Impact of High-Temperature Exposure on Reinforced Concrete Structures Supported by Steel Ring-Shaped Shear Connectors

Abstract

Ensuring the structural integrity of reinforced concrete (RC) components in nuclear facilities exposed to extreme conditions is essential for safe decommissioning. This study investigates the impact of high-temperature exposure on RC pedestal structures supported by steel ring-shaped shear connectors—critical elements for maintaining vertical and lateral load paths in containment systems. Scaled-down cyclic loading tests were performed on pedestal specimens with and without prior thermal exposure, simulating post-accident conditions observed at a damaged nuclear power plant. Experimental results show that thermal degradation significantly reduces lateral stiffness, with failure mechanisms concentrating at the interface between the concrete and the embedded steel skirt. Complementary finite element analyses, incorporating temperature-dependent material degradation, highlight the crucial role of load redistribution to steel components when concrete strength is compromised. Parametric studies reveal that while geometric variations in the inner skirt have limited influence, thermal history is the dominant factor affecting vertical capacity. Notably, even with substantial section loss in the concrete, the steel inner skirt maintained considerable load-bearing capacity. This study establishes a validated analytical framework for assessing structural performance under extreme conditions, offering critical insights for risk evaluation and retrofit strategies in the context of nuclear facility decommissioning.

1. Introduction

A magnitude 9.0 megathrust earthquake struck the Tohoku region on 11 March 2011, causing extensive damage to both structural and non-structural components [1,2]. The nuclear accident triggered by this event—known as the Great East Japan Earthquake—led to the ongoing decommissioning of the Fukushima Daiichi Nuclear Power Plant, spearheaded by the Japanese government. Addressing a complex and globally significant engineering challenge and ensuring the safe and systematic progress of the decommissioning effort requires the collective expertise of all involved organizations.

Decommissioning is a prolonged process, requiring the structure to endure various factors that impact long-term safety, including durability concerns and external forces. Japan, situated on the Pacific Ring of Fire, is one of the most earthquake-prone countries in the world. As such, particular attention must be given to horizontal forces, especially when combined with the residual dead load of the structure. Of particular concern is the structural integrity of internal components within the Primary Containment Vessel (PCV). Recent internal inspections have revealed significant deterioration of the reinforced concrete (RC) pedestal that supports the Reactor Pressure Vessel (RPV), including concrete section loss and exposed reinforcement at its base, as detailed in Figure 1 and the subsequent chapter [3]. To ensure the safe progression of decommissioning, it is essential to verify the pedestal’s capacity to support vertical loads from the upper structure and to resist lateral forces, particularly those induced by seismic activity.

Previous studies have extensively investigated the degradation of steel [4,5,6,7,8,9,10,11,12,13,14,15,16] and concrete [17,18,19,20,21,22,23] after exposure to elevated temperatures. For steel, a representative study by Smith et al. [4] reported laboratory-based experiments demonstrating the reduction in strength across various steel grades subjected to high temperatures. Similarly, Qiang et al. [5,6] examined the mechanical property changes in several Chinese steel grades after fire exposure. More recently, Suzuki et al. [16] conducted cyclic material tests on steel specimens subjected to heating and cooling protocols. The results indicated that the effects of thermal exposure persist even under cyclic loading conditions, highlighting the need for accurate material characterization for seismic design following high-temperature events.

With respect to concrete, Abe et al. [17] carried out a comprehensive investigation into the high-temperature behavior of high-strength concrete, compiling data on compressive strength, strain at peak stress, fracture energy, and Young’s modulus as functions of experienced temperature. Li and Li [19] explored the recovery of concrete properties after thermal exposure, while Mahmoud et al. [22] examined the post-heating compressive strength of recycled concrete. They applied various machine learning models to predict performance, effectively incorporating parameters such as exposure time and peak temperature. In a subsequent study, Mahmoud et al. also analyzed the behavior of concrete incorporating waste powder, identifying the optimal replacement ratio and characterizing strength development mechanisms from a chemical perspective under temperatures up to 800 °C [23].

Despite these advances in material-level studies, only a limited number of researchers have extended this understanding to system-level structural behavior [24,25,26]. Ozaki and colleagues have conducted extensive work on the redundancy and collapse modes of steel frames under fire conditions [24,25,26]. Their simulation models incorporate temperature-dependent material degradation and successfully captured global failure mechanisms, including column buckling, beam failure in fire-exposed zones, and column failure in non-fire zones due to stress redistribution.

In the context of nuclear power plant structures, Kontani et al. [27,28,29] have contributed valuable insights by integrating material degradation and structural response to evaluate post-accident performance. Additionally, Nagasawa et al. [30] examined the seismic behavior of nuclear facility structures. While these studies have improved our understanding, they have not addressed critical regions at the base of pedestal structures where loads are transferred to the foundation through complex interactions between concrete and embedded steel. This region is particularly vulnerable to degradation from elevated temperatures.

To fill this gap, the present study develops a detailed finite element model and conducts cyclic loading tests on scaled pedestal specimens that reflect various thermal histories. By elucidating the damage mechanisms near the pedestal base, this research aims to support a more robust analytical framework for evaluating collapse risk and informing seismic retrofitting strategies. It is important to emphasize that this study does not seek to predict the actual temperatures experienced during the Fukushima Daiichi nuclear accident, nor does it attempt to assess the specific damage sustained by the plant. Although substantial research efforts have been devoted to estimating thermal conditions, the presence of numerous unknown or unmeasurable factors continues to hinder reliable reconstruction of thermal histories [3]. Even with sophisticated material models and heat transfer simulations, the underlying assumptions must still be critically evaluated.

In light of these challenges, this study adopts a parametric approach, simulating a range of thermal conditions from an intact state up to an upper-bound temperature of 1200 °C. While this upper limit represents an extreme and unlikely scenario, it is selected to conservatively bound the structural behavior. The authors consider this a pragmatic approach to respond to societal demands for safety evaluation, particularly when definitive thermal data remain unavailable. Furthermore, the analytical framework presented here offers flexibility and can be refined as more detailed information becomes available through the ongoing decommissioning process.

2. Current Condition of Damaged Power Plant by Nuclear Accident

As illustrated in Figure 1, the pedestal was subjected to elevated temperatures due to fuel melting, necessitating a structural performance evaluation that accounts for thermally induced material degradation. As illustrated in Figure 2, when lateral forces, such as seismic loads, act on the structure, load transfer to the foundation occurs primarily through a steel ring-shaped member embedded in the concrete—referred to as the inner skirt. In this configuration, damage to either component can significantly impair structural performance. In particular, when concrete’s material properties deteriorate due to heat exposure, its capacity to transmit stress may be significantly reduced, potentially leading to unforeseen failure modes not considered in the original design. According to the guidelines of the Architectural Institute of Japan (AIJ) [31], the compressive strength vanishes at temperatures approaching 1200 °C, while the tensile strength disappears around 800 °C. These forms of degradation reduce the overall bearing capacity of the structure and diminish its ability to redistribute stress to the inner steel plate.

Based on the preceding discussion, this study proposes failure mechanisms corresponding to vertical bearing force and horizontal lateral force, as illustrated in Figure 2a,b, respectively. The vertical load is assumed to be resisted by the concrete and inner steel skirt acting in parallel, thereby contributing jointly to the overall bearing capacity of the structure. In contrast, the horizontal resistance mechanism is modeled as a series system, where the lateral force may lead to the punching or dislodgement of heat-degraded concrete. The following chapters provide experimental and analytical validation of these proposed mechanisms.

3. Cyclic Loading Test of Pedestal Specimens with Steel Inner Skirt

To investigate the damage mechanism at the lower part of the pedestal, a series of cyclic loading tests were conducted on scaled specimens. Figure 3 presents the drawings of the specimen and a photograph taken prior to mortar casting. The specimen was manufactured by casting a cylindrical mortar body around an inner steel skirt welded to the base plate. The inner skirt was equipped with multiple shear keys, arranged in a stepped configuration to simulate actual structural conditions, and designed to resist vertical tensile forces. The reinforcement was fabricated in a mesh configuration using longitudinal and transverse bars. The longitudinal reinforcement bars were anchored to the base plate using adhesive at every second pitch to ensure adequate fixation and load transfer. This anchorage condition was designed to replicate that of the actual structure. The adhesive used is a two-component epoxy resin with an adhesive strength of 34.5 N/mm². The pitch spacing is approximately 20 mm.

The specimen is scaled down to 1/30 of the actual structure, and thus, scale effects may influence the observed structural performance. It is important to note that the primary objective of this study is to investigate the damage mechanisms of the RC pedestal supported by the inner skirt, as well as the influence of heat exposure. Therefore, caution should be exercised when interpreting the experimental results, as they may not directly represent the actual damage conditions of the full-scale nuclear power plant structure.

Figure 4 illustrates the loading apparatus and specimen setup. A loading jig was mounted on the top of the specimen, to which both horizontal and vertical hydraulic jacks were connected via biaxial pins. The height of the horizontal jack was determined based on the ratio between the pedestal height and the stabilizer height in the actual structure. Due to the offset of the horizontal jack, the specimen is subjected primarily to a couple rather than a direct shear force. This setup reflects a scenario in which the stabilizer no longer functions properly due to explosive events or long-term degradation.

The horizontal load was controlled by the rotational deformation (δ/h), calculated as the horizontal displacement δ measured at the top of the specimen divided by the height h of the displacement transducer, as shown in Figure 4b. The installment of displacement transduces is indicated in Figure 4c. The loading protocol consisted of symmetric cyclic displacement control with increasing amplitude levels of δ/h = 0.00125, 0.0025, 0.005, 0.01, 0.015, 0.02, 0.025, and 0.03 radians. Each amplitude level was applied for two loading cycles. The axial compressive force was applied such that the axial force ratio matched the one derived from the dead load in the actual structural system.

Two specimens were tested, with the presence or absence of heat exposure being the primary variable. After casting, the specimens were air-cured until the day of testing. For specimens subjected to heat exposure, preheating was conducted at 100 °C for 24 h to prevent explosive spalling, followed by main heating [32]. The used air furnace and placement of specimen are visualized in Figure 5a,b. The heating protocol increased the temperature at a rate of 1.0 °C/min up to 20 °C below the target temperature T_c and at 0.33 °C/min thereafter until reaching the maximum temperature of 600 °C. The target temperature was determined based on accident scenario estimations [3]. The selected value of 600 °C corresponds to the estimated temperature at the outer side of the structure, where the concrete section is expected to remain, in contrast to the inner side. Furthermore, this temperature level prevents excessive cracking or deformation during heating, thereby ensuring the integrity of the specimen during subsequent cyclic loading tests. The specimens were held at the target temperature for 72 h before undergoing natural cooling. The measured history of the inside temperature is illustrated in Figure 5c, whose horizontal and vertical axes represent the time t and the measured temperature T, respectively. The temperature was controlled successfully. Although thermocouples were not attached to the specimens during heating, preliminary analyses confirmed that the selected heating rate would result in a reasonably uniform temperature distribution. In addition, the interior of the furnace was lined with heat-insulating materials to minimize temperature gradients. Consequently, it was assumed that the specimens were exposed to a uniform temperature distribution throughout the heating process.

Both the specimens and material test samples were heated in the same furnace. Figure 6a illustrates the sample after experiencing heat exposure. The material test specimens and testing procedures followed the Japan Industrial Standards (JIS) [33,34]. The cylindrical specimens had a diameter of 50 mm and a height of 100 mm. Subsequent compressive and splitting tensile tests were conducted. The stress–strain relationship obtained from the compression tests is shown in Figure 6b. Thermal loading led to a noticeable reduction in strength, and Young’s modulus decreased by approximately 7%.

Figure 7 presents the obtained hysteresis curves, where the horizontal axis represents the rotational deformation angle (δ/h) of the specimen, and the vertical axis corresponds to the applied horizontal load P. Although individual crack progression was not traced cycle by cycle, the surface condition of the specimen was recorded every one second using a camera. The observation focused particularly on the region corresponding to the inner skirt height, where major cracking was concentrated.

Under ambient temperature conditions, as shown in Figure 7a, the specimen experienced initial cracking at the lower part of the pedestal when the horizontal load in the positive direction reached 12.6 kN, followed by a degradation in lateral load-carrying capacity. Upon subsequent loading in the negative direction, cracks also developed on the opposite side of the pedestal due to tensile coupling forces, eventually forming a continuous circumferential crack around the base. As the loading amplitude increased, rotational deformation became prominent, concentrated at the compressive toe located at the upper end of the embedded inner skirt. During this stage, the concrete cover on the compressive side progressively spalled, leading to the exposure of reinforcement bars.

Figure 7b shows the loading test results for the specimen subjected to thermal loading and subsequent cooling. In this case, the initial stiffness is reduced to 1500 kN/rad, in contrast to the unheated specimen, which exhibited an initial stiffness of 7316 kN/rad. When the horizontal load in the positive direction reached 6.0 kN, tensile cracking occurred at the embedded height of the inner skirt, initiating a reduction in strength. Similar to the ambient-conditions case, cracking subsequently occurred on the opposite (tensile) side of the pedestal during the negative loading, leading to the formation of a circumferential crack at the pedestal base. With increasing loading amplitude, progressive spalling of the mortar cover on the compressive side was observed, ultimately exposing the reinforcement at the end of the final cycle.

The failure mode of the specimen is visualized in Figure 7c. A crack propagating circumferentially appeared at the identical height of the inner skirt. As suggested in Figure 2, the horizontal force transfer to the ground mainly occurs by means of the stress transfer between the covering concrete and inner skirt, making the critical section at the top of inner skirt. The failure mechanism observed in this experiment corresponds to this suggestion.

Figure 8, Figure 9 and Figure 10 illustrate the failure process observed from the various locations. The specimen is the one that experienced heat exposure. The specimen does not show notable cracks up to the ultimate strength. The crack propagates particularly during the degradation phase in the force–displacement relationship. Concrete spalling is also visible in Figure 8c,d. After unloading, the crack closed, and, thereby, the detection of structural damage may be difficult during normal operation. The internal view of the specimen also exhibited concrete spalling at the same height as the inner skirt. These characteristics somewhat resemble the present situation of the target structure; still, the experiment could not reproduce the vanishment of concrete. Further investigation is necessary to reveal the damage mechanism of this structural system.

4. Finite Element Analysis of Pedestal Considering the Interaction Between the Inner Skirt and Concrete

To evaluate the vertical bearing capacity of the target structure, a detailed finite element model was developed using ABAQUS (v2024), which is a finite element simulation package software. The constructed simulation model is illustrated in Figure 11. The foundation concrete and inner skirt plate were modeled using eight-node reduced integration solid elements (C3D8R), while the reinforcement was represented by two-node truss elements (T3D2) in three-dimensional space, in conformity to previous analytical studies [35,36,37,38,39]. For interaction definitions, the “Embedded Region” constraint was employed to simulate the composite action between the reinforcement and concrete. While the bond strength between concrete and reinforcement may deteriorate due to heat exposure, modeling such degradation requires a fully detailed three-dimensional representation of reinforcement bars in ABAQUS. Given the high number of reinforcement bars in the model, this level of detail is computationally prohibitive. Moreover, the primary objective of this study is to evaluate the vertical bearing capacity, where reinforcement primarily contributes by confining the encased concrete. Therefore, the bond degradation is assumed to have a limited influence on the simulation results. Based on this consideration, the reinforcement bars were modeled using the embedded region approach.

In the numerical simulation, the adhesive bond giving the anchorage at the rebar edge was not modeled explicitly, as the vertical load-bearing capacity is minimally influenced by the adhesion at the base of the reinforcement. A “Hard Contact” property was applied between the inner skirt plate and the concrete, with frictionless properties. The adoption of a frictionless setting stems from the uncertainty regarding the surface conditions of the steel and concrete interfaces after heat exposure. To ensure a conservative evaluation, the contribution of frictional resistance is intentionally disregarded, rather than assuming a conventional friction coefficient (typically 0.4–0.5). This modeling approach ensures computational accuracy while enhancing convergence efficiency through reasonable simplifications in contact definitions.

Figure 12a–c illustrate the hysteretic constitutive behavior of concrete under compressive stress, tensile stress, and cyclic loading conditions. Young’s modulus is calculated using Equation (1), where γ_c denotes the unit weight of concrete and σ_cm represents the compressive strength.

$$E_{cm}=3.35\times {10}^4\times {\left({\gamma }_c/24\right)}^2\times {({\sigma }_{cm}/60)}^{1/3}$$

(1)

According to EN 1992-1-1 [40], the compressive side depicted in Figure 12a adopts 40% of the compressive strength as the elastic limit. The hysteretic constitutive behavior between the elastic limit and the peak load is defined by Equation (2), while the strain corresponding to the peak load ε_cm is determined using Equation (3) [40]. The post-peak softening branch is modeled as a linear degradation curve based on the previous study by Nguyen and Kim [41], such that the load-carrying capacity reduces to 85% of the peak value at an ultimate strain of ε_cu = 0.01.

$${\sigma }_c=\left({\displaystyle \frac{k\left({\epsilon }_c/{\epsilon }_{cm}\right)-{\left({\epsilon }_c/{\epsilon }_{cm}\right)}^2}{1+\left(k-2\right)\left({\epsilon }_c/{\epsilon }_{cm}\right)}}\right){\sigma }_{cm}$$

(2)

$${\epsilon }_{cm}={0.07{\sigma }_{cm}}^{0.31}\le 0.28\%$$

(3)

On the tensile side, as shown in Figure 12b, the hysteretic constitutive behavior is defined using Equation (4), which incorporates crack width as a variable, based on an earlier report by Hillerborg [42]. The tensile strength is calculated using Equation (5) [43], and the fracture energy G_F is obtained from Equation (6) [44].

$${\sigma }_t={{\sigma }_{ctm}\left(1+0.5{\sigma }_{ctm}/G_F{w}_t\right)}^{-3}$$

(4)

$${\sigma }_{ctm}=0.291\times {{\sigma }_{cm}}^{0.637}$$

(5)

$$G_F=10\times {d_{max}}^{1/3}{{\sigma }_{cm}}^{1/3}$$

(6)

Here, σ_t represents the tensile stress in the concrete, w_t is the crack width, σ_ctm is the tensile strength, G_F is the fracture energy, and d_max is the maximum aggregate size.

Figure 12c presents the unloading and reloading stiffness under compressive and tensile cyclic loading, governed by the damage variables d_c and d_t. The following equation proposed by Goto et al. [45] is also adopted in this study.

$$d_t=1.24k_t/{\sigma }_{cm}{w}_t\left(\le 0.99\right)$$

(7)

$$d_c=\left\{\begin{array}{c}{\displaystyle \frac{k_{ci}{\epsilon }_c}{{\left[1+{\left({\epsilon }_c/{\epsilon }_0\right)}^n\right]}^n}}\left({\epsilon }_c\le 0.0814\right)\\ 0.3485\left({\epsilon }_c>0.0814\right)\end{array}\right.$$

(8)

The parameters k_ci, ε₀, n, and k_t are constants with values of 155, 0.0035, 1.08, and 386.65 N/mm²/m, respectively.

The degradation of concrete mechanical properties after thermal cycling (heating and cooling) is based on the report by the Architectural Institute of Japan (AIJ) [31], as shown in Figure 13. The stress–strain relationships reflecting the effect of elevated temperatures are obtained by computing the material properties of concrete at room temperature according to empirical temperature-based formulas and substituting them into the mentioned constitutive equations.

For the reinforcement, a tri-linear constitutive model is adopted, with yield stress and ultimate tensile strength used as breakpoints. The post-peak stiffness is set to 1/1000 of the initial stiffness. After thermal cycling, both the yield stress and ultimate tensile strength are reduced based on the experimental data reported by AIJ [31], as illustrated in Figure 14.

The list of simulation models is summarized in

Table 1. List of finite element analysis models.

№	Inner Skirt	Experienced Temperature	Rebar	Section Loss at Bottom
1	None	20 °C	None	0%
2	None	20 °C	Reinforced	0%
3	None	20 °C	Reinforced	50%
4	None	600 °C	Reinforced	50%
5	Standard	20 °C, 600 °C, 1200 °C	Reinforced	0%
6	Standard	20 °C, 600 °C, 1200 °C	Reinforced	50%
7	Thick	20 °C, 1200 °C	Reinforced	0%
8	Thick	20 °C, 1200 °C	Reinforced	50%
9	Low	20 °C, 1200 °C	Reinforced	50%
10	High	20 °C, 1200 °C	Reinforced	50%

, with the inner skirt shape, experienced temperature, and presence of section defects as variables. Uniaxial compression loading was applied, and the force was continuously increased until a deterioration in load-carrying capacity was observed.

Figure 15a presents the analysis results for № 1 (without inner skirt plate, 20 °C, no section loss). The gray solid line represents the result without mesh reinforcement, while the black solid line represents the result with mesh reinforcement. The predicted load-carrying capacity of 836 kN, shown in the figure, was calculated based on the concrete strength and base cross-sectional area. In the absence of mesh reinforcement, the load-carrying capacity closely matched the predicted value. Conversely, when mesh reinforcement was present, the confining effect increased the maximum capacity to 1140 kN. Therefore, under conditions where rebar anchorage remains effective, a conservative evaluation of load-carrying capacity can be made based on the concrete strength and cross-sectional area alone.

Figure 15b shows the results for № 5 (standard inner skirt plate, 20 °C, no section loss). The black solid line represents the total reaction force obtained from the analysis, the gray solid line indicates the load carried by the concrete, and the gray dashed line indicates the load carried by the inner skirt plate. At peak load, the axial force distribution was 81.4% carried by the concrete and 18.6% by the inner skirt plate.

Figure 15c shows the results for № 5 (standard inner skirt plate, 600 °C, no section loss). The maximum load-carrying capacity was 616 kN, a 54.0% reduction compared to the one at 20 °C. However, considering the concrete strength reduction factor of 0.35 at 600 °C, the overall decrease in structural capacity was smaller than the reduction in material strength. At peak load, the axial force distribution was 76.1% in the concrete and 23.9% in the inner skirt plate. Compared to the 20 °C case, the contribution of the inner skirt plate increased, but the overall ratio remained roughly 8:2.

Figure 15d presents the analysis results for № 5 (standard inner skirt plate, 1200 °C, no section loss). Under the condition of 1200 °C, the residual compressive strength of the concrete is nearly zero, resulting in a significantly reduced load-carrying capacity compared to the two previously discussed cases. However, considering that the base has not lost its vertical support capacity, this condition is considered to have limited correspondence with actual structural behavior.

Figure 15e shows the results for № 3 (without inner skirt plate, 20 °C, with section loss). Due to the reduced bearing area caused by internal section loss, the vertical load-carrying capacity was lower than that in the case without section loss. The predicted value based on the remaining area and concrete strength was 472 kN, and the analysis result was in good agreement with this value.

Figure 15f presents the results for № 6 (standard inner skirt plate, 20 °C, with section loss). In the presence of section loss, the inner skirt plate carried 46.6% of the axial force at peak load, resulting in a vertical load-carrying capacity approximately twice that of the case without the skirt plate. This capacity is comparable to that predicted based on the concrete strength and cross-sectional area under unheated and undamaged conditions, indicating the significant contribution of the inner skirt plate to structural performance.

Figure 15g shows the results for № 4 (without inner skirt plate, 600 °C, with section loss), where the vertical capacity further declined due to the degradation in concrete strength. In contrast, Figure 15h presents the results for № 6 (standard inner skirt plate, 600 °C, with section loss), showing that the inner skirt plate carried 47.7% of the axial load, resulting in a notable enhancement of the overall vertical capacity.

Figure 15i–l show the results for № 7 (thick inner skirt plate, 20 °C, no section loss), № 8 (thick inner skirt plate, 20 °C, with section loss), № 9 (short inner skirt plate, 20 °C, with section loss), and № 10 (tall inner skirt plate, 20 °C, with section loss), respectively. Even with variations in the geometry of the inner skirt plate, its impact on the vertical load-carrying capacity was relatively small. Therefore, for existing structures equipped with inner skirt plates, a certain level of contribution to load capacity can be expected.

5. Evaluation of the Vertical Load-Bearing Capacity of Pedestal

Figure 16 shows a comparison between the evaluated and analytical values of bearing capacity. The horizontal axis represents the evaluated value, and the vertical axis represents the analytical value. The evaluation formula for the vertical bearing capacity is as follows: the calculated formula is the product of the concrete strength and the cross-sectional area.

$$P=\sigma \left(T\right)\times A_c$$

(9)

Here, σ(T) represents the concrete strength considering the temperature effect, and A_c represents the concrete cross-sectional area considering defects. The calculation of σ(T) adopts the relationship between heating temperature and residual strength shown in AIJ’s guideline [31].

As can be seen from Figure 16a, under the combined conditions of no inner skirt and no reinforcement, the data points are distributed along the diagonal line, indicating that the predicted bearing capacity values are roughly consistent with the actual analytical values. Furthermore, when the temperature rises to 1200 °C, the data points are concentrated near the origin because the concrete has almost no residual strength. However, as mentioned earlier, considering the current actual condition of the pedestal, the possibility of this condition matching the actual phenomenon is low.

Under other conditions, this evaluation formula can provide a conservative side evaluation of the analytical results. When there is a cross-sectional defect and no inner skirt, the safety factor is reduced, but since the inner skirt still exists in actual structures, this situation does not match reality. In the analysis considering the contribution of the inner skirt to the bearing capacity, the actual bearing capacity values exceed the evaluated values by more than two times, and the evaluation results are still on the sufficiently conservative side.

Figure 16b illustrates the safety margin plotted against the maximum experienced temperature. The average values and standard deviations are also indicated to reflect the statistical variation in the results. The safety margin shows a general increasing trend as the temperature rises. This is mainly due to the increased structural contribution of the inner skirt as the concrete strength diminishes.

Figure 16. Evaluation of vertical bearing capacity of pedestal.

Figure 16. Evaluation of vertical bearing capacity of pedestal.

6. Conclusions

This study presented a comprehensive investigation into the cyclic behavior and structural performance of pedestals subjected to thermal exposure through both experimental testing and finite element analysis. The failure mechanisms under cyclic lateral loading and the associated load-bearing capacities were systematically examined. The key findings are summarized as follows:

(1)

Thermal exposure led to a significant reduction in concrete strength, disrupting the balance with the strength of the inner skirt. As a result, the cracking capacity decreased to 47.6% after exposure to 600 °C.

(2)

Cyclic loading tests on scaled-down pedestal specimens revealed that the lower region of the pedestal, particularly the height corresponding to the inner skirt, acts as a critical section. Exposure to elevated temperatures led to a measurable reduction in horizontal load-bearing capacity.

(3)

The influence of the inner skirt geometry on axial compressive strength was found to be minimal within the geometric variants tested. Instead, the thermal history of the concrete emerged as the dominant factor governing compressive resistance.

(4)

Even in the presence of sectional loss, the redistribution of axial forces to the inner skirt can compensate for the loss in load-bearing performance, provided that the section loss is limited to 50% at the inner surface. Additionally, the confinement effect provided by the reinforcement bars contributes to maintaining the structural capacity. Under this scenario, the inner skirt is also not expected to buckle or yield under the anticipated axial force.

(5)

A conservative estimation of the load-bearing capacity was achieved using the thermally degraded concrete strength and effective cross-sectional area, ensuring structural safety under post-heating conditions. The resulting safety margin was found to be 1.5 at ambient temperature and increased to 15.6 at an experienced temperature of 1200 °C.

This study demonstrated cyclic loading tests on two specimens, which limits statistical confidence. The simulation was conducted considering only the vertical loading cases. However, as mentioned in the manuscript, cyclic horizontal forces are also a critical concern that warrant further analytical investigation. This issue will be addressed in a subsequent study. In future studies, this trend will be validated through comprehensive experiments that account for heat exposure, pedestal configuration, and inner skirt geometry. Furthermore, shaking table tests will be conducted to reflect the factors contributing to performance degradation.