Improvement in Reactive Power Consumption of Line Commutated HVDC Converters for Integration of Offshore Wind-Power

Transcription

1 Improvement in Reactive Power Consumption of Line Commutated HVDC Converters for Integration of Offshore Wind-Power Muhammad Jafar, Marta Molinas Abstract--This work relates to improvement in line commutated converter technology for it to consume less reactive power. Reactive power supply requires heavy compensation through the use of capacitors. In this work a new topology for pulsed injection of series voltage is suggested. It is observed that this scheme will enable commutation at a time prior to that in a conventional converter to reduce reactive power consumption of the terminal. Simulation results verify the approach. This would help in the reduction of capacitor number and size at the converter stations. Also, the source voltages required to achieve rated power flow decrease resulting into smaller voltage ratings of the ac equipment. Index Terms--HVDC, Line Commutated Converters, Offshore Wind, Renewable Energy L I. INTRODUCTION INE commutated converter (LCC) technology [] is the leading technology for high voltage direct current transmission (HVDC) systems in terms of power transmitted when compared to the self-commutated converter technology (SCC) [2]. There are however numerous inherent problems related to LCC which pose a threat to the costs and reliability of the energy supply. These include (to name a few) requirement of commutation voltages, heavy reactive power consumption, lower order harmonics, commutation failures etc. [3]. The world has seen many medium sized offshore wind energy projects in recent times and there are more which are in the pipeline. The majority of these projects are in shallow waters meaning that the foundations are solidly bonded to the ocean floor. However, there are many projects foreseen in deep waters including many sites in the North Sea off the coast of Norway. The associated water depth is of the order of 2 meters [4]. Engineers have come up with the idea of floating turbines and allied structures. These will be bonded to the ocean floor through mooring systems [4]. The shear distance from the coast to these suggested sites means that HVAC is technically not viable. Therefore, HVDC is a must for carrying power from these deep sea wind-farms to shore. M. Jafar is with Department of Electric Power Engineering, Norwegian University of Science & Technology, 749, Trondheim, Norway ( muhammad.jafar@elkraft.ntnu.no). M. Molinas is with Department of Electric Power Engineering, Norwegian University of Science & Technology, 749, Trondheim, Norway ( marta.molinas@elkraft.ntnu.no). The rectifier terminal will have to be placed close to the turbines and will be built on a floating structure held in place with the help of mooring systems. The cost of the support structure and its mooring systems is proportional to the volume and weight of the equipment that it has to support. With the presence of a large number of capacitors for reactive power compensation, this cost will be higher. Due to limited accessibility of these sites owing to harsh environmental conditions for most part of the year it is vital that equipment on the converter station should be reduced to a minimum. This places demand on innovation for LCC technology to reduce the equipment. Static compensator (STATCOM) aided LCC [] is one of the solutions which is essentially a shunt device. The principle is to inject a current in such a manner so that the source current should be in-phase with respective phase voltages. This work attempts to introduce a new concept for reactive compensation through the use of series injection of voltage pulses. The concept of series injection of voltages is applied in many applications. Fixed capacitors for enhancement of power system transmission capability have been employed successfully [6]. Concept of switched series capacitors were presented as early as 966 [7]. Thyristor-controlled series compensation (TCSC) [8] in power systems took this technology one step further. Static phase-shifters [9] are employed for enhancement of stability limits of a power system as well as for steady-state power flow control. Static synchronous series capacitors (SSSC) [] are also used for transmission line compensation. Dynamic voltage restorer (DVR) [] is a device which injects series voltage dynamically to cancel out the effect of voltage sags or swells and thus enhances the power quality to end consumers. The application of magnetic energy recovery switch (MERS) [2] is suggested for power factor correction [3] as well as power flow control in power systems [4], []. For HVDC, series injection schemes are quite limited. Capacitor commutated converters (CCC) [6], [7] employ series capacitors between the converter transformer and the valve arms and is well established. As these are connected in series, current variation also causes a variation in the reactive power contribution of the series capacitors. This reduces number of switching operations in other reactive power compensation equipment. Another technique by the

2 name of Controlled Series Capacitor Converter (CSCC) was introduced in [8] which employs thyristor controlled series capacitors in a manner similar to TCSC. This however has not seen much interest in academia and industry. The scheme suggested in this paper is similar to CCC and CSCC as far as connection is concerned, however, the strategy and control is totally different. Voltage is injected in the form of pulses rather at specific times rather than a continuous set. The methodology and the effects are presented in the form of MATLAB /Simulink simulations. Fig. : Schematic for case I (ideal reference case). II. METHODOLOGY As we are presenting the concept only, the system under test is very simple. Rectifier ac side with three-phase ac source voltages and a six-pulse, single-pole converter (made up of diodes) will be part of the simulation in all cases. The rest of the link is represented on the DC side by a smoothing reactor L d, the line resistance R d and a voltage source E inv representing the inverter. There are three cases that have been simulated. A. Case I This represents the ideal situation where there is no source inductance between the commutation bus and the converter terminals as shown in Fig.., V b, and V c are the phase voltages at the source. I a, I b, and I c are the line currents., and are the three-phase active and reactive powers supplied by the source., and are the active and reactive powers consumed by the converter. V d, I d, and P d are the DC side voltage, current, and power respectively. This will help establish the reference values for commutation bus voltages, line current, active and reactive powers, DC side voltage, and DC side current. B. Case II This case represents the real-world condition where there is source inductance between commutation bus and converter terminals as shown in Fig. 2. Source inductances are represented by L a, L b and L c and are equal in magnitude. The value of source inductance is 9. H. This is based on a base impedance of.6 and the source reactance chosen to be % of the base impedance. Q ind is the reactive power consumed by source inductance. This case will demonstrate the effects of source inductance on reactive power consumption by itself and the converter terminal. C. Case III This represents the suggested solution where nonsinusoidal series voltages are injected and would help in determining the effects of the suggested scheme. The schematic is shown in Fig. 3. V acomp, V bcomp, and V ccomp are the series voltages injected by compensator. P comp, and Q comp are the active and reactive power supplied by the series compensator. In all the above cases, for purpose of comparison, the source voltages have been adjusted so as to keep the value of the mean DC side quantities (i.e. voltage, current, and - V c Fig. 2: Schematic for case II (practical case including source inductance). I a -V b I c I b L a L b L c -V acomp -V bcomp -V ccomp power) at the same level. Also, the DC side inductor, resistor and voltage source representing the inverter are the same. All the cases are simulated for.2 s i.e. cycles to let the system quantities settle to steady-state values. Source resistances as well as converter losses have been neglected (except losses in RC series snubbers, values of snubber resistance and capacitance have been selected to avoid numerical oscillations in simulations only). Positive values for active and reactive powers from the source and the compensator mean supply whereas positive values in case of converter and source inductances mean consumption of active and reactive power. III. SIMULATION & RESULTS Major simulation results for case I are summarized in Table I. These show that the set input voltage values yield mean DC side voltage and current of pu., and thus the mean DC power is pu. Time domain plots for, V b, V c, and I a are shown in Fig. 4. These help us visualize the current commutation from phase c to a during the positive and negative half cycles of phase voltages. As there is no inductance between the converter and the source, there is instantaneous commutation of current from phase c to a making the fundamental component of I a in phase with V a. Fig. shows the harmonic spectrum of I a which is predictable with significant harmonics appearing at harmonic number kp, where k is any positive integer and p is the pulse number which is 6 in this case. The reverse voltage appearing across one switching device during its off period in a single cycle is shown in Fig. 6. Table II gives the steady-state values of major parameters V d - L d P d Rd I d E inv Rectifier P comp Q ind Q comp Fig. 3: Schematic for case III (suggested case for reactive power compensation).

3 TABLE I MAJOR PARAMETER VALUES FOR CASE I.646 pu pu pu.3 pu.3 pu TABLE II MAJOR PARAMETER VALUES FOR CASE II.622 pu.3 pu.3 pu.2273 pu.729 pu.44 pu Q ind Vb Vc Fig. 4: Time domain plots showing current commutation from phase c to a for case I. Only I a is displayed for clarity. Vb Vc Fig. 7: Time domain plots showing current commutation from phase c to a for case II. Only I a is displayed for clarity. % of Fundamental) 2 Fundamental (Hz) =., THD= 3.7% Fig. : Harmonic analysis of I a for case I. % of Fundamental) 2 Fundamental (Hz) =.99, THD= 24.6% Fig. 8: Harmonic analysis of I a for case II Fig. 6: Voltage across one switching device for case I. for simulations performed for case II. To keep the DC side quantities equal to those for case I, the phase voltage peak required is.622 which is 2.88% higher than the phase voltage for case I. Though the converter does not have any inductive reactance, it is still consuming reactive power due to delayed current commutation. The reactive power consumption of the converter is more than thrice that of the source inductance. Fig. 6 shows, V b, V c, and I a. It is clear that the current commutation lags source voltage crossovers and hence the fundamental component of I a lags the source voltages thus causing a consumption of reactive power by the converter as well as the source inductance as mentioned above. Fig. 8 shows the harmonic spectrum of I a. A comparison with Fig. reveals that harmonics have reduced somewhat due to the inductance which inherently blocks higher frequencies. Fig. 9 shows the reverse voltage appearing across a single switching device during its off Fig. 9: Voltage across one switching device for case II. state and it shows sharp edges if compared to the smooth waveform of the ideal case shown in Fig. 6. However the peak reverse voltage across any switch during its off period is not very different from case I. After careful examination of the Fig. 4 and Fig. 7 it is evident that if current commutation is forced to be in-phase with crossover of the phase voltages, line current will almost be in phase with respective phase voltage, and reactive power consumption will be reduced to minimum. There are two possible ways to do that. One is injection of an opposing voltage in the outgoing phase to reduce the magnitude of converter terminal voltage for the outgoing phase so that the incoming phase takes over prematurely. This opposing voltage will be negative during crossover in the positive half-cycle and positive during negative halfcycle. However, the polarity suggested in this case will be such that to absorb active power all the time and to put it in

4 another way, it means power loss. In physical terms this is equivalent to switching of resistors alternatively into conducting phases of the system which cause power loss. The other possibility is the injection of a supporting voltage to the incoming phase to enhance the incoming phase voltage so that it takes over before source voltage crossover. This scheme has been implemented for simulation of case III. This voltage will be positive during crossover in positive half-cycle and negative during crossover in negative half-cycle of the concerned source voltages. Fig. elaborates this control logic for voltage injection in phase a implemented in Simulink. The magnitude of the difference between V c (the outgoing phase) and (the incoming phase) is compared to a constant which determines how early the voltages at the converter terminals should cross. If this difference is less than or equal to the constant (.2 in this case), a switch is enabled passing the sign of V c to the output where it is again multiplied by the same constant to get the injected voltage value. Thus the injected voltage into phase a will be positive during crossover in the positive half cycle and negative during negative half-cycle. The value of the constant to which the voltage difference is compared can be adjusted using feedback taken from the measurement of reactive power consumption of the converter. However, this value cannot be increased indefinitely as the converter would have an upper voltage injection limit. The magnitude and timing of injection voltage determines how early the commutation will occur prior to source voltage crossover, and hence will control the reactive power supplied to the system. Table III summarizes the major parameters for simulation of case III. It becomes evident that this injection strategy has reduced the reactive power consumption of the converter terminal from.729 pu to.428 pu, a reduction of 7%. The reactive power consumed by source inductance is the same. Thus this scheme can be considered as a reactive power requirement limiter for converter as well as a supplier of reactive power. The reactive power supplied by the source also reduces significantly from that in case II. Consequently, the system is able to supply the same active power at a lower source phase voltage than case I (i.e..874 which is 2.84% lower than case I). In addition to the reactive power, the compensator also supplies a small fraction of the total power. Fig. shows the commutation of current from phase c to phase a. V inja (voltage injected into phase a ) is also plotted to understand the reason for early current commutation. Another thing worth noting is that due to this injection strategy, current in the incoming phase rises steeply but the rate of decay when the concerned phase is going out of conduction phase is not similar (slow decay in I a observable between.88 s and.9 s in Fig. ). Due to this, the current waveform is fatter and fundamental component of current still lags the respective source voltage causing a small consumption of reactive power by the converter. Fig. 2 shows the harmonic spectrum of I a and Fig. : Control logic for injection of voltage in phase a for commutation of current from phase c to a. shows that harmonic content in the current is smaller when compared to case I and II indicating lighter filtering requirements. Fig. 3 gives the plot of voltage across one switch in the converter. Peak reverse voltage value is not different when compared to Fig. 6 and Fig. 9. As we have been working with pu values throughout this work, it is suitable to get an idea of what the actual voltage. -. TABLE III MAJOR PARAMETER VALUES FOR CASE III.874 pu.9682 pu P comp.322 pu.4 pu.22 pu Q comp.72 pu.428 pu.44 pu Vinja Q ind Vb Vc Fig. : Time domain plots showing current commutation from phase c to a for case III. Only I a is displayed for clarity. (% of Fundamental) Fundamental (Hz) =.99, THD= 23.% Fig. 2: Harmonic analysis of I a for case III Fig. 3: Voltage across one switching device in converter for case III.

5 ratings of different system parameters will be. As an example case, assuming the DC link voltage to be kv, the peak source phase voltage to obtain this comes out to be 32 kv (.646 kv) corresponding to an rms value of 23 kv. The converter injected voltage for compensation will be kv peak (.2 kv). One of the advantages of this injection strategy is low switching frequency of the compensator and lower active time leading to lower switching and conduction losses. This makes this scheme more efficient than STATCOM which is operational all the time and switching at a high frequency leading to higher switching and conduction losses. A foreseeable problem might be the presence of high common mode voltages at the compensator terminals demanding higher insulation protection. If the compensator arrangement is connected at the base of the star connected converter secondary, the problem of common mode voltages could be solved. System protection from over-current during faults is another issue worth tackling like all series devices. IV. CONCLUSIONS An effort has been made to understand the reactive power consumption phenomenon in the conventional LCC technology. A series injection topology along with its control logic has been suggested for forced-commutation of conventional LCC technology to improve upon the reactive power consumption. It has been argued with simulation results that the suggested scheme helps reduce the reactive power consumption and improves the active power delivery capability of the ac system connected to the converter. This would help reduce (or eliminate) the need for capacitor banks for reactive power compensation. A converter setup to achieve required voltage injection is being worked upon and will be the topic of another publication. V. REFERENCES [] C. Adamson and N. G. Hingorani, High voltage direct current power transmission. London: Garraway, 96. [2] B. R. Andersen, L. Xu, P. J. Horton, and P. Cartwright, "Topologies for VSC transmission," Power Engineering Journal, vol. 6, pp. 42, 22. [3] J. Arrillaga, Y. H. Liu, and N. R. Watson, Flexible Power Transmission : The HVDC Options. Chichester, England: John Wiley & Sons, 27. [4] S.-P. Breton and G. Moe, "Status, plans and technologies for offshore wind turbines in Europe and North America," Renewable Energy, vol. 34, pp , 29. [] B. R. Andersen and X. Lie, "Hybrid HVDC system for power transmission to island networks," in Transmission and Distribution Conference and Exposition, 23 IEEE PES, 23, pp. -6 Vol.. [6] G. D. Breuer, H. M. Rustebakke, R. A. Gibley, and H. O. Simmons, Jr., "The use of series capacitors to obtain maximum EHV transmission capability," IEEE Transactions on Power Apparatus and Systems, pp. 9, 964. [7] E. W. Kimbark, "Improvement of system stability by switched series capacitors," IEEE transactions on power apparatus and systems, vol. pas-8, p. 8, 966. [8] J. J. Paserba, N. W. Miller, E. V. Larsen, and R. J. Piwko, "A thyristor controlled series compensation model for power system stability analysis," IEEE transactions on power delivery, vol., pp , 99. [9] M. M. R. Iravani and D. Maratukulam, "Review of semiconductor-controlled (static) phase shifters for power systems applications," IEEE transactions on power systems, vol. 9, pp , 994. [] L. Gyugyi, C. D. Schauder, and K. K. Sen, "Static synchronous series compensator: a solid-state approach to the series compensation of transmission lines," Power Delivery, IEEE Transactions on, vol. 2, pp , 997. [] Y. H. Zhang, D. M. Viathgamuwa, and S. S. Choi, "An experimental investigation of dynamic voltage restorer (DVR)," in Power Engineering Society Winter Meeting, 2. IEEE, 2, pp vol.4. [2] T. Takaku, N. Iwamuro, Y. Uchida, and R. Shimada, "Experimental demonstration of 2 V IGBT module for a magnetic energy recovery switch application," Transactions of the Institute of Electrical Engineers of Japan, Part C, vol. 28C, pp , 28. [3] T. Takaku, T. Isobe, J. Narushima, H. Tsutsui, and R. Shimada, "Power factor correction using magnetic energy recovery current switches," Transactions of the Institute of Electrical Engineers of Japan, Part D, vol. 2-D, pp , 2. [4] J. A. Wiik, F. D. Widjaya, T. Isobe, T. Kitahara, and R. Shimada, "Series connected power flow control using magnetic energy recovery switch," in Fourth Power Conversion Conference, Piscataway, NJ, USA, 27, pp [] J. A. Wiik, F. D. Wijaya, and R. Shimada, "An innovative series connected power flow controller, Magnetic Energy Recovery Switch (MERS)," in Power Engineering Society General Meeting, 27. IEEE, 27, pp. -7. [6] J. Reeve, J. A. Baron, and G. A. Hanley, "A technical assessment of artificial commutation of HVDC converters with series capacitors," IEEE transactions on power apparatus and systems, vol. pas-87, pp. 8384, 968. [7] Y. Kazachkov, "Fundamentals of a series capacitor commutated HVDC terminal," Power Delivery, IEEE Transactions on, vol. 3, pp. 76, 998. [8] K. Sadek, M. Pereira, D. P. Brandt, A. M. Gole, and A. Daneshpooy, "Capacitor commutated converter circuit configurations for DC transmission," IEEE transactions on power delivery, vol. 3, pp , 998.