratory Analysis of Behavior & Influence of Interfacial Conditions Amit Varma, Jungil Seo Chandu Bolisetti, & Justin Coleman

NLSSI Hypothesis: Nonlinear interface between soil and structure modifies the seismic input to the structure Objective � To investigate the reduction in maximum acceleration between the free field and basemat at Fukushima Daichii unit 1 or 6 Tasks 1. 2. Building simple structural model of either unit 1 or unit 6 in accordance with ASCE 4 and ASCE 43 Perform NLSSI sensitivity studies to determine the physical reason for reduction in motion

DATA GATHERING EFFORT AND COMPARISON • • • Data was gathered from Fukushima Daichii and Daini, and Kaswazaki-Kariwa. The data was used to determine if seismic energy is dissipated across the soil-structure interaction, interface and if so why A comparison was made between free field recorded motion (no SSI) and In-Structure recorded acceleration (see red circles)

DATA GATHERING EFFORT AND COMPARISON • A report was issued February 27 th Coleman, J. , and Chandu, Bolisetti, “Seismic Data Gathering and Validation” • Results from the report are: • Reduction in maximum accelerations between the free field soil and concrete basemat for Fukushima unit’s 1 and 6. • Reduction in maximum accelerations indicates that seismic energy is being dissipated between the soil-structure contact interface. • Could be both nonlinear soil behavior and geometric nonlinear behavior due to gapping and sliding. • Current numerical analyses activities are determining if energy is being dissipated due to gapping and sliding at the soil-structure interface.

NLSSI Report Outline (Work in Progress) � Introduction � NLSSI LS-DYNA Model Soil domain Boundary conditions NL soil model NPP structure Soil-to-Foundation NPP structure contact interaction � Analysis results Site response analysis Acceleration response history Acceleration response spectrum � Conclusions

NLSSI LS-DYNA Model � NLSSI LS-DYNA Model The same approach on soil domain, boundary condition, and NL soil model as stated in INL/EXT-15 -35687 “Light Water Reactor Sustainability Program Advanced Seismic Soil Structure Modeling”

NLSSI LS-DYNA Model � NLSSI LS-DYNA Model NPP structure – Simplified Fukushima Daichii reactor unit 6 based on schematic drawings Tur bin e Bui ldin g r cto g a Re ildin Bu oil S r N a ine l on S tic s Ela oil

NLSSI LS-DYNA Model � NLSSI LS-DYNA Model Soil-to-NPP structure interaction Surface-to-surface – friction coefficient Gap elements using cohesive material – not supported in LS-DYNA implicit Tiebreak_nodes_surface – currently being investigated � Tiebreak_nodes_surface A constraint-based tied contact with a failure criterion Failure criteria in shear – the max shear strength of the NL soil Failure criteria in tension – typical soil tensile strength

NLSSI Analysis Results � Multiple cases will be compared Elastic soil – NPP structure – tied Elastic soil – Nonlinear soil – NPP structure – tie break_nodes_surface � Site response analysis will be provided � Acceleration response history of each case will be provided at different locations (basemat and floor levels) and compared with recorded data � Acceleration response spectrum of each case will be provided at different locations (basemat and floor levels) and compared with recorded data

INTERFACE MODELING FOR SOLIDS: A FUNDAMENTAL STUDY WITH APPLICATION TO NLSSI Amit Varma, Vikas Tomar, Dhrubajyoti Dutta

EXISTING CONSTITUTIVE MODELS u Thin layer element (Desai et al 1984; Boulon et al. 1989): Translational freedoms are considered at interface nodes and sliding is considered by considering relative motions between surrounding soil elements and independent DOF. Stress/strain relations are correlated by a constitutive matrix (the shear stiffness found from direct shear tests and normal stiffness from participation of thin layer elements and adjoining solid elements u Disturbed State concept (Desai et al. (1992): u The disturbed response of the material is described through a reference or intact state considering induced anisotropy, microcracking or softening behavior. u The fully disturbed state is characterized by the materials incapacity to carry further shear stress or the deformation at constant shear without change in volume, averaged over a planar projected area.

EXISTING CONSTITUTIVE INTERFACE MODELS u Critical state soil mechanics (CSSM approach): States that the critical state at large shearing, shear deformation continues without dilatancy and change of stress ratio. The critical void ratio is inversely affected by confining pressure. The behavior of soil at any state depends on the distance between the current and critical state. u Elasto-plastic model (Fakharian et al. 2000): Increments of normal and shear stresses related to increments of corresponding displacements by a elasto plastic constitutive matrix

PROPOSED: J-INTEGRAL FOR DUCTILE -QUASI BRITTLE BIMATERIAL INTERFACES u u u The interface is modelled assuming negligible interface thickness such that fracture energy released into process zone of crack tip is a result of overcoming adhesive forces When an interfacially cracked controlled volume experiences surface tractions along it’s boundaries; the strain energy is proportional to differential potential energy released with respect to crack length elongation. Variation of external work is equivalent to the strain energy density (which is related to applied body force, surface traction, and variations in displacement) Stress fields near the crack tip are difficult to compute using linear elastic fracture mechanics due to presence of oscillating singularities and plastic zone Hence, the J-Integral proves to be beneficial as it is path dependent; which allows us to compute strain energy release rate, assuming energy contours far from the crack tip.

MODIFIED J-INTEGRAL FOR PLASTICITY EFFECTS AND HARDENING/SOFTENING u u u Since one of the surfaces is ductile in nature, plastic zone formation is imminent. Hence, it is quintessential to consider an integral path about the plastic zone for a Dugdale type plasticity model The energy conservation is no longer valid, however the strain energy release rate can be linked to another essential parameter the Crack Tip Opening Displacement; assuming yield boundary conditions @crack tip For cracks impinging or deviating from interface, fracture parameters are uni-materialistic in nature as propagation takes place through a single media For mechanical anchors at the interfaces, there is corresponding crushing at the crack face resulting in strain hardening (if interface becomes rough) or softening (if interface becomes smooth) This is accounted for by using appropriate strain hardening/softening function for the crack tip

APPLICABILITY OF PROPOSED MODEL FORSSI u Compacted soil is assumed to be elasto plastic. The crack propagation through soil-structure interface is a multi-phase fracture problem. u The interface element is a thin finite element whose stress and strain field are linked by a constitutive matrix composed of normal and shear stiffnesses Intact phase (LEFM): During this phase the crack formation can be assumed to be a sharp Griffith crack and asymptotic stress fields can be predicted from far field boundary conditions assuming Airy Stress function Critical phase (EPFM): The controlled volume model for the interface needs to be modified for oedometric correction. Crack tip is elasto-plastic in nature and the J-Integral is applicable. Softening or hardening may result from roughening or smoothening of the interface Disturbed phase(PFM): The phase where the crack tip becomes completely plastic. By increasing the normal compressive stress on the soil-structure interface the cohesive properties and angle of friction plane can be u u

VALIDATION AND CALIBRATION USING MATERIAL TEST DATA Cyclic testing and modelling of SS interfaces u CYMDOF: capable of performing load and displacement controlled test. Focused mainly on normal stress, rel. density of sand, displacement amplitude and shear stress, no. of loading cycles under slip/no-slip u Static behavior of interface described with Mohr-Coulomb stress envelop u The displacement (strain) controlled tests were used to parameterize the modified Ramberg-Osgood model. Constitutive behavior of interface is non-linear elastic for loading-unloading-reloading Desai, C. , Drumm, E. , & Zaman, M. (1985). Cyclic Testing and Modeling of Interfaces. Journal of Geotechnical Engineering, 111(6), 793 -815.

EXPERIMENTAL FINDINGS u u u Experimental trends: u Peak shear stresses increases with number of cycles; more rapidly for higher densities u Interface exhibits increased stiffness as the normal stress is increased u Mobilized shear stresses are influenced by higher displacement amplitudes and rates Stress path effects during debonding and rebonding could not be deciphered Strain hardening or softening depending on dilatancy or contraction varies for soil-types

MODIFIED DIRECT SHEAR TESTS AND CONCLUSIONS u Uesugi et al. (1986): steel-sand: Steel-sand roughness influence frictional resistance. Skin friction coefficient correlated to roundness of sand particles u Tejchman et al. (1995): treated friction angle vs. local shearing zone as a BVP. Plain strain and parallel glided apparatus showed friction angle is bound by internal friction angle of sand quartz/quartz. u Fakharian et al. (1996) (3 D cyclic sand-steel): Shear stress-disp curves varied for dense and loose sand. Modelled monotonic stress behavior by plasticity theory. Shear-normal stress ratio decreases with increasing normal stress

MODIFIED DIRECT SHEAR TESTS AND CONCLUSIONS u u Frost (2004): Shearing failure occurred in shear zone near interface even when continuum material sticks to granular media. Peak and residual friction angle increase with surface roughness and decrease with hardness(coupling effect) Porcino et al. (2003): did the constant normal stiffness tests between sand Al-plates: Mobilized shear stress was caused by evolution of normal stress. Hu, Pu (2001): used charged coupled camera device to record particle movement at the interface. Critical relative roughness affects strain localization causing a shear band formation leading to large inter-particle relative motions Zhang et al(2003): gravel-rough steel interface

PROPOSED TESTS u Interface element and material model calibration / validation using fundamental tests between soilconcrete interface u Parameters: Material, geometric, loading u Setup and instrumentation: defined by the computational model needs u Full-scale dynamic validation – using large-scale tests at UB. 20