Dissolved Gas Analysis (DGA) is an important
analysis for fault diagnosis and condition monitoring of
power transformer. The various techniques, such as conventional
methods, Artificial Intelligence (AI), Artificial Neural
Network (ANN), Fuzzy Expert Systems (FES), Genetic Algorithm
(GA), Bayesian Network (BN), Extended Relation Functions
(ERF) and Self-Organizing Map (SOM) algorithm, can be used
to increase the efficiency and accuracy of the diagnostics
system. This paper presents the systematic review of various
techniques with their relative advantages and disadvantages.
Transformer is one of the most important and costly apparatus
in a power system. The reliable and efficient fault-free
operation of the high voltage transformer has a decisive
role in the availability of electricity supply. In transformer,
oil and paper insulation degrade under a combination of
thermal, electrical, chemical, mechanical and environmental
stresses during its operation (Saha and Purkait, 2004).
When power transformer is subject to electrical and thermal
stresses, characteristic gases such as H2, CH4, C2H2, C2H4,
C2H6, CO, and CO2 will be generated in the transformer oil
(Limin et al., 2005). These stresses also change the properties
of the paper and oil, which age the insulation system (Mitchinson
et al., 2006). Pardhan (2006) assessed the status of insulation
during thermal and electrical stresses on transformer prototype.
A number of properties such as dissipation factor, capacitance,
Breakdown Voltage (BDV) of oil and paper, Degree of Polymerization
(DP), Total Combustible Gases (TCG), furan contents, etc.,
have been identified as being reasonably sensitive indices
of degradation. According to Mitchinson et al. (2006), the
characteristics of the aged oils were also determined by
using various analytical techniques, such as Ultra-Violet/Visible
(UV/Vis) spectroscopy, Infra-Red (IR) spectroscopy, acid
number and water content test and dielectric spectroscopy.
A significant factor in ageing is chemical action through
moisture, metal contaminants and oxidation (Mitchinson et
al., 2006). Dissolved Gas Analysis (DGA) is widely used
to detect incipient faults in oil-filled electrical equipment.
But according to Duval and Dukarm (2005), some inaccuracy
is always associated with laboratory DGA measurements of
transformer oil, which may affect the gas ratios and other
calculations. Consequently, techniques based on artificial
intelligence are proposed by various researchers.
A number of conventional methods are available for interpretation
of DGA results (Verma et al., 2004). The conventional methods
may be classified as characteristics key gases, gas ratios
(such as Dorenberg ratio, Rogers ratio method, IEC ratio,
Duval triangle (Thang et al., 2001), etc.), and regression
analysis (Saha and Purkait, 2004). Key gas diagnostics criterion
is qualitative in nature and gives an indication of the
nature of fault in the transformer. The gas ratio is a semi-quantitative
technique and helps to identify the faults more specifically
than the characteristics gases. This method applies generally
to the relatively common conservator type units with expansion
tank (Verma et al., 2004). The ratio methods have an advantage
that they are independent of the volume of gases involved.
But the main drawback of the ratio methods is that they
fail to cover all ranges of data (Morais and Rolim, 2006).
Regression method is a quantitative technique in which the
rate of gas production is correlated with operation parameters,
such as load duration, load current, load current squared,
etc., (Verma et al., 2004).
|
