JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

Appendix A.

THE DQ TRANSFORMATION

A.1. Definition

The constant-power dq transformation for a three-phase system is:

 cos(θ ) cos(θ − 2π ) cos(θ + 2π ) 

 r r
3
r
3 
2  2π 2π 
Tdq = − sin(θ r ) − sin(θ r − ) − sin(θ r + ) , (A.1)
3  3 3 
 1 1 1 
 2 2 2 

where θ r is the rotating coordinate angle defined as:


θ r = ω dt +θ 0 ,
0
(A.2)

ω : coordinate angular frequency.

θ 0 : initial position angle of the coordinate axes.

When ω is constant, such as in applications with direct connection to the utility,

this transformation can be expressed as follows

 cos(ω t + θ ) cos(ω t + θ − 2π ) cos(ω t + θ + 2π ) 

 0 0
3
0
3 
2  2π 2π 
Tdq = − sin(ω t + θ 0 ) − sin(ω t + θ 0 − ) − sin(ω t + θ 0 + ) . (A.3)
3  3 3 
 1 1 1 
 2 2 2 

APPENDIX A: THE DQ TRANSFORMATION Page 193

JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

A.2. Power Conservative Transformation

The dq transformation is an orthogonal transformation since it accomplishes the

−1
propriety Tdq = Tdq
T
. The internal product remains invariable under such kind of

transformation. Thus, given x1r = Tdq x1 and x 2r = Tdq x 2 , the internal product

< x1r , x 2r > is:

< x1r , x 2r > = < Tdq x1, Tdq x 2 > = < x1, Tdq
T
Tdq x 2 > = < x1, Tdq
-1
Tdq x 2 > = < x1, x 2 > . (A.4)

Similarly, the instantaneous power of a set of three-phase voltages and currents is:

i a 
 
p = v a ia + v b i b + v c i c = [v a v b v c ] i b  = v Tph i ph , (A.5)
i 
 c

and applying the dq transformation to the variables:

p = v Tph i ph = v Tph (Tdq

−1
Tdq ) i ph = (v Tph Tdq
−1
)(Tdqiph ) = (vTphTdqT )(Tdqiph ) = vTphr iphr . (A.6)

In addition, when the sum of the three currents is zero, the homopolar current
component (io) is also zero. Therefore, calculation of the instantaneous power is
simplified as follows:

i d 
 
p= v Tphr i phr = [v d vq v o ] i q  = v d i d + v q i q + v o i o = v d i d + v q i q . (A.7)
i 
 o

A.3. Transformation of a State-Space-Formulated System

A.3.1. General Application

A three-phase system in the state-space representation can be transformed into

the dq components as in the following. The standard state-space notation is:

d
x = A x + Bu , (A.8)
dt

and the relationship with the transformed variables is:

−1
x r = Tdq x and x = Tdq x r = Tdq
T
xr . (A.9)

Substituting x from (A.9) into (A.8):

APPENDIX A: THE DQ TRANSFORMATION Page 194

JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

d
x r = A r x r + B r ur , (A.10)
dt

in which:

−1 d −1 −1
A r = Tdq A Tdq − Tdq Tdq , and Br = TdqBTdq . (A.11)
dt

d −1
Solving for the term Tdq Tdq with the definition of Tdq given in (A.3):
dt

0 − ω 0
d −1 d T  
Tdq Tdq = Tdq Tdq = ω 0 0 . (A.12)
dt dt
0 0 0


The new matrices of the transformed system are:

 0 ω 0
 
Ar = T
Tdq ATdq + − ω 0 0 and Br = TdqBTdq
T
. (A.13)
 0 0 0
 

A.3.2. Application to the Three-Level System

The three-level system formulated in (2.22) is:

 R 0   1 0 0  − 1 
− 0  −  
i a   L  i a   L
 ea − v a 0   L 
d  
ib =  0 −
R
0  i  +  0 − 1 0  e − v  + − 1  v , (A.14)
dt     b   b b0  N0
L  L   L
ic   R ic   1 ec − v c 0   1 
0 0 − 0 0 − −
 L   L   L 

in which the matrices can be identified as follows:

− R 0 0  − 1 0 0 
i a   L  L
   
x = i b  ,
R 1
A= 0 − 0  , B1 = B2 =  0 − 0 ,
   L   L 
ic   0 R  0 1
0 − 0 −
 L  L 
(A.15)
ea − v a0  v N 0 
u1 = eb − v b0  , and u2 = v N 0 .
   
ec − v c 0  v N 0 

APPENDIX A: THE DQ TRANSFORMATION Page 195

JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

Applying (A.9) and (A.13), the vectors and matrices become:

− R ω 0  − 1 0 0 
i d   L  L
   
x r = iq  ,
R 1
Ar =  − ω − 0  , B1r = B2r =  0 − 0 ,
   L   L 
i o   0 R  1
0 − 0 0 −
 L   L 
(A.16)
ed − v d   0 
 
u1r = eq − v q , and u2r =  0  .
   
eo − v o   3 v N 0 

As a result, the new state-space representation of the system is as follows:

 R   1 
− ω 0  − 0 0   
i d   L   d  
i  L
 ed − v d   0 
d   R  1 e − v  +  0  v . (A.17)
iq =  − ω − 0  i q  +  0 − 0 
dt     q q
L   L   3  N0
 i o   R  i o   1 eo − v o  − 
0 0 − 0 0 −  L 
 L   L 

APPENDIX A: THE DQ TRANSFORMATION Page 196

JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

Appendix B.

THE TWO-DIMENTIONAL SPACE-VECTOR

TRANSFORMATION

B.1. Definition

The two-dimensional SV transformation is defined as follows:

2π 2π
& j −j
x = xa + x b e 3 + xc e 3 , (B.1)

in which the variables xa, xb and xc are a set of three-phase components, either
voltages or currents.

This transformation is sometimes multiplied by the coefficient 2/3 in order

normalize the length of the resulting vector. The coefficient 2 3 is also used.

B.2. Influence of Harmonics in the Vector

A balanced set of three-phase components with addition of a generic h order

harmonic and a common component f(t) is given as:

 x = Xˆ cos(θ + ϕ ) + Xˆ cos(hθ + ϕ ) + f (t ) ,
 a 1 1 h h

 2π 2π
 xb = Xˆ 1 cos(θ − + ϕ1) + Xˆ h cos 
h(θ − ) + ϕ h  + f (t ) , and (B.2)
 3  3 

 x = Xˆ cos(θ + 2π + ϕ ) + Xˆ cos h(θ + 2π ) + ϕ + f (t ) ,

c 1
3
1 h
 3
h


APPENDIX B: THE TWO-DIMENSIONAL SPACE-VECTOR TRANSFORMATION Page 197

JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

which can be also expressed as follows:

j (θ +ϕ 1 )
 ˆ e + e − j (θ +ϕ1 ) ˆ e j (hθ +ϕ h ) + e − j (hθ +ϕ h )
 a x = X 1 + Xh + f (t ) ,
 2 2
 j  θ −
2π
+ϕ 1  − j  θ −
2π
+ϕ 1  j 
h  θ −
2π  − j 
h θ −
2π 
 +ϕ h   +ϕ h 
 e  3 
+ e  3 
e   3  
+ e   3  
 bx = Xˆ + Xˆ + f (t ) , and
1 h
 2 2
 2π 2π 2π  2π 
j  θ + +ϕ 1  − j  θ + +ϕ1  j 
h  θ +
 +ϕ h  − j 
h  θ +
 +ϕ h 
  3 
+e  3    3  
+e   3  
 xc = Xˆ 1 e + Xˆ h
e
+ f (t ) .
 2 2

(B.3)

Applying (B.1) to this set of variables the following expression is obtained:

& & & & 3 Xˆ 1 j (θ +ϕ1 )

x = x1 + xh , with x1 = e and
2
(B.4)
& Xˆ 2π  e j (hθ +ϕ h ) + 1 + 2 cos (h + 1) 2π
xh = h  
1 + 2 cos (h − 1) 
 
 e − j ( h θ +ϕ h )  .
 
2   3   3  

It can be observed that the common component f(t) has disappeared in the
obtained vector. As a result, in the case of transforming voltages, the reference
potential of va, vb and vc does not affect the obtained vector. Therefore, the point
taken as a reference for the three variables can be changed with no effect on the
voltage-vector representation.

Triplen order harmonics, h={3, 6, 9, 12, …}, have neither influence in the vector,
&
since xh = 0 for these order of components.

Triplen-plus-one order harmonics, or h={4, 7, 10, 13, …}, produce direct

rotational vectors with the same frequency as the harmonic, such that:

& 3 Xˆ h j (hθ +ϕ h )
xh = e , (B.6)
2

while triplen-minus-one order harmonics, or h={2, 5, 8, 11, …}, produce reverse

rotational vectors, as follows:

& 3 Xˆ h − j (hθ +ϕh )

xh = e . (B.7)
2

APPENDIX B: THE TWO-DIMENSIONAL SPACE-VECTOR TRANSFORMATION Page 198

JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

B.3. Equivalences in the First Sextant

Assuming no distortion in the scalar components Xˆ h = 0 ∀ h ≠ 1, and ϕ1 = 0 , the

resulting vector is represented by:

2π 2π
& j −j 3 Xˆ 1 jθ
x = xa (θ ) + xb (θ ) e 3 + xc (θ ) e 3 = e . (B.8)
2

An equivalent vector in the first sextant can be found for any vector. These
equivalences are very useful for processing all of the calculations in the first sextant.
The relationships are described in Table B.1.

The equivalent vector in the first sextant can be obtained just interchanging the
scalar variables when applying the two-dimensional transformation. The rotational
direction of the equivalent vector depends on the number of variables that have been
interchanged; thus, if the reference vector is in the 3rd sextant or 5th sextant, the
equivalent vector in the first sextant is also a direct rotational vector. In contrast, for
2nd, 4th and 6th sextants, the equivalent vector is a reverse rotational vector.

The equivalences given in Table B.1 have the reflexive propriety. Thus,
interchanging the three-scalar components of the equivalent vector in the first sextant
in the opposite direction than in Table B.1, the original vector in the corresponding
sextant is obtained.

APPENDIX B: THE TWO-DIMENSIONAL SPACE-VECTOR TRANSFORMATION Page 199

JOSEP POU TECHNICAL UNIVERSITY OF CATALONIA

Table B.1. Equivalent vector in the first sextant

2nd Sextant
π 2π
xb

Reference Vector in the 2nd Sextant: ≤θ ≤

3 3 3rd Sextant 1st Sextant

x → xb
Translation from the 2nd Sext. into the 1st Sext.:  a
 xb → xa
xa

st
Equivalent Vector in the 1 Sextant:
2π 2π 2π
& j −j 3 Xˆ 1 j ( 3 −θ )
x1 = xb + xae 3 + xc e 3 = e 4th Sextant 6th Sextant

2 xc
5th Sextant

2π 2nd Sextant
≤θ ≤π
xb
Reference Vector in the 3rd Sextant:
3 1st Sextant
3rd Sextant
 xa → xc

Translation from the 3 Sext. into the 1 Sext.:  xb → xa
nd st
xa
 xc → xb
Equivalent Vector in the 1st Sextant:
2π 2π 2π
& j −j 3 Xˆ 1 j (θ − ) 4th Sextant 6th Sextant
x1 = xb + xc e 3 + xa e 3 = e 3
2 xc
5th Sextant
2nd Sextant
4π
xb

Reference Vector in the 4th Sextant: π ≤ θ ≤

3 3rd Sextant 1st Sextant

 x → xc
Translation from the 4nd Sext. into the 1st Sext.:  a
 xc → xa
xa

st
Equivalent Vector in the 1 Sextant
2π 2π 4π
& j −j 3 Xˆ 1 j ( −θ )
x1 = xc + xbe 3 + xa e 3 = e 3 4th Sextant 6th Sextant

2 xc
5th Sextant

4π 5π 2nd Sextant
≤θ ≤
xb
Reference Vector in the 5th Sextant:
3 3
3rd Sextant 1st Sextant
 xa → x b

Translation from the 5 Sext. into the 1 Sext.:  xb → xc
th st

 xc → xa
xa

Equivalent Vector in the 1st Sextant:

2π 2π 4π
& j −j 3 Xˆ 1 j (θ − 3 ) 4th Sextant 6th Sextant
x1 = xc + xae 3 + xbe 3 = e
2 xc
5th Sextant
2nd Sextant
4π
xb

Reference Vector in the 6th Sextant: ≤ θ ≤ 2π

3 3rd Sextant 1st Sextant

x → xc
Translation from the 6nd Sext. into the 1st Sext.:  b
 xc → x b
xa

st
Equivalent Vector in the 1 Sextant:
2π 2π
& j −j 3 Xˆ 1 j ( 2π −θ )
x1 = xa + xc e 3 + xbe 3 = e 4th Sextant 6th Sextant

2 xc
5th Sextant

APPENDIX B: THE TWO-DIMENSIONAL SPACE-VECTOR TRANSFORMATION Page 200