MV Power Cable Diagnostics by Frequency Domain Spectroscopy. Peter Werelius Programma Electric AB

Transcription

1 MV Power Cable Diagnostics by Frequency Domain Spectroscopy Peter Werelius Programma Electric AB

2 Frequency Domain Spectroscopy Measurements of insulation capacitance and losses in a frequency interval

3 Frequency Domain Spectroscopy Accurate measurements of capacitance and loss at several frequencies in a selected frequency range Ageing in insulating materials affects the polarisation process, i.e. the capacitance and loss frequency characteristics A wide frequency range gives more information and makes it possible to separate different types of insulation materials different types of ageing influence of accessories

4 Application PILC Cables Dielectric response methods, e.g. frequency domain spectroscopy, and partial discharge measurements are complementing each other Dielectric response methods detect aging products such as moisture while partial discharge measurements finds local defects

5 Application PILC Cables Moisture is generated by aging accelerates the aging process Moisture is detectable by frequency domain spectroscopy measurements To detect moisture in paper insulation and other cellulose materials (e.g. transformer pressboard) low voltage measurements is preferably

6 IDA 200 Measurement at 120 Vrms 1 A 11 kv Paper Insulated Power Cable Tanδ Cable H17, ph.2 0,1 0,01 0,001 0,001 0,01 0, Frequency [Hz]

7 Impregnated Cellulose Paper Moisture content Temperature Type of paper High density pressboard Low density pressboard Type of Kraft paper and more

8 Effect of Moisture 10 1 Paper with different moisture contents Tanδ Temperature ca. 22C W.C.= 0.18% W.C.= 1.52% W.C.= 2.12% W.C.= 2.60% W.C.= 3.51% 0,1 50 Hz 60 Hz Loss minimum 0,01 0,001 0,001 0,01 0, Frequency [Hz]

9 Effect of Temperature 10 Dry paper (< 0.5%) at different temperatures Tanδ 20 o C 40 o C 1 60 o C 80 o C 0,1 50 Hz 60 Hz 0,01 Loss minimum 0,001 0,001 0,01 0, Frequency [Hz]

10 Determination of Moisture Content 10 1 Paper with different moisture contents, Tanδ and Cable H17, Ph.2 W.C.= 0.18% W.C.= 1.52% W.C.= 2.12% W.C.= 2.60% W.C.= 3.51% Cable H17, ph.2 0,1 0,01 0,001 0,001 0,01 0, Frequency [Hz]

11 Determination of Moisture Content Moisture content = Approximately 1 % (OK) Observations made: The loss minimum increases with moisture content The loss minimum seems to be almost independent of temperature The loss minimum seems to be almost independent on type of Kraft paper

12 Summary The frequency domain spectroscopy technique is a good tool to estimate the average moisture content in PILC cable By use the loss minimum loss value, the influence of temperature and type of Kraft paper is minimized The frequency domain technique, that record the characteristic frequency dependence in capacitance and loss make it possible to separate moisture from accessories or leakage currents

13 Application MV XLPE Cables Water trees are growing in the insulation and lower the electrical withstand The water tree aging process is very slow A heavily aged cable fail if the insulation stress is increased (Lightning impulses, faults, etc)

14 Application MV XLPE Cables Voltage tests shorten the cable life significantly Non-destructive diagnostics saves money Preferably measured at high voltage levels Non-destructive diagnostics (no voltage test) IDA200 with an external high voltage unit

15 Measurement Procedure (Medium Voltage XLPE Cables) Measurements of short frequency sweeps around 0.1 Hz at several voltage levels up to service voltage level, Uo Preliminary judgement of the cable If the cable is judged good, a slightly higher voltage levels can be used in order to detect aging in some cable designs

16 The Response of Water Tree Deteriorated XLPE Cables VDP Response (Voltage Dependent Permittivity) A voltage dependent increase of loss and capacitance. TLC Response (Transition to Leakage Currents) A VDP response at initial low voltage levels. At a higher voltage level, the response changes characteristics. Leakage currents are added. LC Response (Leakage Currents) Leakage currents through water trees are present already at low voltage levels.

17 VDP response: A voltage dependent increase of loss and capacitance Capacitance part Loss (Tanδ) part kv 4,5 kv 6 kv 3 kv 6 kv 3 kv ε'' ε'' ε'' ε'' ε'' ε'' ε'' 3 kv 4,5 kv 6 kv 3 kv 6 kv 3 kv ,01 0, Frequency (Hz) ,01 0, Frequency (Hz)

18 TLC response: The response change characteristics. A leakage current is added Capacitance part Loss (Tanδ) part ,5 kv 3 kv 4,5 kv 6 kv 3 kv ε" ε" ε" ε" ε" ε" at 1,5 kv at 3 kv at 4,5 kv at 6 kv at 3 kv ,01 0, Frequency (Hz) ,01 0, Frequency (Hz)

19 LC response: Leakage current through water trees present at low voltage levels Capacitance part at 3 kv at 6 kv at 9 kv ε" Loss (Tanδ) part ε" ε" ε" at 3 kv at 6 kv at 9 kv ,01 0, Frequency (Hz) ,01 0, Frequency (Hz) The loss curves have -1 slope in the log-log diagram

20 Medium Voltage XLPE Cables Relatively low voltage levels are used in order to ensure non-destructive measurements. By non-destructive measurements cable replacement can be scheduled and delayed. The response of water trees are identified and classified into three different groups By recognise the response of water trees, influence of accessories can be separated IDA200 and an external high voltage unit

21 12 Cable Faults in North Botkyrka Faults/100 km/year Year

22 Measured Response versus ACBD Level ε at U, 1 Hz *) See below Loss versus Breakdown Voltage Leakage at U<Uo Leakage at Uo ε at U, 1 Hz *) See below Voltage Dependence in Loss versus Breakdown Voltage ε nonlin at U 0, 1 Hz Leakage at U<Uo Leakage at Uo at U, 1 Hz 0 ε nonlin U bd /U 0 U bd /U 0 *) Cable samples with leakage currents are placed in the upper part of the graph without any loss value associated