Issuu

International Research Journal of Engineering and Technology (IRJET)

e-ISSN: 2395-0056

Volume: 11 Issue: 10 | Oct 2024

p-ISSN: 2395-0072

www.irjet.net

Coherence-based numerical differential current relay for synchronous generator stator windings protection R. A. Mahmoud1, E. S. Elwakil2 1Misr University for Science and Technology (MUST), College of Engineering Science & Technology, Department of

Electrical Power and Machines Engineering (PME), 6th of October City-Giza-Egypt.

2Power Electronics and Energy Conversion Department, Electronics Research Institute, Cairo -Egypt.

------------------------------------------------------------------------***------------------------------------------------------------------------Abstract: A novel protective scheme for stator windings of synchronous generators is presented. The proposed method utilizes coherence coefficients for currents at both stator winding terminals to detect the fault condition in the AC power system, to determine which phase is faulty and to specify whether the fault is internal or external. The suggested scheme has developed closed-tripping characteristics to identify the relay tripping and blocking zones. Extensive fault situations are simulated using ATP software; then the protective scheme is verified using MATLAB package. In this method, the fault detection and classification are not affected by the fault types, fault locations, fault resistances, fault time span, fault inception times, and initial operating conditions. Moreover, it is smart and reliable; it can be practically implemented and gives fast time response (about half a cycle). In addition, the settings of the coherence coefficients can be selected in a narrow boundary. Besides, the proposed scheme can detect the ground faults near the neutral terminal of SG windings with high grounding impedance. Keywords: Synchronous generator; Fault detection; Fault classification; Internal fault; External fault; Coherence coefficient. 1. Introduction Due to the fact that Synchronous Generator (SG) is the most critical equipment in power networks, fault detection and classification are essential to avoid outages and blackouts [1-2]. As a result, many protection methods were proposed to protect SGs against various types of faults. Several researches and applications were processed for SGs protection in the literature. Faults in the SG are commonly classified as internal and external faults. The common internal faults may be phase to phase and phase to ground faults in stator windings, whereas external faults are those which occur outside the generator’s protection zone, which may be caused by short circuits, over loading or unbalanced loads [3]. The conventional protection schemes consist of differential and stator earth-fault relays; those are known to be slow in clearing the faults unless the fault develops and

current reaches the operating value [4]. Faults near neutral point, cannot be detected by earth-fault relays, especially for SG with high impedance grounding neutral due to insufficient voltage to drive the fault current [4]. To overcome these limitations, several techniques were proposed as in [5-7]. The paper in [7] presented a protection technique for stator windings against earth faults using the third harmonic voltage at the neutral and the terminals of the SG. Fault detection and classification schemes using statistical approaches such as correlation and alienation techniques were presented in [8-10]. SG protection methods based on signal processing techniques such as wavelet transforms were developed in [11-12]. Recently, Artificial Neural Network (ANN) technique was proposed for fault detection in SG stator windings [13]. The ANN technique is provided for detecting various internal, external and ground faults, which are close to neutral point. Other algorithms using a fuzzy logic controlled neural network were introduced in [14-17]; the algorithms utilized the fault generated transients for internal and external faults discrimination. The ANN and fuzzy logic were used for turn-to-turn fault detection, fault type classiﬁcation and fault location identiﬁcation [18-19]. In [20], the authors presented support vector machine-based fault detection and classification for electrical AC machines using vibration waves. Fault detection in Synchronous Generators (SGs) based on independent component analysis and machine learning was introduced in [21]. This paper presents a novel protection scheme for SG stator windings based on coherence analysis. The coherence coefficients are estimated for current signals measured at both terminals of three phase stator windings, which can be used for fault detection, faulty phase(s) selection, and fault location discrimination with respect to the generator protection zone. The proposed algorithm presents new tripping characteristic curves, utilizing coherence coefficients, to identify the relay tripping and blocking regions. 2. Synchronous generator modeling Modeling of internal faults in synchronous generator cannot be built directly in the ATP program. When dynamics of the machine is not required, sinusoidal

Page 317