Adaptive Control of Starting Point Based on System Data Degradation SCR

In this paper, the prediction starting point detection algorithm based on the real-time change trend of the state is used to divide and control the health status of the system data, and the adaptive control of the starting point threshold of life decay is carried out by analyzing the change trend of the SCR (State Change Rate) of the system data degradation process, so as to improve the accuracy of the prediction starting point detection. The time-domain and frequency-domain features of the processed signal are extracted, and the evaluation index is constructed through the monotonicity and predictiveness of the features, and the suitable features are selected as the basis for the remaining life prediction of the system data. High-dimensional feature processing and screening of system data. In this paper, through a large number of simulation experiments, the monotonicity law of system data features and the trend law of system data characteristics are obtained. The validity of weighting the monotonicity and trend of the system data features is verified, and the weight of monotonicity is 0.6 and the weight of trend is 0.4 in the verification system, and the calculation results of the evaluation index of each feature of the system data are obtained. It is concluded that the peak, variance, root square amplitude, root mean square value, kurtosis value and peak factor are used as inputs for the remaining life prediction for the degradation of the system data.