Sweep Frequency Response Analysis

  • June 17, 2020
  • Posted by: Dominique
  • Category: Transformer Testing
No Comments

A valuable tool for transformer diagnostics is sweep frequency response analysis (SFRA), which consists of measuring the impedance of transformer windings over a wide range of frequencies. Faults, which change either the winding capacitances or the winding inductances, are detectable. Fault simulation programs have indicated that the following faults are detectable using SFRA:

  • Additional turns on yoke or limbs
  • Shorted turns
  • Open windings
  • Faulty core grounds
  • Winding collapse
  • Broken or loosened clamps
  • Mechanical damage to windings
  • Mechanical damage to core
  • Loose turns
  • “Normal” ageing

SFRA is an off-line test, meaning the transformer must be de-energized and completely isolated. The corresponding high frequency circuit is a complex R-L-C network and the measured frequency response (transfer function) of this network is unique, like a “fingerprint”.

The frequency response describes the frequency behavior of a two-port network. To determine the frequency response of a transformer, the assembly is considered as a linear, complex, time-resistant and passive network. 

The frequency response method is normally a relative method of diagnosis. Actual measurements have to be compared to reference results (previous measurements). Results of SFRA measurements are conveniently displayed in modulus-argument form. The modulus is usually referred to as the gain or the amplitude and the argument as the phase.

The basis of SFRA measurements is the comparison between past measurements and repeatability of accurate tests, which requires:

  • High quality, high accuracy device, such as the Haefely FRA 5311
  • The same applied voltage in all SFRA measurements
  • Good connections on the measurement terminals
  • Shortest braided ground method
  • Good documentation, photographs of connections

IEC 60076 standardized the measurement of frequency response in power transformers and called for data to be recorded in XML format. The Haefely FRA 5311 adheres to this format and the included analysis software can analyze measurements taken by other manufacturer’s devices as well, even if they’re not in the XML format.

Comparison Methods

SFRA essentially consists of measuring the impedance of the transformer windings over a wide range of frequencies and comparing the results of these measurements with a reference measurement. The reference measurement is done previously on the same winding of the same transformer using the same setup. If no measurements from the same winding are available, they can be obtained from another phase, which is known or assumed to be unchanged, or from a sister transformer. Results of this kind of comparison should be taken with great care as different manufacturing processes, or geometrical differences between phases, can end in FRA differences without indicating a transformer failure.

The comparison is usually made by superimposing the curves within the software and observing the differences. Both sets of measurements are plotted on the same axes. If additional measurements are available, these can be plotted as well. In comparing the sets of measurements, the key indicators of damage are:

  1. Changes to the overall shape of the graph
  2. The creation of new resonant frequencies or the elimination of existing resonant frequencies (maximum and minimum curves)
  3. Large shifts in existing resonant frequencies.

Below is a screenshot of the Haefely FRA 5311 analysis software. This software can be used even when a measurement device is not connected.  Up to 10 curves can be plotted at once for convenient comparison. The tests below were all done on the same transformer on different windings and different parts of the transformer. Different problems can be seen in different parts of the curve. There is no exact frequency range for each part of the transformer, but in general:

  • Low Frequencies
    • Core problems and shorted/open windings
  • Medium Frequencies
    • Winding deformations
  • High Frequencies
    • Tap connections and other winding connection problems

Time-based Comparison:

Actual results of measurements are compared to fingerprints of former tests. This method is excellent to use for fast detection of transportation damages.

Type-based Comparison*:

The results of an identically constructed transformer (aka “Sister Unit”) are taken as a reference. Distribution transformers are specially constructed according to standards, so the results of different distribution transformers can be compared.

Construction-based Comparison*:

Comparing the results of separately tested phases or coils of a transformer. Sometimes, the frequency behavior of the windings U, V and W of a 3-phase transformer are quite similar, so the results can be compared.

* With these types of comparisons, design-based differences in the compared curves will be seen and are sometimes hard to classify. A previous measurement (fingerprint) of the exact same winding configuration (time-based comparison) is always preferred. There is no such thing as a “typical” measurement, so measurements made on “similar” transformers cannot be used as a basis for detailed comparison.

Test Setup:

Two measuring channels and one variable frequency power supply are required to perform the test, connected as below.

  1. The variable frequency sinusoidal power supply connected to a point (generator)
  2. One measuring input (Channel 2 – Source) measures the voltage at the connecting point (four wires) to eliminate the cable length influence
  3. Second input (Channel 1 – Receiver) measures the output voltage.

Both amplitude and phase of the reference and measured voltage are measured.

Parameters to Measure:

One input signal UIN and one output signal UOUT or IOUT have to be defined. The division of the two signals represent the frequency response.  The following three transfer functions can be measured and calculated:

  • Transfer Voltage Function: UOUT/UIN (f)
  • Transfer Impedance Function: UIN/IOUT (f)
  • Transfer Admittance Function: IOUT/UIN (f)

Types of SFRA Measurements:

Open Circuit: Measurement made from one end of a winding to the other end of the same winding with all terminals left open (floating). Core influence will be dominating in this impedance measurement.

Short Circuit: Measurement made from one end of a winding to the other end of the same winding with one or all other terminals shorted together. (Free neutrals should be left floating). Windings influence will be dominating in this impedance measurement.

Inter-Winding type A: Measurement made from one winding to another winding with all other terminals left open (floating), with grounded neutral and transfer impedance measurement with inductive coupling.

Inter-Winding type B: Measurement made from one winding to another winding with all other terminals left open (floating), with open neutral and transfer impedance measurement with capacitive coupling.

Advantages of the FRA 5311

  • Smallest sweep freqeuncy response analyzer on the market
  • Rugged and lightweight
  • One single USB connection for data and power to a laptop
  • Measures magnitude (dB), phase angle (°), impedance (Ω), admittance (S), and ratio
  • Comliant with IEC 60076-18 and IEEE C57.149
  • Includes Windows based software with easy to use graphical interface
  • Software includes analysis mode for curve comparisons and automatic reporting capabilities
  • Predefined test routines for the most common transformers
  • XML file format compatible to import, export, and share test results with other SFRA devices

If you would like to learn more, feel free to contact one of our sales managers at +1 (703) 365-2330 or hvsales@hvtechnologies.com.