IEEE TRANSACTIONS ON INSTRUMENTATIONAND MEASUREMENT, VOL. 54, NO. 4, AUGUST 2005 1471
An Easy VI Program to Detect Transient Disturbances
in the Supply Voltage
Alessandro Ferrero, Fellow, IEEE, and Simona Salicone, Member, IEEE
Abstract—The assessment of the “power quality” and “voltage
quality” in a deregulated electricity market has become a more
and more critical point. As far as the voltage quality is concerned,
the first step toward its assessment is the availability of measure-
ment systems able to detect and classify the possible disturbances
affecting the voltage itself. This task is quite difficult when tran-
sient disturbances are involved, and very complex algorithms have
been proposed, mainly to classify these disturbances. This paper
proposes a very simple detection algorithm, whose aim is to act as
a real-time “trigger” for subsequent, more complex classification
algorithms. A VI has been realized implementing this detection al-
gorithm in real time, and the experimental results collected during
several months of operation are reported.
Index Terms—Disturbance detection, virtual instruments,
voltage quality.
I. INTRODUCTION
T
HE detection of the disturbances affecting the line volt-
ages is one of the most qualifying points in the estimation
of the “voltage quality” or “supply quality.” The correct assess-
ment of the quality of the supplied voltage has become one of
the key issues in the deregulated electricity market. Ensuring
a “high quality” of the supply voltage is the main requirement
for ensuring a high “power quality.” Great attention is therefore
paid to the definition of suitable indexes of voltage quality and
the definition of suitable measurement procedures to evaluate
these indexes.
As far as the definition of voltage quality is concerned, sev-
eral recommendations have recently been issued by the stan-
dards organizations to define the acceptable voltage character-
istics [1]–[4], also in terms of the acceptable number and level
of the disturbances superimposed to the voltage waveform and
their characteristics.
Among the possible disturbances that may affect the voltage
quality, the most difficult to detect are the transient ones, such as
spikes and sags. Several methods have been proposed in the lit-
erature to detect these disturbances [5]–[10], and many of them
are based on the wavelet analysis of the voltage signal [9], [10].
The wavelet transform is a complex, time-consuming algo-
rithm very useful when the disturbance parameters have to be
detected or the disturbances have to be classified into families,
according to their characteristics. However, the most critical
part of the whole detection and characterization process is the
Manuscript received June 15, 2004; revised April 21, 2005.
The authors are with the Dipartimento di Elettrotecnica, Politecnico di
Milano, 20133 Milan, Italy (e-mail: alessandro.ferrero@polimi.it; simona.sali-
cone@polimi.it).
Digital Object Identifier 10.1109/TIM.2005.851078
detection part, since it must be performed on-line and in hard
real time. On the other hand, all further processing can be done
off-line, by employing a suitable algorithm for postprocessing
the acquired disturbed signal.
The most effective solution to this problem is therefore to use
different methods for:
1) the preliminary detection part of the disturbance;
2) the subsequent analysis of the detected disturbance.
Fast and simple algorithms should be used in the detection stage
[5], while more complex algorithms are allowed in the charac-
terization part [8]–[10].
This paper proposes a very simple algorithm for transient dis-
turbance detection in the line voltage signals of a three-phase
system and shows the result of the experimental work done on
the low-voltage distribution network of Politecnico di Milano.
II. DETECTION ALGORITHM
According to the above introductory notes, the detection al-
gorithm is required to be fast, robust, and simple, since it must
execute its detection function on-line and in real time. No ac-
curate measurements are required for this algorithm, since its
only function is to act as a real-time trigger for subsequent, more
complex classification algorithms, if needed.
However, in many practical situations, the subsequent anal-
ysis is very often left to the operator for diagnostic purposes in
order to investigate on the possible reasons that caused the injec-
tion of the disturbance itself. In this situation, the only additional
function required to the detection algorithm is to “capture” the
portion of the voltage waveform affected by the disturbance and
store it for an off-line postanalysis.
For these reasons, with measurement accuracy not being the
main requirement, a very simple detection algorithm has been
analyzed, which appears to be simpler than the simplest ones
available in the literature [5].
This algorithm is based on the consideration that the supply
voltage of a distribution network is quite stable, in the short
term, unless a disturbance occurs.
Therefore, the difference between the voltage signal in a pe-
riod and the signal in the previous period
(1)
where is the period, should be zero for every value of time ,
unless a transient disturbance occurs.
The operating principle of the proposed detection algorithm
is hence based on (1). One period of the signal is acquired,
converted into digital, and stored in the system memory. Each
0018-9456/$20.00 © 2005 IEEE