A Generalised Harmonic Linearisation Method for Power Converters

Input/Output Impedance Calculation

Valerio Salis, Alessandro Costabeber, Pericle Stephen Cox

Zanchetta School of Mathematical Sciences
Power Electronics, Machines and Control University of Nottingham
Group University Park, NG7 2RD
University of Nottingham Nottingham, United Kingdom
University Park, NG7 2RD
Nottingham, United Kingdom
Tel.: +44 / (0) – 115 9515592.
E-Mail: [email protected]
URL: https://www.nottingham.ac.uk

Keywords
«Frequency-Domain Analysis», « Impedance measurement», « Impedance-based interactions », «
Modelling», « Harmonics».

Abstract
A general method for the evaluation of small-signal input impedances of power converters is presented in
this paper. The method is based on harmonic linearisation and Fourier analysis techniques and a single-
phase boost PFC is considered as case study. The analysis is based on the average model of the converter
but no further approximations have been made and both voltage and current loops are included. A
comparison between time domain and numerical simulation is presented to validate the proposed model.

Introduction
DC and AC distributed power systems are complex and highly non-linear because both the converters and
the loads interconnected often show a non-linear behaviour. Nevertheless they cover a wide range of
applications [1] and a theory for analysis and design of these networks is required. One of the most used
power system analysis techniques is the impedance-based approach. This frequency-domain method
requires that each load and source is described by its input and output impedance. For the simplest case
where a single load is connected to a single source, a transfer function from input to output voltage can be
defined: ܸ௢௨௧ ሺ‫ݏ‬ሻ ൌ ܼ௅ை஺஽ ሺ‫ݏ‬ሻܸ௜௡ ሺ‫ݏ‬ሻΤሺ ܼௌை௎ோ஼ா ሺ‫ݏ‬ሻ ൅ ܼ௅ை஺஽ ሺ‫ݏ‬ሻሻ ൌ ܸ௜௡ ሺ‫ݏ‬ሻΤሺ ͳ ൅ ܼௌை௎ோ஼ா ሺ‫ݏ‬ሻΤܼ௅ை஺஽ ሺ‫ݏ‬ሻሻ.
This function requires the knowledge of the impedances of source and load. It is now possible to analyse
the stability of the system by applying the Nyquist Stability Criterion to the ratio ܼௌை௎ோ஼ா ሺ‫ݏ‬ሻΤܼ௅ை஺஽ ሺ‫ݏ‬ሻ
[6],[7],[9].
For DC distributed power systems the standard linearisation techniques can be applied because of the DC
operating point and from the linearised model the impedances can be obtained. But for AC systems, since
there is a steady state time-periodic trajectory, new linearisation techniques are required. If an AC system
is a balanced three phase system, it can be linearised in the d-q domain [11], but this is no longer true for
unbalanced or single phase systems. In this paper Harmonic Linearisation will be used in order to obtain
the input impedance of the analysed converter. In [2] and [3] the technique has been used to derive the
input impedance of a single and three-phase line-frequency rectifier. In [4] the single-phase boost PFC
converter has been analysed considering only the current loop (model valid only above the crossover
frequency of the output voltage loop and below half of the switching frequency) and in [5] a model for the
range of frequency below the crossover has been developed.

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on February 25,2021 at 07:27:40 UTC from IEEE Xplore. Restrictions apply.
 This paper proposes a general method that can be applied to the average model of any AC-DC converter
that is valid for frequencies from DC to the limit of validity of the average model, i.e. §1/3 of the
switching frequency. Such a method will be used to analyse a single-phase boost PFC converter with
resistive load.

PFC model
Figure 1 depicts a single-phase boost PFC converter with standard two-loop control. The inner loop
controls the inductor current and the outer loop regulates the output voltage. The reference current,
݅௥௘௙ ሺ‫ݐ‬ሻ, is given by a multiplier with three inputs: the rectified sinusoidal input voltage, ȁܸ௜௡ ሺ‫ݐ‬ሻȁ, the
voltage generated by the outer control loop, ܸ௔ ሺ‫ݐ‬ሻ, the input feedforward voltage signal, ‫ݒ‬௙௙ ሺ‫ݐ‬ሻ, which in
our analysis we can consider as a constant signal, ‫ݒ‬௙௙ . This multiplier has a gain given by ݇௠ . The load of
the converter will be assumed to be pure resistive. The current and voltage regulators are defined as
following:

௦ ௦
݇௖ ቀͳ ൅ ቁ ݇௩ ቀͳ ൅ ቁ
௪೥೎ ௪೥ೡ
‫ܪ‬௖ ሺ‫ݏ‬ሻ ൌ Ǣ‫ܪ‬௩ ሺ‫ݏ‬ሻ ൌ ሺͳሻ
௦ ௦
‫ ݏ‬൬ͳ ൅ ൰ ‫ ݏ‬൬ͳ ൅ ൰
௪೛೎ ௪೛ೡ

il(t) Io(t)
L
Iin(t)
d(t)
+

Load
Vin(t) |Vin(t)| Co Vo(t)

current
sensor

il(t)
- Vc(t) d(t)
Zin(s) Hc(s) PWM
+
Iref(t) |Vin(t)|
km|Vin(t)|Va(t) Va(t) Hv(s) -

Vff^2 +
Vff^2
Vref
Figure 1: Single-phase boost PFC converter with an inner current loop and an outer voltage loop.

ܸଵ ൌ ͳͳͷξʹܸ Input voltage ܸ௠ ൌ Ͷ PWM gain

݂ଵ ൌ ͷͲ‫ݖܪ‬ Input frequency ݇௖ ൌ ͳ͹ͲͲ͹
Current regulator
‫ݒ‬௙௙ ൌ ܸଵ Τξʹ Feedforward voltage ‫ݓ‬௭௖ ൌ ͸ͻͷͶ
parameters
‫ ܮ‬ൌ ͳ݉‫ܪ‬ Inductor ‫ݓ‬௣௖ ൌ ͵͹ͺͲͺ
‫ܥ‬௢ ൌ ʹͲͲߤ‫ܨ‬ Capacitor ݇௩ ൌ ͲǤʹͻ
ܸ௥௘௙ ൌ ͵ͺͷܸ
Voltage regulator
Reference voltage ‫ݓ‬௭௩ ൌ ͵ǤͲ͵
parameters
ܴ ൌ ͳͶͺʹߗ Load resistor ‫ݓ‬௣௩ ൌ ͸ͻǤ͸ͻ
݂௦௪ ൌ ͳͲͲ‫ݖܪܭ‬ Switching frequency ݇௠ ൌ ͸Ͷ Multiplier gain

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on February 25,2021 at 07:27:40 UTC from IEEE Xplore. Restrictions apply.
 Table 1: Converter parameters

This system can be fully described by the following set of equations:

݀‫ܫ‬௅ ሺ‫ݐ‬ሻ ܸ஼ ሺ‫ݐ‬ሻ

‫ܮ‬ ൌ ȁܸ௜௡ ሺ‫ݐ‬ሻȁ െ ቈͳ െ ቉ ܸ௢ ሺ‫ݐ‬ሻሺʹሻ
݀‫ݐ‬ ܸ௠
ܸ݀௢ ሺ‫ݐ‬ሻ ܸ஼ ሺ‫ݐ‬ሻ ܸ௢ ሺ‫ݐ‬ሻ
‫ܥ‬ ൌ ቈͳ െ ቉ ‫ܫ‬௅ ሺ‫ݐ‬ሻ െ  ሺ͵ሻ
݀‫ݐ‬ ܸ௠ ܴ
ͳ ݀ ଶ ܸ஼ ሺ‫ݐ‬ሻ ܸ݀஼ ሺ‫ݐ‬ሻ
൅
‫ݓ‬௣௖ ݀‫ ݐ‬ଶ ݀‫ݐ‬
݇௖ ݇௠ ȁܸ௜௡ ሺ‫ݐ‬ሻȁܸ௔ ሺ‫ݐ‬ሻ ݇௖ ݇௠ ݀ȁܸ௜௡ ሺ‫ݐ‬ሻȁ ܸ݀௔ ሺ‫ݐ‬ሻ
ൌ ଶ ൅ ଶ ቈ ܸ௔ ሺ‫ݐ‬ሻ ൅ ȁܸ௜௡ ሺ‫ݐ‬ሻȁ ቉ െ ݇௖ ‫ܫ‬௅ ሺ‫ݐ‬ሻ
‫ݒ‬௙௙ ‫ݓ‬௭௖ ‫ݒ‬௙௙ ݀‫ݐ‬ ݀‫ݐ‬
݇௖ ݀‫ܫ‬௅ ሺ‫ݐ‬ሻ
െ ሺͶሻ
‫ݓ‬௭௖ ݀‫ݐ‬
ͳ ݀ ଶ ܸ௔ ሺ‫ݐ‬ሻ ܸ݀௔ ሺ‫ݐ‬ሻ ݇௩ ܸ݀௢ ሺ‫ݐ‬ሻ
ଶ
൅ ൌ ݇௩ ܸ௥௘௙ െ ݇௩ ܸ௢ ሺ‫ݐ‬ሻ െ ሺͷሻ
‫ݓ‬௣௩ ݀‫ݐ‬ ݀‫ݐ‬ ‫ݓ‬௭௩ ݀‫ݐ‬

This is a system of four equations in the four variables ‫ܫ‬௅ ሺ‫ݐ‬ሻ, ܸ௢ ሺ‫ݐ‬ሻ, ܸ஼ ሺ‫ݐ‬ሻ and ܸ௔ ሺ‫ݐ‬ሻ. In order to evaluate
the small-signal input-impedance we apply the harmonic balance techniques. Hence we add a small signal
perturbation to the input voltage:

ܸ௜௡ ሺ‫ݐ‬ሻ ൌ ܸଵ ‘•ሺ‫ݓ‬ଵ ‫ݐ‬ሻ ൅ ܸ௣ ‘•൫‫ݓ‬௣ ‫ݐ‬൯ , ܸ௣ ‫ܸ ا‬ଵ . ሺ͸ሻ

In this way we are not altering the behaviour of the system, since this perturbation is negligible compared
to the main signal. The injection of such a signal implies the generation of several harmonics in the input
current, with negligible magnitudes compared to the main input current, at frequencies ݂௣ ൅ ݂݊ଵ , ݊ ൌ
Ͳǡ േͳǡ േʹ ǥ By the evaluation of the input current at the perturbation frequency, ݂௣ , is then possible to
calculate the small-signal impedance as the ratio of voltage and current components at this perturbation
frequency.

From mathematical analysis we know that the solutions of the system, under the above assumption for
ܸ௜௡ ሺ‫ݐ‬ሻ, are given by:

ାஶ ାஶ
௝ଶ௡௪భ ௧ ା ௝ቀሺଶ௡ାଵሻ௪భ ା௪೛ ቁ௧ ି ௝ቀሺଶ௡ାଵሻ௪భ ି௪೛ ቁ௧
‫ܫ‬௅ ሺ‫ݐ‬ሻ ൌ ෍ ߙ௡ ݁ ൅ ෍ ൤ߙ௣௡ ݁ ൅ ߙ௣௡ ݁ ൨ ሺ͹ሻ
௡ୀିஶ ௡ୀିஶ
ାஶ ାஶ
௝ቀሺଶ௡ାଵሻ௪భ ା௪೛ ቁ௧ ௝ቀሺଶ௡ାଵሻ௪భ ି௪೛ ቁ௧
ା
ܸ௢ ሺ‫ݐ‬ሻ ൌ ෍ ߚ௡ ݁ ௝ଶ௡௪భ ௧ ൅ ෍ ൤ߚ௣௡ ݁ ି
൅ ߚ௣௡ ݁ ൨ ሺͺሻ
௡ୀିஶ ௡ୀିஶ
ାஶ ାஶ
௝ቀሺଶ௡ାଵሻ௪భ ା௪೛ ቁ௧ ௝ቀሺଶ௡ାଵሻ௪భ ି௪೛ ቁ௧
ା
ܸ஼ ሺ‫ݐ‬ሻ ൌ ෍ ߛ௡ ݁ ௝ଶ௡௪భ ௧ ൅ ෍ ൤ߛ௣௡ ݁ ି
൅ ߛ௣௡ ݁ ൨ ሺͻሻ
௡ୀିஶ ௡ୀିஶ
ାஶ ାஶ
௝ቀሺଶ௡ାଵሻ௪భ ା௪೛ ቁ௧ ௝ቀሺଶ௡ାଵሻ௪భ ି௪೛ ቁ௧
ା
ܸ௔ ሺ‫ݐ‬ሻ ൌ ෍ ߜ௡ ݁ ௝ଶ௡௪భ ௧ ൅ ෍ ൤ߜ௣௡ ݁ ି
൅ ߜ௣௡ ݁ ൨ ሺͳͲሻ
௡ୀିஶ ௡ୀିஶ

When we substitute these expressions in the equation system (2)-(5), we find three kind of terms: zero
order, first order (proportional to ܸ௣ ) and second order terms (proportional to ܸ௣ ଶ ) . It is then possible to

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on February 25,2021 at 07:27:40 UTC from IEEE Xplore. Restrictions apply.
 split the original system into two subsystems: a zero order system and a first order system. As usual in
small signal linearisation, the contribution of the second order system can be ignored. Solving the zero and
first order systems allows us to evaluate the inductor current which will be shown later. Before starting the
evaluation of the zero and first order systems it is useful to calculate explicitly the Fourier expression of
ȁܸ௜௡ ሺ‫ݐ‬ሻȁ, taking into account the small perturbation component at ݂௣ :

ାஶ
ʹܸଵ ሺെͳሻ௡ ݁ ௝ଶ௡௪భ ௧
ȁܸ௜௡ ሺ‫ݐ‬ሻȁ ൌ ෍
ߨ ሺͳ െ Ͷ݊ଶ ሻ
௡ୀିஶ
ାஶ
ܸ௣ ሺെͳሻ௡ ௝ቀሺଶ௡ାଵሻ௪భ ା௪೛ ቁ௧ା௝ఝ೛ ௝ቀሺଶ௡ାଵሻ௪భ ି௪೛ ቁ௧ି௝ఝ೛
൅ ෍ ൤݁ ൅݁ ൨ሺͳͳሻ
ߨ ሺʹ݊ ൅ ͳሻ
௡ୀିஶ

Zero order system

The zero order system includes all the main contributions and it is described by the following equations:

ାஶ
ʹܸଵ ሺെͳሻ௡ ͳ
݆ʹ݊‫ݓ‬ଵ ‫ߙܮ‬௡ െ ൅ ߚ௡ െ ෍ ߚ௠ ߛ௡ି௠ ൌ Ͳ
ߨሺͳ െ Ͷ݊ ሻ
ଶ ܸ௠
௠ୀିஶ
ାஶ
ͳ ߚ௡
݆ʹ݊‫ݓ‬ଵ ‫ߚܥ‬௡ െ ߙ௡ ൅ ෍ ߙ௠ ߛ௡ି௠ ൅ ൌͲ
ܸ௠ ܴ
௠ୀିஶ
ାஶ
Ͷ݊ଶ ‫ݓ‬ଵଶ ߛ௡ ݇௖ ݇௠ ʹܸଵ ሺെͳሻ௠
െ ൅ ݆ʹ݊‫ݓ‬ଵ ߛ௡ െ ଶ ෍ ߜ
‫ݓ‬௣௖ ‫ݒ‬௙௙ ߨሺͳ െ Ͷ݉ଶ ሻ ௡ି௠
௠ୀିஶ
ାஶ ାஶ
݇௖ ݇௠ ʹܸଵ ሺെͳሻ௠ ݆ʹ݉‫ݓ‬ଵ ʹܸଵ ሺെͳሻ௠ ݆ʹሺ݊ െ ݉ሻ‫ݓ‬ଵ
െ ଶ ൥ ෍ ߜ ௡ି௠ ൅ ෍ ߜ௡ି௠ ൩
‫ݓ‬௭௖ ‫ݒ‬௙௙ ߨሺͳ െ Ͷ݉ଶ ሻ ߨሺͳ െ Ͷ݉ଶ ሻ
௠ୀିஶ ௠ୀିஶ
݇௖ ݆ʹ݊‫ݓ‬ଵ ߙ௡
൅ ݇௖ ߙ ௡ ൅ ൌͲ
‫ݓ‬௭௖
Ͷ݊ଶ ‫ݓ‬ଵଶ ߜ௡ ݇௩ ݆ʹ݊‫ݓ‬ଵ ߚ௡
െ ൅ ݆ʹ݊‫ݓ‬ଵ ߜ௡ െ ݇௩ ܸ௥௘௙ ‫ݖ‬ሺ݊ሻ ൅ ݇௩ ߚ௡ ൅ ൌ Ͳሺͳʹሻ
‫ݓ‬௣௩ ‫ݓ‬௭௩

with ‫ݖ‬ሺ݊ሻ ൌ ͳ for ݊ ൌ Ͳ and ‫ݖ‬ሺ݊ሻ ൌ Ͳ for ݊ ് Ͳ. This system of equations (12), with unknown quantities
ߙ௡ , ߚ௡ , ߛ௡ and ߜ௡ , can be solved numerically in Matlab using fsolve.

First order system

The first order system includes all the perturbation contributions and it is described by the following
equations:

ାஶ ାஶ
ା
ܸ௣ ሺെͳሻ௡ ା
ͳ ା ା
݆ ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ‫ߙܮ‬௣௡ െ ൅ ߚ௣௡ െ ൥ ෍ ߚ௠ ߛ௣௡ି௠ ൅ ෍ ߚ௣௠ ߛ௡ି௠ ൩ ൌ Ͳ
ߨሺʹ݊ ൅ ͳሻ ܸ௠
௠ୀିஶ ௠ୀିஶ
ାஶ ା ାஶ
ା ା
ͳ ା ା
ߚ௣௡
݆ ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ‫ߚܥ‬௣௡ െ ߙ௣௡ ൅ ൥ ෍ ߙ௠ ߛ௣௡ି௠ ൅ ෍ ߙ௣௠ ߛ௡ି௠ ൩ ൅ ൌͲ
ܸ௠ ܴ
௠ୀିஶ ௠ୀିஶ

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on February 25,2021 at 07:27:40 UTC from IEEE Xplore. Restrictions apply.
 ଶ
ା
ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ߛ௣௡
ା
െ ൅ ݆ ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ߛ௣௡
‫ݓ‬௣௖
ାஶ ାஶ
݇௖ ݇௠ ܸ௣ ሺെͳሻ௠ ʹܸଵ ሺെͳሻ௠ ା
െ ଶ ൥ ෍ ߜ௡ି௠ ൅ ෍ ߜ ൩
‫ݒ‬௙௙ ߨሺʹ݉ ൅ ͳሻ ߨሺͳ െ Ͷ݉ଶ ሻ ௣௡ି௠
௠ୀିஶ ௠ୀିஶ
ାஶ
݇௖ ݇௠ ݆ ቀሺʹ݉ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ܸ௣ ሺെͳሻ௠
െ ଶ ቎ ෍ ߜ௡ି௠
‫ݓ‬௭௖ ‫ݒ‬௙௙ ߨሺʹ݉ ൅ ͳሻ
௠ୀିஶ
ାஶ ାஶ
݆ʹ݉‫ݓ‬ଵ ʹܸଵ ሺെͳሻ௠ ା ݆ʹሺ݊ െ ݉ሻ‫ݓ‬ଵ ܸ௣ ሺെͳሻ௠
൅ ෍ ߜ௣௡ି௠ ൅ ෍ ߜ௡ି௠
ߨሺͳ െ Ͷ݉ଶ ሻ ߨሺʹ݉ ൅ ͳሻ
௠ୀିஶ ௠ୀିஶ
ାஶ
݆ ቀሺʹሺ݊ െ ݉ሻ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ʹܸଵ ሺെͳሻ௠
ା ା
൅ ෍ ߜ௣௡ି௠ ቏ ൅ ݇௖ ߙ௣௡
ߨሺͳ െ Ͷ݉ଶ ሻ
௠ୀିஶ
ା
݇௖ ݆ ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ߙ௣௡
൅ ൌͲ
‫ݓ‬௭௖
ଶ
ା
ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ߜ௣௡ ା
݇௩ ݆ ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ߚ௣௡
ା ା
െ ൅ ݆ ቀሺʹ݊ ൅ ͳሻ‫ݓ‬ଵ ൅ ‫ݓ‬௣ ቁ ߜ௣௡ ൅ ݇௩ ߚ௣௡ ൅
‫ݓ‬௣௩ ‫ݓ‬௭௩
ൌ Ͳሺͳ͵ሻ
ା ା ା ା
The solutions of this system of equations (13), ߙ௣௡ , ߚ௣௡ , ߛ௣௡ and ߜ௣௡ , can be found again using the
command fsolve in Matlab. It is worth nothing that we are considering only the positive perturbation part,
since the negative part is simply the complex conjugate of the positive one. This allows us to simplify
numerical solution.

Small-signal input impedance

Solving the zero and first order systems provides an analytical expression for the inductor current. Such
current is related to the input current by the following formula:

‫ܫ‬௜௡ ሺ‫ݐ‬ሻ ൌ ‫݊݃ݏ‬൫ܸ௜௡ ሺ‫ݐ‬ሻ൯‫ܫ‬௅ ሺ‫ݐ‬ሻǡሺͳͶሻ

where the sign function ‫݊݃ݏ‬൫ܸ௜௡ ሺ‫ݐ‬ሻ൯ ൌ ͳ when ܸ௜௡ ሺ‫ݐ‬ሻ ൒ Ͳ and ‫݊݃ݏ‬൫ܸ௜௡ ሺ‫ݐ‬ሻ൯ ൌ െͳ when ܸ௜௡ ሺ‫ݐ‬ሻ ൏ Ͳ. It
can be expressed as a Fourier Series as:

ାஶ ାஶ
ܸ௣ ʹሺെͳሻ௡ ݁ ௝ሺଶ௡ାଵሻ௪భ ௧
‫݊݃ݏ‬൫ܸ௜௡ ሺ‫ݐ‬ሻ൯ ൌ ෍ ሺെͳሻ௡ ݁ ௝൫ଶ௡௪భ ା௪೛ ൯௧ ൅ ෍ ሺͳͷሻ
ߨܸଵ ߨሺʹ݊ ൅ ͳሻ
௡ୀିஶ ௡ୀିஶ

Evaluating the product in (14) and considering only the perturbation components gives us the part of the
input current that is function of the perturbation input voltage:

ାஶ ାஶ
ା௪೛ ܸ௣ ௝௪ ௧ ା
ʹሺെͳሻ௡ ߙ௣௡
‫ܫ‬௜௡ ൌ ݁ ೛ ෍ ሺെͳሻ௡ ߙ௡ ൅ ܸ௣ ݁ ௝௪೛ ௧ ෍ ሺͳ͸ሻ
ߨܸଵ ܸ௣ ߨሺʹ݊ ൅ ͳሻ
௡ୀିஶ ௡ୀିஶ

The input voltage at the perturbation frequency is given by:

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on February 25,2021 at 07:27:40 UTC from IEEE Xplore. Restrictions apply.
 ܸ௣ ௝௪ ௧
ܸ௜௡ ା௪೛ ൌ ݁ ೛ ሺͳ͹ሻ
ʹ

Eventually it is possible to evaluate the small-signal input impedance as the ratio between the input
voltage (17) and current (16) at the perturbation frequency:

ାஶ ାஶ ିଵ
ܸ௜௡ ା௪೛ ߨ ͳ ௡
ା
ʹሺെͳሻ௡ ߙ௣௡
ܼ௜௡ ൫݆‫ݓ‬௣ ൯ ൌ ା௪೛ ൌ ൥ ෍ ሺെͳሻ ߙ௡ ൅ ෍ ൩ ሺͳͺሻ
‫ܫ‬௜௡ ʹ ܸଵ ܸ௣ ߨሺʹ݊ ൅ ͳሻ
௡ୀିஶ ௡ୀିஶ

This formula gives the small-signal input impedance as magnitude and phase for a particular value of
frequency, i.e. ݂௣ . It is then necessary to make an evaluation for a set of values of the perturbation
frequency in order to obtain a full description for the frequency range of interest. It is worth to remark that
this model is valid from the low frequency up to one third of the switching frequency of the circuit, due to
the limitation introduced by the average model.

Simulation results
In this section we present simulation results in order to validate the proposed model (18). The small-signal
input impedance of the converter is first evaluated through a time domain simulation. First a set of values
for the perturbation frequency is chosen. Then for each of these values a time domain simulation is
performed and the FFT is applied to the recorded input current vector in order to obtain the component at
the perturbation frequency. Then the ratio of the voltage to the current components at perturbation
frequencies gives the small-signal input impedance. This is a typical approach that can be used when no
models are available. To validate the model, (18) has been numerically simulated for the same set of
perturbation frequencies. In both cases we set ܸ௣ ൌ ͳܸ. Figure 2 shows the comparison between the two
results. In Figure 2a the model has been implemented with a number of harmonics for (7)-(10) equals to
݊ ൌ ͷ; in Figure 2b we used ݊ ൌ ͻ. A really good agreement is achieved, which gives us the validation of
the proposed model.

Conclusions
In this paper a general method based on harmonic linearisation and Fourier analysis techniques to evaluate
small-signal input impedance of power converters is proposed. A case study of a single-phase boost PFC
rectifier has been used to show the practical application of the method. A numerical simulation has been
performed to validate the proposed model. A detailed description of the relevant voltages and currents of
the converter is given. So, not only the harmonics at the perturbation frequency, but a complete spectrum
of these signals is given. This will be a starting point in order to define a new stability–analysis approach,
which will consider not only the components at the perturbation frequency but will consider the full
spectrum of signals, leading to the concept of harmonic transfer functions.

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on February 25,2021 at 07:27:40 UTC from IEEE Xplore. Restrictions apply.
 70
Magnitude [dB]

60

50

40

30
100 101 102 103 104 105

100

50
Phase [°]

-50

-100
100 101 102 103 104 105
Frequency [Hz]
(a) (b)
Figure 2: Small-signal input impedance of single-phase boost PFC converter: (a) ݊ ൌ ͷ; (b) ݊ ൌ ͻ;
(circles) impedance from time domain FFT calculations, (stars) impedance from formula (18)

References
[1] Jian Sun, "Small-Signal Methods for AC Distributed Power Systems–A Review," in Power Electronics,
IEEE Transactions on , vol.24, no.11, pp.2545-2554, Nov. 2009
[2] Jian Sun; J. Colon, "Input Impedance Modeling of Line-Frequency Rectifiers by the Method of Impedance
Mapping," in Computers in Power Electronics, 2006. COMPEL '06. IEEE Workshops on , vol., no., pp.69-
75, 16-19 July 2006
[3] Zhonghui Bing; K.J.Karimi; Jian Sun, "Input Impedance Modeling and Analysis of Line-Commutated
Rectifiers," in Power Electronics Specialists Conference, 2007. PESC 2007. IEEE , vol., no., pp.1981-1987,
17-21 June 2007
[4] Jian Sun; Zhonghui Bing, "Input impedance modeling of single-phase PFC by the method of harmonic
linearization," in Applied Power Electronics Conference and Exposition, 2008. APEC 2008. Twenty-Third
Annual IEEE , vol., no., pp.1188-1194, 24-28 Feb. 2008
[5] M. Chen; Jian Sun, "Low-Frequency Input Impedance Models for Boost Single-Phase PFC Converters,"
in Power Electronics Specialists Conference, 2005. PESC '05. IEEE 36th , vol., no., pp.1062-1068, 16-16
June 2005
[6] R.D. Middlebrook, "Input filter considerations in design and application of switching regulators", in Proc.
IEEE Industry Applications Soc. Annu. Meeting, 1976, pp. 366-382.
[7] P. Kundur, Power System Stability and Control. New York: McGraw-Hill, 1994.
[8] C.W. Taylor, Power System Voltage Stability. New York:McGraw-Hill, 1994.
[9] M. Belkhayat, "Stability criteria for AC power systems with regulated loads", Ph.D. dissertation, Purdue
University, Washington, DC, 1997.
[10] P.C. Todd, "UC3854 controlled power factor correction circuit design", Unitrode Application Note U-134,
2004.
[11] Bo Wen; D.Boroyevich; R. Burgos; P. Mattavelli.; Zhiyu Shen, "Analysis of D-Q Small-Signal Impedance
of Grid-Tied Inverters," in Power Electronics, IEEE Transactions on, vol.31, no.1, pp.675-687, Jan. 2016

Authorized licensed use limited to: King Mongkut's University of Technology North Bangkok. Downloaded on February 25,2021 at 07:27:40 UTC from IEEE Xplore. Restrictions apply.