Fast fault detection scheme for series-compensated lines during power swing

Sequence‐space‐aided SVM classifier for disturbance detection in series compensated transmission line

“…In [13], the variation in compensation level is limited to 0-15% and it is also not validated for high resistance faults. Travelling-based schemes as outlined in [15,19] show better performance as compared to proposed SVM-based classifier but it is found that scheme outlined in [15] is validated for 20% fix compensation and the one outlined in [19] is validated for only two levels 30 and 70% fix compensation. Moreover, it imposes high computational burden leading delayed protective actions.…”

“…However, the presented scheme is not investigated for line with compensation subjected to wide variation in fault contexts as well as power swing scenario. An innovative travelling wave-based fast fault detection scheme for SCTL during power swing has been reported in [15]. The voltage and current samples are applied for model transformation to calculate mathematical morphological gradient (MMG) applied to SVM.…”

An unblocking assistance to distance relays protecting TCSC compensated transmission lines during power swing

“…The methods based on the cumulative sum of the negative-sequence current, 13 the rate of change of zero-sequence current and magnitudes of voltage and current in each phase and relative phase angle between them, 23 the superimposed component of apparent power, 24 the morphology gradient-based technique, 25 the tangents perpendicular to the axis of current signal on the Lissajous figure, 26 and the cumulative sum of the covariance indices of the current system 27 are reported for detection of faults in series compensated transmission lines during power swing. A data-mining approach is proposed in Dubey et al 28 to detect faults in both compensated and uncompensated lines during power swing.…”

Differential voltage‐based fault detection during power swing

“…The machine learning techniques are popular for classification of different power system events because of their high accuracy. Different intelligent tools such as adaptive neuro‐fuzzy interface system [21], neural network [22, 23], support vector machine (SVM) [24, 25], decision tree (DT) [26] and random forest (RF) [26] are applied to isolate faults from power swing. Features are extracted from the fault signals by different techniques for training these tools.…”

“…In [26], the performances of DT, RF and SVM are compared for distinguishing the fault from power swing. ST [22, 23] and MM [25] can also extract quality features from fault signals. The main drawbacks of machine learning techniques are the requirement of exhaustive off‐line training and fault signals of long duration in most cases, which enhances fault detection delay.…”