On-line monitoring and diagnosis of transformers have been investigated and discussed significantly in last decade. This study has concentrated on issues arising while on-line transformer winding deformation diagnosis is going to be applied on transformers with various kinds of techniques. From technical perspective, before replacing off-line methods by on-line methods and eventually by intelligent approaches, practical challenges must be addressed and overcome. Hence, available off-line transformer winding deformation diagnosis methods are discussed precisely. Mathematical calculation in on-line short circuit impedance measurement is investigated. On-line transformer transfer function measurement setup is presented. A profound insight to the problems pertaining on-line transformer winding deformation recognition methods, characterizes existing online methods, explains the concepts behind online measurements and striving to open the discussion doors towards challenges are discussed. In the end a 400 MVA step up transformer has been taken as a case in order to clarify the capability of Frequency Response Analysis (FRA) method in fault detection while short circuit impedance could only demonstrate some rough understanding about transformer condition.

Advanced transformer winding deformation diagnosis: moving from off-line to on-line

Partial discharge (PD) is one of the sources of insulation failures in Power transformers. Knowledge of the PD source locations are important for transformer maintenance and repair. Due to the complex structure of the transformer, it is very difficult to locate the PD accurately in transformer winding. Correlation methods are applied to locate the PD source across the winding. The PD across parts of the winding are simulated by applying voltage pulse across the sections. The winding response for a reference PD impulse across transformer sections are taken as base for correlation and correlated with winding response of PD impulse across any section with different PD pulse widths. In this paper the limitation of time domain correlation method in locating PD in transformer windings is found out. To overcome the limitation of time domain correlation method, frequency domain correlation with Gram Schmidt orthogonalization is proposed. Simulation and experiments are performed on layer, continuous disc and interleaved windings to prove the feasibility of the proposed method.

Frequency domain correlation technique for PD location in transformer winding

The ability of a transformer to withstand the dynamic effects of a short circuit test is conventionally evaluated by a measurement of short circuit reactance before and after the test. We propose new techniques for fault detection based on continual assessment during the test. The methods are based on the comparative null method for accuracy measurement in instrument transformers. They include voltage comparison, current comparison and real/reactive power measurements. High resolution sampling based acquisition systems are used to compute changes on a cycle-to-cycle basis. The proposed methods are validated on a number of models including a voltage transformer and a distribution transformer winding. They would be useful for designers in assessing events leading to failure in a more transparent manner.

Real-Time Techniques to Measure Winding Displacement in Transformers During Short-Circuit Tests

This paper introduces a novel fault diagnosis approach for transformer based on self-powered radio-frequency identification (RFID) sensor and deep learning technique. The exploited RFID sensor tag with functionalities of signal collection, data storage, and wireless transmission employs surrounding electromagnetic field as power source. A customized power management circuit, including ac–dc converter, supercapacitor, and its corresponding charging circuit, is presented to guarantee constant dc power for the sensor tag. The measured vibration signal contains miscellaneous noises and is characterized as nonlinearity and nonstationarity, so it is difficult to extract robust and useful features by using traditional feature extraction approaches. As one of the deep learning techniques, stacked denoising autoencoder (SDA) shows satisfactory performance in learning robust features from complex signal. Hence, in this paper, SDA approach is employed to learn robust and discriminative features from measured signals. The experimental results show that the presented power supply can generate 2.5-V dc voltage, which is the rated operating voltage for the rest of the sensor tag. The developed sensor tag can achieve a reliable communication distance of 17.3 m in the test environment. Furthermore, the SDA approach shows satisfactory performance in learning robust and discriminative features. Experimental results indicate that the presented approach is effective and time-saving in terms of fault diagnosis for transformer winding and core.

Transformer Fault Diagnosis Using Self-Powered RFID Sensor and Deep Learning Approach

Monitoring and diagnosis of transformers have been investigated and discussed for many years While transformer routine tests indicate electrical and also insulation weakness or defects, Frequency Response Analysis (FRA) as a modern test can indicate mechanical defects in transformer active part This paper introduces the concepts behind online measurements in frequency domain and is striving to open the discussion doors towards advanced investigations which are all aligned with smart grid concept This study has also concentrated partly on issues rising while online transfer function measurement techniques are going to be applied on transformers with various kinds of winding connections and accessories In the end, crucial issues from theoretical and also practical perspectives for online transfer function implementation are discussed

Practical challenges in online transformer winding deformation diagnostics

We propose an objective formulation of the impulse analysis problem from a signal analysis viewpoint. The winding response is quintessentially that of a deterministic network to a finite energy signal, with breakdown and partial discharge being inherently nonlinear events. A significant improvement to the acquisition of waveforms is demonstrated by a virtual instrument approach. It retains the advantages of the time- and frequency-domain methods. The drawbacks of the transfer function method are highlighted and a new piecewise linear approach is proposed for analysis. Experiments on a discrete lumped parameter model of the winding are used to validate the PXI based instrument.

Virtual Instrument for Lightning Impulse Tests

Experimental results on fault simulations during impulse tests across windings and other specifically designed networks are reported. Both breakdown and partial discharge studies are conducted with digital acquisition of fault records.

A. Palani

Papers

Real-Time Techniques to Measure Winding Displacement in Transformers During Short-Circuit Tests

Virtual Instrument for Lightning Impulse Tests

Simulation of faults during impulse tests