A fast-executing single ended fault location method using transmission line characteristics for wind parks integrated HVDC network

https://doi.org/10.1016/j.epsr.2021.107552Get rights and content

Highlights

  • In this paper a fast-executing single ended travelling wave based method for pointing fault location has been proposed. Contribution of these papers can be summarized as-.

  • This method is very fast and can be executed within few microseconds for fault detection.

  • The proposed method maintains suitable accuracy in fault location with average relative error less than 1%.

  • This method doesn't require whole wavelet of fault surge signal but just a part of it, which results in less sample requirement and ultimately rapid execution of algorithm.

  • Despite being a single ended method, it's doesn't require reflected wave front arrival, which is remarkable feature and really useful for trigger detection if terminal distances are large (>200 km) and fault impedance in significantly high (>1000 ohms).

  • The proposed algorithm works as standalone method for fault finding and further can be integrated with other time domain fault finding algorithms. The method is flexible in nature and can be a part of real time protection system for HVDC networks.

Abstract

Protection programs against DC faults in the HVDC Networks play a critical role in the development and expansion of integrated HVDC networks. When DC fault happens in HVDC network, every passing second counts huge power loss which makes a fast executing fault detection system necessary. Traditional fault detection methods either lacks accuracy or consume significant time in execution. This paper proposes fast-executing single ended method computing fault location within a few microseconds just after arrival of the first wave front of fault surge at the terminal of the HVDC network. The proposed method utilizes frequency dependent transmission line characteristics and travelling wave wavefront information for determining location of fault. The proposed method's functionality is tested in PSCAD simulation with realistic models. The results have been further compared with existing fault location methods and the proposed method shows suitable accuracy and significantly faster execution.

Introduction

High Voltage DC (HVDC) networks have proven themselves quite useful for integrating far distant solar farms and wind farms. They offer more controllability of power delivery and flexibility in the structural assembly compared to High Voltage AC (HVAC) networks [1]. Voltage Source Converter (VSC) based HVDC networks are even more popular due to their capability of integrating weak AC grids [2]. However, VSC based HVDC networks are susceptible towards DC faults.

In a Multi Terminal HVDC networks when a fault occurs, it creates a surge of discharging voltage which travels rapidly towards terminals. Due to the fact that VSC has a capacitor bank connected directly to the network, the incoming surge hits the capacitor bank and a huge amount of current get discharged very rapidly through DC fault. Further surge hits adjacent terminals resulting additional current flowing towards fault, and simultaneously surge hits AC networks which are connected to the converters [3]. Breakers connected at AC side can easily disconnect AC networks as smoothing reactors are connected at AC side just before converters and therefore the surge is not very rapid. However, for DC Circuit Breakers (DCCB) fault isolation is hard owing to the very short time requirement for fault current to achieve peak point [4, 5]. Hence, it's necessary to have a very fast fault detection and fault location method.

There are multiple methods available for the detection of fault and estimation of fault location like telemetry-based pulse or signal methods [6], Artificial Intelligence (AI) based fault detection methods [7, 8] and wavelet transform methods both double and single ended type [9], [10], [11], [12], [13], [14], [15], [16], [17]. Off-line pulse methods are suitable for past fault computations, when fault moves over or get diminished. Whereas AI-based fault detection methods are computationally very expensive and take long time in the processing and detection of fault. Wavelet Transform (WT) methods are Transmission Line (TL) characteristics dependent and generally faster than above mentioned methods. However, WT methods sometimes can be computationally expensive [17] and can take a significant time in the processing if the algorithm is running on a low-end processor. WT methods are either Double Ended (DE) type in which data collected from terminals located at both ends of Line segment for fault determination or Single Ended (SE) type in which data collection rely on primary and reflected wavelets.

Double Ended (DE) fault location method is widely used where availability of communicating methods are easily available like fiber optic channels or global positioning systems for synchronizing all the terminals. Fault detection method utilizing Continuous Wavelet Transform (CWT) has been proposed in [10] but requires synchronized terminals in HVDC network. In [18], fault location on the basis of least distance method has been proposed but requires a proper calibration of TL segments at regular interval to achieve sufficient accuracy. In order to achieve higher accuracy, Digital Wavelet Transform (DWT) proposed in [15] and further involvement of various signal processing methods like Empirical Mode Decomposition (EMD) [19], Vector Mode Decomposition (VMD), Ensemble EMD, Hilbert Huang Transform (HHT) [20] has been proposed. These methods do improve accuracy, but at the cost of high complexity and significant processing time in fault location.

Single ended fault location methods are popular for the locations where availability of communicating medium is scarce or reliability issues are present. Fault is located in this method by observing the primary wave front of the fault surge and wave front which is reflected by the terminal of the other end of the transmission line [9]. If the distance between terminals of the transmission line is good enough and the fault resistance is significantly high, it is hard to pick up the reflected signal. Further issues like distorted waveform, noise interference put up significant hindrance in execution of this method [13]. Attempts have been made to rectify these problems in the form of inclusion of advance signal processing methods such as EMD, EEMD, VMD, HHT but these methods are computationally expensive and can take significant time in the processing and hence increased time taken in the prediction of fault location [16, 17, 19].

The proposed method in this paper includes following advantages-

  • 1

    Although the method is single-ended based but it does not require reflected wave front for the calculation of fault location moreover the proposed method can be executed within milliseconds from the arrival of primary wave front of fault surge.

  • 2

    The proposed method is computationally less expensive than above-mentioned methods and doesn't require heavy preprocessing of incoming fault surge signal.

  • 3

    Inherently being single ended based method, it does not require additional communication channels or GPS for prediction of fault location.

The structure of proposed paper is as follows. Section II discusses about the theory behind proposed method. Section III represents the modeling of test system with specifications of each of the component. Section IV shows the results for the verification of the proposed solution finally Section V concludes the theory and discusses about future options.

Section snippets

Proposed solution

This paper utilizes Fourier Transform theory to detect the dominating frequency of incoming primary wavefront of Travelling Wave (TW). Theory proposed in the paper simply says that the rise time of the fault current is inversely proportional to the dominating frequency and hence latter can be calculated by detecting rise time. Dominating frequency further can be translated into fault location.

The proposed solution in first part talks about method of inception of fault and next part derives

Test system

In this paper a four terminal Bipolar HVDC system has been modeled to support the proposed solution. The model four terminal HVDC system is structured as a square with each terminal at the corner of square. A star configuration has been used keeping Terminal 1 at the center and other terminals on the edges. As mentioned in Fig. 3 distance between edge terminals is 200 km and distance between T1 and T3 is 282 km.

Each terminal of the test system includes Wind farm as a weak AC source with limited

Results and discussion

A fault is introduced at 4.0 second into the system at distance Dffrom Terminal T2 and algorithm is tested for the fault response. Maximum time window for result observation has been kept to 1 ms. Following results has been obtained from the tests

Conclusion

In this paper a fast-executing single ended fault location method has been proposed which utilizes TW wavefront information. Contributions of this paper can be summarized as-

  • 1

    This method has relatively a smaller number of iterations for a given time window and hence takes time as less as 10  µs in fault detection.

  • 2

    The proposed method is fairly accurate with average error less than 1 km and relative error less than 1%.

  • 3

    The proposed method is not affected by Network Configuration (radial or ring

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Govind Kant Mishra received the B.Tech. degree in Instrumentation and Control Engineering from U.P.T.U. in 2012 and M.Tech. degree in Power Systems from NIT Patna in 2014. Currently he is a Research Scholar in Dr. A.P.J. Abdul Kalam Technical University (AKTU) His-research interest includes Design and Modelling of power system, Distributed generation systems, Protection of HVDC Systems, Power electronics and design of instrumentation systems.

References (23)

  • T. Bi et al.

    Single pole-to-ground fault location method for Mmc-Hvdc system using active pulse

    IET Gener. Transm. Distrib.

    (2018)
  • Cited by (0)

    Govind Kant Mishra received the B.Tech. degree in Instrumentation and Control Engineering from U.P.T.U. in 2012 and M.Tech. degree in Power Systems from NIT Patna in 2014. Currently he is a Research Scholar in Dr. A.P.J. Abdul Kalam Technical University (AKTU) His-research interest includes Design and Modelling of power system, Distributed generation systems, Protection of HVDC Systems, Power electronics and design of instrumentation systems.

    Prof. (Dr.) Yaduvir Singh obtained M.E. in Control and Instrumentation from MNNIT Allahabad in 1993 and Ph.D. degree in Industrial Electronics from Thapar Institute of Engineering and Technology, Patiala in 2004. Presently he is Professor in Department of Electrical Engineering at School of Engineering, HBTU, Kanpur. His-research interest includes intelligent systems design, soft computing, automated control systems, artificially intelligent systems, power system modeling and identification, industrial electronics and design of instrumentation systems.

    View full text