A dynamic state evaluation method for current sensors based on reference values and time series

Abstract

Due to the complex and ever-changing power environment, when designing dynamic state evaluation methods for current sensors, there is often a problem of inaccurate calculation of state evaluation values, leading to poor evaluation accuracy of the methods. A dynamic state evaluation method for current sensors based on reference values and time series is proposed. Firstly, we normalize the time series data of the current sensor, construct a covariance matrix, calculate the degree of correlation between the time series data, and construct a higher quality current sensor state dataset. Then, we select multiple indicators as state evaluation indicators and quantify them. Then, the feature vector of the evaluation indicators and the state evaluation value are calculated. Under the influence of the reference value, we set the threshold of the current sensor in different states to divide the current sensor into different states. Through the above design, the dynamic state evaluation method for current sensors is completed. The experimental results show that the designed dynamic state evaluation method for current sensors based on reference values and time series has an average coefficient of determination of 0.95 in practical applications, with high evaluation accuracy.