Investigation on Full-Converter-Based Wind Power Plant Behavior During Short Circuits

Abstract: This paper presents an investigation on the behavior of a full-converter-based wind power plant (FCWPP) during short-circuits. To do so, an actual power network is taken into account, in which a transmission line that connects a FCWPP to a traditional system fed by typical high inertia generation units is studied. The test system was modeled and tested by means of Electromagnetic Transients Programs (EMTP), namely Alternative Transients Program (ATP) and Power System Simulator (PS SIMUL). Single-phase fault scenarios at different positions on the studied line were tested and then, voltage and current characteristics at both terminals were investigated. The obtained results show that the FCWPP control leads voltages and currents to present different patterns from those verified when typical synchronous machines are considered, which can impair monitoring functions or even interfere disturbance analysis studies.

I. INTRODUCTION

The gradual reduction of fossil fuels usage in power generation systems has been a topic of global interest. Discussions and solutions to problems such as environmental pollution emitted by fossil fuel based power generation units have dictated the trend of exploring renewable energies [1].

Along with demographic growth, electricity demand has substantially increased over the years. In accordance with the social and environmental aspirations that call for clean energy sources, Brazil has invested in alternative power generation sources, so that 86% of its generated electric power will derive from renewable sources by the end of 2027 [2]. Regarding the wind power generation capacity, it is expected to increase by up to 2000 MW per year mainly through the exploration of Brazil’s wind potential. As a result, challenging scenarios will arise in the Brazilian power grid, since wind power has different technical and operational characteristics when compared against commonly used power sources, such as non electronically coupled rotating machines.

Concerns arise about possible impacts that the growth of wind sources connections may have on steady and transient state. These complications emerge due to its intermittent operation and the use of electronic converters. The combination of those factors can change the traditional power system operation features, also influencing in voltage and current behavior during short-circuits [3], [4].

Based on the aforementioned context, researches on possible impacts caused by wind power plants on power networks operation have attracted the attention of utilities. Indeed, studies need to include mechanical and system control modeling and also short-circuit analysis in order to cover different paths in power systems operation procedures that may be affected by converter-interfaced generation units. Thus, this paper investigates the voltage and current signals behavior in a 500 kV/60 Hz transmission network that connects a typical power system with a FCWPP. In order to do so, a power system with wind power integration was reproduced in EMTPs using data taken from [4]. As a result, modeling aspects are firstly addressed and then short-circuit studies are carried out by means of the comparison between voltage and current signals measured over the test system. Obtained results demonstrate that the FCWPP can result in atypical voltage and current waveforms, which differ from those observed when traditional power generation is considered.

II. FULL CONVERTER-BASED WIND GENERATION

According to the wind turbine types described in [5], there is a number of different possible configurations that may be used to assemble the generation units. In this paper, the investigated wind power unit is defined as Type 4, whose mathematical details can be found in [4]. It is composed by a mechanical turbine, a synchronous generator (SG), and a full-converter with its associated controls, as depicted in Fig. 1.

The turbines play the role of converting the kinetic wind 978-1-7281-2920-4/19/$31.00 ? 2019 IEEE power into mechanical power, which in turn drives the synchronous generator. This type of mechanical turbine depends on a power coefficient, which expresses the relation between the available wind power and the one indeed extracted by the wind turbine itself [6].

The synchronous machine works as in other typical power generation applications. The rotor winding is fed by direct currents, inducing a magnetic field which varies in time when it passes through the stator coils. As a result, voltage signals are induced in the generator terminals, supplying power to connected loads. The amplitude and frequency of terminal voltages depend on the machine mechanical speed, which may figure as an issue due to the variant wind speed. Therefore, power converter-based interface between the machine and the power system is taken into account, making the machine terminal voltages to have well behaved frequency and amplitude, so that the generator can operate over a range of speeds [7]. Despite the challenges that may arise when using converter interfaced wind power plants, one should bear in mind the use of power electronics is crucial for the proper operation of this kind of power generation [6].

Regarding the power conversion solutions, several options have been reported in the open literature and by manufacturers. In the evaluated FCWPP, the converter on the generator side consists of a passive rectifier bridge and serial boost DC/DC converters. The rectifier unit is associated with another buck type DC/DC converter, through which the field winding of the synchronous machine is fed. The grid side converter is characterized by a two-level voltage source inverter, which is responsible for delivering all the power generated by the machine to the grid respecting the applicable standards. The control associated with the generator side converter is expected to extract the maximum electrical power of the SG given the conditions of mechanical turbine supplied power, consisting of the so-called maximum power point tracking [8]. The machine excitation control works by relating the turbine mechanical speed to its field current, producing a variable excitation which depends on the mechanical speed in the axis of engagement between turbine and generator. Finally, the control associated with the inverter unit aims to control the machine power factor and the power flow given the operating network conditions. By doing so, the FCWPP becomes capable of controlling its stability under different network abnormalities.

III. STUDIED POWER SYSTEM

The power system presented in Fig. 2 was modeled by using the ATP [9] and PS SIMUL [10] platforms, which are EMTP software that allow the detailed modeling of the FCWPP control units. By doing so, differences between the EMTP simulations could be analyzed, attesting that the major short-circuit behavior of the FCWPP is properly represented by both simulation tools despite of slight deviations.

The test power system consists of a 500 kV/60 Hz transmission line 239 km long that connects three wind power plants (local bus) and Thévenin equivalent circuits, which represent the power grid around the analyzed system (remote bus). The transmission line was modeled using the distributed parameter line model, constant in frequency, as fully transposed line. The wind power units are composed of full-converter wind turbine generators, which were modeled in detail as reported in [4]. The FCWPP behavior during short-circuits is evaluated by monitoring the voltage and current signals at the local and remote terminals while applying faults at different points on the analyzed transmission line. Aiming to evaluate cases that yield adverse asymmetrical currents, phase-A-to-ground (AG) faults were simulated, which in turn consist in the most common type in power networks. Assuming that typical protective relays operate in average times of about one power cycle and that the electromechanical opening of traditional circuit breakers usually occur in about two power cycles after the protection operation [11], three cycles before and after the fault inception were analyzed in the presented results.

ATP and PS SIMUL simulations were carried out using 1.0 μs time step, emulating real analog signals. A third order low-pass anti-aliasing Butterwoth filter with cutoff frequency at 480 Hz was applied and then the filtered signals were resampled to 32 samples per cycle of 60 Hz in order to proceed with the phasor-estimation process. In this paper, the modified cosine filter was utilized [12].?

IV. ANALYSIS AND RESULTS

In order to evaluate occasional differences caused by discrepancies in the transients computation methodology in ATP and PS SIMUL, an specific comparison analysis was performed considering an AG fault at 10% of the monitored line length. Voltage and current signals disregarding the antialiasing filter for each phase are displayed on Fig. 3.

As one can see, the compared waveforms conduct themselves quite similarly. Regarding the voltages signals, it was observed a sag on phase A, as expected, at both line terminals, while the sound phases experience a smooth swelling. It should be noted that vL,A oscillates with high frequencies after fault inception, as expected due to the proximity of the fault in relation to the local bus (wind power plant terminal).

The local current signals show in all phases a slight increasing at the wind power plant side, which are much smaller than those observed at the remote terminal, i.e., at the traditional power network side. This is an interesting aspect to be investigated especially because the fault occurs close to the local line end and this abnormal characteristic could lead protective functions to misoperate.

To provide a more thorough comparison between ATP and PS SIMUL simulations, Fig. 4 depicts a scatter plot that shows the correlation between the signals obtained by means of both programs. In other words, the scatter plot creates a Cartesian point with samples taken from ATP and PS SIMUL, thereby plotted points are expected to concentrate themselves on the coincidence central line highlighted in the figures.

From the obtained scatter plots, despite slight differences, it is noted that most samples take place around the coincidence line, demonstrating that the major FCWPP behavior under short-circuit conditions is properly represented by both ATP and PS SIMUL platforms. It proves that the use of different EMTP softwares should not result in significant differences in the FCWPP operation during faults, specially when only fundamental component analysis is considered. Thus, from now on, short-circuit contributions will be evaluated considering only PS SIMUL data.

Aiming to demonstrate the different behavior of voltages and currents at local and remote buses at which FCWPP and traditional power network are connected, respectively, Fig. 5 illustrates scatter plots obtained by considering the absolute value of voltage and current fundamental components at both line ends. Results for faults at 10%, 30%, 50%, 70% and 90% of the line length (taking the local line terminal as reference) are tested.

From Figs. 5, it is concluded that significant voltage and current variation at the traditional power network terminal are verified, which vary their values depending on the fault location. On the other hand, at the wind power plant side, it is noted that voltages and currents converge for very similar values, irrespective of the fault distance, which differs in relation to the expected behavior for traditional generation sources. This behavior is due to the associated control schemes of power electronic converters on wind power plants. Indeed, one can see that the areas with greater density of scatter plot samples are vertically aligned, proving that the FCWPP electrical quantities do not significantly vary when compared to the variations observed at the traditional power grid.

From Fig. 5(a), it can be observed that all points take place above the coincidence line, demonstrating that very low fault current contributions are expected from the modeled FCWPP. A similar behavior is verified in Fig. 5(b), attesting that voltage sags at the FCWPP terminal were greater than those observed at the traditional power network line side.

It is worth to mention that the obtained results bring up questions on the atypical behavior of converter-interfaced wind power generators under short-circuit conditions. Indeed, significant voltage sags and specially the low fault current contributions may compromise the system protection operation, so that appropriated schemes should be developed. In future works, it is intended to analyze the impact of the modeled FCWPP on protective relays under different fault types, detailing the relation of obtained voltage and current waveforms with turbine control aspects. In addition, it is expected to evaluate other wind power generator topologies, indicating potential solutions for challenging scenarios that traditional protective devices may face.

V. CONCLUSIONS

In this paper, an actual power system was modeled and tested with the aim to compare the behavior of voltage and current signals considering short-circuits on a 500 kV/60 Hz transmission line that connects full-converter wind power plants to a power network with traditional generation units. Firstly, different EMTP programs were used to evaluate differences between transient simulations, demonstrating that the major system behavior is properly represented despite slight differences. Then, scatter plots were plotted in order to compare voltages and currents measured at the converter-interfaced wind power plant line side and those taken from the power network terminal. The obtained results show that voltage sags are greater at the wind power plant side, where very low current contributions are verified. These operational features consist in challenging scenarios for protective devices which must be considered in actual systems.

REFERENCES

[1] M. R. Miveh, M. Gandomkar, S. Mirsaeidi, and M. R. Gharibdoost, “A review on protection challenges in microgrids,” in 2012 Proceedings of 17th Conf. on Electrical Power Distribution. IEEE, 2012, pp. 1–5.

[2] Decennial Energy Expansion Plan 2027 (in Portuguese), Empresa de Pesquisa Energética, Rio de Janeiro, march 2018.

[3] S. De Rijcke, P. S. Pérez, and J. Driesen, “Impact of wind turbines equipped with doubly-fed induction generators on distance relaying,” in IEEE PES General Meeting. IEEE, 2010, pp. 1–6.

[4] A. Reis, “A contribution to the operational control of wind units: modeling, voltage regulation and minimization of harmonic distortions (in Portuguese),” Ph.D. dissertation, UFU, 2015.

[5] IEEE PES Wind Plant Collector System Design Working Group, “Characteristics of wind turbine generators for wind power plants,” in 2009 IEEE Power & Energy Society General Meeting. IEEE, 2009, pp. 1–5.

[6] M. Rashid, Power electronics handbook. Butterworth-Heinemann, 2017.

[7] R. Walling, E. Gursoy, and B. English, “Current contributions from type 3 and type 4 wind turbine generators during faults,” in PES T&D 2012. IEEE, 2012, pp. 1–6.

[8] Z. Chen, J. M. Guerrero, and F. Blaabjerg, “A review of the state of the art of power electronics for wind turbines,” IEEE Transactions on power electronics, vol. 24, no. 8, pp. 1859–1875, 2009.

[9] E. C. LEUVEN, “Alternative transients program rule book,” Leuven EMTP Center, Belgium, 1987.

[10] PS SIMUL: Software for Power System Modeling and Simulation of Electromagnetic Transients (in Portuguese), Conprove Engenharia, Industry and Commerce, Uberlandia, June 2019. [Online]. Available: https://www.conprove.com.br/pub/i_ps_simul.html.

[11] E. O. Schweitzer, B. Kasztenny, A. Guzmán, V. Skendzic, and M. V. Mynam, “Speed of line protection-can we break free of phasor limitations?” in 2015 68th Annual Conference for Protective Relay Engineers. IEEE, 2015, pp. 448–461.

[12] D. G. Hart, D. Novosel, and R. A. Smith, “Modified cosine filters,” Nov. 28 2000, uS Patent 6,154,687.