An efficient Detection Algorithm for CT Saturation using linear predictive coding

This paper proposes an algorithm for CT saturation by behavior of distorted secondary current. The purpose of current transformer is to convert the primary current in the secondary circuit in a most accurate way. However, during fault conditions CT gets saturated. Some reasons due to current transformer (CT) can saturate like: attempting to mal operation of CT, heavy burden apply on the secondary side or due to transient DC offset fault current. Fortunately, due to the ever changing nature of the alternating current, even if the CT is pushed into saturation, in each cycle of the waveform, there are times during which magnetic core comes out of saturation. During these times the transformation of CT is correct but as soon as the magnetic core enters in to deeper saturation the current transformer will start exhibiting enormous errors. Thus, the key to correct these errors is to detect the instant when the saturation sets in and the instant when saturation ends. The idea of linear predictive coding (an autoregressive model) is to express the nth sample of signal as a sum of the past K known weighted samples. When analyzing disturbance signals the optimal prediction coefficients will change for each system state (e. g. pre-fault, fault and post fault) and therefore must be estimated continuously.The method for CT saturation detection that has been simulated is, CT saturation detection using linear predictive coding, in this method can successfully detect the onset of CT saturation. The test results clearly indicate that the proposed algorithm successfully detects the interval of saturation