Fault diagnosis and prediction of accelerometer servo circuit based on machine learning algorithm

In view of the numerous uncertain factors in the manufacturing of quartz flexible accelerometer servo circuit, the tolerance effects of internal component parameters and fuzzy failure location, this paper proposed a network structure algorithm based on machine learning. The algorithm learnt and was trained according to the output function parameters of the circuit, in order to accurately determine the cause of the circuit fault and locate the fault through quantitative analysis. Through the measurement data in the production process, our algorithm has an accuracy rate of 99.6% of circuit failure factor prediction. The application of the algorithm can effectively improve circuit testing efficiency and manufacturing yield. Furthermore, it can be extended to other electronic equipment manufacturing fields showing a high engineering applicable significance.