Keywords

1 Prologue

Compared to other countries, Korea has many stone cultural heritages. As of 2018, there are 755 state-designated architectural heritages, of which stone cultural heritages are 573, making up 75.9% of the total. Many historically important stone pagodas have outstanding material durability, with simpler structures than those made of wood. Such stone cultural heritages are exposed to the atmosphere for a long time and deteriorated due to chemical, physical and biological reasons. They are also exposed to damage caused by natural disasters like floods, earthquakes and typhoons.

To date, the management of stone cultural heritages has mainly focused on damage and reinforcement. However, we must now establish a structural system based on structural stability in combination with preparation of stability assessment criteria based on scientific methodology.

This paper presents the results of research conducted by NRICH on Cheomseongdae, a stone cultural heritage from 2009 to 2013. A non-destructive investigation, seismic hazard assessment, and structural stability assessment are conducted. Notably, the scientific measurement results from the analysis of the earthquake behavior in Cheomseongdae and a dynamic centrifugal model experiment showed very similar results as those of the earthquake in 2016. Hence, a basic guideline for the conservation of stone cultural heritages is given as the conclusion, based on a comparison between the study on structural characteristics of the stone cultural heritages and the actual damage by an earthquake.

2 Historical Background

Cheomseongdae is the oldest astronomical observation facility in the East that was built in A.D. 632–647. It has been designated as a national treasure that represents Korea. Though many earthquakes are recorded on historical records, but so far no history of interventions since the construction of Cheomseongdae has been found. (see Table 1) Standing approximately 9 m in height, this observatory consists of a cylindrical body, a platform, and a square top. The cylindrical body includes 27 layers of fan-shaped stones, and the outer face is trimmed smooth whereas the interior wall face is not even. At the upper part, long ends of stone material that is geared into inside part of hash shape are extended to the outside. This arrangement can be seen in layers 19–20 and 25–26, and so it assumed that a ladder was easily placed on those steps in the inside. (see Fig. 1a–d) Around the southeast opening, the lower part is filled with rubble, while the upper part to the top is open and hollow. At the top, there are 2 stone layers with square-shaped (or lattice shaped). As their shape looks after the shapes of Chinese character, they are called as ‘jeongjaseok’ which means ‘a well-character stone’. The upper stone layer is made up of 4 stones that compose each side. Each edge of these stone is halved and combined like a prominence and depression, but some part of SE and NW corner are damaged. Cheomseongdae tilted to the north by approximately 1° and there is a large window at the south side. (see Fig. 1e–g).

Table 1. Historic records on seismic damage in Gyeongju
Full size table
Fig. 1.
figure 1

(a) Cheomseongdae; (b) South elevation; (c) Section (N-S); (d) Plan; (e) East side; (f) Degree of tilting to north; and (2) Displacement of center axis

Full size image

In Gyeongju, where Cheomseongdae is located, multiple earthquakes have occurred to date, and the results of the damage have been recorded in a few historical documents like “The Chronicles of the Three State (Samguksaki)” and “History of Goryeo (Goryeosa)” [1]. According to the recent research result, there were more than 2,000 earthquakes in Korea from A.D. 2 to 1904. Modern seismic observation in Korea began in 1905, and also a modern seismometer was first installed in 1963. As of December 2017, 156 seismic observation stations set up all over the country.

3 Scientific Analysis

3.1 Research on the Characteristics of Stone and Deterioration

The characteristics of the material that forms Cheomseongdae were identified, and a quantitative deterioration assessment on weathering and damage was performed. For this purpose, rock fragments exfoliated from the surface and rubble from the inside were collected, and their mineralogical compositions and organizational characteristics were observed using microscopes for both polarization and reflection (Nikon, Eclipse E600 W). Further, the movements of the rock-forming minerals were observed by using X-ray diffraction analyzer (Rigaku, D/MAX-IIB). Moreover, the magnetic susceptibility of the entire rock was measured using a magnetic susceptibility meter (ZH Instruments, SM 30). The damage states on each stone were combined by type to compile a weathering map and the percentage shares by damage type were calculated. An ultrasonic velocity meter (CNS FARNELL, Pundit Plus) was used to calculate the quantitative weathering degree by estimating the material strength and modeling was done using 2D graphics specialized program. Furthermore, the scaling deterioration and actual damaged area were analyzed by combining infrared camera (FLIR, P640) and acoustic tests.

Therefore, it was confirmed that Cheomseongdae is deteriorated by the combined effect of physical damage, such as cracks, delamination, exfoliation, separation, discoloration, and biological contamination. (see Fig. 2a) Except for the top part, every column showed separations between stones (1–135 mm, 19 mm on average). The stone weathering degree shown was in a wide range from the 2nd grade that is a less weathered stage to the 5th grade that is substantially weathered grade. Most parts are distributed over 3rd and 4th grade. (see Fig. 2b) As structural inequities caused by the uneven settlement of the ground was identified, it was verified that close monitoring of structural stability and step-by-step measures were necessary.

Fig. 2.
figure 2

(a) Biological weathering map; and (b) Ultrasonic speed survey map

Full size image

Therefore, in addition to the periodic stability inspection for structural displacement, vibration measurements, visual inspection, and compressive strength performed once or twice a year since 2004, a real-time monitoring system to closely monitor the structural displacement has been installed and operated. Primarily, since the earthquake occurred on 09.12. 2016, an earthquake accelerometer, accelerometer and underground water level meters were installed and operated.

3.2 Structural Stability Analysis

Cheomseongdae is currently tilted northward at approximately 1°, and many researchers have conducted 3D scanning, model tests, and ground geophysical surveys in this regard. In general, there is a possibility that the ground vibration at an adjacent main road north of Cheomseongdae in the past caused the coarse-grained ground subsidence and resulted in the inclination.

3.2.1 Ground Survey

The ground survey is essential to determine the seismic force applied to Cheomseongdae at the time of an earthquake. For Cheomseongdae, a point each at the south and north was drilled, and down-hole, a seismic wave test and the surface wave test on neighboring ground were performed at each exploratory hole. Samples were collected from the two exploratory holes, and the drilling log was secured. Analyzing the samples revealed that the ground of Cheomseongdae is made up of layers of sand mixed sand or silty sand which was mixed with gravel to 15–17 meters deep.

The supporting bedrock is located in 17 meters deep and it is composed of soft rock weathered from granite. The one-dimensional ground response analysis of Cheomseongdae using drilling log and seismic columnar section showed that the maximum ground acceleration increased to 1.85 times at the south exploratory hole 1 and to 1.69–1.73 times at the north exploratory hole 2 [2]. Based on the results of the drilling investigation, a resistivity survey was conducted to create a 3D image of the geological structure. The result shows that a low-resistivity anomaly zone appeared NW and NE of Cheomseongdae at the lower part at 8–9 meters. The ground in this area is judged to be less rigid than in other areas. The reasons for such results are probably caused by artificial ground improvement work or differentiated amount of construction pledges at the time of initial construction, and discriminatory weathering. (see Fig. 3a–b).

Fig. 3.
figure 3

(a) Linear detonation method seismic survey, and (b) 3-dimensional electric resistivity survey

Full size image

3.2.2 Analysis of Ground Response to a Scenario Earthquake

Currently, there are no seismic performance standards for cultural heritage in Korea. However, considering the importance of architectural cultural heritage and their impact on decay, seismic accelerations of the special class and first class, which respond to anti-collapse level of earthquake recurrence periods of 2,400 or 1,000 years, respectively, were used for the analysis. The ground strata of the southern and northern area of Cheomseongdae are not homogeneous, and their characteristics are different to a certain extent. The analysis was carried out using elastic wave velocity columnar section and drilling columnar section extracted through a ground drilling survey. The natural frequencies of Cheomseongdae were analyzed to be 4.52 Hz in the S-N directions and 4.59 Hz in the E-W directions. When converted to ground natural period, it was estimated that at 0.22 s. Therefore, it was estimated that an earthquake at this site is likely to cause resonance and amplification. However, due to the roughness between the stones and the internal packing materials, it is considered that the Cheomseongdae have a self-resistivity to horizontal dynamic loads such as earthquakes.

3.2.3 Evaluation of Seismic Behavior Characteristics by Dynamic Centrifuge Test

In 2011, the NRICH made a miniature model of Cheomseongdae and conducted a dynamic centrifugal model test with Chungnam National University and KAIST. The evaluation of the seismic hazard was conducted based on the results of the ground response analysis on the lower ground. Through this process, the maximum acceleration at the surface, the shear strain, the shear stress and the response spectrum by structure cycle were derived.

The experimental model of centrifugal force of the vibration table for evaluating structural performance against earthquakes was constructed at a scale of 1/15 of the actual model with granite, which is the same material as the stones of Cheomseongdae. Based on the load and the ratio of similarity, a centrifugal force of 15 g, which is 15 times of the gravitational acceleration, was applied vertically to this model.

The granite specimens for use in the model were fabricated and the elastic wave velocity, elastic modulus and density were determined through FFRC (Free-Free Resonant Column) test and density test. The granite used in Cheomseongdae was estimated to have a shear wave velocity of 2295 m/s. The density and stiffness of the stones used in the model were not significantly different from those of the stones used in Cheomseongdae. Therefore, the actual seismic behavior simulated in the centrifugal model test is reliable.

Two Cheomseongdae test models were constructed: model 1 has a square-shaped stone while model 2 does not have. The earthquake simulated with a level of collapse prevention 2nd class (recurrence period 500 years) was applied to the models. Model 1 showed 1.4 mm of displacement in the 15 g gravitational earthquakes over 1 s. Under same conditions, model 2 showed a displacement of 3.7 mm. Meanwhile, the displacement of the east sided stone of square-shaped stone on top of Cheomseongdae was significantly moved compared to the west one. Based on the results of this test, the actual displacement calculated as the actual size of Cheomseongdae were 21 mm and 55 mm, respectively, for earthquake duration of 15 s. In addition, the rectangular stones joined across the middle of structure at the 19th, 20th, 25th and 26th layer of the body were found to be critical structural elements for the seismic capacity of Cheomseongdae. Through this test, it was confirmed that stone structures such as a square shaped stone of Cheomseongdae which is composed of masonry relatively vulnerable to earthquake, were needed to ensure structural performance against lateral loads, namely earthquake shaking. As such, the repair based on the identification of the principle of the structural originality and the construction technique of traditional architecture, can be regarded as an effective and authentic method for the response to earthquake. (see Fig. 4a–f).

Fig. 4.
figure 4

(a) Geo-centrifugal test; (b) Model of Cheomseongdae; (c) Model 1, before test; (d) Model 1, after test; (e) Model 2, before test; (f) Model 2, after test.

Full size image

4 9.12 (Gyeongju) Earthquake and Its Characteristics of Damage

On September 12, 2016, at 19:44 (KST)., an earthquake measured 5.1 on the Richter scale occurred 8.2 km southwest of Gyeongju (35.77° N, 129.19° E). Subsequently, at 20:32 (KST), the main earthquake measured 5.8 on the Richter scale occurred 8.7 km southwest of Gyeongju. (35.76° N, 129.19° E) A week later, on September 19, at 20:33, an aftershock measured 4.5 on the Richter scale occurred. The earthquake, named as 9.12-earthquake, is a classic strike-slip fault, presumed to have been affected by the Great East Japan earthquake, which measured 9.0 on the Richter scale and occurred in the Tohoku region of Japan on March 11, 2011 [3]. The earthquake was the most powerful earthquake ever measured in the history of Korean earthquake observation. This earthquake caused a nationwide alert and caused minor damages to 97 cultural heritages including Cheomseongdae. And most damages were minor ones, such as damages on fences, roof-tile sliding and partial corruption of non-structural walls.

Cheomseongdae was tilted by about 0.1° to the north due to this earthquake. The stones on south-east corner of the top square-shaped stone were separated by about 5 cm. And south side stone moved to the north by about 3.8 cm. This damage pattern is partially similar to the damage pattern caused by the earthquake with collapse prevention 1st grade (recurrence period 1,000 years) in the above-mentioned dynamic centrifugal test.

Before the earthquake, the center axis of Cheomseongdae was tilted towards the north by 20 cm. Furthermore, the southeast and northwest corner had partially lost their original joint before the earthquake. Particularly, the displacements of the stones in the corners of square shaped stones were almost the same as that observed in test. Given that the structure of the square shaped stones displaced, it can be assumed that the damages of corners were caused by previous earthquakes. Moreover, a precise analysis of the results of the 3D survey conducted immediately after the Gyeongju earthquake and the results of previous inspections revealed that the upper square-shaped stones of Cheomseongdae was twisted and rotated clockwise at the same time. Given the seismic response capacities of the upper square-shaped stones, this analysis indicates that restoration of the damaged corners can improve its seismic capacity. (see Fig. 5a–c, Table 2)

Fig. 5.
figure 5

(a) Additional displacement; (b) Displacement of upper stone; and (c) Separation on SE corner to Gyeongju earthquake (in red color - approx. 5 cm)

Full size image
Table 2. Measurement result on Cheomseongdae, before & after 9.12-Gyeongju Earthquake
Full size table

5 Conclusions

The one of goal of cultural heritage management can be said to conserve its original characteristics. So, obtaining response capacity to resist to disasters such as earthquake is essential. In case of earthquake, from structural point of view, it is need to obtain the response capacity based on the premise of conserving the original characteristics.

This study conducted various scientific studies on Cheomseongdae and analyzed the actual damage. Based on the results of this study, the following conclusions can be made.

  • The structural originality of cultural heritage should be respected, and research on the inherent structural principles is necessary.

  • The similarity between the results of a comprehensive study on ground/structures and actual damages caused by earthquake was verified. Therefore, it is necessary to apply such methodologies broadly to other cultural heritage and improve the disaster response capacity.

  • Cheomseongdae has very good response capacity to earthquakes structurally, and has been found to have a structural performance that is close to the basic seismic design standard.

  • Management of stone cultural heritage should focus not only on the conservation of the original characteristics but also on improving disaster response capacity in terms of structure and construction. So, it is necessary to conduct researches on traditional techniques and modern scientific techniques.