A Sodium-cooled-Fast-Reactor, which is characterized by operation at high temperature and low pressure, employs liquid sodium as a coolant. Since components of the SFR have thin-walled structures to reduce thermal stress, components such as the reactor vessel have disadvantageous characteristics against seismic loading. Thus, to ensure structural integrity against seismic events, application of seismic isolation technology is essential. Conceptual design and validations of isolated seismic response performance for the SFR have been performed through parametric response analysis concerning specification of the isolation system such as horizontal and vertical natural period. The results of previous study show that structural integrity of main components against seismic events was ensured by employing thick rubber bearings with a horizontal natural period of 3.0 s or more, and a vertical natural period of 0.125 s or more.

The thick rubber bearings, which have a rubber layer roughly two times thicker in comparison with conventional rubber bearings, have been developed by the authors to ensure seismic safety margins for components by reducing the seismic response for the reactor building in the horizontal and vertical direction. The thick rubber bearings, 1600 mm in diameter at full-scale have been designed to provide a rated load of about 10000 kN with a horizontal natural period of 3.4 s and a vertical natural period of 0.133 s. Moreover, a linear shear strain limit of the thick rubber bearings was designed to accept a horizontal displacement of 700 mm or more, corresponding to shear strain of 226%, to ensure a two-fold safety margin for the response displacements due to the design basis ground motion which has a maximum input acceleration of 8.0 m/s^{2}.

The first shape factor of thick rubber bearings which satisfies the above design requirement is out of JIS K 6410. Additionally, variations of shear and compression on stiffness, under designed response region have been not clarified yet since previous studies concerning the thick rubber bearings are few. Moreover, ultimate characteristics are required, and must be utilized for PRA which is performed in order to reduce residual risk.

Therefore, the purposes of this paper can be divided into three main groups. The first is to clarify variations of shear and compression on stiffness, for the thick rubber bearing using 31 half scale models. The second is to investigate, through the breaking tests, the linear strain limit, tensile yield stress and breaking stress or strain as ultimate characteristics, and these results are defined as breaking surface. The third is to evaluate the ultimate characteristics, which are horizontal hardening characteristics and vertical softening characteristics, using two types of analytical model, 3D-FEM model and modified macroscopic mechanical model which is modified to capture the tensile breaking under offset shear strain. The primary results are summarized as follows:

1) Variation of horizontal and vertical stiffness for provided design values and provided design formulas is small. Even if variation in stiffness is considered to be 95% confidence interval, the confidence interval in horizontal and vertical direction is -4.3% to +5.8% and -3.9% to +4.2%, respectively. Thus, the design formulas have sufficient accuracy.

2) The average breaking shear strain under compressive region was 415.4%, which is a sufficient margin against basis design ground motion. In addition, thick rubber bearings have comparable breaking capacity, compared to conventional rubber bearings.

3) The skeleton curves obtained by two types of analytical models for ultimate characteristics are consistent with measured values under each test condition. These analytical models are useful in order to evaluate the ultimate restoring force characteristics for thick rubber bearings.