Performance Comparison of Conventional STATCOM and STATCOM with Energy Storage in a Low Voltage Induction Motor Application

Transcription

1 Performance Comparison of Conventional STATCOM and STATCOM with Energy Storage in a Low oltage Indction Motor Application Antti irtanen Department of Electrical Energy Engineering Tampere University of Technology Tampere, Finland antti.a.virtanen@tt.fi Abstract This paper presents a performance comparison between a conventional STATCOM and a STATCOM with an integrated energy storage (ESTATCOM), within a low voltage stone crsher application. The systems are compared in terms of spply grid crrent rating mitigation capabilities, line voltage reglation and voltage flicker compensation characteristics. In addition, the reqired crrent ratings for the compensator inverter bridges, as well as the system total energy consmptions are stdied. The fnctionality and power losses of the ESTATCOM have been stdied with a smallscale laboratory test setp and the fllscale systems by compter simlations with Matlab Simlink software. According to the research reslts the ESTATCOM shows best performance within line voltage reglation, voltage flicker compensation, and spply crrent mitigation properties. In addition the ESTATCOM inverter bridge cold be rated significantly smaller than the one of the conventional STATCOM. Load compensation increases the system energy consmptions when the spply grid is strong, bt when the spply grid is weak the ESTATCOM consmes slightly less energy than the other stdied cases. Overall, load compensation becomes most beneficial within weak spply grids with the STATCOM devices stdied. I. INTRODUCTION Power qality isses sch as harmonic distortion, voltage flicker and voltage sags may be compensated with flexible ac transmission system devices (FACTS) sch as active filters and static synchronos compensators (STATCOM) [1], []. Typically active filters are sed for highorder harmonics compensation and STATCOMs for reactive power compensation and point of common copling (PCC) voltage reglation [1], []. The crrent trend of increasing distribted generation and the concept of smart grid will presmably increase the nmber of FACTS devices connected to the grid. Local energy storages (ES), sch as batteries and spercapacitors (SC), are also seen as an important part of the smart grid, since peaks in prodction and consmption can be flattened by temporarily storing electrical energy [3]. Heikki Tsa Department of Electrical Energy Engineering Tampere University of Technology Tampere, Finland heikki.tsa@tt.fi The above mentioned facts highlight the concept of integrating ES into STATCOMs, which enables the STATCOM to provide also active power spport to the grid in addition to the conventional STATCOMs reactive power. This paper presents a performance comparison between a conventional STATCOM and a STATCOM with integrated energy storage (ESTATCOM) in a low voltage indction motor (IM) application. The systems are compared in terms of spply grid crrent rating mitigation, line voltage reglation and voltage flicker compensation characteristics. In addition, the reqired crrent ratings for the compensator inverter bridges, as well as the system total energy consmptions are stdied. The noncompensated system (NC) consists of a directly gridcopled IM driving a stone crsher, which is known to case heavy flctation in the power drawn from the grid [4]. Stone crshers are typically sed in the mining indstry and in rral areas where the strength of the spply grid may be very weak. Ths, the variance of the spply grid strength is also taken into accont by stdying the systems within three different strength spply grids. The research has been divided into two parts: experimental tests and compter simlations. The experimental tests inclde verification of the ESTATCOM operation in practice with a smallscale laboratory test setp, and power loss measrements of the system components dring operation. The compter simlations consist of fllscale system modeling and simlation with the power loss models developed on the basis of the smallscale laboratory measrements. Section II presents the compensation systems stdied and their control principles. Section III presents the ESTATCOM laboratory test setp and its simlation model. Section I presents the fllscale system simlation models and the analysis of the simlation reslts. Finally, the conclsions are drawn in Section.

2 II. THE COMPENSATION SYSTEMS A. STATCOM with Integrated Energy Storage (ESTATCOM) Fig. 1 presents the main circit and control system of the ESTATCOM. It consists of a shntconnected voltage sorce inverter with an SC bank connected to the system dc link via a bidirectional DCDC converter. An LCLfilter is inclded to damp the switching freqency ripple of the inverter crrent i inv. The aim of the control is to draw steady average active power from the grid and flly compensate the reactive power of the load. Dring system operation the SC is discharged when the instantaneos active power of the load is greater than the average, and charged when the active power is lower than the average. The control system can basically be divided into three parts: one consisting of the inverter side to control i inv, and the second to reglate the dc link voltage dc with the combination of the DCDC converter and the SC. The third control branch is sed to keep the SC voltage SC within its desired operating range nder longterm operation. The inverter control system is based on the pq theory implemented in a stationary (α,β) reference frame. First the instantaneos active and imaginary powers (p l, q l ) of the load are calclated on the basis of load crrent i l and PCC voltage PCC measrements in the PQblock [1], [4] 3 pl = ( PCCαilα PCCβilβ ) (1) 3 ql = ( PCCβilα PCCαilβ ). The active power reference p f,ref is fond by inverse lowpass filtering p l (1 st order LPF with time constant τ =.5 s) and adding the SC voltage controller otpt p f0. The imaginary power reference q f,ref is fond by inverting q l, since the aim is to flly compensate the load imaginary power. After these operations the crrent reference i inv,ref is calclated in the G(p,q, PCC )block as follows [1], [4]: i invα,ref i invβ,ref p = 3 p = 3 f,ref f,ref PCCα PCCα PCCβ PCCα q q PCCβ f,ref PCCβ PCCβ PCCα Finally the crrent controlloop prodces the LCLfilter voltage reference LCL,ref, and the inverter voltage reference inv,ref is obtained by sbtracting LCL,ref from the measred PCC. The DCDC converter control system is based on feedback control of the measred dc. The objective is to keep dc arond its reference vale dc,ref. B. Conventional STATCOM The main circit and control system of the conventional STATCOM is presented in Fig.. The difference with the ESTATCOM is that the system incldes a PCC voltage controller, which otpts an imaginary power reference q PCC,ref. With positive q PCC,ref the system injects imaginary power to the grid increasing PCC, and with negative q PCC,ref the system draws imaginary power from the grid and PCC decreases [5], [6]. The time constant of the PCC voltage controller LPF is τ =.5 s and the time constant of the LPF for p l filtering is τ =.7 ms in order to compensate the possible load harmonics. III. f,ref. ESTATCOM LABORATORY TEST SETUP AND SIMULATION MODEL This section presents the strctre of the ESTATCOM laboratory test setp. In addition, the development of the simlation model inclding power loss models for the system components is presented and the experimental reslts are compared with the simlation reslts. () αβ, Transformer 0k/400 L grid R grid i s PCC i l Indction motor M 3~ LOAD i l p l PQ PCC αβ, R damp C f αβ, L f1 L f i inv inv Inverter C dc DCDC converter i SC C SC SC PCC LPF PCC i inv,ref i inv LCL,ref G( pq,, PCC ) PID Modlator dc p f,ref Crrent controller p l0 p lh pf0 Pe Spercapacitor voltage controller Figre 1. ESTATCOM main circit and control system. inv,ref i SC,ref PI Dc link voltage controller dc,ref SC SC,ref

3 αβ, Transformer 0k/400 L grid R grid i s PCC i l Indction motor M 3~ LOAD Mains 400 UPM IM 11 kw HBM T30FN Load DC Machine 6 kw Mains 400 R grid L grid l, i l p l, q l T l n l i l p l M 3~ M DC PQ LPF Thyristor converter 100 ka G( pq,, PCC ) PID p f,ref Crrent controller Modlator p l0 p lh p f0 Pe Dc link voltage controller Figre. STATCOM main circit and control system. s, i s p s, q s PCC, i PCC p PCC, q PCC LPF Inverter 18 ka PCC α β PI PCC voltage controller UPM UPM Tektronics TCP303 Figre 3. Strctre and measrement devices of the ESTATCOM laboratory test setp. dc i dc Tektronics P500 dspace ds1103 and LeCroy oscilloscopes 9301A and LT 354M for data acqisition DCDC Converter 90 kw PCC Tektronics TCP303 i SC αβ, A. Test Setp Strctre and Components The control system of the ESTATCOM laboratory test setp was similar to the one presented in Fig. 1. The strctre of the setp and the measrement devices sed are presented in Fig. 3, and the system component parameters in Table I. A photograph of the testsetp is presented in Fig. 4. The nominal power of the IM was 11 kw, and a 6 kw separately exited DC machine fed by a threephase forqadrant thyristor rectifier was sed to emlate the stone crsher load. The spply grid impedance was composed of an indctor of L grid = 1.7 mh and R grid = 13 mω, in order to artificially weaken the spply grid. The ESTATCOM inverter was a threephase twolevel inverter with rated power of 18 ka. It was bilt by modifying a commercial freqency converter into an active power filter [7]. The inverter IGBTswitches were of type SKM 75GB176D. The SC i inv,ref Tektronics P500 i inv SC F R damp C f αβ, L f1 L f LCL,ref i inv PCC inv,ref inv Inverter C dc dc dc,ref DCDC converter was a commercial bidirectional converter of type Msc 00DCDC750 with rated power of 90 kw. The spercapacitor was a Maxwell BMOD0063 P15 modle with a nominal capacitance of 63 F and nominal voltage of 15. A Motorola MPC555 microcontroller was sed to implement the inverter control. The dc link and SC voltage control systems, as well as spplying the load torqe reference for the thyristor rectifier, were carried ot with dspace ds1103 realtime simlation platform. The measrements were carried ot with the following instrments. For measring i SC and i dc Tektronics TCP303 crrent probes together with TCPA300 amplifiers were sed, and SC and dc were measred with Tektronics P500 high voltage differential probes. For the threephase ac qantities three analog niversal power meters (UPM), each consisting of LEM LA 50P crrent transdcers and differential voltage measrement nits for each phase, were sed. The load torqe T l and rotational speed n l were measred with a HBM T30FN torqe transdcer and a HBM DA 3418 amplifier. Figre 4. Photograph of ESTATCOM laboratory test setp.

4 TABLE I COMPONENTS AND SYSTEM PARAMETERS OF ESTATCOM LABORATORY TEST SETUP Indction Motor Nominal shaft power P N 11 kw Nominal voltage U N 380 Nominal crrent I N 3 A Spply Grid Impedance Grid indctance L grid = 1.7 mh (0.051 p) Grid resistance R grid = 13 mω (0.001 p) ESTATCOM (IGBTs: SKM 75GB176D) Rated power S N 18 ka Switching freqency f s 10 khz LCL filter indctor L f1 0.6 mh (0.0 p) ESR of L f1 0 mω (0.00 p) LCL filter indctor L f 6 mh (0. p) ESR of L f 100 mω (0.01 p) LCL filter capacitor C f 30 μf (0.08 p) Damping resistor R damp 4.5 Ω (0.53 p) Dc link capacitance C dc 3.35 mf (8.96 p) Dc link voltage reference dc,ref 730 DCDC Converter (Msc 00DCDC750) Rated power 90 kw Maximm continos crrent 10 A Maximm peak crrent 00 A, 1 min every 10 min Spercapacitor Modle (Maxwell BMOD0063 P15) SC nominal capacitance C SC 63 F ( p) SC nominal voltage SC 15 SC voltage reference SC,ref 115 B. Laboratory Test Setp Simlation Model for Component Power Loss Modeling A simlation model of the ESTATCOM laboratory test setp was developed simltaneosly with the experimental tests. The main objective was to accrately model the power losses of the system components in order to se the power loss models in the fllscale system simlations. The following power losses were taken into accont in the simlations: 1) inverter condction and switching losses, ) LCLfilter iron and copper losses, and 3) DCDC converter losses and SC losses. The parameters of the power loss models are presented in Table II. 1) Inverter Condction and Switching Losses In the calclation of both the inverter condction and switching losses it is assmed that the inverter operates with sinsoidal crrents, and that the calclated losses are achieved as an average over the 50 Hz fndamental period. The average condction losses for one active switching device P C SW and one antiparallel diode P C D in a threephase twolevel inverter, can be calclated from [8], [9] P P C SW C D 1 = iˆ 1 = iˆ inv inv 1 M cos ϕ ˆ t i π 4 1 M cos ϕ ˆ f i π 4 inv inv R R CE AK M cos ϕ (3) π 3π 3 M cos ϕ,(4) π 3π TABLE II PARAMETERS OF ESTATCOM LABORATORY TEST SETUP POWER LOSS MODELS IGBT zerocrrent voltage drop t 0.9 Diode zerocrrent voltage drop f 0.9 IGBT onstate resistance R CE 31 mω Diode onstate resistance R AK 18 mω IGBT loss energy determination voltage CC 100 Inverter ventilation & control losses P vent 75 W LCLfilter iron losses P loss,fs,lcl 150 W DCDC converter minimm losses P loss,min,dcdc 300 W SC one way efficiency η SC where î inv is the peakvale of the sinsoidal inverter crrent, t and f the voltage drops at zerocrrent condition for the IGBT and diode, R CE and R AK the resistive elements of the IGBTs and diodes, cos φ the displacement power factor, and M is the modlation index calclated from inv, ref M =. (5) dc / Parameters t, f, R CE, and R AK can be obtained from manfactrer s datasheets. The inverter average switching loss model consists of IGBT trnon and trnoff losses P ON/OFFSW, and diode reverse recovery losses P RRD, respectively. These are calclated from [10] [ E (ˆ i ) E (ˆ i )] 1 dc P ON/OFFSW = fs ON inv (6) OFF inv π CC 1 dc PRR = fs ERR(ˆ iinv), (7) π CC as a fnction of the switching freqency f s for one active switching device and one antiparallel diode. In (6) and (7) E ON, E OFF and E RR are the respective trnon, trnoff and reverse recovery loss energies as a fnction of the inverter crrent, and CC is the collectoremitter spply voltage sed in determination of the loss energies. These parameters can all be obtained from the manfactrer s datasheet. Finally the total power losses for the six inverter bridge IGBTs and antiparallel diodes can be calclated from P LOSS TOT = 6 ( PC SW PC D PON/OFF SW P RR ) P,(8) vent where P vent is an additional power loss of 75 W of the inverter ventilation fan and control electronics. ) LCLfilter Power Losses In the modeling of the LCLfilter power losses the copper and iron losses of the filter indctors were taken into accont. The copper losses were modeled with appropriate eqivalent series resistances (ESR) for indctances L f1 and L f. The iron losses cased by the switching freqency ripple of the compensator crrent were approximated as constant P loss,fs,lcl = 150 W by comparing the simlation reslts with the measrements.

5 3) DCDC Converter and SC Power Losses The power losses for the DCDC converter were modeled with efficiency maps for the charge and discharge operations, presented respectively in Fig. 5. The maps were obtained by charging and discharging the SC modle with crrents of 5 10 A, and the efficiency is fond as a fnction of the SC crrent and the conversion ratio SC / dc. The minimm power losses of the DCDC converter were limited to P loss,min,dcdc = 300 W in order to cover the losses de to the ventilation and control electronics and de to slight inaccracies in the maps dring low crrent operation. The power losses of the SC were modeled with a fixed oneway efficiency of η SC = 0.985, on the basis of measrements for a Maxwell BMOD0063 P15 modle presented in [11]. C. Measrement and Simlation Reslts The measrement reslts of the ESTATCOM laboratory test setp and the NC system, as well as the respective simlation reslts dring a 60 s load cycle are presented in Figs. 6a i. In the figres the ESTATCOM measred waveforms are presented in ble, ESTATCOM simlations in red, and NC measrements in black. Figs. 6a,b present the spply transformer active and imaginary powers for the systems stdied. With the ESTATCOM the active power becomes smoother than with the NC, and steady average power is drawn from the grid. The imaginary power becomes also compensated to zero as anticipated in Section II. Fig. 6c presents the system PCC voltages. Dring the load cycle PCC decreases becase of the voltage drop over the spply grid impedance. The ESTATCOM is still able to maintain higher and more leveled PCC than the NC, de to the compensated active and imaginary powers. Figs. 6d f present the ESTATCOM dc link voltage, SC crrent and SC voltage, respectively. The dc control system keeps dc arond dc,ref, by discharging the SC when the load active power is higher than the average, and by charging the SC when the active power is lower than the average. Dring longterm operation SC settles into a voltage vale in which the system losses become compensated by the SC voltage controller otpt p f0. Figs. 6g h present the measred ESTATCOM total power losses, combined losses of the inverter and LCLfilter, and DCDC converter losses compared against the respective simlated power losses. The measred and simlated average power losses P loss,avg of the ESTATCOM components dring the measrement period are presented in Table III. Despite some inaccracy between the measred and simlated power losses of the inverter and LCLfilter dring the idle period at the beginning and end of the measrement period, the power loss models give a reasonable approximation for the system losses, and can be sed in fllscale system modeling. I. FULLSCALE SYSTEM SIMULATIONS A. Simlated Systems The simlation models of the ESTATCOM and the NC system developed in Section III were scaled p to match the power levels of the fllscale 50 kw systems. In addition, a simlation model of the conventional STATCOM of Fig. was developed. The parameters of the fllscale systems are presented in Table I. The nominal power of the fllscale system IM was 50 kw, and the rated powers of both ESTATCOM and conventional STATCOM inverters were 300 ka. The per nit vales of the LCLfilter passive elements were designed to match those of the ESTATCOM laboratory test setp. The type of the IGBTmodles sed in the inverter bridge power loss model was SKM 900GA1E4 and the losses P vent as well as P loss,fs,lcl were mltiplied by a factor of 300/18, according to the ratio of the laboratory test setp and fllscale system rated powers. The DCDC converter power losses were modeled with the efficiency maps presented in Fig. 7, and P loss,min,dcdc was approximated to correspond to three Msc 00DCDC750 converters. The SC bank was composed of two parallel nits, both comprising of a series connection of five Maxwell BMOD0063 P15 SC modles. The systems were simlated in spply grids of three different strengths: strong grid, medim grid, and weak grid. The strong grid consisted of a 500 ka transformer and an AMMK cable of 00 m, the medim grid of a 500 ka transformer and an AMMK cable of 500 m, and the weak grid of a 500 ka transformer and an AMMK cable of 1000 m, respectively. The per nit indctance of the strong grid matched the grid indctance of the laboratory test setp. (a) (b) Figre 5. DCDC converter efficiency maps for (a) charge efficiency, (b) discharge efficiency.

6 (a) (b) (c) (d) (e) (f) (g) (h) (i) Figre 6. ESTATCOM test setp measrement reslts. (a) Spply active powers, (b) spply imaginary powers, (c) PCC voltages PCC,rms, (d) Dc link voltage DC, (e) SC crrent i SC, (f) SC voltage SC, (g) ESTATCOM total power losses, (h) inverter and LCLfilter combined power losses, (i) DCDC converter power losses. ESTATCOM measred (ble), ESTATCOM simlated (red), NC measred (black). TABLE III MEASURED AND SIMULATED AERAGE POWER LOSSES OF ESTATCOM LABORATORY TEST SETUP System P loss,avg Measrement Simlation DCDC converter 370 W 34 W Inverter LCLfilter 590 W 599 W Total 960 W 941 W B. Simlation Reslts The system performances were analyzed in the following categories: 1) Spply transformer and cable crrent ratings ) PCC voltage reglation and system flicker indices 3) Compensator crrent ratings Figs. 8a f present the simlation reslts dring a 60 s stone crsher load cycle. In the figres the ESTATCOM waveforms are presented in red, the conventional STATCOM in green, and the NC system in ble. A smmary of the reslts is presented in Table. 1) Spply Transformer and Cable Crrent Ratings Fig. 8a presents the spply phasea crrents I s,rms for the stdied systems, in the medim grid. The ESTATCOM effectively smoothes the I s,rms waveforms in comparison with the other systems. The reslts of Table predict that in the weak grid the rated crrent of the spply transformer and cables cold be approximately 0 % lower with the ESTATCOM, and 10 % lower with the conventional STATCOM than with the NC system, depending of the grid strength. 4) Total energy consmption

7 (a) (b) Figre 7. Efficiency maps of fllscale ESTATCOM DCDC converter. (a) Charge efficiency, (b) discharge efficiency. ) PCC oltage Reglation and System Flicker Indices Fig. 8b presents the PCC,rms in the medim grid for the stdied systems, and Table the PCC minimm vales and percental voltage drops of U N (400 ) respectively. Both compensated systems are able to maintain the PCC drop at an acceptable level below 15 %, de to power compensation and the PCC control. The maximm voltage drop of the NC system in the weak grid is 31 %, which is significantly higher than the voltage drops of the compensated systems. The voltage flicker indices P st of the PCC and spply transformer voltages were analyzed with a flicker meter simlation model implemented in line with the standard IEC [1]. With the ESTATCOM all of the cases stdied prodce P st vales below the threshold level for irritating voltage flicker of 1 p (Table ). With the conventional STATCOM the flicker levels are acceptable in the strong grid, bt in the medim and weak grids the PCC voltage P st vales exceed the threshold vale. With the NC system the P st threshold level is exceeded in all cases stdied, ths load compensation significantly improves the voltage qality. 3) Compensator Crrent Ratings Fig. 8c presents the crrents I inv,rms of the both compensator inverters in the medim grid. The crrent of the conventional STATCOM has clearly higher peak vales than the crrent of the ESTATCOM. The peak vales of the crrents determine the crrent ratings for the IGBT switches of the compensator inverter bridges. It can be observed from Table that in the case of the strong grid IGBTs with approximately the same rated crrent cold be sed in both systems. However, in the medim grid the ESTATCOM IGBTs cold be downsized to abot 75 %, and in the weak grid to abot 63 % compared to the conventional STATCOM IGBTs. 4) Total Energy Consmption Figs. 8d f present a comparison of the systems energy consmption as a fnction of the grid strength. The energy consmptions were calclated assming that the systems wold be in the steadystate mode of the time period of s for one hor. In the case of the strong grid the total energy consmption of the ESTATCOM and conventional STATCOM becomes higher than with the NC system, becase of the power losses in the compensator power electronics and passive elements. However, in weaker grids TABLE I SYSTEM PARAMETERS AND COMPONENTS OF SIMULATED FULLSCALE SYSTEMS Indction Motor Nominal shaft power P N 50 kw Nominal voltage U N 400 Nominal crrent I N 450 A Spply Grid Definitions Strong grid (00 m cable) L grid = 81 μh (0.050 p) R grid = 15 mω (0.09 p) Medim grid (500 m cable) L grid = 131 μh (0.081 p) R grid = 33 mω (0.065 p) Weak grid (1000 m cable) L grid = 15 μh (0.13 p) R grid = 6 mω (0.1 p) ESTATCOM & STATCOM Rated power S N 300 ka Switching freqency f s 10 khz LCL filter indctor L f1 37 μh (0.0 p) ESR of L f1 1 mω (0.00 p) LCL filter indctor L f 370 μh (0. p) ESR of L f 5 mω (0.01 p) LCL filter capacitor C f 130 μf (0.0 p) Damping resistor R damp 0.8 Ω (0.53 p) Dc link capacitance C dc 53.8 mf (8.96 p) Dc voltage reference dc,ref 800 Inverter Bridge Power Loss Model (SKM 900GA1E4) IGBT zerocrrent voltage drop t 0.7 Diode zerocrrent voltage drop f 0.9 IGBT onstate resistance R CE 1.7 mω Diode onstate resistance R AK 1.6 mω IGBT energy determination voltage CC 600 Other losses Inverter ventilation & control losses P vent 150 W LCLfilter iron losses P loss,fs,lcl 500 W DCDC converter min losses P loss,min,dcdc 900 W Spercapacitor Bank SC capacitance C SC 5. F (400 p) SC nominal & reference voltage SC,ref 65 SC one way efficiency η SC the compensator losses become less dominant in proportion to the increasing grid losses. The grid losses increase the most with the NC system, and in the weak grid the total energy consmption with the ESTATCOM becomes approximately one percentage nit lower than with the NC system.

8 (a) (b) (c) (d) (e) (f) Figre 8. Fllscale system simlation reslts. (a) Spply crrents i s,rms, (b) PCC voltages PCC,rms, (c) compensator crrents i f,rms, (d) total energy consmptions, (e) grid energy losses, and (f) compensator energy losses. ESTATCOM (red), STATCOM (green), NC (ble).. CONCLUSIONS A performance comparison was presented between the ESTATCOM, conventional STATCOM and the NC system within a stone crsher application. First the fnctionality of the ESTATCOM was verified with a smallscale 11 kw laboratory test setp and the power losses of the system components were measred. Second, simlation models for modeling the power losses in the ESTATCOM inverter bridge, LCLfilter, DCDC converter and SC were developed on the basis of the laboratory measrements. Finally, the smallscale simlation models were scaled p to match the power level of a fllscale 50 kw process, and the systems were simlated in three different strength spply grids. The system performances were analyzed on the basis of the spply grid crrent ratings, line voltage reglation capabilities, voltage flicker indices, compensator crrent ratings, and the total energy consmptions. The best performance in line voltage reglation, flicker redction, and spply grid crrent mitigation were achieved with the ESTATCOM. In addition the ESTATCOM inverter bridge cold be rated significantly smaller than that of the conventional STATCOMs, especially in the weak grid. Load compensation increases the system energy consmptions in the stronger spplies de to power losses in the compensator components. However, in the weak grid the ESTATCOM consmes slightly less energy than the other stdied cases. Overall, load compensation becomes most beneficial in the weak spply grid with the STATCOM devices stdied. REFERENCES [1] H. Akagi, E. H. Watanabe, M. Aredes, Instantaneos power theory and applications to power conditioning, John Wiley & Sons, USA 007, 379 p. [] N. Hingorani, L. Gygyi, Understanding FACTS, John Wiley & Sons, USA, 000, 43 p. [3] M. Bollen, The smart grid: adapting the power system to new challenges, Morgan & Claypool, USA, 011, 164 p. [4] A. irtanen, H. Tsa, Power compensator for high power flctating loads with a spercapacitor bank energy storage, in proc. IEEE PeCon, 008, pp [5] P. S. Sensarma, K.R. Padiyar,. Ramanarayanan, Analysis and performance evalation of a distribtion STATCOM for TABLE PERFORMANCE COMPARISON OF FULLSCALE SYSTEMS Spply Crrent (Arms / % of NC) Case ESTATCOM STATCOM NC Strong grid 33 / / / 100 Medim grid 334 / / / 100 Weak grid 35 / / / 100 Minimm PCC oltage (rms / % drop of U N) Strong grid 388 / / / 9.0 Medim grid 375 / / / 18.0 Weak grid 351 / / / 31.3 oltage flicker indices P st (Spply / PCC) Strong grid 0.04 / / / 3.08 Medim grid 0.11 / / / 6.6 Weak grid 0.11 / / / Compensator Peak Crrents (Arms / % of STATCOM) Strong grid 50 / / 100 Medim grid 505 / / 100 Weak grid 506 / / 100 Total Energy Consmption (kwh/h / % of NC) Strong grid / / / 100 Medim grid 9 / / / 100 Weak grid 41 / / / 100 compensating voltage flctations, IEEE Trans. Power Delivery, vol. 16, no., April 001, pp [6] M. Bongiorno, J. Svensson, oltage dip mitigation sing shntconnected voltage sorce converter, IEEE Trans. Power Electron., vol, no 5, Sept. 007, pp [7] P. Parkatti, M. Rotimo, H.Tsa, Modification of a commercial freqency converter to an active power filter in 690 power system, in proc. PCIM Erope, 007, 6 p. [8] J. S. Lai, R. W. Yong, J. W. McKeever, Efficiency consideration of DC link soft switching inverters for motor drive applications, in proc. IEEE PESC, 1994, vol., pp [9] L. M. Tolbert, F. Z. Peng, T. G. Habetler, Mltilevel converters for large electric drives, IEEE Trans. Ind. Appl., vol. 35, no. 1, Jan/Feb, 1999, pp [10] U.Nicolai, T.Reimann, J.Petzhold, J. Ltz, P.Martin, SEMIKRON Application Manal, ISLE, Germany, 1998, 70 p. [11] A. irtanen, H. Haapala, S. Hännikäinen, T. Mhonen, H. Tsa, Calorimetric efficiency measrements of spercapacitors and lithimion batteries, in proc. APEC, 011, pp [1] IEC , Testing and measrement techniqes Flickermeter Fnctional and design specifications, 003, 47 p.