Introduction
The behaviour of a three-phase machine is usually described by its voltage and current equations:
Mathematically modelling these components tends to be complex as we must model each individual phase component as it varies with time. This can make undertaking analyses or building a machine control system difficult.
It is therefore beneficial to manipulate these equations into an alternative format that is easier to work with. To do this we go through a process of decoupling variables so that the behaviour of the three-phase machine can be fully described using a small number of constants and variables. For electrical machines two common transformations are used: The Clarke Transformation and the Park Transformation.
Note that these transformations can also then be reversed so that that if we control some aspect of machine behaviour in its simplified form, we can then transform this mathematically to directly control the machine parameters.
Unit Content
In this module, we shall work through the process of manipulating the time-variant, per phase components into the dq frame of reference through the following chapters:
-
Chapter 1: Specification of the per phase components individually in the time frame.
-
Chapter 2: Representation of the per phase components in the phase space (the abc reference frame).
-
Chapter 3: Transformation from the abc reference frame to the αβ reference frame.
-
Chapter 4: Transformation from the αβ reference frame to the dq reference frame.
-
Coursework: Transformation, control and inversion of the Clarke and Park transforms.
Revision of the Three-Phase Machine Model
This chapter presents a recap of the fundamental operating principle of the three-phase rotating electrical machine model.
As illustrated in the animation, the magnetic field of the rotor sweeps through the electric field of the stator within the core of the generator. This process induces a voltage in the stator coils and thus gives rise to a phase current.
In the three-phase machine model, the generator is designed to produce three individual phase voltages that are each offset by 120° ( i.e.: ).
Considering Phase A
Figure 1 presents the induced voltage for Phase A in the time frame.
Note that for simplicity this waveform has a voltage amplitude of 1 Volt, starts at time = 0 seconds and is an ideal 50 Hz sine wave. This waveform can be described by the following formula:
where:
Figure 1: Phase A Voltage (Time Frame)
Considering Phase B
Figure 2 includes the induced voltage for Phase B in the time frame.
Note that Vb is offset from Va by 2π/3 rads. This second waveform can be described by the following formula:
where:
Figure 2: Phase A & B Voltages (Time Frame)
Considering Phase C
Figure 3 includes the induced voltage for Phase C in the time frame.
Note that Vc is offset from Va by 4π/3 rads. This second waveform can be described by the following formula:
where:
Figure 3: Phase A, B & C Voltages (Time Frame)
In chapter 2, we shall see how Va, Vb and Vc can be represented in the phase space (the abc frame).
Introducing the ‘abc’ Reference Frame
In the previous chapter, we saw how Va, Vb and Vc would be typically represented in the time frame. At any given time each phase voltage can also be represented purely as a magnitude value on a single axis graph, i.e., at t = 3 ms as shown on Figure 4:
Figure 4: Phase A (Phase Space)
Representing 3 Phases
All phase voltages could therefore be illustrated in a vector space representation simultaneously if we rotate each axis by 120° (i.e.: 2π/3 rads) and set the zero crossing point of each axis at the centre of the abc frame.
In the following pages, examples of three phase voltage values at various times have been plotted in both the time frame and in the abc frame.
Figure 5: The abc Frame
Examples
Please click on the headings to expand the examples.
ωt = 0° = 0 rads
In the lower right illustration, the resultant vector (V ̅) is calculated through vector addition.
ωt = 90° = π/2 rads
ωt = 120° = 2π/3 rads
Summary
At any given time, the resultant vector (V ̅) is of constant magnitude and sweeps around the circle of the abc frame as illustrated in the animation.
Note that in this animation, time (t) is increasing.
Representation in abc frame allows us to easily visualise the vector of the resultant voltage.
However, this representation remains complex due to the requirement for any analyses to be done using three axes. In the next chapter we shall see how Va, Vb and Vc might be represented in a dual axis reference frame (the αβ frame).
Original image: Mathworks
Introducing the ‘αβ’ Reference Frame
In the previous chapter, we saw how Va, Vb and Vc could be represented in phase space (using the abc frame) and that the resultant vector () of each of the three phase voltages was of constant magnitude as it rotated around the frame. In this chapter, we shall define two new axes in that same space to simplify the definition of the vector ().
The Clarke Transformation achieves this by defining two new axes: alpha (α) and beta (β). We shall use arithmetic with complex numbers to reference the voltage vector to this new frame in phase space.
Note that in this chapter, students should be comfortable manipulating complex numbers using rectangular, polar or exponential form such that the following conversions can be assumed:
A given vector can be expressed as:
Rectangular Form:
Polar Form:
Exponential Form:
where r is the magnitude of the vector .
Further Reading
Further reading and revision can be found at: Khan Academy
Converting from ‘abc’ to ‘αβ’
In this figure, the new αβ frame is superimposed upon the abc frame. It should be noted that the centre point of the αβ frame is at the centre point of the abc frame and that for simplicity the alpha (α) axis is aligned with the a axis.
Any given vector can therefore be written as a vector sum of each of its [α and β] components or its [a, b and c] components (and these are equal) such that:
thus:
Figure 6: The αβ Frame
Combining these equations:
gives:
and expanding the right-hand side:
calculating our sine and cosine values:
If we set the arbitrary vectors ra, rb and rc to our voltages Va, Vb and Vc and collect real and imaginary components we can derive two equations.
Writing these two equations in matrix form, we have the Clarke Transformation:
Summary
This animation illustrates the behaviour of Vα and Vβ as they vary with time (t).
The Clarke Transform converts the individual components of a three-phase system (the abc frame) to two components defined in a new orthogonal stationary system (the αβ frame). The αβ frame representation is a simplification in that the behaviour of the three-phase system is now fully described using only two variables, Vα and Vβ each of which vary with time (t).
In the next chapter, we shall see how Vα and Vβ can be represented in a rotating dual-axes reference frame (the dq frame) such that the resultant components are constant rather than variable with regards to time.
Introducing the ‘dq’ Reference Frame
In the previous chapter, we used the Clarke Transform to derive Vα and Vβ such that the three-phase machine characteristics could be described using only two time dependent variables. In this chapter we shall define a rotating orthogonal reference frame such that the resultant components are constant with regards to time.
The Park transformation achieves this by defining yet another axis system: direct (d) and quadrature (q), which are at an arbitrary offset angle to the αβ frame and by rotating this axis at the synchronous speed of the electrical machine. In doing so the complex components of the resultant vector (V ̅) remain constant when referenced to the rotating frame.
Now we can say that the for any given vector in phase space, the reference frame is 'arbitrary', i.e., it does not matter that the alpha (α) axis of the αβ frame happened to be aligned with the a axis of the abc frame. The alignment of these frames was selected because it simplified the arithmetic and was visually appealing (in that the αβ frame appears to us like a typical xy frame). However, we could have referenced the resultant vector to any other frame (centred at zero) and defined the relationship between this new frame and the αβ frame by the angle rho (ρ) as illustrated:
As we have seen in the previous chapter that this xy frame relates to the αβ frame by the following relationship:
where:
and
Figure 7: An arbitrary offset reference frame
Converting from ‘αβ’ to ‘dq’
Expanding both sides of this equation:
and equating the right-hand sides:
finally if we expand the brackets and collect real and imaginary components:
If we set the arbitrary vector rα and rβ to our voltages Vα, and Vβ and equate the real and imaginary components we can derive two equations.
Writing these two equations in matrix form, we have:
In the Park transform, we let rho (ρ) be equal to the synchronous rotating speed of the machine (i.e., ρ = ωt) and name this rotating set of axes, the ‘dq’ frame as illustrated:
Figure 8: The dq Frame
Summary
As this animation illustrates, when we set rho (ρ) to the synchronous speed; Vd and Vq are constant with regards to time (t) as the frame itself is rotating.
The Park transform can therefore be expressed as:
where:
The Park transform therefore converts the two time-variant components in the αβ frame to an orthogonal rotating reference frame (dq) and in doing so simplifies computations by producing two time-invariant components.
Coursework Exercise
Students should now undertake the ‘Clarke & Park Transformation Coursework’.
In this exercise, students shall be required to implement the Clarke and Park transformation in MATLAB to assess the output of a small-scale generation site that is synchronised to the grid via an AC-AC power converter. Students shall then demonstrate how voltage control can be achieved by using the inverse Clarke and Park transformations.