Dr. Viral Rathod - Research

Research

Ph.D. Work - High Voltage Laboratory - Experimental Setups

PD Measurement Setup

A High-Voltage Source Transformer of 100 kV is used to generate Partial Discharge (PD) in the Transformer Prototype.
The various PD defect models are used to generate artificial PD, such as Corona, Surface, and Void Discharges.
Electrical PD measurement in transformer winding using High-Frequency Current Transformers (HFCTs).
Electrical PD measurement in transformer winding using a Coupling Capacitor (1 nF).
Acoustic PD measurement in the transformer using piezo-electric Acoustic Sensors (150 kHz) mounted on the transformer tank.
The recording of electrical and acoustic PD signals using a Digital Storage Oscilloscope (DSO) and Computer.

FRA Measurement Setup

The experimental setup was used to conduct Frequency Response Analysis (FRA) measurements of a high-voltage continuous disk winding model.
The winding model has 24 disks, and each disk has 16 turns.
In this winding model, various tappings are extracted from every four disks, and the resultant seven terminals are connected to the front terminal box of the winding model.
An aluminum cylinder is positioned within the winding to replicate the effect of the core and simultaneously serve as the ground path of the winding.
The impedance analyzer (Bode 100, OMICRON Lab) is used to conduct FRA measurements of the winding model.
The impedance responses measured from the winding model are stored in the computer for subsequent analysis.

PD Localization in Transformer - Acoustic Time Reversal

Journal Publication - Performance Analysis

Acoustic Time Reversal Technique:

An acoustic time reversal (ATR) technique for 3D PD localization in a power transformer has been proposed.
The ATR technique comprises three steps:

Measurement of the acoustic PD signal using at least one sensor.
Time reversal of the measured signal and back injection into the transformer model through numerical simulation
Determining the focusing point of the back-injected signal using the maximum acoustic pressure field criterion that indicates the location of the PD source.

The traditional TDOA technique fails to provide accurate localization in the presence of transformer windings and core.
Conversely, the proposed ATR technique can demonstrate precise PD source localization, even with a single sensor.

Experimental Setup for PD Measurement:

The experimental setup to verify the proposed technique comprises a small transformer tank, a winding model, a point-plane electrode system, a fly-back transformer-based high-voltage supply, and an acoustic measurement system.
The air-filled transformer tank, made of mild steel, has dimensions of 60 cm×60 cm× 50 cm, and the wall thickness of the tank is 4 mm.
The winding model, which has 8 numbers of disks made of copper strips separated by a 13 mm thick pressboard spacer, is placed in the middle of the tank.
The winding has an outer diameter of 20 cm, an inner diameter of 15 cm, and a height of 20 cm.
The adjustable point-plane electrode system with variable height is used to generate the artificial PD inside the tank.
The point-plane electrode made of brass is supported by an acrylic structure placed over a 10 mm thick acrylic top cover on the tank, with eight holes to change the horizontal position of the electrode system.
The electrode system is connected to the 30 kV high-voltage supply generated by a flyback transformer.
The four piezo-electric acoustic sensors (R15a, Physical Acoustics Corporation) are mounted on the transformer tank through a magnetic holder, with silicone grease applied between them to ensure better acoustic contact and minimize reflections.
These sensors have a bandwidth of 50 to 400 kHz and a resonance frequency of 150 kHz.
The sensors capture acoustic signals resulting from PD events, amplified using signal amplifier units, and then transmitted to a data acquisition system for further analysis.

PD Localization Result:

The acoustic PD signals measured by the four acoustic sensors are reversed in time and back-injected into the numerical model of the transformer prototype.
The time-reversed simulation is conducted for 2000 μs, and the maximum acoustic pressure is calculated for all time steps.
The peak value of the maximum acoustic pressure calculated for all time steps provides the focusing point of the back-injected signals.
The normalized acoustic pressure distribution at the focusing time is shown in the xy-plane and xz-plane.
The location coordinates of the estimated PD source are (44.83, 45.00, 9.91) cm, with a location error of 0.19 mm relative to the actual PD source.

PD Localization in Transformer - Electromagnetic Time Reversal

Journal Publication - 1 : Simulation Study

Journal Publication - 2: Experimental Validation

Electromagnetic Time Reversal Technique:

An electromagnetic time reversal (ATR) technique for PD localization inside a transformer winding has been proposed.
The EMTR technique comprises three steps:

Measurement of PD signals from the line-end and neutral-end terminals of the transformer winding.
Then, a ladder network-based winding model is constructed by time reversal of the measured PD signals and back-injecting the time-reversed PD signals into the model.
Thereafter, the energy of the PD current signal is computed by short-circuiting the capacitance at various locations in the winding model.

The winding location, associated with the highest energy concentration, indicates the actual

Experimental Setup for PD Measurement:

The experimental setup comprises a PD-free high-voltage (HV) source transformer capable of generating voltages up to 100 kV, a real-scale 11 kV winding model with tappings, a coupling capacitor (Ck) of 1 nF, a measuring impedance (Zm) of 1000 Ω, a signal generator with a minimum pulse width of 20 ns, a PD calibrator capable of generating a charge range of 5–5000 pC, a test cell housing various PD defects, and an oscilloscope with a bandwidth of 100 MHz.
In this setup, the line-end (L) of the winding model is connected to Ck in series with Zm, while the neutral-end (N) of the winding model is directly connected to the ground.
The experiment involves the injection of the PD signal generated using a signal generator, PD calibrator, and live PD into the various taps of the winding model.
Subsequently, the resulting PD signals are recorded at the measuring impedance via an oscilloscope.
The live PD signal is generated by applying a high-voltage supply generated using the HV transformer to the test cell containing various PD defect models.

PD Defect Models:

Artificial PD defects are created using five distinct electrode configurations made of brass and insulation pressboard.
Corona discharge is created using point-plane or plane-point electrode configurations with a 30 mm gap between them.
When HV is applied to the point electrode and the plane electrode is grounded, it produces corona discharge at HV, as shown in Fig (a).
Conversely, when HV is applied to the plane electrode and the point electrode is grounded, it produces corona discharge at LV, as shown in Fig (b).
Surface discharge is generated by inserting a 4 mm thick insulation pressboard between a point-plane or sphere-plane electrode configuration.
Applying HV to the point or sphere electrode while grounding the plane electrode results in surface discharge - I and surface discharge - II, as shown in Fig (c) and (d).
Void discharge involves placing three layers of 12 mm thick insulation pressboard with a 4 mm diameter void in the middle layer between plane-plane electrodes.
Applying HV to the top plane electrode while grounding the bottom electrode produces void discharge, as shown in Fig (e).

PD Localization Result:

The effectiveness of the proposed EMTR technique is assessed using live PD signals generated by applying high-voltage to different PD defect models.
These models produce five types of live PD signals: corona discharge at HV, corona discharge at LV, surface discharge - I, surface discharge - II, and void discharge, with applied voltages of 10 kV, 12.5 kV, 5 kV, 8.5 kV, and 9 kV, respectively.
These signals are then injected into Tap-2, Tap-4, and Tap-6 of a real-scale winding model.
Subsequently, the PD current signal energy (PDCSE) is calculated for each guessed PD location (GPDL) through reversed time simulations normalized relative to the maximum PDCSE value.
The PD localization results, expressed in terms of normalized PDCSE for injected live PD signals, are presented in Table, in which the highest normalized PDCSE value indicates the real PD location.
It is observed from the Table that the suggested approach effectively pinpoints the real location of the PD source in the winding model for injected live PD signals of various PD defects.
The proposed EMTR technique does not rely on reference signals, which makes it self-sufficient for PD localization.
Further, it does not involve intricate calculations or entail extensive data processing and training datasets for PD localization, which makes it more suitable for practical implementation.

Page updated

Google Sites

Report abuse