Accurate seismic vulnerability assessment requires high quality and large amounts of ground motion data. Ground motion data generated from time series contains not only the seismic waves but also the background noise. Therefore, it is crucial to determine the high-pass cut-off frequency to reduce the background noise. Traditional methods for determining the high-pass filter frequency are based on human inspection, such as comparing the noise and the signal Fourier Amplitude Spectrum (FAS), f2 trend line fitting, and inspection of the displacement curve after filtering. However, these methods are subject to human error and unsuitable for automating the process. This study used a deep learning approach to determine the high-pass filter frequency. We used the Mel-spectrogram for feature extraction and mixup technique to overcome the lack of data. We selected convolutional neural network (CNN) models such as ResNet, DenseNet, and EfficientNet for transfer learning. Additionally, we chose ViT and DeiT for transformer-based models. The results showed that ResNet had the highest performance with R2 (the coefficient of determination) at 0.977 and the lowest mean absolute error (MAE) and RMSE (root mean square error) at 0.006 and 0.074, respectively. When applied to a seismic event and compared to the traditional methods, the determination of the high-pass filter frequency through the deep learning method showed a difference of 0.1 Hz, which demonstrates that it can be used as a replacement for traditional methods. We anticipate that this study will pave the way for automating ground motion processing, which could be applied to the system to handle large amounts of data efficiently.

PURPOSES : The tire-pavement interaction noise (TPIN) comprises four sources, among which the tire tread vibration noise (TTVN) and air pumping noise (APN) are known to be the most influential. However, when evaluating TPIN, the noise level is estimated based on the overall noise, because general noise measurement methods cannot separate TTVN and APN. Therefore, this study aims to develop a method to separate TTVN and APN in TPIN for quantitative assessment of pavement noise. METHODS : Based on the results of our literature review and frequency band noise data measured in our study, we identified the dominant frequency ranges for TTVN and APN. Additionally, we evaluated TTVN and APN across various pavement types. RESULTS : TTVN was found to be dominant in frequency bands below 800 Hz, while APN was dominant in frequency bands above 800 Hz. Additionally, regardless of the vehicle type, vehicle speed, or pavement type, APN exhibited higher levels compared to TTVN. This result shows that APN has a more significant impact on TPIN than TTVN. CONCLUSIONS : The separation method of TTVN and APN proposed in this study can be utilized to quantitatively assess the relationship between the primary noise sources in TPIN and the characteristics of pavement texture in future research. Furthermore, it is anticipated that characteristics of low TPIN and optimal texture conditions can be proposed to mitigate TPIN, thus contributing to the development of lownoise pavements.

Recently, domestic fishing production of Japanese horse mackerel has been continuously decreasing. To achieve sustainable fishing of this species, it is essential to acquire its target strength (TS) for accurate biomass estimation and to study its ecological characteristics. To date, there has been no TS research using a broadband echosounder targeting Japanese horse mackerel. In this study, for the first time, we synchronized an underwater camera with a broadband frequency (nominal center frequency of 200 kHz, range: 160-260 kHz) to measure the TS according to the body size (16.8-35.5 cm) and swimming angle of the species. The relationship between Japanese horse mackerel length and body weight showed a general tendency for body weight to increase as length increased. The pattern of the frequency spectra (average values) by body length exhibited a similar trend regardless of body length, with no significant fluctuations in frequency observed. The lowest TS value was observed at 243 kHz while the highest TS values were recorded at 180 and 257.5 kHz. The frequency spectra for the swimming angles appeared to be flat at angles of –5, 0, 30, 60, 75, and 80° while detecting more general trends of frequency spectra for swimming angle proved challenging. The results of this study can serve as fundamental data for Japanese horse mackerel biomass estimation and ecological research.

본 논문에서는 볼트로 체결된 구조체에 대하여 초기 볼트풀림 상태에서의 볼트 체결력 예측 합성곱 신경망 훈련 방법을 제시한다. 8개의 볼트의 체결력이 변경된 상태에서 계산한 주파수응답들을 완전 체결된 상태의 초기 모델과의 크기 및 모양 유사성을 표현하는 유사성 지도로 생성한다. 주파수응답 데이터들의 생성에는 크리로프 부공간법 기반의 모델차수축소법을 적용하여 효율적인 방법으 로 수행할 수 있도록 한다. 합성곱 신경망 모델은 회귀 출력 계층을 사용하여 볼트의 체결력을 예측하도록 하였으며, 훈련 데이터의 개 수와 합성곱 신경망 계층의 개수를 다르게 준비하여 훈련시킨 네트워크들을 비교하여 그 성능을 평가하였다. 주파수응답에서 파생되 는 유사성 지도를 입력 데이터로 사용하여 초기 볼트풀림 영역에서 볼트 체결력의 진단 가능성과 유효성을 제시하였다.

Controller modeling is essential for the design. It allows various control techniques to be simulated in advance, and various interpretations can be performed. If this is not the case, we need to reverse engineering in the real system developed by others. In this paper, controller modeling was reversely designed using the frequency test results of the target system. First, the characteristic equation of the target equipment was based on and a block diagram was assumed. Thereafter, controller variables were estimated using the frequency test results for each of the four control loops. In addition, time response simulations were performed using the estimated controller modeling. This method is thought to be of great help to reverse engineering in situations where there is completed equipment but no controller modeling.

In September and October 2020, combined acoustic and trawl surveys were conducted in the northwestern sea of Jeju Island. In the survey area, a region, so called Jeju region, was designated to esimate the biomass of chub mackerel and jack mackerel using a trawl survey method and frequency difference method. In the September survey, the weight ratios of jack mackerel and chub mackerel to the total catch were 24.6% and 2.8%, respectively, and in the October survey, those ratios were 24.9% and 20.7%, which were used to calculate their biomass (trawl survey). Using the frequency difference range (–8 to –3dB) corresponding to two species in 120 and 200 kHz, their biomass was estimated (frequency difference method). As a result, the biomass of two species from the trawl method was 3252.3 tons in September and 5777.0 tons in October. The estimated biomass by the frequency difference method was 4926.6 tons in September and 7521.5 tons in October. It was the first trial to estimate the biomass of two species using the trawl and frequency differencing methods in South Korea although there were some differences between two methods. In addition, horizontal distributions of acoustic backscattering strength over the entire survey area were mapped.

PURPOSES : The purpose of this study is to identify the dynamic behavior of a cement concrete paving machine (paver) by measuring its response using accelerometers. This is because the dynamic behavior of pavers affects the quality of data from various applications of IoT sensors, such as laser, ultrasonic, optical sensors and so on. Therefore, it is believed that the understanding of dynamic behaviors can contribute to the effective use of various IoT sensors for the acquisition of real-time quality control data in pavement construction. METHODS : Dynamic signals are obtained using accelerometer sensors to identify the dynamic characteristics of paving machines. The main parameters for acquiring dynamic signals are the status of the machine’s operating or standby conditions, and available locations for attaching various IoT sensors. Time domain data are logged at a particular sampling speed using a low-pass filter, subsequently, they are converted to digital data, which are analyzed on three rectangular axes. In addition frequency analysis is conducted on the measured data for identifying the peak frequencies, via FFT (Fast-Fourier-Transform) using MATLAB. RESULTS : The magnitude of the x-directional vibration is higher than that of any other direction under the paver’s operating or standby condition. However, signals from the smoother beam show that the z-directional vibration is more significant in the operating status. It means that the primary vibration depends on the location. Furthermore, the peak frequencies are quite various depending on the status of a paver and its sensing location. CONCLUSIONS : The magnitude of machine vibration and peak frequencies at each status or location are identified from time- and frequency-domain data. When using IoT sensors for quality control or monitoring pavements in construction, the dynamic characteristics of a paver should be considered to mitigate the interference of signals from the paver body or its elements.

The governing equation for a dome-type shallow spatial truss subjected to a transverse load is expressed in the form of the Duffing equation, and it can be derived by considering geometrical non-linearity. When this model under constant load exceeds the critical level, unstable behavior is appeared. This phenomenon changes sensitively as the number of free-nodes increases or depends on the imperfection of the system. When the load is a periodic function, more complex behavior and low critical levels can be expected. Thus, the dynamic unstable behavior and the change in the critical point of the 3-free-nodes space truss system were analyzed in this work. The 4-th order Runge-Kutta method was used in the system analysis, while the change in the frequency domain was analyzed through FFT. The sinusoidal wave and the beating wave were utilized as the periodic load function. This unstable situation was observed by the case when all nodes had same load vector as well as by the case that the load vector had slight difference. The results showed the critical buckling level of the periodic load was lower than that of the constant load. The value is greatly influenced by the period of the load, while a lower critical point was observed when it was closer to the natural frequency in the case of a linear system. The beating wave, which is attributed to the interference of the two frequencies, exhibits slightly more behavior than the sinusoidal wave. And the changing of critical level could be observed even with slight changes in the load vector.

This paper examined the dynamic instability of a shallow arch according to the response characteristics when nearing critical loads. The frequency changing feathers of the time-domain increasing the loads are analyzed using Fast Fourier Transformation (FFT), while the response signal around the critical loads are analyzed using Hilbert-Huang Transformation (HHT). This study reveals that the models with an arch shape of h = 3 or higher exhibit buckling, which is very sensitive to the asymmetric initial conditions. Also, the critical buckling load increases as the shape increases, with its feather varying depending on the asymmetric initial conditions. Decomposition results show the decrease in predominant frequency before the threshold as the load increases, and the predominant period doubles at the critical level. In the vicinity of the critical level, sections rapidly manifest the displacement increase, with the changes in Instantaneous Frequency (IF) and Instant Energy (IE) becoming apparent.