Advancements in Fault Detection Techniques for Transmission Lines: A Literature Review

I. Introduction

Table 1. Abbreviations that appear in the study.

• To compare and critically assess the current methods for fault detection used in transmission line systems.
• To list the main obstacles and restrictions related to the existing techniques for transmission line fault detection and classification.
• To investigate current developments as well as new directions in transmission line systems fault detection technology.
• To evaluate the suitability and practicality of various fault detection techniques in actual transmission line situations.
• To make suggestions on possible areas for future research and advancements in transmission line fault detection techniques.

II. Methodology

III. Results and Discussion

Table 2. Literatures for Fault Detection on Transmission Lines Using Signal Processing Techniques.

B. Machine Learning (ML) and Artificial Intelligence (AI) Techniques:

Table 3. Literatures for Fault Detection on Transmission Lines Using Machine Learning (ML) and Artificial Intelligence (AI) Techniques.

Table 3. Literatures for Fault Detection on Transmission Lines Using Machine Learning (ML) and Artificial Intelligence (AI) Techniques.

Ref	Literature	Algorithms Used	Strengths	Limitations
[24]	Deep learning through LSTM classification and regression for transmission line fault detection, diagnosis and location in large-scale multi-machine power systems - Belagoune, S., Bali, N., Bakdi, A., Baadji, B., Atif, K. (2021)	DL (Presents three innovative DL classifier regression models based on DRNN for FRI, FTC, and FLP) Current and voltage signals are monitored using PMUs at various terminals and used as input characteristics for DRNN models SDL with LSTM to model spatiotemporal sequences of high-dimensional multivariate characteristics	Accurate region classification in large-scale multi-machine power systems Excellent performance in fault location prediction Excellent categorization and prediction accuracy	The dependency on synchrophasors indicates a possible constraint in application to systems lacking access to modern synchrophasor technology Scalability and generalizability
[25]	A deep learning based intelligent approach in detection and classification of transmission line faults - Fahim, S., Sarker, S., Muyeen, S., Das, S., Kamwa I. (2021)	Capsule Network with SF model	Accurately manage limited system information and develop resistance to system sounds Extracts faulty features into a single characteristic, making the fault identification procedure easier Does not call for manual labeling of data during training and testing, increasing its scalability and applicability to varied datasets Accuracy rates surpassing 99% in identifying and classifying faults, even under difficult conditions such as the presence of system disturbances, high impedance faults, and line parameter fluctuations	The evaluation focuses primarily on simulated datasets Does not discuss the possible computational or resource needs involved in putting the suggested model into practice in practical applications Does not offer the model’s generalizability to other transmission line layouts and operating situations outside of those that were evaluated
[26]	Self-attention convolutional neural network with time series imaging based feature extraction for transmission line fault detection and classification - Fahim, S., Sarker, Y., Sarker, S., Sheikh, M., Das, S. (2020)	SAT-CNN model DWT	Ability to adjust to various operating environments Imperceptibly concentrates on the output data from the hidden layer, improving the system’s classification precision Operates well with a variety of sampling frequencies and signal kinds, demonstrating its robustness and adaptability Withstand noise interference	It does not discuss the possible computational or resource needs involved in putting the suggested model into practice in practical applications The evaluation focuses primarily on simulated datasets Future deployment of the classifier utilizing actual data gathered from equipment in real-world power grids may necessitate careful consideration of data quality, calibration, and measurement errors It does not offer the model’s generalizability to other transmission line layouts and operating situations outside of those that were evaluated
[27]	End to end machine learning for fault detection and classification in power transmission lines - Rafique, F., Fu, L., Mai, R. (2021)	‘End to end’ ML model employing LSTM	Removal of the requirement for intricate feature extraction procedures Differentiate between several states, such as fault and non-fault, and between various kinds of faults Flexibility to different problem scenarios and system specifications Capable of handling voltage and current signals, providing a variety of data sources Ability to recognize and function in situations of power swings, increasing its usefulness in practical situations Strong performance across a range of fault scenarios, including as changes in fault impedance, loading circumstances, distance from measurement nodes, and signal noise levels	Further examination is necessary in order to determine its efficacy and dependability in real-world applications using data from real power systems in both recorded and real-world settings Necessary to carefully evaluate the computational load that comes with reconfiguring operational data using a Timesteps strategy to reduce computational load, particularly for large-scale power systems with substantial data requirements The degree of noise in the signals, the intricacy of the power system structure, and the availability of measuring instruments could affect the robustness of the model
[28]	Detection and classification of transmission line transient faults based on graph convolutional neural network - Tong, H., Qiu, R., Zhang, D., Yang, H., Ding, Q., Shi, X. (2021)	GCNN	Permits building a model for graph classification Integrating topological data into the network to deliver temporal and spatial data for fault identification and categorization Show the suggested method’s high accuracy and excellent generalizability in identifying a variety of transient defects in a variety of settings Demonstrates sensitive and steady performance with respect to robustness and response speed	The suggested solution uses a spectral convolution for graph convolution, which is theoretically sound but not very flexible It is considered that the edge weight is essential for locating faults, indicating that more investigation is required to improve the weighing system Recommended investigation in the application of dynamic GNN in order to overcome the drawbacks of spectral convolution and facilitate the identification, categorization, and localization of malfunctions in dynamic power grids
[29]	Transmission Line Fault Classification Using Hidden Markov Models - Freire, J., Castro, A., Homci, M., Meiguins, B., Morais, J. (2019)	ANN, SVM, KNN, and RF	When it came to fault classification, the HMM algorithm performed better and had reduced error rates About 90% faster processing times were demonstrated by the HMM method than by any of the FBSC architecture’s classifiers Direct fault event classification is made possible by the HMM method, which streamlines the classification procedure and may even result in less processing cost	The study used a particular dataset, UFPA Faults, and concentrated mostly on short-circuit faults As the stated findings were produced on a workstation with particular hardware requirements (i7 CPU, 16GB memory), the algorithm’s application to systems with varying processing capacities may be limited Its usefulness in practical applications has not yet been confirmed
[30]	High Impedance Single-Phase Faults Diagnosis in Transmission Lines via Deep Reinforcement Learning of Transfer Functions - Teimourzadeh, H., Moradzadeh, A., Shoaran, M., Mohammadi-Ivatloo, B., Razzaghi, R. (2021)	TF method CNN and the hybrid model of DRL	Exhibited greater performance in identifying and precisely finding single-phase to ground short circuit problems in power networks Attained strong correlation values during the training and testing phases, demonstrating the models’ efficacy in fault classification The DRL approach demonstrated its effectiveness in identifying subtle fault situations and perhaps averting catastrophic failures by outperforming the CNN method in the early identification of high-impedance faults (7000 and 9000 ohms)	Mainly concerned with single-phase to ground short circuit problems by simulating data from an IEEE transmission line Depends on local data gathered from the transmission line, which may restrict its application in circumstances where access to extensive or centralized data is restricted Deep learning technique skill and substantial computational resources may be needed for its installation and training
[31]	Detection and Evaluation Method of Transmission Line Defects Based on Deep Learning - Liang, H., Zuo, C., Wei, W. (2020)	Faster R-CNN, an end-to-end and high recognition accuracy deep learning algorithm	Identification of transmission line defects using automation, which decreases labor intensity and the necessity for manual inspection The absence of accessible and standardized datasets in the field of transmission line components is addressed by the development of the Wire_10 dataset, which consists of aerial photos taken by UAVs Based on the Wire_10 dataset, the defect detection network achieves a low false detection rate of 0.68% and a mean Average Precision (mAP) of 91.1%	The scope of the dataset collected by UAVs may be constrained, which could result in missed and false detections in some circumstances, especially in the winter and in non-rural locations The present dataset may not fully capture the fine features of transmission line components since it classifies problems into broad categories In actuality, the precise categorization of transmission line components can be complicated, necessitating close examination of a number of variables
[32]	Detection and classification of internal faults in bipolar HVDC transmission lines based on K-means data description method - Farshad, M. (2019)	KMDD method	Exhibits excellent precision and dependability while identifying and categorizing internal DC faults in bipolar HVDC transmission lines Demonstrates resilience to external errors and standard operating circumstances Demonstrates its flexibility and adaptability by being able to handle a variety of fault situations, including ones that weren’t taken into account during creation Contributes to its accuracy and stability by being less sensitive to elements including measurement noise, fault resistance, and fault location Lessens reliance on connected equipment and communication channels Appropriate for integration into current systems without requiring substantial hardware modifications	Although the scheme’s low sample frequency makes it more applicable in current systems, it might make it less capable of capturing high-frequency transient events or intricate waveform data Although the scheme’s low sample frequency makes it more applicable in current systems, it might make it less capable of capturing high-frequency transient events or intricate waveform data
[33]	Fault Detection and Classification in Power Transmission Lines using Back Propagation Neural Networks - Teja, O., Ramakrishna, M., Bhavana, G., Sireesha, K. (2020)	BPNN	The Π-modeling of the three-phase medium power transmission line system simplifies representation and analysis, making fault detection algorithms easier to construct Compatibility and simplicity of use when MATLAB/Simulink® is used Creating training data from transmission system simulated values guarantees a regulated training environment for neural networks By utilizing feedforward BPNN techniques, faults can be accurately classified and detected. This is because neural networks can identify intricate patterns from training data	The transmission line system may become overly simplistic if it is converted to a Π-model. The analysis was limited to fault scenarios including AG and ABG, which may not accurately reflect the range of faults that might arise in real-world transmission line systems. The system complexity, dataset size, and computational resources are only a few of the variables that may influence the optimization algorithm selection In order to evaluate the efficacy of fault detection algorithms in practical applications, performance analysis mostly concentrated on MSE, epochs, and training time, ignoring other critical metrics like accuracy, precision, and recall There are no specifics given about the hybridization criteria, optimization method, or anticipated performance gains
[34]	Component identification and defect detection in transmission lines based on deep learning - Zheng, X., Jia, R., Aisikaer, Gong, L., Zhang, G., Dang, J. (2020)	Faster R-CNN	By eliminating the need for extra specialist hardware, using power grid video surveillance technology for transmission line component classification and fault detection lowers installation costs The use of SSD, Mask R-CNN, and Faster R-CNN algorithms shows how successful deep learning methods are in target identification and semantic segmentation Understanding the advantages and disadvantages of the Faster R-CNN, SSD, and Mask R-CNN algorithms is possible through comparison and study Contextual and semantic information are better combined when an object detection framework built on FPN-SSD is introduced Attaining an average accuracy of 89.3% on the dataset indicates how well the suggested algorithm works for identifying transmission line components and detecting defects	Limiting the suggested algorithm’s application to situations involving different kinds of transmission line components by concentrating on just five different kinds of transmission widgets The complete variety of transmission line environments may not be captured if sample images are just derived from drone aerial inspection shots of the power grid Lacks in-depth investigation into component failure detection in favor of focusing mostly on target detection of aerial photography components

This table shows previous studies proposing various novel deep learning-based approaches for FDC in power transmission systems. The first method utilizes DRNNs with LSTM units within a SDL framework. This approach demonstrates exceptional performance in fault region identification, type classification, and location prediction. However, its reliance on synchrophasor data may limit its applicability to systems lacking this infrastructure, raising concerns regarding scalability and generalizability across diverse grid configurations.

Conversely, a CapsNet with SF exhibits remarkable performance, surpassing 99% accuracy in FDC tasks even under challenging conditions. However, the evaluation primarily focused on simulated datasets, and the discussion surrounding practical implementation hurdles and generalizability remains limited. Similarly, a SAT-CNN combined with DWT showcases promising adaptability and robustness. Nonetheless, a comprehensive evaluation of its real-world deployment challenges and generalization to various system configurations is currently lacking.

Furthermore, an End-to-End ML approach offers flexibility and delivers strong performance across diverse fault scenarios. Nevertheless, further investigation is necessary to assess its efficacy in real-world applications, particularly regarding computational load considerations.

Moreover, the GCNN offers a new method by enabling the creation of a model for graph classification and incorporating topological information into the identification and classification of faults. This model exhibits potential for reliable and quick fault detection, as evidenced by its excellent accuracy and generalizability across a range of problem scenarios and settings. However, the need for more research indicates possibilities for future study and development in the fields of edge weighting systems and graph convolution flexibility.

Additionally, when it comes to detecting internal DC faults in bipolar HVDC transmission lines, the KMDD method excels in terms of accuracy and dependability. With less dependence on linked devices and communication lines, its accuracy and stability are enhanced by its flexibility, adaptability, and resistance to outside mistakes. The necessity for additional improvement and validation in practical environments is highlighted by restrictions on sample frequency and the ability to record high-frequency transient occurrences.

Finally, the Faster R-CNN algorithm solves the issues of human inspection and dataset standardization, providing great opportunities for automated identification of transmission line defects. This model shows promise for real-world implementation with the creation of specific datasets and reduced false detection rates. However, there is a need for continued study to improve applicability and reliability due to limits in accurate component categorization, representativeness, and dataset coverage.

IV. Conclusion

References

1. S. Bhattacharyya, P. Debnath, A. Deyasi, and N. Dey, Eds., Contemporary Developments in High-Frequency Photonic Devices: in Advances in Computer and Electrical Engineering. IGI Global, 2019. [CrossRef]
2. D. Bai, X. Lv, J. Wang, and W. Zheng, “Application of Directional Transmission Data Technology for Transmission Lines,” J. Phys.: Conf. Ser., vol. 1584, no. 1, p. 012040, Jul. 2020. [CrossRef]
3. N. O. Mogaru, “The Basics of Transmission Line Protection,” in 2020 IEEE PES/IAS PowerAfrica, Nairobi, Kenya: IEEE, Aug. 2020, pp. 1–5. [CrossRef]
4. J. Q. Fahad and R. K. Chillab, “Design aspects of high voltage transmission line,” Bulletin EEI, vol. 9, no. 3, pp. 853–861, Jun. 2020. [CrossRef]
5. A. Mukherjee, P. K. Kundu, and A. Das, “Transmission Line Faults in Power System and the Different Algorithms for Identification, Classification and Localization: A Brief Review of Methods,” J. Inst. Eng. India Ser. B, vol. 102, no. 4, pp. 855–877, Aug. 2021. [CrossRef]
6. H. V. Patil, “Three-Phase Fault Analysis on Transmission Line in Matlab Simulink,” IJRASET, vol. 9, no. VI, pp. 5386–5391, Jun. 2021. [CrossRef]
7. S. A. Kumar and D. Chandramohan, “Fault test analysis in transmission lines throughout interfering synchrophasor signals,” ICT Express, vol. 5, no. 4, pp. 266–270, Dec. 2019. [CrossRef]
8. F. M. Gebru, A. O. Salau, S. H. Mohammed, and S. B. Goyal, “Analysis of 3-Phase Symmetrical and Unsymmetrical Fault on Transmission Line using Fortescue Theorem,” WSEAS TRANSACTIONS ON POWER SYSTEMS, vol. 17, pp. 316–323, Oct. 2022. [CrossRef]
9. Z. Zhiqiang, C. Fuyong, Z. Han, G. Jieting, J. Peng, and W. Wenzhen, “Modeling and Technology Application for Transmission Line Fault,” in 2018 2nd IEEE Conference on Energy Internet and Energy System Integration (EI2), Beijing, China: IEEE, Oct. 2018, pp. 1–5. [CrossRef]
10. A. Reda, I. Al Kurdi, Z. Noun, A. Koubyssi, M. Arnaout, and R. Rammal, “Online Detection of Faults in Transmission Lines,” in 2021 IEEE 3rd International Multidisciplinary Conference on Engineering Technology (IMCET), Beirut, Lebanon: IEEE, Dec. 2021, pp. 37–42. [CrossRef]
11. K. Nandhini and C. A. Prajith, “Review on Fault Detection and Classification in Transmission Line using Machine Learning Methods,” in 2023 International Conference on Control, Communication and Computing (ICCC), Thiruvananthapuram, India: IEEE, May 2023, pp. 1–6. [CrossRef]
12. T. S. Kumar et al., “Deep Learning based Fault Detection in Power Transmission Lines,” in 2022 4th International Conference on Inventive Research in Computing Applications (ICIRCA), Coimbatore, India: IEEE, Sep. 2022, pp. 861–867. [CrossRef]
13. B. J. Mampilly and S. V. S, “Transmission Lines Fault Detection using Empirical Mode Decomposition in a Grid-Connected Power System,” in 2020 International Conference on Power Electronics and Renewable Energy Applications (PEREA), Kannur, India: IEEE, Nov. 2020, pp. 1–6. [CrossRef]
14. C. Zhuang, S. Xie, H. Yang, H. Yu, Y. Geng, and R. Zeng, “Flexible Noncontact Approach for Fault Location of Transmission Lines Using Electro-Optic Field Sensors,” IEEE Trans. Electromagn. Compat., vol. 63, no. 6, pp. 2151–2158, Dec. 2021. [CrossRef]
15. H. Fasihipour and S. Seyedtabaii, “Fault detection and faulty phase(s) identification in TCSC compensated transmission lines,” IET Generation, Transmission & Distribution, vol. 14, no. 6, pp. 1042–1050, Mar. 2020. [CrossRef]
16. A. Abu-Siada and S. Mir, “A new on-line technique to identify fault location within long transmission lines,” Engineering Failure Analysis, vol. 105, pp. 52–64, Nov. 2019. [CrossRef]
17. B. R. Kumar, A. Mohapatra, S. Chakrabarti, and A. Kumar, “Phase angle-based fault detection and classification for protection of transmission lines,” International Journal of Electrical Power & Energy Systems, vol. 133, p. 107258, Dec. 2021. [CrossRef]
18. S. K. Mishra and L. N. Tripathy, “A critical fault detection analysis & fault time in a UPFC transmission line,” Prot Control Mod Power Syst, vol. 4, no. 1, p. 3, Dec. 2019. [CrossRef]
19. “Fault detection through discrete wavelet transform in overhead power transmission lines - Ahmed - 2023 - Energy Science & Engineering - Wiley Online Library.” Accessed: May 08, 2024. [Online]. Available: https://onlinelibrary.wiley.com/doi/full/10.1002/ese3.1573.
20. Aker, M. L. Othman, V. Veerasamy, I. bin Aris, N. I. A. Wahab, and H. Hizam, “Fault Detection and Classification of Shunt Compensated Transmission Line Using Discrete Wavelet Transform and Naive Bayes Classifier,” Energies, vol. 13, no. 1, Art. no. 1, Jan. 2020. [CrossRef]
21. V. Ashok, A. Yadav, and A. Y. Abdelaziz, “MODWT-based fault detection and classification scheme for cross-country and evolving faults,” Electric Power Systems Research, vol. 175, p. 105897, Oct. 2019. [CrossRef]
22. U. Saleem, U. Arshad, B. Masood, T. Gul, W. A. Khan, and M. Ellahi, “Faults detection and classification of HVDC transmission lines of using discrete wavelet transform,” in 2018 International Conference on Engineering and Emerging Technologies (ICEET), Lahore, Pakistan: IEEE, Feb. 2018, pp. 1–6. [CrossRef]
23. A. Q. Khan, Q. Ullah, M. Sarwar, S. T. Gul, and N. Iqbal, “Transmission Line Fault Detection and Identification in an Interconnected Power Network using Phasor Measurement Units,” IFAC-PapersOnLine, vol. 51, no. 24, pp. 1356–1363, Jan. 2018. [CrossRef]
24. S. Belagoune, N. Bali, A. Bakdi, B. Baadji, and K. Atif, “Deep learning through LSTM classification and regression for transmission line fault detection, diagnosis and location in large-scale multi-machine power systems,” Measurement, vol. 177, p. 109330, Jun. 2021. [CrossRef]
25. S. R. Fahim, S. K. Sarker, S. M. Muyeen, S. K. Das, and I. Kamwa, “A deep learning based intelligent approach in detection and classification of transmission line faults,” International Journal of Electrical Power & Energy Systems, vol. 133, p. 107102, Dec. 2021. [CrossRef]
26. S. R. Fahim, Y. Sarker, S. K. Sarker, Md. R. I. Sheikh, and S. K. Das, “Self attention convolutional neural network with time series imaging based feature extraction for transmission line fault detection and classification,” Electric Power Systems Research, vol. 187, p. 106437, Oct. 2020. [CrossRef]
27. F. Rafique, L. Fu, and R. Mai, “End to end machine learning for fault detection and classification in power transmission lines,” Electric Power Systems Research, vol. 199, p. 107430, Oct. 2021. [CrossRef]
28. H. Tong, R. C. Qiu, D. Zhang, H. Yang, Q. Ding, and X. Shi, “Detection and classification of transmission line transient faults based on graph convolutional neural network,” CSEE Journal of Power and Energy Systems, vol. 7, no. 3, pp. 456–471, May 2021. [CrossRef]
29. “Transmission Line Fault Classification Using Hidden Markov Models | IEEE Journals & Magazine | IEEE Xplore.” Accessed: May 08, 2024. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/8795474.
30. “High Impedance Single-Phase Faults Diagnosis in Transmission Lines via Deep Reinforcement Learning of Transfer Functions | IEEE Journals & Magazine | IEEE Xplore.” Accessed: May 08, 2024. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/9328357.
31. “Detection and Evaluation Method of Transmission Line Defects Based on Deep Learning | IEEE Journals & Magazine | IEEE Xplore.” Accessed: May 09, 2024. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/9001041.
32. M. Farshad, “Detection and classification of internal faults in bipolar HVDC transmission lines based on K-means data description method,” International Journal of Electrical Power & Energy Systems, vol. 104, pp. 615–625, Jan. 2019. [CrossRef]
33. O. N. Teja, M. S. Ramakrishna, G. B. Bhavana, and K. Sireesha, “Fault Detection and Classification in Power Transmission Lines using Back Propagation Neural Networks,” in 2020 International Conference on Smart Electronics and Communication (ICOSEC), Sep. 2020, pp. 1150–1156. [CrossRef]
34. X. Zheng, R. Jia, Aisikaer, L. Gong, G. Zhang, and J. Dang, “Component identification and defect detection in transmission lines based on deep learning,” IFS, vol. 40, no. 2, pp. 3147–3158, Feb. 2021. [CrossRef]
35. S. Kanwal and S. Jiriwibhakorn, “Advanced Fault Detection, Classification, and Localization in Transmission Lines: A Comparative Study of ANFIS, Neural Networks, and Hybrid Methods,” IEEE Access, vol. 12, pp. 49017–49033, 2024. [CrossRef]
36. R.-A. Tirnovan and M. Cristea, “Advanced techniques for fault detection and classification in electrical power transmission systems: An overview,” in 2019 8th International Conference on Modern Power Systems (MPS), Cluj Napoca, Romania: IEEE, May 2019, pp. 1–10. [CrossRef]
37. K. Swain, S. Mandal, S. s Mahato, and M. Cherukuri, “A Brief Review on Fault Detection, Classification and Location on Transmission Lines Using PMUs,” vol. 8, pp. 2608–2618, Dec. 2018.