Crystal resonator storage life forecasting method based on least squares support vector machine

Abstract

The invention relates to a crystal resonator storage life forecasting method based on a least squares support vector machine. The crystal resonator storage life forecasting method based on the least squares support vector machine includes the following four steps that firstly, a degradation mechanism of a crystal resonator is analyzed when the crystal resonator is stored for a long time, influences on parameter variation by the main degradation mechanism are obtained and degradation sensitive parameters of the crystal resonator are determined; secondly, a crystal resonator accelerated storage degradation test is designed and performed, the selected sensitive parameters are measured and test data are collected regularly; thirdly, the theory of the least squares support vector machine is used for processing the test data, and a degradation model of the sensitive parameters is established under different acceleration stress levels; fourthly, a parameter degradation model of the crystal resonator under normal stress is established, failure criteria are determined and the storage life of the crystal resonator is forecast. According to the crystal resonator storage life forecasting method based on the least squares support vector machine, the practical problems of small samples, non-linearity and the like in forecasting the storage life of the crystal resonator are solved, the calculation complexity is relieved, the rate and precision of convergence are improved and high popularization value is achieved.

Description

technical field [0001] The invention provides a method for predicting the storage life of a crystal resonator, which relates to the theory and algorithm of a least square support vector machine, and belongs to the technical field of accelerated test evaluation. Background technique [0002] Quartz crystal resonators are one of the key components used in the current electronics industry. Quartz crystal resonator, as a high-precision and high-stability oscillation source, is the most important way for various digital sequential circuit systems to generate clock synchronization signals. At the same time, it is also an The core component, a wide range of applications, the use is very common. During the storage of electronic equipment, the internal quartz crystal resonator is easily affected by the storage environment, resulting in performance degradation or failure. Once the quartz crystal resonator fails, the system in which it is located will be completely paralyzed, directly...

AI Technical Summary

Problems solved by technology

The methods used to deal with degraded data usually include time series method, regression analysis method, gray system theory and artificial neural network, etc. However, these methods have obvious defects

Time series method and regression analysis method rely on long-period and large-sample data to establish corresponding linear models. Gray system theory is suitable for fitting smooth data samples, but they are not suitable for small-sample, multi-dimensional, and non-linear relationships. The prediction effect of data samples is poor; the artificial neural network model has the shortcomings of overemphasizing overcoming learning errors and poor generalization performance, and the final weight of the network is greatly affected by the initial value. It is easy to fall into local minimum points and fail to obtain the optimal solution solution, its prediction effect is not very satisfactory

Images