The microelectromechanical systems (MEMS) accelerometer built into a smart meter (SM) has a nominal digital resolution of 16 bits. However, this resolution collapses to 7 bits of information per sample when used in an urban environment. This collapse in resolution limits the sensitivity required to effectively operate the earthquake early warning platform (EEWP). In this study, we evaluate the performance of the MEMS sensor in present SMs with respect to a reference sensor, with a special focus on its poor noise power spectral density (PSD, $~1\mathrm{mg}/\sqrt{\mathrm{Hz}}$). We also explore the general capacity of the SM in an IoT-based EEWP and provide explicit information regarding the 16-bit digital MEMS accelerometer. Then, we investigate the functionality of the sensor in the context of event detection in the presence of background vibration. When the value of acceleration root mean square (RMS) exceeds 20 mg, the meter’s error decreases to <20%, whereas the peak ground acceleration error decreases to <20% for the peak value greater than ~70 mg. The MEMS sensor is unreliable for motions with a peak acceleration of less than 148 mg or those with an RMS value less than 46 mg. However, we note that SMs exhibit reasonable amplitude and phase coherence for frequencies above 1 Hz with respect to the reference accelerometer. To enhance the sensitivity, averaging 1000 coherent accelerometer observations enhances the digital resolution to 14 bits, which allows the efficient usage of the network bandwidth. Since the accelerometer is used as an anti-tampering mechanism, the SM is similar to a tiltmeter. Therefore, it is necessary to reconfigure SMs for early warning systems. Despite the challenges, the use of SM for an IoT-based EEWP is technically feasible.

内置在智能电表（SM）中的微机电系统（MEMS）加速度计的标称数字分辨率为16位。但是，在城市环境中使用时，每个样品的分辨率降低到7位信息。分辨率的下降限制了有效运行地震预警平台（EEWP）所需的灵敏度。在这项研究中，我们评估了目前SM中MEMS传感器相对于参考传感器的性能，并特别关注其差的噪声功率谱密度（PSD，$〜\mathrm{1个}\mathrm{毫克}/\sqrt{\mathrm{赫兹}}$）。我们还将探讨基于IoT的EEWP中SM的一般容量，并提供有关16位数字MEMS加速度计的明确信息。然后，我们在存在背景振动的情况下，在事件检测的情况下研究传感器的功能。当加速度均方根（RMS）的值超过20 mg时，仪表的误差减小至<20％，而对于峰值大于约70 mg的地面加速度峰值减小至<20％。对于峰值加速度小于148 mg或RMS值小于46 mg的运动，MEMS传感器不可靠。但是，我们注意到，相对于参考加速度计，SM对于高于1 Hz的频率表现出合理的幅度和相位相干性。为了提高灵敏度，平均进行1000次相干加速度计观测，可以将数字分辨率提高到14位，从而可以有效利用网络带宽。由于加速度计用作防篡改机制，因此SM类似于倾斜计。因此，有必要为预警系统重新配置SM。尽管存在挑战，但将SM用于基于IoT的EEWP在技术上是可行的。