Temperature rise test of converter transformers
Abstract
Purpose of a temperature rise test is to verify the thermal design of a transformer. The injected losses shall correspond to the losses in real operation at rated load. Converter transformers are subjected to harmonics. Those harmonics contribute to the overall losses and need to be taken into account. Some considerations about the temperature rise of converter transformers are presented in this post.
1. Introduction
Factory testing of a transformer is an essential step to verify the transformer design. Temperature rise test shall specifically check the thermal design of a transformer, its cooling capacity and the effectiveness of the cooling system. It shall prove that the temperature rise is within the limits of the given insulation temperature class and no excessive hot spots are created. Staying within the limits of the respective temperature class is important in the lifetime considerations. For liquid immersed transformers the temperature rise is defined separately for the insulation liquid (mineral or biodegradable oil, synthetic liquid, etc) and for the windings.
Converter transformers are exposed to harmonics in real operation. Those harmonics cause additional losses and must be considered in the design. However, the temperature rise test is performed with sinusoidal (non-distorted) current in line with the IEC standards. The current during the test shall be increased beyond the nominal current to an equivalent value that reflects the thermal effect of harmonics.
This article aims to describe a temperature rise test of a converter transformer and highlight the differences compared to a power transformer.
Remark:
“Temperature rise test” may also be called “heat run test”. In this article we stick to the first term as used by IEC 60076-2.
2. Considerations for converter transformers
Converter transformers differ from power transformers in multiple aspects that shall be reflected in the setup of the temperature rise test. One of them is the non-linear load current, i.e. the presence of harmonics and corresponding additional harmonic losses. The current injected during the temperature rise test must…
3. Applicable standards
IEC standards regarding testing of converter transformers (VFD transformers) are:
- IEC 60076-2: Power transformers – Part 2: Temperature rise test [1]
- IEC 60146-1-3: Semiconductor converters: Transformers and reactors
- IEC 61378-1: Converter transformers – Part 1: Transformers for industrial applications
Temperature rise test is a type test according to IEC (and design test according to IEEE). Therefore, it is performed on a transformer that is a prototype or a representative of other transformers. Such test is not repeated on other products of the same type. Aim of this test is to demonstrate that the transformer complies with specified requirements.
4. Test arrangement
There are principally several methods how to conduct a temperature rise test. One of them is to load the transformer with its rated power, resp. equivalent power (power that considers effect of harmonics). That method is theoretically simple but difficult to conduct practically, especially for large transformers where the infrastructure may not be able to provide so much power. Moreover, the method is costly due to the quantity of energy consumed during the test (unless the setup allows some kind of power in the loop circulation). A modification is a back-to-back configuration using transformer under test and another infrastructure trans
former to create circulating current. In that case the power is not wasted, only the power losses (~ 1% of total power assuming each transformer has about 99.5% efficiency). The standard setup is a test in short-circuit condition [2]. One side of the transformer is shorted while the other is supplied with a reduced voltage*.
The level of voltage is adjusted such that approx. nominal current flows through the shorted winding**. The supply voltage (in % of nominal) corresponds to the short-circuit impedance of the transformer. The IEC standard [1] states following clause:
* In case of multi-winding converter transformers the common practice is to short all secondary windings and supply the primary winding of the transformer with voltage.
** The test current shall correspond to an equivalent current that has identical thermal effect. However, note that the test current may differ a bit. The standards allow certain extrapolation of the results. The actual test current shall be within +/-10% of the determined test current.
5. Two stages of temperature rise test
The temperature rise test is split into two stages. Some literature sources call it even two separate temperature rise tests. Nevertheless, they always follow each other. First of them determines the temperature rise of the insulation with winding current above the nominal current. The second stage verifies the temperature rise test of the winding conductor.
i. Temperature rise of insulation liquid
First stage of the test shall verify that the temperature rise of the internal insulation and cooling liquid (mineral oil, natural or synthetic liquid) is within the limits given by the standard. It implies that total losses shall be injected during this test. In this setup with a short-circuit, there are no losses inside the core. Thus, to simulate the real condition, the test current shall correspond to an equivalent current that produces the same amount of losses as the sum of load losses (incl. harmonics) and no-load losses at rated load.
This stage takes several hours until the transformer reaches a steady state in the sense of thermal equilibrium. The test is finished when the temperatures reach an equilibrium. The steady state is reached when the temperature rise of the top liquid has fallen below 1 K per hour for three consecutive hours. The mean value from the readings during last hour is used as final result.
The overall duration of the test depends on transformer rating, its cooling method and efficiency. Large transformers with optimized efficiency and natural cooling have very long thermal time constant and the test may easily take more than 10 hours.
ii. Temperature rise of windings
Second stage determines the temperature rise of the winding. For the purpose of the test the winding related losses shall be injected. It means that injected losses are somewhat lower than losses injected during temperature rise test of the insulation liquid.
1] calls the second stage of temperature rise test the rated current injection. However, rated current is applicable only in case of a power transformer with sinusoidal load current. In case of converter transformer with non-sinusoidal currents, the additional losses PH need to be calculated and the current shall be higher than the rated current.
This test follows directly after the temperature rise test of the insulation liquid without any break. Duration of this test is one hour and temperature readings (top oil, winding hot spot if measured, external cooling medium) shall be taken at least every 5 minutes.
iii. Overview
For a better understanding, both stages of the temperature rise test are summarized below:
The temperature rise test is also graphically depicted in Figure 1. For simplicity, there is just one curve showing the temperature rise of the winding. In reality, there would be at least two such curves (one for HV winding and one for LV winding).
It shall be also noted that the winding temperatures are mostly measured indirectly using methods such as thermal imaging (winding temperature estimated from current measurement and other design parameters).
Conclusion
Temperature rise test of a transformer is a type test according to IEC 60076-2. The purpose is to verify the thermal design of the transformer. A common practice is to short-circuit one side of the transformer and supply the other side with a voltage adjusted to a level until specified current flows through the shorted winding(s).
For liquid-immersed converter transformers the test is split into two parts: (1) temperature rise of the insulation liquid and (2) temperature rise of the windings. The point is to inject the corresponding losses. Amount of losses is controlled by the current flowing through the windings.
Literature
[1] IEC 60076-2 Power Transformers – Part 2: Temperature rise for liquid-immersed transformers, Edition 3.0, 2011-02.
[2] Testing of converter transformers (article for premium subscribers), MB Drive Services, April 2021
