Site classification Based on Shear Wave Velocity Inversion in The Jlantah Dam Construction Project using the HVSR (Horizontal to Vertical Spectral Ratio) Analysis Method

This research was conducted with the aim to test whether the results of the inversion of the HVSR curve can be used as complementary data for N-SPT values in classifying areas according to site classification. The inversion result in the form of shear wave velocity (VS ) is then compared directly with the N-SPT data in several boreholes. Ellipticity curve method is used for inversion. The data generated in this study are dominant frequency, Amplitude Peak, and H/V curve as a result of the HVSR analysis performed on microtremor data. Furthermore, an inversion was carried out on the H/V curve by taking into account secondary data in the form of borehole, shear wave velocity, density and poisson ratio for each type of lithology. The dominant frequency value obtained has a value range of 0.33 to 17.15 Hz. The Amplitude Peak obtained in the study area has a value range of 1.81 – 11.4. The inversion (VS ) value has a range of 110.5 m/s – 400 m/s, so it is included in the classification of Soft Soil, Medium Soil, and Hard Soil and Soft Rock. The correlation of and N-SPT show that inversion method with HVSR analysis of microtremor data is quite reliable as supporting data for N-SPT. Considering cons of drilling, the method used in this study can be a new alternative for use in construction projects because it can be carried out at a relatively low cost, and is quite simple in terms of data acquisition and processing.


Introduction
Bendungan Jlantah is located in the Jatiyoso district, in the southern part of Karanganyar regency.Like other dams, Bendungan Jlantah serves the common functions of controlling the flow of river water, supplying water for electricity generation, and irrigation purposes.Many dams suffer from structural wall damage due to the strength and quality of the dam, as well as geological and geotechnical conditions [1].Bendungan Jlantah is currently under construction, and given these conditions, it is essential to conduct a soil classification study for construction planning.Before construction can commence, the project area or existing soil must be examined to determine if it can support the structural loads of the building or structure.Buildings, roads, and other structures often fail due to inadequate foundation consolidation on weak soil [2].Failures caused by weak soil are often due to a failure to identify soil expansion early in the design process [3].
In this study, an analysis of Horizontal to Vertical Spectral Ratio (HVSR) was conducted using microtremor data collected in the construction area of Bendungan Jlantah.Microtremor analysis with HVSR spectrum ratios is a cost-effective, environmentally friendly method that can be used to determine subsurface characteristics such as soil type and sediment thickness in a region through the interpretation of the H/V curve as a function of dominant frequency and Amplitude Peak [4].The HVSR approach is a field test used for soil deposit characterization.Subsurface horizontal layers have the ability to capture seismic body waves, which move up and down, while lateral variations in soil composition or thickness can result in the trapping of seismic surface waves, which, in turn, move back and forth, all these wave interferences lead to resonance patterns.The mechanical and geometric characteristics of layers determine the fundamental frequency [5].The HVSR approach is one of the simplest and most cost-effective geophysical methods in data acquisition and processing, thus, it holds priority in seismic microzonation and site effect studies.HVSR represents the most important factor in calculating the fundamental frequency for regions characterized by loose sediments if there is a significant contrast between these sediments and the underlying bedrock layers.Furthermore, this technique has been developed for calculating building frequencies as well [6].One significant benefit of utilizing microtremors in this study region is the infrequent incidence of earthquakes.Consequently, the research solely necessitates ambient noise gathered from the study area surroundings.Moreover, this approach eliminates the need for expensive and time-consuming techniques like seismic methods [7].
The HVSR method was initially introduced by Nogoshi and Igarashi (1971) and further developed by Nakamura (1989), hence it is more commonly known as the Nakamura method.Later, Herak (2008) introduced an inversion model to estimate numerical model parameters based on data using specific models.One modeling method that can be used for HVSR curve inversion is the ellipticity curve of Rayleigh waves.Rayleigh wave ellipticity as a function of frequency represents the shear wave velocity profile and sediment layer thickness.The shear wave velocity (Vs) obtained from this method is not recommended for deep soil values.However, for shallow soil structure profiles, this method can be used to accurately determine velocity profiles [8].
Based on the above description, there is a need for a study related to the construction planning of Bendungan Jlantah in the Jatiyoso district, Karanganyar regency.One approach that can be taken is to classify site classes based on Vs values as supplementary data to N-SPT.

Methodology
The acquisition was made to obtain primary research data.Data acquisition was carried out on the Jlantah Dam construction project located in Tlobo village, Jatiyoso district, Karanganyar regency, Central Java.The data obtained is vibration of frequency spectrum from three components, horizontal components (North-south and East-west) and vertical components (Up-down) [9].The utilization of Pythagorean equation to calculate North-South and East-West spectrum by combining their horizontal components, squared, and then determining the ratio of the average Fourier spectrum between the horizontal and vertical components in the frequency domain, yields the H/V value for that specific point [10].Measurement points of 36 points spread across the Jlantah Dam project area, can be seen in Figure 1 below, After obtaining the H/V curve, the shear wave velocity (Vs) will be determined using the ellipticity curve inversion method approach of the HVSR curve with dinver software on geopsy.Dinver works by iterating over the initial model to match the HVSR curve of the measurement results until the final model is obtained with the least mismatch value (misfit) against the HVSR curve of the measurement results.The initial model is made by entering the soil parameters to be reviewed   ,   , poisson ratio and density.Such parameters are included in secondary data obtained from geotechnical laboratory tests.In addition to requiring secondary data in the form of the 4 parameters mentioned, the inversion process also requires layer thickness.In this study, the determination of layer boundaries was carried out using dam project borelog data as a basis.

Cross Section 𝑉 𝑆
The process of making incisions/profiles,  S .Starting with making an incision topography based on the elevation of the Jlantah Dam project area.The process is done by dividing the incision into several parts, then making an elevation profile on each small part.Furthermore, the small parts are connected so that the elevation of the incision ground surface is obtained.The next process is to place the borelog at the point and elevation adjusted to the incision, then manually interpolate to connect each layer contained in the borelog on the Figure 2. The same applies to the log  S so that an incision is obtained that has lithological information and soil classification based on the inversion value of  S which can be seen in the results of the study.The distribution map of dominant frequency values and Amplitude Peaks are achieved by interpolating the data.The distribution map is shown in Figure 3. Theoretically, the dominant frequency value of the soil is a reflection of the physical condition of the soil.Soil or soft rock will have a long vibration period and vice versa, therefore soft rock has a relatively low dominant frequency value.Meanwhile, wave strengthening is strongly influenced by the thickness of the sediment and lithology in the study area.The dominant frequency is a frequency that is often seen, where this frequency defines the type and characteristics of rock layers in the area [11].The Jlantah Dam construction area has f 0 values which can be grouped into 4 classifications.The area around the Jlantah Dam is dominated by values f 0 < 2.5 Hz which are marked in dark green, this area is included in the type II type IV soil classification proposed by Kanai and Ometer-Nakajima [12] which is described as having a sediment thickness of more than 30 meters.Values of f 0 in the range 2.5 -4 Hz are classified as belonging to type III type III soil which is marked with a yellowish green color on the map.According to the local geology of the construction area, the area with this value range can be described as consisting of a layer of lapilli tuff (volcanic deposits) with a thickness of > 5 meters and a sediment thickness of around 10-30 meters.The area on the map marked in orange has a value in the range 4 -10 Hz which is included in the type IV type II soil classification which is composed of volcanic sediment (lapilli tuff) with a thickness of ±5 meters and surface sediment thickness of around 5 -10 meters.The area on the map in red has a value range of 10 -20 Hz which is classified as type IV type I soil which is composed of volcanic breccia with a very thin surface sediment thickness and is dominated by hard rock.High frequency values indicate that the subsurface structure is a layer of hard rock.Low frequency values indicate that the subsurface structure is a soft layer.The distribution of frequency values with low values is in areas that are not included in the dam construction area, so they still have considerably soft layers of soil.
Amplification is a phenomenon characterized by the strengthening of waves, occurring when seismic waves traverse a less rigid medium.A high amplification factor indicates that the region is undergoing wave reinforcement, rendering it susceptible to earthquake-induced damage.The amplification value is influenced by rock deformations and erosion.Figure 3b represents a map displaying Amplitude Peak values, as determined through HVSR processing.The lower Amplitude Peak values are linked to the elevated shear wave velocity ( S ) attributed to rocky terrain.These rock formations suggest that this area boasts higher rock density compared to other regions.Consequently, seismic waves passing through this area tend to experience reduced wave amplification.Conversely, regions with elevated  0 values due to substantial impedance contrasts exhibit variations in layer densities.This circumstance results in softer soil characteristics within this area, leading to more pronounced wave effects on structures at the surface level.

Inversion
The inversion process is carried out to obtain  S values from the HVSR processing output.The inversion method used is an ellipticity curve with output in the form of ground profiles (presented on Figure 4) which contain information in the form of values for each subsurface layer.The inversion results are largely determined by the parameter constraints used to initiate the initial model, namely  S for each lithology, poisson ratio, density and layer thickness.In this research, inversion parameters were obtained from geotechnical laboratory tests and used dam project borelog data as a basis.
The example of inversion results can be found in Figure 4a contains information such as  S value and layer depth which is presented on graph named 'Ground Profiles'.Based on the distribution of these  S value classifications, three cross-sections (Figure 5) were created containing site classifications based on  S value and N-SPT (main dam, cross-section of spillway and dam bypass tunnel).Next, a correlation is carried out by matching the shear wave velocity classification resulting from the inversion with the N-SPT classification according to the borelog data, with the correlation results in the form of a percentage match between the inversion results and the N-SPT data.The cross-sectional results were then correlated between the classification according to  S and the N-SPT data, the correlation is mainly just to show how similar the inversion results are to the bore log data.This correlation produces a distribution of sliced areas according to site classification.The slice distribution figure is calculated by comparing the area that meets the  S and N-SPT requirements with the total area.The total area is the depth of the borelog point up to 30 meters.The slice distribution number can then be used as a parameter that determines the suitability of the inversion results with the N-SPT data.The calculation of the slice area is presented in Table 1.It can be seen in that the inversion results and N-SPT data have a match of 93.4792% for the main dam cross-section, 89.14% for the cross-section of the spillway structure and 90.1667% for the crosssection of the dam bypass tunnel.With a slice distribution value of 91.4235%, it can be said that the site classification results of the inversion and N-SPT data have a high level of match.The weakness of research using this method is that inversion to determine is difficult to apply in areas with vertical subsoil heterogeneity [13].The inversion process using dinver software cannot carry out inversion with layers that have vertical subsoil heterogeneity, this is influenced by the relatively small range of values for each type of lithology (the difference in values between types of lithology is around 50 m/s) and the small density range causes the software to unable to carry out the inversion process.However, this condition is considered less significant, because the thickness of the soft layer between the hard layers is <5 meters.

Conclusion
Based on the analysis and discussion carried out on the research results, it can be concluded that the results of inversion using the ellipticity curve method carried out on microtremor data with HVSR analysis, obtained a range of values from 110.5 m/s to 400m/s.Based on the classification of SNI 1726-2019 (National Standards Agency, 2019) and NEHRP (Power et al., 2003), the value of inversion results in the Jlantah Dam area is included in the classification of Soft Soil/SE (< 175 m/s), Medium Soil/SD (175 < < 350 m/s), as well as Hard Soil and Soft Rock/SC (350 < < 400 m/s).According to the local lithology that makes up the Jlantah Dam area, the soft soil classification consists of residual soil.Medium soil consists of lapilli tuff, agglomerate, volcanic breccia, and random material (heap) consisting of a mixture of lapilli tuff and volcanic breccia.Meanwhile, hard soil and soft rock are composed of lapilli tuff and volcanic breccia with N-SPT values above 35.
Correlation of inversion results in the form of values with N-SPT values shows that inversion using the ellipticity curve method with HVSR analysis of microtremor data is quite reliable as supporting data for N-SPT, with a match of 89.5963%.Considering that drilling requires quite a large amount of money for each drill point, the method used in this research could be a new alternative for use in construction projects because it can be done at a relatively low cost, and is quite simple in terms of data acquisition and processing.
The ellipticity curve inversion method can reveal variations in shear wave velocity that can be associated with differences in site classification or even differences in lithology and soil compaction in the study area.Meanwhile, the HVSR method can be used as an effective alternative method in determining shear wave velocity at low cost and relatively short measurement time.The results of this research provide an understanding of the characteristics of shear wave velocity and geotechnical properties in the research area environment.The integration of HVSR analysis and the ellipticity curve inversion method provides a comprehensive approach in determining site classification based on shear wave velocity, with significant implications in planning and designing structures in development areas.

Figure 1 . 3 .
Figure 1.Data Acquisition Point Spread Across The Jlantah Dam Project

Figure 4 .
Figure 4b is summary of Borelog BS-10 which contain N-SPT value that can be used as correlation to inversion results to fulfill the objectives of this paper which were explained previously in the introduction.The inversion results are then classified into site classification according to SNI 1726-2019 and modified NEHRP classification based on dam construction area standards.Based on these two classification systems, the  S value of the inversion results in the Jlantah Dam area has a value range of 110.5 m/s -400 m/s, so it is included in the classification of Soft Soil (SE), Medium Soil (SD), and Hard Soil and Soft Rock (SC).According to the local lithology that makes up the Jlantah Dam area, the soft soil classification consists of residual soil.Medium soil consists of lapilli tuff, agglomerate, volcanic breccia, and random material (shell) consisting of a mixture of lapilli tuff and volcanic breccia.Meanwhile, hard soil and soft rock are composed of lapilli tuff and volcanic breccia with N-SPT values above 35.(a) Inversion Results of Number 31 Data Point, (b) Summary of Borelog BS-10

Table 1 .
Calculation of Inversion Results Suitability ( S ) with N-SPT data