Optical Chromatic Monitoring of High-Voltage Transformer Insulating Oils

A. T. Sufian, E. Elzagzoug and D. H. Smith

Introduction

Fligh-voltage transformers are monitored in order to avoid unnecessary electric power interruptions and to extend the transformer’s service life (Lorin, 2005). One approach has been to monitor the electrically insulating oil in which the transformer is immersed using several techniques (Bureau of Reclamation et al., 2005), including optical ones such as colour index (Cl) (Nadkarni, 2007). The use of chromatic techniques for monitoring changes in the optical spectra of polychromatic light transmitted through the oil has been shown to provide a versatile means to yield additional information to that obtained with Cl and other techniques. A number of convenient-to-use, cost- effective systems for such chromatic monitoring have been investigated (Elzagzoug et ah, 2014) which use different combinations of polychromatic light sources (white LED. visual display unit [VDU] screen) and optical detection units (spectrometer, mobile phone cameras). The use of a combination of a VDU screen and a mobile phone camera has been shown to be attractive in providing a tunable polychromatic light source with a user-versatile detection unit.

The optical techniques used are based upon the transmission of polychromatic light through the oils (Elzagzoug et ah, 2014), observation of the fluorescence produced by monochromatic laser light (Lo et ah, 2017) and a combination of transmission and fluorescence techniques (Sufian and Jones, 2017).

Chromatic Optical Transmission

System Structure and Outputs

The form of an optical transmission system for addressing an oil sample in a transparent cuvette is shown in Figure 7.1a. This indicates the interconnection between a polychromatic light source

Optical transmission through transformer oils

FIGURE 7.1 Optical transmission through transformer oils (Elzagzoug. 2013). (a) Schematic diagram of optical transmission system, (b) View of the portable chromatic oil monitoring system (PCOMS). (c) Typical cuvette images (i) with and (ii) without oil.

(e.g., LED, VDU). oil-containing cuvette, wideband optical detector (e.g., spectrometer, mobile phone camera) and data processing computer. A camera-based benchtop form of this portable chromatic oil monitoring system (PCOMS) (Figure 7.1b) has been used for laboratory development tests. A typical image obtained with the arrangement and a cuvette containing oil is shown in Figure 7.1c.

The possible permutations of source (LED, VDU) and detection units (spectrometer, mobile phone) which have been assessed (Elzagzoug, 2013) are shown in Table 7.1.

This has enabled the operation of the VDU-mobile phone camera system to be compared with a conventional measurement system (LED-spectrometer), consistent with the considerations of Chapter 2. The options for the camera operation and so on were based upon the illumination/ modulation/processing/sensor (IMPS) diagram (Chapter 2); see Figure 7.2.

Although the transmission spectra obtained w'ith a VDU source are not continuous, as are those obtained with a LED, the variation with different oil samples is similar. See Figure 7.3.

TABLE 7.1

Permutations of Sources and Detectors Tested

Source

Detector

LED

Spectrometer

LED

Mobile Phone Camera

VDU Screen

Spectrometer

VDU Screen

Mobile Phone Camera

IMPS permutation of chromatic components properties for oil transmission tests (chosen operational parameters highlighted)

FIGURE 7.2 IMPS permutation of chromatic components properties for oil transmission tests (chosen operational parameters highlighted).

Transmission spectra of cuvette and various oil samples (a) LED source; (b) VDU

FIGURE 7.3 Transmission spectra of cuvette and various oil samples (a) LED source; (b) VDU.

Chromatic Processing

Chromatic processing of the R, G, В outputs yields XYZ chromatic maps (Chapter 1), with Z representing relative changes in the short-wavelength regions and X representing changes in the long-wavelength regions (Figure 7.4).

Tests with real transformer oils from Electricity North West (ENW) indicated that chromatic changes were detectable for different levels of oil degradation. The map shows that a convenient chromatic parameter for representing changes in oil degradation can be based upon the gradient of the locus from the origin to the oil sample point, that is, Fn[X(0)/Z(0)] (Elzagzoug et al„ 2014). Validation of this chromatic approach has been proven by comparing Fn[X(0)/Z(0)] values with colour index test results (Elzagzoug et al., 2014) (Figure 7.5).

An indication of trends in short-wavelength optical absorption (A) and long-wavelength light scattering (Sc) have been derived from the optical transmission results (R, G. B) (Sufian and Jones, 2017)

Both parameters have been shown to increase with the oil degradation level (Sufian and Jones, 2017) (Figure 7.6).

Summary

A chromatic system based upon a VDU illumination source and a mobile phone camera has been shown to be capable of monitoring optical transmission changes due to the degradation of transformer oils. The transmission results have been further analysed to provide chromatic insight into short-wavelength optical absorption and long-wavelength scattering which gives an indication of the presence of particles and sludge in the oil.

X. Y, Z Cartesian chromatic map for 14 oil samples from in-service transformers [(Z(O), X(O) are normalised Z, X values] (Elzagzoug et al.. 2014)

FIGURE 7.4 X. Y, Z Cartesian chromatic map for 14 oil samples from in-service transformers [(Z(O), X(O) are normalised Z, X values] (Elzagzoug et al.. 2014).

Comparison of chromatic parameter (Fn[X(0)/Z(0)]) with colour index results for 14 oil samples (Elzagzoug et al., 2014)

FIGURE 7.5 Comparison of chromatic parameter (Fn[X(0)/Z(0)]) with colour index results for 14 oil samples (Elzagzoug et al., 2014).

Variation of transmission-derived parameters with oil degradation

FIGURE 7.6 Variation of transmission-derived parameters with oil degradation (Sufian and Jones. 2017). (a) Short wavelength (absorption) (A); (b) long wavelength (scattering) (Sc). A = [ 1 - (Bt/Bo)/Gt/Go]; Sc = 2{[(Rt/ Ro)/(Gt/Go)] -1).

Chromatic Optical Fluorescence

System Structure and Outputs

An optical system which has been used for monitoring fluorescence produced by oil samples (Sufian and Jones, 2017) consisted of a ultraviolet (UV) laser light source (405 nm, 10 mW), addressing an oil sample contained in a cuvette and addressed with either a spectrometer or electronic camera whose output was fed into a data processing computer (Figure 7.7).

The options for the camera operation and so on were based upon the IMPS diagram (Chapter 2) (Figure 7.8).

The fluorescent emission from the oil sample was captured in a direction orthogonal to the transmission path of the exiting laser beam. Examples of images obtained w'ith a camera are shown in Figure 7.9a, which shows the differences between water and clean and degraded transformer oil

Optical fluorescence through transformer oils

FIGURE 7.7 Optical fluorescence through transformer oils (Sufian and Jones. 2017). (a) Schematic diagram of general system structure for chromatic optical fluorescence monitoring; (b) view of the chromatic fluorescence oil monitoring system.

samples. The spectra of the fluorescence produced by various oils captured with a spectrometer (wavelength range 400-700 nm) also differed from each other (Figure 7.9b) (Lo et al., 2017), but in a complex manner.

Chromatic R, G, В values corresponding to fluorescence from various oils were extracted from the camera images and from the spectra - the latter using R, G. В processors, as shown in Figure 7.9b.

Chromatic Mapping and Calibration

Comparison of the X, Y, Z chromatic maps obtained from the spectra (Rs, Gs. Bs) and the camera outputs (Rc, Gc. Be) (Figure 7.10) showed similar trends, but with the camera having more sensitivity in X and less in Y (Lo et al., 2017).

IMPS permutation of chromatic component properties for oil fluorescence tests (chosen operational parameters highlighted)

FIGURE 7.8 IMPS permutation of chromatic component properties for oil fluorescence tests (chosen operational parameters highlighted).

Oil fluorescence excited via UV laser beam

FIGURE 7.9 Oil fluorescence excited via UV laser beam (405 nm) (Lo et al., 2017). (a) Camera images (water, clean oil, degraded oil), (b) Fluorescence emission spectra for seven transformer oils (S1-S7) and R. G, В processor profiles.

Chromatic X, Y, Z fluorescence maps for seven transformer oils (S1-S7) (Lo et al., 2017). (a) Derived from Rs. Gs, Bs of optical spectra, (b) Derived from Rc. Gc. Be camera outputs

FIGURE 7.10 Chromatic X, Y, Z fluorescence maps for seven transformer oils (S1-S7) (Lo et al., 2017). (a) Derived from Rs. Gs, Bs of optical spectra, (b) Derived from Rc. Gc. Be camera outputs.

Chromatic nominal spread [S = (G - B)/(G + B)] for seven transformer oils (S1-S7) (Lo et al., 2017). (a) Derived from Gs. Bs of optical spectra, (b) Derived from Gc. Be camera outputs

FIGURE 7.11 Chromatic nominal spread [S = (G - B)/(G + B)] for seven transformer oils (S1-S7) (Lo et al., 2017). (a) Derived from Gs. Bs of optical spectra, (b) Derived from Gc. Be camera outputs.

The chromatic spread (S = (G - B)/(G + B)) varied with oil degradation in a similar manner for both the camera and spectra data (Figure 7.11), indicating S has potential as a calibration parameter (Lo et al., 2017).

Summary

The chromatic approach was used with an electronic camera for monitoring laser-induced fluorescence in transformer oils without the need for expensive spectrometers (Lo et al., 2017). The chromatic spread parameter (S) was used as a calibration parameter for detecting oil degradation.

Chromatic Combination of Transmission and Fluorescence Parameters

A compact, portable unit Optical Chromatic Transformer Oil Monitoring (ОСТОМ) unit has been developed which combines addressing optical absorption, scattering and fluorescence by an oil sample and operates automatically, processes the data chromatically and transmits the results wirelessly to a central hub (Sufian et al., 2018).

The optical absorption, scattering system structure (Figure 7.1a) and fluorescence system structure (Figure 7.7a) are combined into a single unit (Figure 7.12) for automatic operation. Illumination for absorption and scattering is produced by a miniature 4-inch visual display unit to provide tuneable, polychromatic light, whilst a monochromatic light beam for fluorescence is provided by an ultraviolet laser diode.

The optical outputs are captured by an electronic camera, and a miniature liquid crystal dis[play (LCD) screen displays preliminary test results. The system is controlled by a Raspberry Pi unit (Figure 7.12b), which also enables data to be transmitted wirelessly.

A typical image is shown in Figure 7.13 of an oil-filled cuvette illuminated simultaneously by the VDU for absorption and scattering evaluation and by the laser for fluorescence evaluation. The main areas in this image addressed for analysis were St for absorption/scattering, Sf for fluorescence and Ref for reference. Also shown are the three black points for image orientation checking.

The monitoring unit extracts R. G, В values from the transmission and fluorescence images and transforms them into reliable chromatic parameters (A, Sc, Flo; Sections 7.2.2 and 7.3.2) for

Combined transmission, scattering and fluorescence chromatic monitoring unit

FIGURE 7.12 Combined transmission, scattering and fluorescence chromatic monitoring unit (ОСТОМ) (Sufian et al„ 2018). (a) View of unit; (b) block diagram of the overall system; (c) schematic diagram of the optical layout of the combined transmission and fluorescence chromatic system (ОСТОМ).

Location of reference and oil sample areas on images for analysis (Sufian and Jones, 2017). (St = oil sample transmission, Sf~oil sample fluorescence. Ref = reference area.)

FIGURE 7.13 Location of reference and oil sample areas on images for analysis (Sufian and Jones, 2017). (St = oil sample transmission, Sf~oil sample fluorescence. Ref = reference area.)

Overall chromatic calibration graphs

FIGURE 7.14 Overall chromatic calibration graphs (Sufian and Jones, 2017). (a) Chromatic oil degradation level (Lv versus transformer oil), (b) XYZ secondary chromatic map [X = Absorption (A).Y = Fluorescence (Flo),Z= Scattering (Sc)].

producing calibration graphs. The chromatic parameters are each normalised empirically to have a range of 0-1. A, Sc, Flo are then treated as three chromatic outputs (R. G, B) to produce X, Y, Z chromatic map parameters with a chromatic effective strength parameter [Lv = (A + SC + Flo)/3] which is indicative of the overall level of oil degradation. An empirically determined calibration graph of Lv versus degradation level is given in Figure 7.14a, which shows that normal oil samples are indicated by Lv < 0.33 and highly degraded oil samples correspond to Lv > 0.66, w'hilst moderately degraded oil samples lie in the range 0.33 < Lv < 0.66. An indication of which optical property dominated any degradation [Absorption (A), Scattering (Sc), Fluorescence (Flo)] is given by the coordinates of an oil on the X, Y, Z chromatic map, an example of which is showrn in Figure 7.14b.

Summary and Overview

The process for combining the outputs from optical absorption, scattering and fluorescence for transformer oil samples using the portable ОСТОМ unit is shown in the flowchart in Figure 7.15 (Sufian and Jones, 2017).

Evaluation of the automated chromatic ОСТОМ unit with 81 real oil samples (TNB. Malaysia) confirmed its capability for distinguishing degraded oil samples when compared with Duval analysis (Singh and Bandyopadhyay, 2010), dissolved gases analysis (Guardado et al., 2001), acidity and so on. The ОСТОМ unit had an accuracy of 80% and a confidence level of high concern of 90% when compared with the combined predictions of Duval and other test methods (Sufian and Jones, 2019).

Further developments of the technique are possible. For example, the system response may be varied by utilising different camera settings (IMPS; Figure 7.8) (Sufian and Jones, 2017) to optimise sensitivity for particular types of oils and for optimising separately with regard to absorption, scattering and fluorescence. In addition, the raw' values of R, G, В rather than the camera’s automatically processed values may be acquired (A. A. Al-Tememy, private communication), whilst the chromaticity of the transmission illumination produced by the VDU screen can be varied via the ОСТОМ software. Fluorescence detection could also be varied by changing the camera setting (e.g., daylight to fluorescent) and incorporating a second laser to provide a laser beam of a different wavelength. Consequently, the ОСТОМ unit has a high degree of versatility.

Chromatic combination decision chart for transformer oil samples using optical transmission and fluorescence (Sufian and Jones. 2017)

FIGURE 7.15 Chromatic combination decision chart for transformer oil samples using optical transmission and fluorescence (Sufian and Jones. 2017).

Acknowledgements

The provision of oil samples by Tenaga National Berhad Research (TNBR) in Malaysia and

Electricity North West Ltd (ENW) in the UK is acknowledged.

References

Bureau of Reclamation, U.S. Department of the Interior, Tech Svcs Group Hydroelectric Research. (April 2005) “Transformers: Basics, Maintenance and Diagnostics.”

Elzagzoug, E. (2013) “Chromatic monitoring of transformer oil condition using CCD camera technology.” PhD Thesis, University of Liverpool.

Elzagzoug. E., Jones, G.R., Deakin, A.G., and Spencer. J.W. (April 2014) “Condition Monitoring of High Voltage Transformer Oils Using Optical Chromaticity.” Institute of Physics, Measurement Science and Technology, 25(6), 9.

Guardado, J. L., Naredo, J. L., Moreno. P.. and Fuerte, C. R. (2001) “A Comparative Study of Neural Network Efficiency in Power Transformers Diagnosis Using Dissolved Gas Analysis.” IEEE Trans. Power Delivery. 16(4), 643-647.

Lo. С. K.. Looe, H. M., Sufian, A. T, Jones, G. R.. and Spencer, J. W. (2017) “Transformer Oil Degradation Monitoring with Chromatically Analysed Optical Fluorescence.” Proceedings of the International Conference on Imaging. Signal Processing and Communication (ICISPC 2017). Association for Computing Machinery, New York, NY, USA, 171-175. doi: 10.1145/3132300.3132329

Lorin, P. (April-May 2005) “Forever Young [Longer-Lasting Transformers].” Power Eng, 19(2), 18-21. doi: 10.1049/pe:20050203

Nadkarni, R. A. (2007) Guide to ASTM Test Methods for the Analysis of Petroleum Products and Lubricants, vol 44 (Philadelphia. PA: ASTM International).

Singh, S.. and Bandyopadhyay, M. N. (2010) “Duval Triangle: A Noble Technique for DGA in Power Transformers.” International Journal of Electrical and Power Engineering, 4(3), 193-197.

Sufian. A. T., and Jones, G. R. (2017) Chromatic Monitoring Methods for High Voltage Transformer Oils Technical Report I. The Center of Intelligent Monitoring Systems, The University of Liverpool.

Sudan, A. T., and Jones, G. R. (2019) Optical Chromatic Transformer Oil Monitoring (ОСТОМ) Unit Performance and Evaluation Technical Report. The Center of Intelligent Monitoring Systems. The University of Liverpool.

Sufian. A. T., Jones, G. R„ and Smith, D. (2018) Optical Chromatic Transformer Oil Monitoring (ОСТОМ) Performance and Evaluation Technical Report. The Center of Intelligent Monitoring Systems. The University of Liverpool.

 
Source
< Prev   CONTENTS   Source   Next >