Academia.eduAcademia.edu

Figure 2 – uploaded by Bonginkosi Thango

See full PDFdownloadDownload figure
To analyse the complete transformer harmonic response (THR), Figure 4 shows the results of the predicted transformer losses by IEC harmonic spectrum and compares them with the simulated harmonic spectrum the transformer is subjected to while in service. The transformer under inspection was designed without Table 2: Transformer losses based on actual DSP plant harmonic spectrum

Figure 4 To analyse the complete transformer harmonic response (THR), Figure 4 shows the results of the predicted transformer losses by IEC harmonic spectrum and compares them with the simulated harmonic spectrum the transformer is subjected to while in service. The transformer under inspection was designed without Table 2: Transformer losses based on actual DSP plant harmonic spectrum

Related Figures (8)

The no-load losses and stray losses in windings and structural parts evaluated at nominal tap position are tabulated in Table 1. In order to compute these losses to account for the presence of harmonics it is necessary to obtain the values of the harmonic loss factors using the IEC 519-2014 [19] standard. Generally, from a manufactures perspective, the case of considering the harmonic spectrum proposed by this standard for a transformer intended to operate in a DSP plant is when there is no full harmonic spectrum supplied by the IPP. The harmonic spectrum recommended by IEC is shown in Figure 2. Figure 2: IEC 519-2014 Harmonic spectrum [2]
The no-load losses and stray losses in windings and structural parts evaluated at nominal tap position are tabulated in Table 1. In order to compute these losses to account for the presence of harmonics it is necessary to obtain the values of the harmonic loss factors using the IEC 519-2014 [19] standard. Generally, from a manufactures perspective, the case of considering the harmonic spectrum proposed by this standard for a transformer intended to operate in a DSP plant is when there is no full harmonic spectrum supplied by the IPP. The harmonic spectrum recommended by IEC is shown in Figure 2. Figure 2: IEC 519-2014 Harmonic spectrum [2]
Figure 1: 2D geometry of the DSP plant transformer are detailed by the authors in [17] and [18]. The advancement of computational power has enabled the development of tools capable of simulating complex geometries, mapping of electromagnetic fields and accurate evaluation of these transformer losses. A software tool embedding analytical methods with Finite Element Method (FEM) is used to estimate the transformer losses. The tool is automated to simulate complete transformer model geometries, magnetic field quantities and computation of eddy current losses in tank walls, flitch plate, core clamps, bus-bar’s and windings. Figure 1 shows the complete model of the transformer 2-D geometry with the main parts that are susceptible to the leakage flux.
Figure 1: 2D geometry of the DSP plant transformer are detailed by the authors in [17] and [18]. The advancement of computational power has enabled the development of tools capable of simulating complex geometries, mapping of electromagnetic fields and accurate evaluation of these transformer losses. A software tool embedding analytical methods with Finite Element Method (FEM) is used to estimate the transformer losses. The tool is automated to simulate complete transformer model geometries, magnetic field quantities and computation of eddy current losses in tank walls, flitch plate, core clamps, bus-bar’s and windings. Figure 1 shows the complete model of the transformer 2-D geometry with the main parts that are susceptible to the leakage flux.
Table 1: Tabulated loss computation under a harmonic spectrum From both the fundamental and harmonic losses above, it can be noticed that the loss percentage error is higher for winding stray losses. Additionally, the core losses are observed to be constant, indicating that the no-load losses are independent of the harmonic currents. The minimum winding stray losses as a result of axial and radial leakage fields are computed by optimizing the selection of conductor dimensions while maintaining the permissible flux and current density of the winding [20] and [21]. The transformer winding subject to harmonic distortion is designed using continuously transposed cable epoxy (CTCE) with minimum axial and radial dimensions of each conductor to reduce the eddy losses [20] and [21]. The tank is designed using mild steel which hold a nonlinear permeability. The quantity of the magnetic field in the tank walls is evaluated by eliminating the impact of nonlinearity. Thereafter, the losses in the LV and HV tank walls are computed taking into account the effect of nonlinearity through repetitive estimation. The flitch plates are also mild steel type and contain slots at the upper and lower positions. While their loss quantity is low compared to winding eddies and tank losses under harmonic effects, the magnetic field quantity is high which results in the flitch plate being more susceptible to HS temperature.
Table 1: Tabulated loss computation under a harmonic spectrum From both the fundamental and harmonic losses above, it can be noticed that the loss percentage error is higher for winding stray losses. Additionally, the core losses are observed to be constant, indicating that the no-load losses are independent of the harmonic currents. The minimum winding stray losses as a result of axial and radial leakage fields are computed by optimizing the selection of conductor dimensions while maintaining the permissible flux and current density of the winding [20] and [21]. The transformer winding subject to harmonic distortion is designed using continuously transposed cable epoxy (CTCE) with minimum axial and radial dimensions of each conductor to reduce the eddy losses [20] and [21]. The tank is designed using mild steel which hold a nonlinear permeability. The quantity of the magnetic field in the tank walls is evaluated by eliminating the impact of nonlinearity. Thereafter, the losses in the LV and HV tank walls are computed taking into account the effect of nonlinearity through repetitive estimation. The flitch plates are also mild steel type and contain slots at the upper and lower positions. While their loss quantity is low compared to winding eddies and tank losses under harmonic effects, the magnetic field quantity is high which results in the flitch plate being more susceptible to HS temperature.
data. Notably, this high harmonic order exceeds the recommended IEEE spectrum presented in the last section. Figure 3: Simulated transformer harmonic spectrum at the PCC
data. Notably, this high harmonic order exceeds the recommended IEEE spectrum presented in the last section. Figure 3: Simulated transformer harmonic spectrum at the PCC
Table 3: DGA oil sample results (ppm) The adverse effect for under-designing this transformer for requirements such as suitable conductor choice and size, and cooling system can be observed from the gas concentration exceeding the permissible IEC limits as shown in red in Table 3. Due to the excessive nature and production rates of hydrogen, a possible solution would be to replace the oil in all the affected units in case there were contaminants in the oil that resulted in the production of the abnormal gassing patterns. If the observed gassing patterns persist after the oil replacement, then a redesign of the transformer units might be needed. A proper harmonic study of the DSP plant as well as the loading conditions should be conducted in order to design for the DSP conditions. For the above
Table 3: DGA oil sample results (ppm) The adverse effect for under-designing this transformer for requirements such as suitable conductor choice and size, and cooling system can be observed from the gas concentration exceeding the permissible IEC limits as shown in red in Table 3. Due to the excessive nature and production rates of hydrogen, a possible solution would be to replace the oil in all the affected units in case there were contaminants in the oil that resulted in the production of the abnormal gassing patterns. If the observed gassing patterns persist after the oil replacement, then a redesign of the transformer units might be needed. A proper harmonic study of the DSP plant as well as the loading conditions should be conducted in order to design for the DSP conditions. For the above
At tender stage, the transformer technical specifications supplied by IPP’s to the manufacture generally include information about the permissible top oil, mean winding and HS flemperatures. Electrical designers are then challenged to optimize transformer designs such that they adhere to the IEC 60076-2 [22] and IEEE C57.91-1995 [23] standards and to design the thermal and cooling requirements at an ambient temperature. During the manufacturing stage, as part of the Factory Acceptance Tests (FAT’s) for newly manufactured transformers, the temperature rise test is conducted with the intent to demonstrate these temperatures, at full load and specified ambient temperature. The loading capability of the transformer is predominantly regulated by the HS temperature. In accordance with the IEC, the HS lemperature is described as the mean winding temperature rise over the mean oil temperature multiply by a HS factor [22] and [23].
At tender stage, the transformer technical specifications supplied by IPP’s to the manufacture generally include information about the permissible top oil, mean winding and HS flemperatures. Electrical designers are then challenged to optimize transformer designs such that they adhere to the IEC 60076-2 [22] and IEEE C57.91-1995 [23] standards and to design the thermal and cooling requirements at an ambient temperature. During the manufacturing stage, as part of the Factory Acceptance Tests (FAT’s) for newly manufactured transformers, the temperature rise test is conducted with the intent to demonstrate these temperatures, at full load and specified ambient temperature. The loading capability of the transformer is predominantly regulated by the HS temperature. In accordance with the IEC, the HS lemperature is described as the mean winding temperature rise over the mean oil temperature multiply by a HS factor [22] and [23].
the provision of a full harmonic study from the IPP, and judging from the results, the indication is that it was under-designed for the environment it was intended to operate, which would explain the reported high temperature rise and oil gassing patterns. Subsequently, this knowledge gap led to insufficient design of the winding conductor sizes and cooling system for this application. Figure 4: Loss comparison (IEC vs. Simulated Harmonic Spectrum)
the provision of a full harmonic study from the IPP, and judging from the results, the indication is that it was under-designed for the environment it was intended to operate, which would explain the reported high temperature rise and oil gassing patterns. Subsequently, this knowledge gap led to insufficient design of the winding conductor sizes and cooling system for this application. Figure 4: Loss comparison (IEC vs. Simulated Harmonic Spectrum)
Table 4: DGA oil sample results after oil replacement (ppm) Air and moisture are generally the main determinants of gas pressure in the transformer oil. The progression of these components are determined by the transformer tank design and loading profile. The above irregular gas bubbling patterns are observed at hot spot temperature [22] far beyond the maximum permissible temperature of 78°C by the IEC. These patterns instigate rapid ageing of oil and cellulose insulation.
Table 4: DGA oil sample results after oil replacement (ppm) Air and moisture are generally the main determinants of gas pressure in the transformer oil. The progression of these components are determined by the transformer tank design and loading profile. The above irregular gas bubbling patterns are observed at hot spot temperature [22] far beyond the maximum permissible temperature of 78°C by the IEC. These patterns instigate rapid ageing of oil and cellulose insulation.
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved