Power transformers being among the power system key components, are very important to protect. Various approaches have so far been proposed for this topic. Artificial intelligence (AI) has provided a concrete basis for numerous power transformer differential protection methodologies. However, the capabilities of AI-based methods have never been put into comparison. This paper provides a comparative analysis of various AI-based approaches for differential protection of power transformers. The performances of AI-based differential protection schemes are investigated in the presence of current transformer (CT) saturation condition, series capacitors compensation, and superconductor fault current limiter (SFCL), which possibly deteriorate the capability of differential protections to correctly tell inrush currents and internal faults apart. The PSCAD/EMTDC simulation tool is applied in producing the necessary test dataset for performance evaluation of the suggested novel strategy. The attained outcomes from the assessment of various methods have been analyzed to introduce the most superior AI-based differential protection scheme.@en

Photovoltaic (PV) as one of the most promising energy alternatives brings a set of serious challenges in the operation of the power systems including PV system protection. Accordingly, it has become even more vital to provide reliable protection for the PV generations. To this end, this paper proposes two-stage data-driven methods. In the first stage, a feature selection method, namely t-distributed stochastic neighbor embedding (t-SNE) is implemented to select the optimal features. Then, the output of t-SNE is directly fed into the strong data-driven classification algorithm, namely robust soft learning vector quantization (RSLVQ) to detect PV array fault and identify the fault types in the second stage. The proposed method is able to detect the two different line-to-line faults (in strings and out of strings) and open circuit fault and fault type considering partial shedding effects. The results have been discussed based on simulation results and have been demonstrated the high accuracy and reliability of the proposed two-stage method in detection and fault type identification based on confusion matrix values.@en

The conflicting issues of growing demand for electrical energy versus the environmental concerns have left the energy industries practically with one choice: to turn into renewable energies. This duality has also highlighted the role of power transmission systems as energy delivery links in two ways, considering the increased demand of load centers, and the integration of large-scale renewable generation units connected to the transmission system such as wind power generation. Accordingly, it has become even more vital to provide reliable protection for the power transmission links. The present protection methods are associated with deficiencies e.g., acting based on a predefined threshold, low speed, and the requirement of costly devices. A two-stage data-driven-based methodology has been introduced in this paper to deal with such defects, considering wind power generation. The proposed approach utilizes a powerful feature extraction technique, namely the t-distributed stochastic neighbor embedding (t-SNE) in the first stage. In the second stage, the extracted features are fed to a robust soft learning vector quantization (RSLVQ) classifier to detect and locate transmission line faults. The WSCC 9-bus system is used to evaluate the performance of the proposed data-driven method during various system operating conditions. The obtained results verify the promising capability of the proposed approach in detecting and locating transmission line faults.@en

One of the most challenging issues in protecting power transformers is to discriminate internal faults from inrush currents. This paper proposes a new approach for differential protection of power transformers based on the robust soft learning vector quantization (RSLVQ) method. Statistical features from the normalized differential current gradient are extracted in order to train the RSLVQ classifier. Furthermore, the performance of the proposed differential protection scheme is investigated in the presence of superconductor fault current limiter (SFCL), which can greatly affect the ability of differential protection schemes in correctly discriminating inrush from internal fault currents. The PSCAD/EMTDC software is utilized to generate sampled data in order to evaluate the performance of the proposed approach. The results obtained from the evaluation of the proposed method verified the promising performance of the RSLVQ-based differential protection scheme.@en