TSOGTBAYAR NYAMTUYA

Preliminary Assessment of ADXL355 for Vibration and Inclination Measurement in Earthquake Damage Analysis

Masataka SHIGA, Takaaki IKEDA

When conducting on-site inspections of infrastructure damaged by earthquakes is challenging, we focus on developing an affordable system for remote structural health monitoring. Health monitoring using MEMS (Micro-Electro-Mechanical Systems) accelerometers is effective, especially when on-site access is impossible or poses safety risks.
Some background studies highlighted the shift from traditional, expensive seismographs or vibrometers to more miniature, more affordable MEMS accelerometers. However, there are issues concerning the precision of MEMS accelerometers and their proper system management and operation. These issues have been under-researched and are addressed in this study, focusing on the measurement accuracy of the MEMS accelerometers, particularly the vibration and inclination characteristics of the ADXL355 accelerometer.
The experiments were carried out in two stages. First, a small shaking table was used to compare the ADXL355 with existing accelerometers (ASW-5A) and velocimeters (VSE-15D) under different shaking conditions. Details of the equipment, sensors, vibration table settings, experimental conditions, data organization, and analysis procedures are meticulously documented. In particular, we conducted experiments under different amplitude and frequency conditionm, and analyzed the results. The accelerometer data was organized, and discrete Fourier transformations were used to evaluate the vibration characteristics and frequency components and investigate the performance of the MEMS accelerometers.
Next, for future assessments of the building's inclination of remote monitoring using MEMS accelerometers, we set angles on the vibration table using a protractor. We analyzed the results, finally discussing the maximum detectable tilt angle.
In conclusion, we improved the system by separating the ADXL355 and a microcontroller onto different boards. This allowed for relatively accurate measurements, but there are still significant challenges for real-world application, including improving the accuracy of individual modules and evaluating their performance in real-world environments.