PRACTICAL APPLICATION OF TAN DELTA DIAGNOSTIC TESTING IN CABLES

GJ Catlin, HV Diagnostics Sarl

Abstract Very Low Frequency (VLF) Tan Delta (loss angle) diagnostic testing is a popular tool used to evaluate the condition of a cable. Careful attention to detail must be made during the setup of the equipment in order to reduce the possibility of incorrect measurement. The analysis of the measured data can, if performed correctly, reveal valuable information about the condition of the cable. A number of operators of the equipment are not aware of the correct procedures that should be implemented when utilizing this type of equipment. Operators may also be unaware of the techniques applicable to analysis of the test data as well as some of the limitations of this diagnostic technique. The principles of Tan Delta (TD) testing are explained and the benefits of using this diagnostic tool are presented. The practical considerations of the measurement of the Tan Delta values are detailed and demonstrated. Actual sample test results are presented and analysed in order to demonstrate the application and benefits of this technique. The limitations of diagnosis using the only TD measurements are also considered. Unless otherwise stated the equipment referred to in this document is the HV Diagnostics VLF and TD equipment. Introduction Tan Delta (also referred to as loss angle) is the ratio of the resistive current to the capacitive charging current flowing in a cable. The resistive current is in phase with the applied voltage but the capacitive current leads the applied voltage by 90o. The vector sum of the two currents is the total current flowing.

Fig 1 : Tan Delta vectors

Conference on Power Cables CABLE TECH 2010 25 & 26 Nov. 2010

S IV - 246

In the context of cables in good to medium condition, the charging current is many times larger than the resistive current which translates into small TD ratios. The TD values are dimensionless and expressed as a percentage (%) or as thousandths (10-3). For PILC cables the expected ratios can be of the order of 50 times higher than that for XLPE cables. For a purely capacitive load the phase angle between the current and voltage waveforms would be 90o degrees ( /2 radians) and for a largely capacitive load the phase difference ( ) is slightly less than 90o. Measurement of the phase angle can allow the TD value to be computed and this is the more common technique in modern digital signal processor based equipment. Measurement bridges with standard capacitors are still commonly used for 50Hz Tan Delta measurement. Because the TD is a ratio, the resulting measurement is independent of the length of the cable. As the length of the cable increases so does the charging current and the DC loss current in a fixed ratio for a healthy cable or homogeneous unhealthy cable. TD equipment is generally capable of measuring resistive currents in the nA range. The Tan Delta measurement is therefore extremely sensitive to small changes in the dielectric health.Tan Delta readings are a global reading and cannot indicate the position of the damage along the length of the cable. The magnitude of the TD is not an indication of the magnitude of the damage to a cable in a single position but is an aggregate measurement of the total effect of damage in the cable. Concentrated damage in a single area will reflect similarly to distributed damage over a wider area although the former is more likely to cause short term failure of the equipment; if severe. Because TD is an aggregated variable, measurements of mixed-technology cables require more detailed analysis of the gradient and standard deviation of the TD readings at different voltages rather than purely a consideration of the magnitude.The effects of cable accessories such as joints and terminating cones are also aggregated into the TD reading and it is not trivial to separate the effects of the accessories from the cable. Trending of cable TD data over time can assist in the diagnosis of accessory damage. Some work has been published wherein it has been demonstrated that TD measurements taken at various frequencies (spectroscopic analysis) can be analysed and the effects of the accessories separated from that of the cable. Although promising there is no evidence of the success of this in the field and the laboratory experiments appear only to be successful for short lengths of cable. TD has been shown to be sensitive to many types of cable damage such as water ingress (the most commonly experienced damage) as well as to electrical damage in the form of partial discharge. TD features have also been shown to correlate well to VLF breakdown performance [10] TD testing can be performed at all stages of the lifecycle of a cable. TD testing after delivery to site of the cable and after installation can confirm that the cable and its accessories are healthy and will record a fingerprint useful for future comparison and trending. TD results at VLF cannot be compared to those obtained at 50Hz at the factory.
Conference on Power Cables CABLE TECH 2010 25 & 26 Nov. 2010

S IV - 247

TD testing can be performed on cables that are performing poorly but cannot be performed on cables that have been damaged with a short-circuit and are ineffective on cables with open-circuit fault. For cables with corroded shields the TD is dramatically affected, and although it is then unclear where the elevated TD is coming from - the insulation or the shield. The conclusions though often remain the same cable needs to be replaced. Safety Remember that you are working with potentially lethal test equipment and test conditions. Suitably certified operators should be used to switch the cable out of service and make the cable safe. Ensure that unused phases are earthed and that a proper station earth is available and used. During testing the HVDSA TD equipment is live at the test voltage hence the design of the equipment to be elevated from the surrounding earth using the collapsible tripod stand or elevated in air. For the HVDSA TD equipment the connection from the TD unit to the DUT is not designed to insulate at the test voltages. The test cable should therefore be treated as live and should not be routed close to any possible discharge paths more from a measurement than a safety perspective. Preparation and connection Prior to testing make sure that the batteries are charged and that you have a spare set of batteries. Also check that the Bluetooth communications on the laptop is setup and communicating with the TD set. Obtain the cable voltage ratings and prepare the test procedure understanding what voltages are to be tested, how many readings need to be taken and what the expected results would be for a good cable and for a damaged cable. Typically take the final test voltage and divide by 4 to get the 4 step voltages. So example if going to 3Vo for an 11kV cable, the test voltages 3Vo would be 18kV and then steps 4.5kV, 9kV, 13.5kV and 18kV. The TD testing should always be performed with an awareness of the factors influencing the test results. Remember that any small external DC charge movements that occur during the application of the voltage will distort the TD readings so care must be taken to eradicate all possible capacitive/charge/leakage paths. Corona and other partial/surface discharges will influence the measurements. Corona rings are supplied to be attached to both the ends of the phase being tested. Different grounding topologies have not been demonstrated to affect TD measurements from HVDSA equipment. Equipment where the current measurements are made on the low-side/current return path may be sensitive to earth return paths. It is good practice to always consider the effects of the earthing efficiency and configuration that you employ when making sensitive measurements.
Conference on Power Cables CABLE TECH 2010 25 & 26 Nov. 2010

S IV - 248

Where the Tan Delta measurement is made some distance from the DUT the effect of the test cable may need to be compensated for using a guarding circuit or equivalent technique. With the HVDSA equipment the TD equipment is remote from the voltage generator and close to the DUT - therefore there is no need to provide such compensation. The HVDSA Tan Delta equipment must be properly cleaned to ensure that no leakage currents can flow down its insulation. Ensure that the test cable from the TD to the DUT is suitably routed with a decent separation from any possible discharge/leakage points. Only use the lead provided and ensure that the distance between the TD unit and the DUT is as short as possible and take up the slack by coiling the HV lead. The cable terminations need to be thoroughly cleaned with appropriate solvents and free of any dirt or contamination to eliminate any leakage currents. The terminations should be suitably separated from each other and the surroundings and free of any sharp points that could cause partial breakdown. Prior to making this connection it is recommended to test the TD equipment by making a TD reading at 0.5U0 and at 1.5U0 in free space. For both voltages the TD reading should be very low, typically software limited to 0.1x10-3 (0.1%) for the HVDSA equipment. If possible repeat these measurements with a known test capacitance (and known TD) and ensure that the measured values are as expected. Calibration Tan Delta equipment should be calibrated at the recommended manufacturer intervals but should also be checked before testing to confirm the setup and DUT preparation. Setup The test can be performed using the manual or an automated test program from the HVA series VLF generators. Allow sufficient dwell time (typically 3 minutes) at each test voltage to obtain at least 6-10 readings. The first reading may take a little longer that the others. The VLF units should be set to generate sinewave shaped waveforms at 0.1Hz/auto. HVDSA recommends testing cables at least at three voltages 0.5 U0, U0 and 1.5U0 beginning at voltages lower than the operating voltage and proceeding to the next higher voltage only if the TD readings indicate that the cable is undamaged and therefore capable of withstanding increased voltage stresses. This greatly reduces the possibility of further damaging an already damaged cable during the testing. In some cases though, it may be useful to test the cable to fault. For longer cable lengths the frequency of the VLF signal will be reduced in order to supply the load current. When trending ensure that the TD readings are taken at the same frequency although values obtained over the VLF range have been found to be comparable.
Conference on Power Cables CABLE TECH 2010 25 & 26 Nov. 2010

S IV - 249

Testing When testing using the HVDSA TD system the measurements reported on the VLF unit LCD should be ignored. The testing is performed automatically by the TD and HVA units. The TD unit communicates wirelessly with the connected laptop running the TD control centre software. As the VLF unit applies the various test voltages these are automatically recorded by the TD software and the recordings are grouped accordingly. It is good practice to observe the measurements as they are reported. Consider the capacitance and the charging currents measured and equate this to the expected values for the DUT and compare the measurements from the different phases. As the test is quick and easy to perform there is always the possibility to make modifications to the test setup and repeat the process. As with all testing it is good advice to treat results with suspicion. Especially readings that look too good or are too bad could be due to poor measurement setup or technique rather than the DUT. Analysis

Fig 2 : Sample TD results showing voltage dependence of TD indicating onset of PD. Ideally TD values should be below an expected threshold for the cable type and age. The TD measured values should be consistent with voltage and the multiple TD values measured at the same voltage should not vary. The measured TD data should be analysed according to the following criteria

Conference on Power Cables CABLE TECH 2010 25 & 26 Nov. 2010

S IV - 250

(1) Absolute value of TD the TD readings taken at all voltages should be within the range expected for the cable type and age being tested. Type XLPE PILC PVC ** EPR * New < 1 x 10-3 < 50 x 10-3 No Data < 50 x 10-3 Aged > 4 x 10-3 > 50 x 10-3 No Data > 50 x 10-3

* EPR cable types have their own unique TD fingerprints ** Where no absolute values are available it is still possible to trend data from similar cables or from the same cable over time. (2) Stability of reading obtained at each voltage the TD measurements obtained at each voltage should be within < 0.04% standard deviations. Any wildly varying TD readings are indicative of damage to the cable. (3) Voltage dependence of TD a. the mean TD readings obtained at the different voltages should be within +- 20% of the average for all voltages. A steady increase of TD with test voltage is indicative of water damage. New XLPE cables will show a consistent TD for typical recommended measurement voltages. b. A sudden increase of TD with voltage often described as a tip up is associated with the onset of Partial Discharge activity. The PD can then be investigated with appropriate equipment. (4) Trending and comparison It is often useful method of analysing changes in the health of a cable. Consistent measurement techniques enable the comparison of TD values over time and within similar classes of cable. This is especially useful where the impact of accessories or hybrid sections needs to be accounted for. Analysis of TD data can lead to conclusions such as (1) Cable healthy return to service (2) Cable deteriorated monitor (3) Cable unhealthy investigate further (4) Cable damaged repair/replace as imminent failure likely. Conclusions Tan Delta measurements made on cables can be an accurate and thorough indicator of the dielectric health of a cable. Tan Delta measurements are automated and simple to perform but care must be taken during the preparation and setup in order to achieve accurate, repeatable results. The most common cause of XLPE cable failure is due to water ingress and TD is an effective tool that can be used to detect this. TD measurements are able to detect a wide variety of damage and with the application of feature trending techniques TD can be considered as an almost universal diagnostic tool.
Conference on Power Cables CABLE TECH 2010 25 & 26 Nov. 2010

S IV - 251

References [1] HV TEST, A handy guide to the safe over-voltage pressure testing and diagnostic monitoring of cable installations, HV Test Pty Ltd, South Africa, 2009 [2] HV Test, A handy guide to the safe on-site testing of electrical equipment, HV Test Pty Ltd, South Africa, 2009 [3] HV Diagnostics Sarl, TD30 Operators Manual, HV Diagnostics Sarl, Switzerland, 2010 [4] IEEE, IEEE Guide for Field Testing and Evaluation of the Insulation of Shielded Power Cable Systems, IEEE Std. 400-2001, Apr. 2002. [5] IEEE, Draft: IEEE Guide for Field-Testing of Laminated Dielectric, Shielded Power Cable Systems Rated 5kV and above with High Direct Current Voltage, IEEE Std. 400.1, Draft 5 D5, Oct. 2004. [6] IEEE, IEEE Guide for Field Testing of Shielded Power Cable Systems Using Very Low Frequency (VLF), IEEE Std. 400.2-2004, Feb. 2005. [7] TOSHIKATSU, T. and GREENWOOD, A. Advanced power cable technology: present and future, CRC Press Inc., Vol. I and II, Boca Raton, FL, 1983, ISBN 08493-5165-0 (Vol. I) and 0-8493-5166-9 (Vol. II). [8] ORTON, H. and R. HARTLEIN, R. Long-life XLPE-insulated power cables,Internal Publication by Dow Wire and Cable and Borealis, Somerset, NJ, 2006. [9] Brugg Kabel AG, HIGH VOLTAGE XLPE CABLE SYSTEMS Technical User Guide, Brugg Kabel AG, 2006 [10] JC Hernndez Meja CHARACTERIZATION OF REAL POWER CABLE DEFECTS BY DIAGNOSTIC MEASUREMENTS, A Thesis Presented to The Academic Faculty In Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the School of Electrical and Computer Engineering, Georgia Institute of Technology, December 2008

Conference on Power Cables CABLE TECH 2010 25 & 26 Nov. 2010

S IV - 252