A Review of the Monitoring & Diagnostic Methods of Oil Immersed Transformers

Transformer is vital equipment in a power system and to ensure reliable service and to plan for timely preventive maintenance suitable monitoring & diagnostic methods are required. This paper gives a review of the on-line monitoring & diagnostic and off-line diagnostic methods used for oil immersed transformers. Keywords—Distribution Transformer; Dissolved gas Analysis (DGA); Furan analysis; Degree of polymerization of transformers; Sweep frequency response analysis


I. INTRODUCTION
The modern power system is mandated to provide reliable uninterrupted power supply and therefore, it is important to ensure the health and longevity of the transformer. There are several effective on-line and off-line diagnostic tools available today for the diagnostic the condition of the transformer which can identify incipient faults, estimate the remaining life, and ascertain maintenance strategy or to schedule the retirement & replacement of the transformer. This paper lists out the diagnostic / monitoring methods and reviews some of the commonly used techniques.

II. ON-LINE AN OFF-LINE DIGNOSTIC METHODS
The online monitoring / diagnostic methods measure the inputs continuously without affecting the operation of the transformer whereas the off-line methods need the planned shutdown of the transformer. Maintaining the Integrity of the Specifications I.
On Line monitoring Table I shows the commonly used on line data monitoring methods. The measurement applicable for a particular transformer is selected based on the design, the application, the rating etc.  Gases dissolved in oil are analyzed by gas chromatography which separates each gas form other and directly measures their concentrations individually. The interpretation is carried out by the guidelines given in IEC 60599-2007-5 "mineral oil Impregnated Electrical Equipment in service: guide to the interpretation of Dissolved and Free gas Analysis" or IEEE C57.104.2008 -"IEEE guide for the Interpretation of gases generated in Oil Immersed Transformer". or CIGRE guidelines or Dual triangle or a combination. The concentration of various gases and the ratios of the concentration give conclusion regarding the type of fault/condition existing in the transformer. When the values of CO > 500 PPM and Co2 > 5000 PPM CO2/CO Ratio of <3 indicates thermal cellulosic Degradation. (However, this ratio is not reliable for sealed type Transformers). A ratio 3 to 11 is considered healthy cellulose insulation.
• Interpretation by CIGRE SC15 The guidelines for interpretation of DGA in oil filled transformers as per CIGRE SC-15 is given in Table IV  imperfections, conductive particles etc create localized stresses leading to partial breakdown of the dielectric. These discharges are very fast electrical processes and radiate electromagnetic waves of high frequency and ultra High frequency range. If such partial discharges are continuously increasing, it is a sure sign of a weak insulation, high stress concentration etc. These are signs of incipient faults. A healthy oil filled transformer can have PD magnitude of 10-50 PC or lower at operating voltage. The increasing PD levels signify the following: The different methods of PD detection are shown in Table VII

B. Furan Analysis and Degree of Polymerization (DP)of Transformers
When the transformer operates the insulation ages (becomes weak) and it would ultimately lead to a breakdown causing the failure of the transformer. The number of glucose rings in the cellulose is reduced when the kraft paper ages and this is called Degree of Polymerization (DP). New Cellulose will have 1000-1200 glucose rings when a transformer is manufactured, what it is processed under drying some of the rings are broken down and hence the DP value of new transformer is about 950. A DP value of 250 is considered as the end point of insulation life and at this point the insulation becomes highly brittle and the tensile strength of the cellulose becomes very low. [If left without disturbance some transformers may give extended useful service after reaching this point].
Measuring the DP value of the insulation paper can be done by taking sample from the transformer. However, this has several drawbacks like (a) The transformer will have to be switched off and opened to take a sample and (b) The sample collected may not be from the area of maximum aging i.e. the hot spot The disintegration of the cellulose inside the transformer generates Furan Compounds which are dissolved in the transformer oil. Hence, it is convenient to analyze the transformer oil sample for Furan content to determine the aging (DP) of the insulation. This can be done without switching off the transformer.
The Furan compounds generated by various types of reactions are different. The concentration of Furaldehyde (2FAL) in the oil is due to the overheating of the paper and the maximum aging is at the location of hot spot. The drawback of this method is that the measurement gives an average of the Furan Concentration and therefore, the same at the hot spot may be much higher. The concentration of Furans is measured by High Pressure Liquid Chromatography. The concentration of various Furans indicate different types of stresses as shown in Table VIII  The concentration of 2 FAL value is used by different authors to estimate the DP value and the results of DP values corresponding to the concentration of 2 FAL values as per "chengdong" are given in Table IX  part results in change of this RLC circuit. When different frequencies are applied to the transformer, the RLC network offers different impedance path and the transfer function at each frequency is a measure of the effective impedance of the network. Any geometrical change alters the RLC network which in turn changes the transfer function at different frequencies thereby highlighting the area of concern. When a new transformer is manufactured and passes the tests, the SFRA of the transformer can be taken for using as a reference (signature). This can be compared with SFRA measurement in future, for example, after transportation to site, after installation, after an incident like a tripping, earthquake etc. If SFRA of a transformer is not available, detection of abnormalities can be tired in any one of the following methods.
• Checking the SFRA of another transformer of the same design from the same batch of production, if available. • Comparing SFRA of other phases of the transformer SFRA test is done typically between 20 Hz to 1 MHz The number of tests for various types of transformers is as per Table 10  The response of SFRA for different range of frequencies can pin-point the region / component of concern / problem. A guideline for the interpretation of probable reasons is given in Table XI.

B. Dielectric Frequency Response Analysis
The traditional method of measuring the moisture content of the oil in the transformer does not give the oil absorbed in the cellulose insulation. This is because when the transformer is loaded, and the winding temperature rises, the moisture comes out from the cellulose and absorbs in oil. When the load comes down the temperature comes down and the moisture returns back to the cellulose. Though the moisture equilibrium curves can be used to determine the moisture in cellulose, the results are highly dependent on temperature of the oil and it may give inaccurate results. The dielectric frequency response is an advanced technique to ascertain the insulation of moisture. The frequency response of a dielectric is a unique characteristic of a particular insulation system. The increased moisture content changes the dielectric model and the frequency response. By measuring the dielectric response over a wide frequency range, the moisture content of the insulation can be determined and the condition assessed. The test is normally done between the High Voltage and Low Voltage winding of a transformer by applying frequencies from 0.001 Hz to 1000 Hz to measure the dissipation factor. The voltage used for testing is up to 200 volts. The moisture content and aging is revealed at higher frequencies.
The shape of the curve shows the relative moisture levels and a typical response pattern is shown in Figure-

C.
Recovery Voltage Measurement (RVM) A step voltage is applied to the transformer and a charging current flow. Then it is short circuited during the discharge time and then it is open circuited. A recovery voltage appears across the terminals. This is caused by the unfinished polarization. The recovery voltage is measured by changing the charging time. The discharging time is kept as 2 times charging time for all measurements. These measurements at different charging times give the polarization spectra.

Optical Spectroscopy
The absorbance of ultra violet (UV) light of fresh oil is low. When the oil ages, the absorption of UV light goes up. The absorption is the highest between the wavelengths of 200-380 nm. Figure 2 shows the pattern of absorbance of UV light in fresh oil and service aged oil.

E. Search Coil based on-line diagnostics of transformer internal faults
Internal faults related to transformer windings such as inter turn faults and winding displacement changes the inductance of the winding when compared to the inductance at healthy conditions. If search coils are installed during manufacture, the induced voltages in the search coils can be analyzed to monitor the changes of the inductance. The faulty winding and location of the defect can be detected.

F. Polarization and Depolarization Current Test (PDC)
The polarization current of a transformer winding is a function of the geometry, oil properties & aging. Initial value of polarization current is determined by the oil conductivity. The depolarization current comes down with the elapse of time when compared to polarization current depending on the moisture.
A typical pattern of the polarization and depolarization current is shown in Figure 3

G. Embedded wireless monitoring and Fault diagnostic system
On line monitoring is possible for diagnostic methods including dissolved gases in oil, oil temperature, winding temperature, oil level, core temperature, cooler operation etc. Embedded system is used and the data is transmitted by RF transmitter to the control room. Schematic diagram of a typical wireless monitoring system is shown in figure 5 International

H. Frequency Domain Spectroscopy (FDS)
When a dielectric material is subjected to an Electric field, the electrical dipoles are oriented in the direction of the field. This orientation has a time delay which depends on the material, the moisture content and the geometry of the insulation. The frequency range used for analysis is normally 0.001 Hz to 1000 Hz.

I.
Monitoring of temperature Monitoring of the ambient, winding, oil and tank body temperatures can identify any excessive temperature above the permissible values. If the temperature exceeds the designed value applicable for the class of insulation, the aging becomes rapid and also this can be an indication of defects / faults leading to failure. The top oil temperature and winding temperature can be monitored on-line and tools are available for remaining life estimation & load management. On-line thermograph can monitor and detect temperature gradients on tank surface. The gradients can be compared to the tested values under defined operating conditions to take suitable action.

J.
Load Monitoring The loads of each phase of the primary, secondary, and other windings (if applicable) is monitored for load management by including the data of ambient temperature, oil temperature and winding temperature. Where Fans and pumps are used for transformer cooling, the input currents to these devices and the operation can be monitored.

K.
Vibration Monitoring Excessive vibration is often an indication of abnormal operation of fan, pump, loose coils, loose clamping, shorted turns etc. which can lead to a failure. For comparison, vibration measurement results of new transformer after installation is necessary.

L. Monitoring of the functioning of bushings & on Load Tap Changer
The operation of condenser bushing is monitored by measuring the current through the test tap in operation. Change in tan delta and the charging current are monitored to detect the abnormality / fault of the bushing.
• In oil insulated on-load tap changer, (OLTC) the contact temperature can be indirectly monitored by measuring the temperature of the oil in the OLTC compartment. This temperature is compared with the oil temperature of the main tank.

III. FUZZY INFORMATION APPROACH TO INTEGRATE RESULTS OF DIFFERENT DIAGNOSTIC METHODS
Good experience is essential to interpret the results of many of the diagnostic techniques with accuracy. Artificial Intelligence (AI) tools can be applied to arrive at the most probable interpretation using the Fuzzy information approach. Information about the relationship between different fault types and diagnostic results is made into a fault tree.
• This approach consists of the following: • Select the diagnostic method.