Abstract:The traditional fault detection method based on time-series data mining cannot perform sensitivity analyses of the high-sensitivity electronic earthquake early-warning device, and its fault detection rate is low. In this paper, we propose a new fault detection method for the high-sensitivity electronic earthquake warning device. Based on the device structure, we constructed a mechanical model using a mass block, hinge, and extended beam to obtain the sensitivity expression of the device. According to this expression, we established a multi-signal model correlation matrix, and obtained the fault source cost of the fault preemption probability. We adopted an improved multi-signal model detection method based on the fault probability of the fault mode in detecting faults of the electronic earthquake warning device. The experimental results show that the fault detection rates of the two indicators G and I in the TE process of the electronic earthquake warning device are 0.989 and 0.905, respectively, the fault detection time of the PL process is 180 s, and the fault detection rate is 0.412 8. These values are all higher than those obtained by the traditional fault detection method based on time-series data mining, which shows that the proposed method has higher fault detection performance.