Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Control of Indirect Matrix Converter Under Unbalanced Source Voltage and Load Current Conditions

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Fourth National Power Electronics Conference, IIT Roorkee, June - 2010

CONTROL OF INDIRECT MATRIX CONVERTER UNDER UNBALANCED SOURCE VOLTAGE AND LOAD CURRENT CONDITIONS
Mr. V. Y. Vekhande, Mr. B. B. Pimple and Prof. B. G. Fernandes Department of Electrical Engineering, Indian Institute of Technology Bombay, Mumbai ABSTRACT Doubly Fed Induction Generator (DFIG) is widely used along with AC-DC-AC converter for wind power generation. Merits and demerits of Indirect Matrix Converter (IMC) over AC-DC-AC converters are discussed in this paper. Detailed analysis is carried out for operation of IMC when unbalanced voltages are applied to IMC input and unbalanced voltages and currents derived at its output. Simulation results are presented for control strategy which eliminate sub-harmonic components in output and input currents caused due to unbalanced input voltages. INDEX TERMS Indirect Matrix Converter (IMC), Doubly Fed Induction Generator (DFIG), Unbalanced Grid Voltage INTRODUCTION DC-link capacitor occupies significant volume of conventional AC-DC-AC converter. Capacitor size depends on current rating and ripple contents. DC-link capacitor have shortest lifetime of any element used in power electronic converters. As the ambient temperature increases, its lifetime further decreases [1]. Indirect Matrix Converter (IMC) is an AC-DC-AC converter without DC-link capacitor. This converter topology is free from above mentioned problems associated with DC-link capacitor. It has features like bidirectional power flow capability, sinusoidal input currents with only high frequency switching harmonics and unity displacement power factor [2]. However, maximum voltage transfer ratio is limited to 0.866 [3]. Special care needs to be taken to overcome commutation problem. Variable speed generation scheme utilizing Doubly Fed Induction Generator (DFIG) is widely accepted solution for wind power generation. In DFIG, stator is directly connected to grid and rotor is connected to grid through AC-DC-AC converter. Controlled voltages are applied to rotor terminals to control active and reactive power flow between DFIG and grid [4]. For operation of DFIG up to 20-25% of slip, small voltages are required for rotor injection. So, IMC with limited voltage transfer ratio can be used for this particular application. IMC BASIC OPERATIONAL THEORY IMC Circuit Topology The rectifier section consists of six fully controlled bidirectional switches while inverter consists of six fully controlled unidirectional switches as shown in Fig. 1. Bidirectional switches are required to obtain a sinusoidal input current [5]. Rectifier section is capable of carrying current in both direction and produce positive or negative DC-link voltage. However, rectifier operation is restricted to produce positive average and instantaneous DClink voltage to avoid forward biasing of inverter section freewheeling diodes.

Figure 1. Indirect Matrix Converter Rectifier Operation Rectifier section is operated such that it produces maximum possible voltage with unity displacement factor at input. Only two switches will conduct at a time. Three phase input voltages are,

Fourth National Power Electronics Conference, IIT Roorkee, June - 2010


ua = ui cos(i t ) (1)

ib = d ab i ic = dac i

(7) (8)

ub = ui cos(i t

2 ) 3

(2) (3)

2 uc = ui cos(i t + ) 3

A phase input is clamped to the positive or negative bus in 3 -wide (60o) intervals when corresponding phase voltage has the highest absolute value. The other bus is toggled between remaining two phases. For example, if phase a is at its peak, then instantaneous value of DC-link voltage is u = uab or u = uac . Fig. 2 shows resultant DC-link voltage. Average value of DC-link voltage over pulse period TP is,
u = 3 1 ui 2 cos in

To achieve unity displacement factor, there should be proportional relationship between average value of an input phase current over pulse period TP and corresponding input phase voltage [2], [3]. i.e. voltage having higher magnitude will conduct for longer time. So,
dab = ib ub = ua ia

(9)

dac =

ic uc = ua ia

(10)

(4) (5)

For pulse period TP absolute conduction times are,

umin

3 = ui 2

ab = dab ac = d ac

TP 2 TP 2

(11) (12)

Average value of DC-link voltage over pulse period TP can also be written as,

u = uab d ab + uac dac


Inverter Operation

(13)

Inverter stage can be analyzed as normal VSI. It operates on pulsating DC-link voltage as shown in Fig. 2, formed by rectifier. Fig. 3 shows the operation of inverter in sector-I. The instantaneous values of active space vectors are,
u(100) =

2 u 3

(14)

Figure 2. DC-link Voltage

u(110) =

2 j ue 3 3

(15)

Let, dab = duty ratio of uab and d ac = duty ratio of uac In given pulse period, dab + dac = 1 If ia , ib , ic and i are average values of input phase a, b, c and DC-link currents respectively over pulse period TP . Then,

Let, (100) = absolute time period of u(100) and (110) = absolute time period of u(110) Output phase space vector is given by,
Tp 2 uoref = u(100) (100) + u(110) (110) (16)

Each active vector of inverter has two levels depending on rectifier switching condition.

ia = (dab + dac )i

(6)

Fourth National Power Electronics Conference, IIT Roorkee, June - 2010


Commutation

There are two methods to transfer rectifier current from one phase to the other.

First connect incoming phase and then open outgoing phase - which will momentarily short circuit the input phases. First open the outgoing phase and then connect the incoming phase - inductive current needs to be interrupted.

Figure 3. Sector-I of Inverter Space Vector Diagram [5]

Tp

2 2 uac (100),ac + uab (100),ab 2 3 3 j j 2 2 + uac e 3 (110),ac + uab e 3 (110),ab 3 3 uoref =

Both the conditions are undesirable and can be avoided by proper coordination between rectifier and inverter switching. 0 voltage vector is applied to inverter while transferring rectifier current from one phase to the other [3], [6]. It means, all output phases are connected to same bus which results in zero DC-link and rectifier currents. Now, rectifier current can be transferred safely using second method. Let, inverter vectors u1, u0 and u2 have time periods

(17)

1, 0 and 2 respectively. These time periods are


calculated from average DC-link voltage u over pulse period TP and reference output voltage requirement. r1 and r2 are time periods calculated for rectifier line voltages ur1 and
ur2 from unity displacement factor requirement of

Inverter active vectors u(100) and u(110) should be distributed in time periods ac and ab in same ratio as dac and dab . So,

(100),ac

(100),ac (100),ab = = (100),ab = = (100) (18) ac ab

rectifier. Time periods u1 , u0 and u2 are split into ratio such that,

(110),ac

(110),ac (110),ab = = (110),ab = = (110) (19) ac ab

u r u r

11

u r u r

0 1

u r u r

2 1

u u

r1

(23)

12

0 2

2 2

r2

So, output space vector expression is also given by, uoref = 2 2 u (100) + u (110) 3 3 (20)

These time periods are utilized to generate switching patterns for rectifier and inverter over one pulse period TP , as shown in Fig. 4.

Overall voltage transfer ratio is [3], M=

uoref 3 2 ui

(21)

Maximum voltage transfer ratio also depends on input power factor angle i [3],
uoref = M cos i ui (22)

It means, maximum voltage transfer ratio (0.866) is available only when input rectifier operates at unity power factor.

Figure 4. Rectifier and Inverter Switching Pattern in Pulse Period TP

Fourth National Power Electronics Conference, IIT Roorkee, June - 2010


Variable Power Factor at Rectifier Input

Let rectifier switching function is,


sw i = sw i e j (i t +i )

Average DC-link voltage can also be written as [5],


u=

(27)

3 ui m1 cos i 2

(24)

So, expression for input current in terms of DClink current is,


i1 = sw i i = e j (i t +i ) i ,(assume sw i = 1) (28)

If i is increased beyond 6 , then DC-link voltage has negative values for some instants and average DC-link voltage further reduces. Finally, when i = 2 , average DC-link voltage is zero [5]. However, negative DC-link voltage will forward bias the anti-parallel diodes across inverter IGBT switches. So, maximum possible adjustable input power factor angle can be achieved is 6 .
EFFECTS OF UNBALANCED INPUT/ OUTPUT SOURCE/ LOAD CONDITIONS

The instantaneous input power and DC-link power should be equal. Pi = 3 Re ui ii * = ui 2

{ }

(29) (30)

3 Re ui e j (i t +i ) i = ui 2

Substituting (26) in (30) and simplifying it further, gives the expression for DC-link voltage as, u= 3 u pi cos i + uni cos ( 2i t + i + nui ) 2

The IMC is to be used with DFIG for wind power generation. Wind power plants often installed in remote, rural areas. These remote areas usually have weak grids. The size of wind generator determines the PCC grid voltage level at which it will be connected. Single or group of small plants of 100-500 kVA are connected at low voltage grid at the consumer supply points. Generators in the range of 1-5 MVA are connected at distribution level (11-33 kV) [7]. At this level, the permissible limit of PCC voltage unbalance is upto 3% [10]. Under unbalanced grid voltage condition electromagnetic torque, stator currents, active and reactive powers have second harmonic pulsations. These pulsations cause mechanical stress, noise in turbine structure and unbalanced heating in stator windings of DFIG. Unbalanced voltages or currents are to be injected into DFIG rotor to compensate for effects of unbalanced grid voltages. So, it is necessary to analyze IMC operation under such conditions.
Effects of Unbalanced Input Voltages

(31)

This shows second harmonic component in DClink voltage. The inverter output voltage space vector can be expressed in terms of DC-link voltage as, uo = sw o e j (o t +o ) u 2 (32)

Substituting (31) in (32) and simplifying, it gives output voltage expression, uo = + + 3 4 3 4 3 sw o u pi cos i e j (o t +o ) 4 u sw o ni e j (2i t +o t +i +o +nui ) 2 u sw o ni e j (o t 2i t i +o nui ) 2

(33)

So, output voltage has fo , 2fi + fo

and 2fi fo

frequency components. Output current will have components same as output voltage. So, expression for output current space vector is,

For this analysis, it is assumed that input voltages at the rectifier terminals are unbalanced and load connected at the inverter terminals is balanced. Input voltage space vector can be written as, 2 ui = [ua + ub + 2uc ] 3
ui = u pi e j (i t ) + uni e j (i t +nui )

io = + +

u pi 3 sw o cos i e j (o t +o +o1 ) Z o1 4 (34)

uni 3 sw o e j (2i t +o t +i +o +nui +o 2 ) 4 2 Zo 2 uni 3 sw o e j (2i t o t +i o +nui +o 3 ) 4 2 Zo 3

(25)

In terms of +ve and -ve sequence components, (26)

From instantaneous power balance of inverter output and DC-link power,

Fourth National Power Electronics Conference, IIT Roorkee, June - 2010


Po = 3 Re uo io* = ui 2

(35) (36)

3 u Re sw o e j (o t +o ) io* = ui 2 2

Substituting (34) in (36) and simplifying, the expression for DC-link current is obtained as, i= + + 9 sw o 16
2

u pi Zo 1

cos i cos o1 (37)

uni 9 2 sw o cos(2i t + i + nui + o 2 ) 16 2 Zo 2 uni 9 2 sw o cos(2i t + i + nui + o 3 ) 16 2 Zo 3

Substituting in Equation (28) and simplifying, the expression for input current space vector is, ii = + + + 9 sw o 16
2

u pi Zo 1

cos i cos o1e j (i t +i )

uni 9 2 sw o e j (3i t + 2i +nui +o 2 ) 16 4 Zo 2 uni 9 2 sw o e j (i t +nui +o 2 ) 16 4 Zo 2 uni 9 2 sw o e j (3i t + 2i +nui +o 3 ) 16 4 Zo 3 (38)

Figure 5. Frequency Spectrum of Output Current: (a) Without Unbalance Compensation and (b) With Unbalance Compensation Effects of Unbalanced Output Currents

uni 9 2 sw o e j (i t +nui +o 3 ) + 16 4 Zo 3 This shows input current has frequency components of fi and 3fi . So, with unbalanced input voltages, DC-link has second harmonic and input current has harmonic component thrice the input frequency [8]. In above analysis, it is assumed that the rectifier and inverter operate with balanced switching function. However, balanced output voltages can be produced by using unbalanced inverter switching function. For this, an average DC-link voltage u is calculated for each pulse period. At the start of each pulse period the modulation index is calculated for that pulse period based on u of previous pulse period. This variable modulation index control strategy eliminates 2fi + fo and 2fi fo frequency components in output current as shown in Fig. 5 and 3fi component in input current as shown in Fig. 6.

Inverter section of IMC may have to produce unbalanced currents if unbalanced output voltages are applied to balanced loads or balanced output voltages are applied to unbalanced loads. So, output current space vector consists of +ve and -ve sequence components [9].
io =

2 [ i A + i B + 2 iC ] 3
j ( o t +po )

(39) (40)

io = i po e

+ i no e j (o t +no )

The instantaneous output and DC-link powers should be equal.


Po =

3 Re uo io* = ui 2

(41)

The inverter output voltage can be represented in terms of DC-link voltage and inverter switching function.

Fourth National Power Electronics Conference, IIT Roorkee, June - 2010


This shows that IMC input currents have frequency components fi , 2fo + fi and 2fo fi .
CONCLUSION

IMC could be an alternative for AC-DC-AC conversion. It has bidirectional power flow capability. Analysis of IMC is carried out for unbalanced input voltages and unbalanced output currents. Effects of unbalanced input voltages on output and input currents are eliminated by dynamically varying modulation index of inverter section.
REFERENCES

[1] (2001). Life-Limiting Factors in Electrolytic Capacitors (and how to improve them) [Online]. Available: http://www.evoxrifa.com/technote\_pdf/electrolyti c\_life\_factors.pdf [2] L. Wei and T. A. Lipo, A Novel Matrix Converter Topology with Simple Commutation, in Conf. Rec. 2001 IEEE, Industrial Applications Conference, 2001, 36th IAS Annu. Meeting, vol. 3, pp. 1749 - 1754. [3] J. W. Kolar, F. Schafmeister, S. D. Round and H. Ertl, Novel Three-Phase AC-AC Sparse Matrix Converters, IEEE Trans. Power Electronics, vol. 22, pp. 1649 - 1661, Sept. 2007. [4] F. Poitiers, T. Bouaouiche and M. Machmoum, Advanced control of a doubly-fed induction generator for wind energy conversion, ScienceDirect Electrical Power System Research, vol. 79, pp. 1085-1096, July 2009. [5] J. W. Kolar, T. Friedli, F. Krismer and S. D. Round, The Essence of Three-Phase AC/AC Converter Systems, in 2008 Power Electronics and Motion Control Conf., pp. 27-42, Sept. 2008. [6] L. Wei, T. A. Lipo and H. Chan, Matrix Converter Topologies With Reduced Number of Switches, in 2002 Power Electronics Specialists Conf., vol. 1, pp. 57-63, June 2002. [7] R. N. Nayak. (2008, Nov. 25). Power Quality Issues and Grid Interconnection Code: An Indian case study [Online]. Available: http://windindia08.wisein.org/wind-india-08presentations.htm

Figure 6. Frequency Spectrum of Input Current: (a) Without Unbalance Compensation and (b) With Unbalance Compensation
uo = sw o e j (o t +o ) u 2

(42)

Substituting (40) and (42) in (41) and simplifying, gives the expression for DC-link current as,
i= 3 sw o i po cos(o po ) 4

3 + sw o i no cos(2o t + o + no ) 4

(43)

The input current is defined with the help of rectifier switching function and DC-link current.
ii = sw i e j (i t +i ) i

(44)

Substituting and simplifying gives expression for rectifier input current as, ii = + + 3 4 3 4 3 sw o i po cos(o po )e j (i t +i ) 4 i sw o no e j (2o t +i t +i +o +no ) 2 i sw o no e j (2o t i t i +o +no ) 2

(45)

Fourth National Power Electronics Conference, IIT Roorkee, June - 2010


[8] P. N. Enjeti and S. A. Choudhury, A New Control Strategy to Improve the Performance of a PWM AC to DC Converter under Unbalanced Operating Conditions, IEEE Trans. Power Electronics, vol. 8, pp. 493-500, Oct. 1993. [9] M. Jussila and H. Tuusa, Illustration of Relation between Load and Supply Current Distortion in Direct and Indirect Matrix Converters, in Power Electronics Specialist Conf., pp. 2522-2528, June 2007. [10] The Central Electricity Authority (Technical Standards for Connectivity to the Grid) Regulations, 2007, 2/X/STD(CONN)/GM/CEA, Feb. 2007. Vishal Y. Vekhande was born in Mumbai, Maharashtra, India in 1983. He received the B.E. degree in electrical engineering from the VJTI, Mumbai, India, in 2005. After graduation he worked with Siemens Ltd. for two years as an Executive engineer. He is currently working towards the M.Tech. degree in Power Electronics and Power Systems at Indian Institute of Technology Bombay, Mumbai, India. His current research interests include applications of Matrix converter and DFIG for wind power generation. B. B. Pimple was born in Buldana, India in 1973. He received the B.E. degree in electrical engineering from Government College of Engineering , Amaravati, India in 1996 and the M. Tech. degree from the Visvesvaraya Regional College of Engineering, Nagpur, India in 1998. He joined as a lecturer at Sardar Patel College of Engineering Mumbai in 1998 and presently he is an Assistant Professor. He is currently working towards the Ph.D. degree in the Department of Electrical Engineering, Indian Institute of Technology Bombay. His current research interests are control of doubly-fed induction generator for wind power generation and matrix converter. B. G. Fernandes received the B. Tech. degree from Mysore University, Mysore, India, in 1984, the M. Tech. degree from the Indian Institute of Technology, Kharagpur, India, in 1989, and the Ph.D. degree from the Indian Institute of Technology Bombay, Mumbai, India, in 1993. He was with the Department of Electrical Engineering, Indian Institute of Technology, Kanpur, India, as an Assistant Professor. Since 1997, he has been with the Department of Electrical Engineering, Indian Institute of Technology Bombay, where he is currently a Professor. His current research interests are permanent-magnet machines, high-performance ac drives, quasi-resonant link converter topologies, and power electronic interfaces for nonconventional energy sources.

You might also like